Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. ANION MOVEMENT I N A STRUCTURED SOI L A thes i s presented i n part ial ful fi lment of the requirements for the Degree of Doctor of Phi l osophy in Soi l Sc i ence at Massey Uni versity P IMPAN KANCHANASUT 1980 tn . _.ta ABSTRACT Anion movement in so i l was studied both i n the l aboratory and in the fi e l d , us ing structured "undi sturbed" soi l and s i eved aggregates . The movement of ch l oride , bromide and phosphorus was investigated . Ch loride and bromide being non-react ive in soi l , were used to indi cate whether the f low was more uni form or preferential , whi l e phosphorus was used to indicate the behaviour of strongly adsorbed anions . Experiments invo lving the movement of ch l oride and phosphorus through columns of 0 . 5 - 1 mm soi l aggregates provided data on phosphate adsorption during misc ib le d i sp lacement . Chl oride breakthrough curves were described we l l by convent ional convective -d i spersive theory . For phosphorus , l inear adsorption i sotherms were determined i ndependent l y , using so lut ion concentrat ions and equ i l ibrium t imes s imi l ar t o those pertaining in the aggregate columns . Conventional theory using these data predi cted reasonab ly we l l the ear l y part of the breakthrough curves , but did not predict the observed "tai l ing" of phosphorus breakthrough curves . The movement of anions through art ific ia l soi l channe l s and p l anar cracks was studied . The breakthrough curves showed the movement of both chloride and phosphorus was highly preferential through 0,5 mm diameter channel s and 0 . 1 7 mm wide p lanar cracks . The resu l t s agreed quite we l l with mode l predict ions . The movement of anions through 2 . 4 litre "undisturbed" soi l ? cores was also studied . Under saturated cond it ions , both chl oride and phosphorus moved preferential l y . Dye studies i nd i cated the maj or pathways were worm channe l s , root channe l s , and soi l cracks . Under unsaturated condit ions when the pressure potent ia l was maintained at -0.02 bar (at whi ch channe l s larger than 0 . 15 mm diameter and cracks wider than 0 . 0 7mm wou ld be drained) , the breakthrough curves for bromide were much less preferential than under saturated condit i ons . The experimental set -up for this experiment was des igned so that the b lockage of natural iii f l ow paths was min imi zed and the effects of porous p lates at e ither end of the cores were avoided . Two fie ld experiments were conducted at a mol e - t i l e drained s i te on Tokomaru silt l oam (a Fragiaqual f) . One experiment investigated the movement of ch loride and phosphorus so l ut ion ponded on the soi l surface . The breakthTough curves for both chl oride and phosphoYus percolat ing from the surface to the mo l e-drains indi cated the movement was very prefe rent ial , both an ions reaching the mo l e-drains located at 400 mm depth within a minute of the ir arp l icat ion to the soil surface . Dye stain ing indicated the movement occurred most l y through worm channe l s and p lant root channe l s assoc i ated w i th p l anar cracks . The other fi e l d experiment investigated the l eaching of bromide under both ponded water and natural rainfal l cond i t ions . When the same amount of water was consi dered , l eaching by rain fal l was more effect ivt than by ponding . Howeve r , under both wate? treatments , relat ive l y large amounts of app l ied bromide remained un l eached near the so i l surface , whjl e some bromide moved deep into the soi l profi l e . I nterception and stem f low appeared to be important factors causing non-uni form l eaching under pasture by natural rainfal l . Very cons iderab l e variat ion in bromide concent rat ion between rep l i cate soil samp l e s was found , with a l og-normal rather than normal di stribut ion . Quite di fferent l each ing patterns were found in soi l under pasture and in a soi l whi ch had been cul t ivat8d and cropped . ACKNOWLEDGEMENT I wish to expres s my sincere thanks to Or D . R ? Scatter , for h i s gui dance , cri t i c i sm , d i scussion , and encouragement throughout th i s study . Also for h is fri endsh ip , interest , and he lp offered to me at a l l t imes . Thanks are al so ext ended to Mr R . W . Ti l lman , and Dr J . K . Syers , So i l Sc ience Department , Massey Un ivers ity , for the ir he lp and valuab l e d iscuss ion . The ach ievement of thi s study would not be pos s i b l e w ithout a Ca l umba P lan Scho larshipoffered by the New Zealand Government , and study l eave g iven hy Khan Kean Un ivers i t y , Thai land , whi ch I very much apprec iate . To Rob in Ti l lman and John Sykes , thanks for their as s i stance w i th soi l samp l ing , and to Tipvanna and P isanu , for their as s i stance with the preparat ion of th is manuscript . To my fami l y and Suvi t , thanks for their cont inual support and encouragement at al l t imes . F inal l y to Mis s Vivienne Mair , for excel l ent typ ing and preparat ion of thi s manuscript . Abstract Acknowledgements Table of Contents List of F i.gures I. i s t of '1';1 h le s L i st o f Symbo l s TABLE OF CONTENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ' . . . . . . . . CHAPTER J GENERAL INTRODUCTION 1 . 1 IMPORTANCE OF SOLUTE MOVEMENT IN SO I LS 1 . 2 MOVEMENT OF SOLUTE IN SO I LS 1 . 2 . 1 Di fferent i a l Equat i ons Used tQ Describe Page ii iv V X XV i i xix 1 2 3 Solute Movement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1 . 2 . 2 So lut i ons of Transport Equat ions 1 . 2 . 2 . 1 Mi s cib l e displacement research 1 . 2 . 2 . 2 Leaching 1 . 2 . 3 Transport Mode l for React ive Solutes a . Equi l ibrium adsorption i sotherms b . Kineti c adsorpt ion mode l c . Combinat i on mode l 1 . 3 FAI LURES OF CONVENT IONAL THEORY 1 . 4 MOD I F I ED SOLUTE TRANSPORT MODELS a . b . Convect i ve- Dispers ive Mode l wi th Lateral D i ffus i on . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Viscous F l ow wi th Latera l D i ffus ion 1 . 5 GENERAL OBJECT I VES CHAPTER 2 ACCOUNTI NG FOR ADSORPT ION DURING PHOSPHORUS MOVEMENT IN SOIL 2 . 1 2 . 2 INTRODUCTION OBJECTIVES 2 . 3 MATERIALS AND METHODS 2 . 3 . 1 Convent iona l Batch Method Adsorpt i on 4 4 5 6 6 7 8 9 1 1 n 1 2 12 1 3 1 4 16 16 I sotherms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 6 v i . Page 2 . 3 . 2 Continuous- Leaching Method I sotherms . . . . . . . 1 7 2 . 3 . 3 Phosphorus and Ch l oride Movement i n Col umns of Soi l Aggregates . . . . . . . . . . . . . . . . . . . . . . . . . 1 7 2 . 4 RESULTS AND D I SCUSSION 19 2 . 4 . 1 Conventi onal Batch Method Adsorption I sotherms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2 . 4 . 2 Continuous- Leaching Method I sotherms 23 2 . 4 . 3 Phosphorus and Ch l oride Movement i n Col umns of Soi l Aggregates . . . . . . . . . . . . . . . . . . . . . . . . . 2 7 2 . 4 . 4 Mode l l ing of Phosphorus Movement i n Soi l Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2 . 5 GENERAL D I SCUSSION 2 . 6 CONCLUSIONS CHAPTER 3 AN ION MOVEMENT THROUGH ART I F I C IAL SOIL CHANNELS AND PLANAR 34 35 CRACKS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3 . 1 3 . 2 INTRODUCTION OBJECTIVE 3 . 3 MATERIALS AND METHODS 3 . 4 COMPUTATIONS 3 . 5 RESULTS AND D ISCUSSION 3 . 6 CONCLUSIONS CHAPTER 4 AN ION MOVEMENT IN SOI L CORES 4 . 1 4 . 2 INTRODUCTION OBJECTIVES 4 . 3 MATER IALS AND METHODS 4 . 3 . 1 Experimental Set-up 4 . 3 . 2 4 . 3 . 3 Saturated F l ow Experiment Unsaturated F l ow Experiment 4 . 4 RESULTS AND D I SCUSSI ON 4 . 4 . 1 4 . 4 . 2 Saturated F low Experiment Unsaturated F l ow Experiment 4 . 4 . 3 Computati ons 4 . 4 . 3 . 1 Convective-d i spersi ve mode l 37 38 39 40 40 46 48 49 5 1 5 1 5 1 5 3 5 4 54 54 60 63 63 4 . 5 4 . 6 v i i. Page 4 . 4 . 3 . 2 Vi scous fl ow with l?teral diffusi on mode l . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 GENERAL D ISCUSSION CONCLUSIONS CHAPTER 5 65 65 WATER AND AN ION MOVEMENT TO MOLE DRA INS . . . . . . . . . . . . . . . 67 5 . 1 I NTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 5 . 2 OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 5 . 3 MATERIALS AND METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 5 . 4 RESULTS AND D ISCUSSION 72 5 . 5 PRACTICAL IMPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 5 . 6 CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ? 79 CHAPTER 6 BROMIDE LEACHING UNDER F IELD CONDITIONS . . . . . . . . . . . . . . . 80 6 . 1 I NTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1 6 . 2 OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 6. 3 MATERIALS AND METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 6 . 4 6 . 3 . 1 6 . 3 . 2 6 . 3 . 3 Experimental Des i gn ...................... . 6 . 3 . 1 . 1 Continuous ponding experiment 6 . 3 . 1 . 2 Natura l rainfa l l experiment Bromide Retent ion by Soi l Bromide Retent ion by P lants 6 . 3 . 4 Bromi de Determinat ion 84 84 85 86 86 86 86 88 88 89' 6 . 3 . 4 . 1 Soi l samp les 6 . 3 . 4 . 2 P 1 ant t i s sues ................... . 6 . 3 . 5 Leaching o f Dye under Natural Rainfa l l 6 . 3 . 6 Soi l Bul k Dens ity ........................ . RESULTS AND DISCUSSION .......................... . 89 6 . 4 . ) Continuous Ponding Experiment . . . . . . . . . . . . . 89 6 . 4 . 1 . 1 Pre- l eaching measurement . . . . . . . . 89 6 . 4 . 1 . 2 Post - leaching measurement . . . . . . . 94 6.4.2 Natural Rainfal l Experiment . . . . . . . . . . . . . . . 9 7 6 . 4 . 2 . 1 Pre- l eaching measurement . . . . . . . . 9 9 6 . 4 . 2 . 2 Post-leach ing measurement . . . . . . . 99 a. F i rst samp l ing . . . . . . . . . . . . . . . 99 b . Second samp l ing . . . . . . . . . . . . . . 102 vi i i . Page 6 . 4 . 3 Variab i l i ty in Bromide D istri but i on 102 6 . 4 . 4 Leaching F l ow Pathways under Natura l Rainfa l l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 06 6 . 4 . 5 Pos s ib l e Factors Tend ing to Cause Low Bromide Recovery Percentages . . . . . . . . . . . . . . . . . . . . . . 1 10 a . Retent ion of bromide b . Retent ion of bromide c . Non-un i form bromide d . Retent i on of bromide e. Soi l samp l ing method 6 . 4 . 6 Computations 6 . 5 CONCLUSI ONS APPEND I X A in soi l by plants . . . . . . . . . . . . . . . . . . app l i cat i on . . . . . . . on p l ants . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 0 1 1 0 1 1 0 1 1 1 1 1 1 1 1 1 1 16 SOIL PROFILE DESCRIPTIONS AND PHYSICAL AND. CHEM ICAL DATA 1 1 8 APPENDI X B DIFFUS ION OF SOLUTE I NTO SPHERICAL SOIL AGGREGATES 1 24 APPENDIX C THEORY OF PREFERENTIAL SOLUTE MOVEMENT THROUGH LARGER SOI L VOI DS AND TYP I CAL COMPUTOR PROGRAMMES . . . . . . . . . . . 1 2 7 C . 1 "Preferent ia l so lute movement through l arger soi l voids . I . Some computati ons us ing s imp l e thoery . " by D . R . Scatter , reprinted from Austra l i an Journal of So i l Research . 16 : 2 5 7 - 267 . . . . . . . . . . . . . . . . . . 1 2 8 C . 2 L ist of Symbols in CSMP Programmes 140 C . 3 Programme for Miscible D i splacement of So l utes in a Soil with Uni form Verti cal Cylindr i cal Channels . . . . . . . 1 4 2 C . 4 Programme for Leaching of Surface App l ied So lutes under Ponded Water in Soi l Column Containing Uni formly Distributed Vert i ca l Cy l indrical Channe l s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 50 APPENDIX D DER I VIATI ON OF EQUAT ION FOR FLOW THROUGH NON- UNI FORM CAP I LLARY TUBES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 57 ix . Page APPENDfX E BROMI DE CONCENTRATION MEASUREMENTSFOR IND IV I DUAL REPL I CATES UNDER CONTI NUOUS POND I NG COND IT IONS 160 APPEND I X F FRACTI LE D I AGRAM CONSTRUCTION 164 RITEI!FNCFS 168 F ig . 2 . 1 F i g . 2 . 2 F i g . 2 . 3 F ig . 2 . 4 F i g . 2 . 5 F i g . 2 . 6 LIST OF F I GURES Page Schematic d iagram of experimental set-up for studying phosphorus and chloride movement i n soi l aggregate co l umns Adsorpti on data and fitted Freundl i ch i sotherms , at two equi l ibrat ion t imes , 3 hours ( -) and 10 hours (- - -) , obtained from the batch method . Arrows indicate the change in phosphorus concentrati on in the s o lut ion due to adsorption . The l inear i sotherms a l so shown were forced through the Freund l i ch i sotherms at 7 and_ 20 ?g/ml . They were used to obtain so lut i on d i stribution coeffic ients (k) for predicting phosphorus movement in the soi l co lumns Phosphorus adsorpt i on rates for d i fferent i n i t ia l so lut ion concentrati ons (C) . The curves shown were fitted to the experimental 18 20 data using a power curve fitting procedure 2 2 Phosphorus and ch l oride breakthrough curves for continuous l eaching experiments . The three leve l s of phosphorus concentrati on are : (a) 1 , (b ) 5 , and (c) 1 0 ?g/ml . . . . . . . . . . . . . 2 4 Comparison of adsorpti on i sotherms obtained from batch and continuous l eaching methods at 6 and 10 hour equi l ibrat ion t imes . For the batch method the curves shown are the Freundl i ch i sotherms 2 5 Dup l icate experimental and calculated break? through curves for (a) chl oride and (b) phosphorus -6 -1 at 9 x 1 0 m s flux dens ity . . . . . . . . . . . . . 2 8 Fi g . 2 . 7 Fig . 3 . 1 F ig . 3 . 2 F i g . 3 . 3 F i g . 4 . 1 F i g . 4 . 2 Dup l i cate experimental and cal cu lated break ? through curves for (a ) ch loride and (b ) phosphorus at 4 . 7 x 1 0-5 m s - 1 f lux dens i ty Breakthrough data for soi l casts containing a s i ng le channe l . (a ) Breakthrough data for s o i l cast w i th exposed soi l surface; Cast I ch l oride ( ?) and phosphorus ( o); Cast I I , chloride ( ? ) and phosphorus? ( !!. ) ? ( b ) Breakthrough data for so i l cast with a lmost c omp lete l y wax-coated surface ; Cast TTI, ch l ori de (?) and phosphorus (o). Al so shown are predi cted breakthrough curves for Cast I and I I , ch loride (--1 and phosphorus (- --) , as suming R = 69 for phosphorus Movement of rhodamine B dye so lut i on in the soi l casts contain ing art i fi c i a l ( a) verti ca l x i . Page 29 4 1 channe l , and (b ) plana r crack . . . . . . . . . . . 4 2 Breakthrough data? for soil casts containing a s ing le crack ; Cast IV, ch loride ( ?) and phosphorus ( o); and Cast V, chloride (!!.) and phosphorus ( ?). Also shown are predi cted breakthrough curves for Cast I V, chlori de (-) and phosphorus (-- -) ass uming R = 69 for phosphorus . The c irc l ed symbol s i nd i cate t ime for one pore volume Genera l experimental set -up for mi s c ib l e d i sp lacement study . . . . . . . . . . . . . . . . . . . . . 4 5 5 2 Breakthrough data for undi sturbed so i l cores ; ch loride (e) and phosphorus ( .A.) for saturated fl ow, and bromide ( ?) for unsaturated f low at - 200 mm pressure potential. Calculated break? through curves for bromide are presented , assuming channel d i ameter of 0 . 15 mm (-) and 0 . 1 mm ( - --). Dup l i cate cores ( a ) and (b ) . . . . . . . 5 7 Fig. 4.3 F i g . 4 . 4 ? i g . 4.5 Fig . 5.1 F i g . 5.2 Cros s - sect ion s 1 4 3 mm in d iameter of a soil core . The b l ue co l our ind i cates the dominant pathways for saturated f l ow and the pi nk co l our i nd icates the dom i nant pathways for unsaturated f l ow at -200 mm pressure potent i al . W = worm xi i . Page channel , R = root channe l , C = planar crack 58 Ver t i ca l s ect ion s o f a soil core . The blue co l our indicates the domi nant pathways of saturated f low and the p i nk colour indicates the dominant pathways of unsaturated f l ow at -200 mm pressure potent i a l . W = worm channel , R = root channe l , C = planar crack Unsaturated breakthrough data for bromide ( ? , o ) and the corresponding pred i cted curves (-,- -) obta i ned from a convective? d i spers ive model Experimental set -up for the miscible d i splacement exper iment above a mole drain . The i nfi l t rometer r i ng was 380 mm in d i ameter 59 62 and the mo l e channe l l ocated at 400 mm depth 71 Breakthrough data for saturat ed so il profile above the mo le d ra in: chloride (.__. ) and phosphorus (? ) for Experiment I w i th infiltration rate of 6.75 x 1 0-5 m s-1, and chloride ( ....... ) and phcsphorus (o-o) for Expe r i ment I l wi th i n filtrat i on rate of 6 . 67 x 10-6 m s- 1? The curves have been v i sually fi tted to the data point s . Times co r re sponding to one pore vol ume were 51 minutes , and 9.5 hours for Experiment I and I I , respect ive l y . C /C . i s the rat i o of effluent e 1 to influent concentration 73 Pi g . 5 . 3 Fi g . 5 . 4 F i g. 5 . 5 Fi g . 6. 1 F i g. 6 . 1 Preferentia l pathways observed in the soi l profi le ahove the mo l e drains, as i nd i cated by methy l ene blue staining . W = worm channe l , R = root channe l . ( a) Preferent ia l pathways in the soil profi l e d i rect l y above the mo l e. x i i i . Page (b) Conducting wonn channels and root channe ls 75 Mo le drain at 400 mm profi l e depth in Tokomaru s i l t l oam soi l i n Dajry Farm No . 4 near Massey Un iversity . Grass roots penetrat i ng into the mole were commonly observed Preferent ia l pathways observed in the soi l profi l e . W = worm channe l , R = root channe l , C = p l anar c rack . (a) Vert i ca l view of 76 conduct ing channe l s in a natura l fracture p l ane . (b) Cross-section view of conduct ing worm channe l s and the interconnected cracks D i stribut ion of surface app l i ed ch loride after l eaching w i th 50 mm of water : (a ) I ni t i a l d i stribution (b ) P is ton f l ow displ acement ( c ) Observed f i e ld ch loride distribut i on under intermi ttent i rri gat ion of bare soi l (Wi l d and Babiker , 1976 ) . ( d ) Ca l cu lated curve fo l l owed convective? di spers i ve theory (Gardner , 1965 ; and equation ( 1 . 5 ) ) , as suming E = 8 . 3 x 1 0- 9 m2 s -1 , C = 1 . 0 , and z = 25 mm . F i e l d 0 0 va l ues for e and v were used (0 . 2 and 7 7 2.2 x 10-7m s - 1 , respect ive ly a s found by Wi l d and Bab i ker , 1976) . . . . . . . . . . . . . . . 8 2 Dai l y rainfa l l ( RF ) and evapotranspirat ion ( ET) data during experiment in 1979 87 Fig . 6 . 3 F i g . 6.4 F i g . 6 . 5 F i g . 6 . 6 Fig. 6 . 7 F ig. 6 . 8 Soil bul k dens i ty d i st r ibut ion in the soi l profi les under pasture and cu l t ivat ion (means and standard devi a tions of 6 rep l i cates at x i v . Page each depth) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Bromide d i s tribut ion in the soi l profile before and after ponding 50 mm of water on the soi l surface in the continuous pasture growing area . The concentrat ion of bromide before pondi ng was the average of 5 core samp l es , and after ponding was the average of 1 3 core samp les . . . . . . . . . 91 Bromide di stribut ion in the soi l profi le before and a fter ponding SO mm of water on the soi l surface in the cultivated area . Each measurement before ponding was the average of 5 core samp les , and a fter ponding was t he average of 1 3 core samp l es . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2 Vo l umetric water content d istribut ion in the sojl profi les before and after pond ing , and poros i t y . I = fi l t ration rate (mm hr- 1 ) i n the first hour after ponding . Under pasture ( a , b , and c ) and under cul t i ? vat i on (J and e ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Bromi de distribut i on i n the soi l profi l e under natural rai n fa l l in the continuous pasture g rowing area Vo lumetri c water content d i s tribut ion in the soi l profi l e under natura l rainfa l l condi t ions at both samp l ing t imes (after 46 and 182 mm excess rainfa l l) , and soi l poros i ty . Means and standard devi at ions of 40 samp les 98 1 00 Fig . 6 . 9 F i g . 6 . 10 F i g . 6 . 1 1 F i g . 6 . 1 2 Fi g . 6 . 1 3 Fi g . B . l F i g . C . l H i stograms show ing frequency distribut i ons of brom idc con centra t ion (C) in soi 1 before and after leach ing by 1 8 2 mm excess rainfa l l . ( a ) Frequency d i str ibut i on of C XV. Page (b) Frequency d i s t ribut ion of ln C . . . . . . . . 1 03 Fract i l e d i agrams showing the re l at i onship of probab i l jty uni t s ( ( x - x) / s ) and bromide concentrat ion in so i l (C or l n C), where x = C or ln C , x mean va lue , s = standard devi at i on , and r = corre lat ion coeffic i ent. Post leach i ng data shown were after 1 8 2 mm excess ra infa l l Photographs i nd i cat ing stem-fl ow respons ib l e for non -un i formity of water i ntake at the soi l surface under natural rainfa l l . W = 104 worm channe l , R = root , and C = crack 109 Computed bromide concentrat ion d i s tributi on after l each ing soi l containing uni form ly spaced,vert i cal , cy l indrical channe l s . I n i t i al l y the soi l so lut ion concentrat ion (C) in the top 10 mm of soi l was C 0 Computed re l at i ve concentration o f bromide l eached from the top 10 mm to be l ow 300 mm depth i n soi l containing vert i ca l cyl indrica l channe l s , 0 . 2 and 0 . 1 5 mm in d iameter . Re l at ionship between Mt /Moo as shown by Crank ( 1956) Geometry of the system and symbo l s used in the programme for miscib l e d i sp lacement of so lutes in a uni form vert ica l cyl indrica l channe l soi l co l umn . Arrows indicate d i rect i on of flow 1 1 3 1 26 1 43 F ig . C . 2 Geometry o f the system and symbol s used i n the programme for l eaching o f the surface app l ied so lutes under ponding water i n a soi l column containing a uniform distributi on o f vertica l cyl indrical channe l s . Arrows indicate xvi . Page predominant d irection of f low .............. 15 1 Tab le 2 . 1 Tab le 2 . 2 Tab le 2 . 3 Tab le 2 . 4 Tab le 2 . S Tab le 3 . 1 Tab le 4 . 1 Tab le 4 . 2 Tab l e 6 . 1 Tab le 6 . 2 L I ST OF TABLES Freund l i ch constants for batch method i sotherms Phosphorus adsorpt ion data for di fferent soi l aggregate s i ze fract ions after 10 hour equi l ibration Cal cu lated t imes for chl oride and phoshphorus to di ffuse into spherical soi l aggregates , assuming a l i near adsorpt ion i sotherm for phosphorus ( see Appendix B for detai l s ) Physica l data for column experiments using O.S-1 mm aggregates Retardat i on factor values ( R) at 7 ( R7) and 20 ( R20 ) ?g/ml s o lut ion concentrat i on obtained from batch method i sotherms,for use Page 2 1 2 1 26 3 1 i n mode l l ing movement through soi l co l umns 32 Phys ica l data for soi l casts containing a channe l or crack Physica l data for saturated and unsaturated fl ow Channel s i zes assumed and the spac ing between them Recovery percentages of appl i ed bromide from the soi l profi les Channe l s i zes and re lated data assuming hydrau l i c conductivity of 6 . 94 x 1 0-7 m s - 1 43 ss 63 93 1 1 2 Tab le A . l Tab le A . 2 Tab le E.l Tab le F.l Tab le F.2 xviii . Page Tokomaru si lt l oam soi l profi l e descriptions (Po l l ok , 19 75) Some physica l and chemical properties of Tokomaru si l t l oam ( Po l l ock , 1975 ) . . . . . . . . Tab le A . 2 . 1 Part i c le s i ze anal ys i s . .. . .. . Tab le A . 2 . 2 Chemica l anal ys i s . . ...... .. . . Tab le A . 2 . 3 C lay minera logy . . .. . . . . . . . . . . Data for bromide concentrati on d istribution i n the soi l profi l es after l eaching by cont inuous ponding : Tab le E.l Under pasture Tab le E . 2 Under cul t ivati on Resul t s from fract i l e diagrams for experiment under natura l ra i nfa l l Data for bromide concentrati on distribution a fter infi l tration o f 46 mm excess rain fa l l over evapotranspirat i on. Four core samp les obtained from 0 . 32 x 0.5 m subp lots were bulked together in each rep l i cate 1 19 1 2 1 1 2 1 1 2 2 1 2 3 1 6 1 1 6 3 1 66 167 a B c c e c. 1 L IST OF SYMBOLS radius of a sphere Brenner number ( B = vd/E) solut ion concentrati on effluent concentrat ion influent concentrat ion C . concentrat ion of s olute in immob i l e lffi region C concentrat ion of so lute in mobi l e regi on m C initia l so lute concentrat ion in the l ayer near 0 the soil surface t o depth z 0 D mo lecu lar diffus ion of so lute in water D mol ecular diffus ion of so lute i n soi l s d soi l co lumn length E di spers ion coeffic ient Eel di spers i on coeffic ient of ch loride ET evapostranspiration g acce lerat ion due to gravity I infi ltrat ion rate i depth of water i nput or i n terger K hydrau l i c conductiv ity k Freund l i ch so lut ion d ist ribut ion coeffi cient ki Freundl i ch solut ion d istribut ion coefficient for adsorpt ion s i te I n Q q R r s adsorption rate coeffi cient desorpt i on rate coeffi cient adsorpt ion rate coefficient de sorpt ion rate coeffi cient for s ite for s i te a constant in the Freund l i ch equat i on number of samp les flow rate Darcy flux dens ity p retardat ion factor (R = ? k) e II rr correl at ion coeffi cient; or radial distance in equat ion ( B . l ) so lute adsorbed per un i t mas s of so i l so l ids ixx . D IMENS IONS L M L - 3 M L- 3 M L- 3 12 T- 1 L 2 T- 1 L L 2 T- 1 L 2 T- 1 L L T- 2 L T- 1 L or none L T-1 L 3M- 1 L 3M- 1 T- 1 T- 1 T- 1 T -1 none none or s1 amount of so lute adso rbed per un i t mass of soi l for s i t e I s11 amount of so lute adsorbed per un i t mass of soi l for s i t e II s standard dev iat ion of x SL t V V 0 V e V standard dev iat i on of ln x t ime cumu lative vo lume of eff luent l iqui d-fi l l ed po re vo lume vo lume of eff luent when C /C. = 0 . 5 e 1 average pore water ve l oc i t y (v = q/8) x vari ab l e (usua l l y C or l nC ) x mean of x z z 0 a mean of ln x z-vt d i stance in d i rect ion of v i n i t i a l depth of so lute near the soi l surface di ffus iona l t ransfer coeffi c ient v i scosity of f luid pressure pot ent i a l vol umetric wat er content i n soi l bu l k density f lu id density L L L T-1 XX M L-1 T-1 1. or ?? 1.-1 T-2 13 L-3 M L-3 M L-3 CHAPTER 1 GENERAL INTRODUCTION 2 . 1 . 1 IMPORTANCE OF SOLUTE MOVEMENT I N SOILS An i mpo rtant p rob l em i n agriculture, hort i cu l ture and forest ry i s the l oss o f fert i l izers , herb i c i des , pest i c ides , and soil nutr i ents in l e?ching water dra in ing below the root zone . Leached fert i l izers and chemi cal s may u l t i mat e l y contaminate ground water , streams and l akes , causing eutrophication of water sources . Leach ing usua l l y occurs when soi l i s re l at ive ly wet during autumn , winter and spring (A l i son , 1968 ; McLean , 1 9 77; Cameron et al., 1978 ) . However l eaching has also been ob served to occur at other times after a large rainstorm or i rr igation, part i cu lar ly i mmed i a t e l y after fert i l i zer app l i cat ion (Ay l more and Karim , 1968; Johnston e t a l . , 1965; Ba l asubraman ian e t a l . , 1 973; K i s s e l et al., 1974). Sign i fi cant l os s of n i t rogen fert i l i zer by l each ing under natura l fie ld cond i t ions has been reported by severa l inve s t i gators , inc l ud ing Wet s e laar ( 1962) , Johr.ston et al. ( 1965); Wil d ( 1972) , Cal vert ( 1975) , Cameron et a l . ( 1 978) , Gast et a l . ( 1974) . In New 7.ea l and , where rainfa l l inten s i ty and d i s tr ibut ion i s re lat ive l y h igh , part icular l y during w i nte r , Sharplcy and Syers (1979h) observed a n i t ro?cn l oss i n mo l e - t i l e drai nage equal l i ng apnroxjmatcly 2 % o f the amount app l i ed w i th i n 4 weeks of application to Tokomaru silt l oam soil under pasture. This percentage of n itrogen ferti 1 i zer 1 ost w?1 s grc?1tcr than that measured in si mi 1 ar stud ies over? seas by Bolton et al. (1970) anJMeek et al. (1969). A l so in Tokoma ru s i l t l oam, but under cultivation, Gandar and Gregg ( 1 979) found n i t rate losses equal to 60% of the fert i l i zer applied in the drainage water ove r a year . A l though phcsphorus is a very s trong ly adsorbed an ion i n soils, a sign i fi cant amount of l eaching has been found in sandy so i l s (Spencer , 1 957; l lumphrey and P r i chett, 1971; Calvert, 1975) , and organ i c s o i l s ( B l ack , 1 968 ; Duxbury and Pever l y , 19 78) . Leached phosphorus i n natural subsurface dra inage and arti f i c ia l drains was observed to be a si gnificant const i tuent of st ream fl ow by Minshal l ( 1 969) , Jackson et a l . ( 1973) , Burwe l l e t a l . ( 1 975 ) , and Sharp ley and Syers ( 1 979 ? and b ) . Sharp l ey and Syers ( 1 979 a) reported that in the watershed area near Mas sey Uni vers i ty , approximate l y 67% and 28% of phosphorus in the s t ream f low are contributed from natura l subsurface drain;1ge and mo l e -ti l e drainage , ?espec t i ve l y . 3 . Leach i ng of phosphorus may occur rap i d l y . Kanchanasut et a l . ( 1 978 ) . have reported that a s i gn i fi cant i ncrease i n phosphorus concent rat i on i n mol e - ti l e drainage was obse rved over-night , after 1 0 mm rain fa l l fol l owing fert i l izer app l i cat i on . llcrbicides and pest i cide s arc usua l l y cons i dered immobile in so i l , due to st rong adsorpt ion and fast degradat ion , however s i gni fi cant amounts have been observed in groundwater as reported by LaF l eur , et a l . ( 1 9 7 2 ) , La F l eur et a l . ( 1 974 ) , l!a l l and Hartwi t ( 1 9 78 ) . Tt i s very important to understand and be ab l e to pred i c t the movement of these so l utes in so i Is , particu lar l y under fie l d cond i t i ons and under d i fferent water management patterns . Thi s thes i s attempt s to make a contribut ion i n t h i s regard . 1.2 MOVEMENT OF SOLUTES IN SO I LS 1 . 2 . 1 Di fferent i a l Equat i ons Used to Describe Solute Movement The two main processes wh i ch are invo l ved in so lute movement i n so i l s are : ( 1 ) mo l e cu l ar d i f?1s i on in response to a concentrat ion grad ient , and ( 2 ) convect i on due to mas s f low of the soi l so l ut i on . When mo l ecu lar d i ffus ion and convect ion occur s imu ltaneous l y , they interact to cause enhanced d i spers ion . Under steady state cond i t i on s , the l ongi tud ina l transport of s o lute i n a uni form s o i l has usua l l y been described by the following partial differential equation (B iggar and Nielsen, 1962) where C ac at ac vaz- so l ut i on concentrat i on (M L- 3 ) E d i spers i on coeffic i ent ( L2T- 1 ) ( 1 . 1) v average pore ve l oc i ty ( L T- 1 ) , obtained from q/8 where q i s the Darcy f l ux densi ty ( L T - 1 ) and 8 the vo lumetric ? water content ( L L -3) z d i stance in d irection of v ( L) and t = t ime ? (T) . 4 . This equat ion , wh ich forms the bas i s of convent ional m i sc ib le d i sp l acement theory in soi l s , w i l l be referred to a s the convect ive? dispersive equat ion in the thes i s . When the f low ve locity i s sufficient l y h igh that the direct effects of mo l ecular di ffusion can be neg l ected , transport of so lute may be described by a s imp ler equation , deri ved for di spers ion about a moving reference p l ane , as (Nie l sen and Bi ggar , 1962 ; Gardner , 1965) where Z ac ? z-vt . 1 . 2 . 2 Solut ions of Transport Equat ions ( 1 . 2 ) 1 . 2 . 2 . 1 M i s c i b l e d i sp l acement research Mi sc ib l e d i spl acement i s the process that occurs when one fluid mixes with and d i sp laces another fluid . Day ( 1 956) , B iggar and Nie l sen ( 1962 , 1 963 , 1967 ) , Nie l sen and Bi ggar ( 1 961 , 1 96 2 ) , and Gardner ( 1 965 ) were among the fi rst to use miscib l e d isp l acement theory to study so lute transport in soi l s . Much mi s cib l e d i sp l acement research has concentrated on observing and anal ysing breakthrough curves , wh ich arc graphs of the rat io C /C . versus the number e 1 of pore volumes of effluent co l l ected (V/V ) , where C. and C are the o 1 e concentrat icn of d i sp lacing and d i sp l aced solut ion respect ive l y , V i s the cumu lat ive volume of effl uent and V the volume of pores occupied 0 by flui d . The pos i t ion and shape of the breakthrough curves g ive infor - mat ion about the pore configurat ion in the porous media . N i e l sen and B i ggar ( 196 1 , 1962 ) and Bi ggar and N i e l sen ( 1 963 ) described genera l i zed breakthrough curves for flow throu?h various pore geometries in re l at ion to f low ve locity and water content . The convect ive-di spersive equat ion ( 1 . 1 ) has been used extens ive l y to predict non -react ive s o lute movement in soi l . The deriviation of the ana lyt i ca l solut ion under the injt i a l and boundary cond i t ions 1 i m z-+?> C ( z , t ) C ( z , t ) C(z , t ) 0 for z > 0 and t = 0 C . for z 0 and t > 0 1 0 for t > 0 5 . ( 1 . 3 ) is given by N ie l sen and Bi ggar ( 1 962 ) and Kirkham and Powers ( 1 972 ) . The so lut i on is C/c k(erfc{ z-?} + exp ( E vz) erfc{2 + vt }) i 2 (4Etri (4 Et ) >l ( 1 . 4 ) Th i s so lut ion i s often used to describe the breakthrough curves from a co lumn of fin i te l ength (d) wi th the exit concentrat ion (C ) approx imated e as C at z = d , and V/V found as vt/d . When p l ot ted, equat ion ( 1 . 4 ) 0 yie lds an approximate l y S-shaped curve , symmetrical about one pore vo lume . The va l ue of E may be obtained e i ther direct ly from the s l ope of the experimental breakthrough curves ( K i rkham and Powers , 1 9 7 2) or by the l og -norma l d i stribut i on method suggested by Rose and Pass ioura ( 1 97 la ) . 1 . 2 . 2 . 2 Leaching D i stribut ion of so l ut e in the soil profi l e after l eaching a thin surface l ayer of non -react ive so lute may be described by equat ion ( 1 . 2) , assuming a constant f low ve locity and d ispersion coeffic i ent . The analyt i ca l so l ut i on for steady state condit ions was g iven by Day ( 1 956) , as c k C z ( 4nEt ) -2 exp ( - ( z -vt ) 2/4Et) 0 0 (1. 5 ) where z i s t he in i t ial depth of so lute near the soi l surface , and C 0 0 the ini t i a l concentrat ion there . Equation ( 1 . 5 ) gives be l l - shaped d i s tribut ion curves , whi ch get deeper and fl atter wi th i ncreas ing t i me . Gardner ( 1 965) found th i s so lut i on adequate l y described some fie l d data for nitrate l eachi ng under natural rainfal l cond i t ions . However , l eaching under fi e l d cond i t ions usua l l y occurs under unsteady f low conditions . Bre s l er and Hanks ( 1 969) and Warrick et a l . ( 19 7 1 ) have attempted to describe chl oride movement in the fi e l d soi l profi l e with unsteady fl ow . To avoi d comp l i cated computat i ons , Bres ler and Hanks ( 1 969) assumed the contribut ion to f low from di ffus ion neg l igib l e 6 . in compar i son with convect ive f low , wh i l e Warrick et a l . ( 197 1 ) as sum?d constant E . 1 . 2 . 3 Transport Mode l s for React ive So l utes So l utes moving through soi l often undergo adsorpt i on -desorpt i on react ions with so l i d surfaces . So lutes usua l l y consi dered non-react ive wi th soi l are nitrate , chl oride , bromide , and tri t ium . However , adsorp? t ion or exc lus i on of these so lutes i n soi l systems has been observed to some extent , depending on soi l chemical propert ies ( Biggar and Ni e l sen , 1962 ; Thomas and Swoboda, 1 970 ; Krupp et a l . , 1 97 2 ) . Phosphorus , most pes t i c i des and herbic i des , and potass ium are examples of react ive so lutes wh i ch arc frequent l y used in agr icu l ture . To account for adsorpt ion -desorption react ions during movement , i t i s necessary t o add an add i t i onal term to the convect ive-di spers ive equation ( 1 . 1 ) , giving (Cho et a l . , 1 970 ; Davi dson and McDouga l , 1 97 3 ; van Genuchten e t a l . , 1 974 ; Se l im e t a l . , 1 9 74 ; Manse l l e t a l . , 19 77) ac ? = ac V -?oz ( 1 . 6 ) where S i s the amount of so lute adsorbed per uni t mass of soi l so l ids (M M-1) and? is the so i l bu l k dens ity (M L-3). The express ion oS/ot , which represents the rate of adsorpt ion by soi l, has been described i n a number of ways , some of wh ich are g iven be low . a . Equi l ibr i um adsorpt i on i sotherms The Freund l i ch equat i on i s the most common l y used t o describe adsorpt ion of phosphorus i n equat ion ( 1 . 6 ) (Cho et al . , 1 9 70 ; Se l im et a l . , 1 974; Manse l l et a l . , l976 ) . . It may be writ ten as: s 3 1 ( 1 . 7 ) where k i s the so lut i on d istr ibution coeffic ient ( L M ) and N i s a constant , wh ich i s usua l l y l ess than un i ty , determined by regress ion ana lys i s from i sotherm data . When N = 1 , the re l ationship becomes l inear . Di fferentiat i on of equation (1.7) with respect to times g ives as at dS dC ac at kNC N- 1 ac at 7 . ( 1 . 8 ) Subs t i tution o f equat i on ( 1 . 8 ) i n equat i on ( 1 . 6 ) and rearranging:gives (1 + R) ?? = p where R = ? kNCN- l e equat ion ( 1 . 9 ) when N and v by v/ ( 1 + R ) . When N = 1 i s V = 1 , ac az R = pb ek . equat ion ( 1 . 4 ) A pos i t ive val ue for R ( 1 . 9 ) The anal yt i ca l so lut i on for wi th E rep l aced by E/ ( 1 + R) moves the breakthrough curve to the right of one pore vo lume . For other N values , numerica l so lut ions must be found . U s ing an equi l ibrium adsorpt ion mode l i n the convective-di spers ive equation tlsual ly resu l t s in an over -est imate of the amount of adsorpt i on occurring during m i s c i b l e d i sp l acement ( Kay and E lrick , 1 96 8 ; Dav idson and Chang , 1 972 ; Manse l l et a l . 1 97 7 ) . Davidson and Chang ( 1 97 2 ) have shown that eaui l i brium adsorpt ion wou l d?not occur over a w ide range of pore -water ve l oc i t ies . Equi l i brium adsorpt i on mode l s can onl y g ive the upper bound for adsorpt ion during the movement of react ive solutes . b . K inet i c adsorpt i on mode l s Phosphorus adsorpt i on by soi l i s i n rea l i ty a k inet i c proces s . The rate o f adsorpt i on i s i n i t i al l y rap i d , fo l l owed by a s l ower sorption react i on wh i ch may cont inue for a l ong period of t ime ( Barrow , 1 974 ; Ryden et a l . , 1 97 7 ) . Lap i dus and Amundson ( 1 952) introduced an equat i on for first order , revers i b l e , k ineti c sorpt i on react i ons wh i ch may be wr i t ten as at k2 (?- c - S) k2pb ( l . 1 0 ) where k1 and k 2 are adsorpt i on and dcsorpt i on rate coeffi c ients (T-1), respect ive l y . The numerica l so lut ion of equat i on ( 1 . 1 0 ) wi th equat i on ( 1 . 6 ) has genera l ly described experimenta l data we l l at l ow pore-water ve l oc i t i e s ( Davidson and McDouga l , 1 973 ; van Genuchten et a l . , 1 9 74 ) . However , Manse l l et a l . ( 1977 ) found that thi s model did not adequately describe phosphate movement through cores of a sandy so i l , the desorpti on rate constant ( k2) appearing to change when pore-water ve l o c i ty changed . 8. c . Comb inat ion mode l Se l i m et a l . (1976) and Cameron and K lute (1977) have devel opment a two-s i te adso rpt ion-desorption mode l for use in describin? t ransport of rea c tive so l utes . The two s i tes are those wh i ch appear to adsorb or reac t with so lutes effect i ve l y instantaneous l y , and t hose wh i ch appear t o adsorb more s l ow l y , resu l ting i n a k i net i c reaction . Th i s d ivision is probab l y fair ly arbi trary , and the physico? chemical s i gn i ficance of the two s i tes i s not c l ear ( Syers et a l . 1973; Ryden et a l . , 1977; Bowden et a l . 1980). The two - s i te l inear adso rpt i on mode l can be expressed as : as I Tt k ac r at ( 1 . 11) ( 1 . 12 ) where s1 and s11 are the amounts of so l ute adso rbed on site and s i t e ll respect i ve l y , ki is the so l ution d i s tribut i on coeffic ient for s i t e 1, and k 3 and k 4 the adsorpt ion and desorpt i on rate coeffi c i ents respec t ive l y for s i te I I . The d i fferentia l equat ion resu l t ing when these expres s ions are subst i tuted into equat i on (1.6) i s where R I ac at ( 1 . 13) pb El ki? Thi s adsorption mode l i s more comprehens i ve than the ones des cribed ear l ier . Howeve r , it i s diffi cul t to determine independent l y the appropr iate adsorption- desert i on parameters , hence they are usua l l y est imated by curve-fit t i ng to the experimenta l breakthrough data . So l v ing equat ion (1.13) numer ica l ly , de Camargo et a l . (1979) found that pore -water ve l oc i ty had a large inf luence on the adsorpt i on and desorpt i on rat e coeffic i ent s (kl and k4) needed to descr i be phospho?us movement through co l umns containing s i eved soi l . For pest i cide 'movement , Rao et a l . (1979) a l so found that two sets of k va lues were requi red to p red i ct the b reakth rough cttrvcs at two d ifferent i nput concentrat i ons . In genera l , i f non- l inear and/or kinet ic adsorpt i on react i ons are assumed , so l ut i on of the resu l t i ng mod i fi ed convect i ve-di spers ive equat ion mus t be found numerica l l y using a computer , as anal yt i ca l so lut ions are not avai l ab l e . 9 . 1 . 3 FAILURES OF CONVENTIONAL TIIEORY Near l y 1 00 years ago , Lawcs et a l . ( 1 8 82) wrote that "A l arge part of water added to soi l prof i les moves immed i ate l y through channe l s and i nteracts s l ight l y with water i n the soi l i t s e l f" . In New Zeal and , Tay lor ( 1956 ) stated that , "It has often been assumed that the moi sture spreads even l y through the soil s , whereas what happen s is far more comp l ex . Soi l wat er percolated down cracks and through spaces between the struc ? tura l surfaces and i s adsorbed l atera l l y into the soi l aggregates" . The concept of rap i d movement of infi l trat ing water described by Lawes et a l . ( 1 882 ) has been ignored until recent l y . For decades , soi l phys i c i st s and soi l sc ient i s t s have treated soi l as a un i form porous medium and app l i ed the theory described in the preced ing sections to pred ict water and solute movement . For examp l e , d i s turbed , homogen i sed and repackcd soil has been use(l extensive l y i n misc i b l e d i sp l acement experiments studying the movement of solutes (Dav idson et a l . , 1 968; L i nstrom and Boesma , 1 97 1 ; Dav idson and Chang , 1 972 ; Rao et a l . ,1 979) . However , some m i sc ib l e d i sp l acement experiments us ing natura l soi l cores have been carried out . The resu l ting breakthrough curves from saturated soi l cores usua l l y show an ear l ier , and i n i t i a l l y steeper , breakthrough than for packed un i form soil columns ( E l rick and French , 1966; Ki sse l et a l . , 1 973 ; Casse l et a l . , 1 974 ; McMahon and Thomas , 1 974) . The d i fferences have been attributed to d i fferent pore connecting patterns . In und i sturbed cores the movement predominan t l y occurs through l arger cont inuous channe l s , wh i l e in the more homogeneous repacked columns the movement i s more un i form , and can be described by the convect ive-d i spers i ve equat ion . Anderson and Bouma ( 1977a) and Bouma and Wosten ( 1979) have shown that the shape of the breakthrough curves i s very sens i t i ve to any d i fferences in f low paths , e i ther among the rep l i cate cores of the same soi l structure , or among different soil st ructura l types . Very often also it has been found that the distribution of surface applied solute after l each ing in the soi l prof i l e i s not be l l - shaped, as wou l d be predicted by convect ive -d i spersive theory ( Zimmerman et a l . , 1 967 ; Boswe l l and Anderson , 1 97 0 ; Wi l d , 1 972 ; Shuford et a l . , 1977 ; Cameron et al . , 1 977 ; Wi l d and Bab iker , 1 976) . The pronounced asymmetry of the ' l eaching pattern 10 . common l y observed has been a t t r i buted to non -un i form movement of soi l water and solutes . Th i s has been confirmed using dye stain i ng techn iques by R itch i e et a l . ( 1 9 72 ) , Anderson and Bouma ( 1 973 ) , Bouman et a l . ( 1 9 77b ) ; Bouma and Dekker ( 1 978) , Omot i and W i l d ( 1 979) . Th i s k i ntl o f b rea k t h rough c u rv e or l each i ng pat tern i s refe r retl to as "preferent ia l f low" in th i s study . The evidence for i ts occurrence , and i t s importance i n soi l and ground water d i s charge and solute move? ment have recent l y been rev iewed by Thomas and Ph i l l i ps ( 1 979) . As we l l as in und i sturbed soi l , dev iat ions from convect ive -d i spersive theory have a l so been observed in packed soi l columns contain ing l arge aggregates ( Bi ggar and N ie l sen , 1 962 ; Green et a l . ? 1972 ; van Genuchten et a l . , 1 974 ; van Genuchten and W i erenga , 1 977 ) . The m i croscop i c f low ve l oc i t i es i n med i a wi th compl ex pore geometries vary from point to poi nt in both magn i tude and d i rect ion . The ve l oc i ty i n a pore i s at a max imum at the centre of the pore , whereas the f luid adj acent to the pore wal l does not move . A l so the ve l oc i ty through l arger pores exceeds the f low through sma l ler pores . B i ggar and N i e l sen ( 1 96 2 ) have shown that when aggregate s i ze i s increased ,the range of m icroscop i c f low ve l oc i t i es increases , and most of the f l ow w i l l oc cur through macropores between aggregates . A l so , as the d i stances between macropores i ncrease , mol ecu l ar d i ffus i on i s much l es s e ffect ive in d i s s i pat ing concentrat i on grad i ents between the soi l sol ut ion in the macropores and w i thin the aggregates . Thus the m i xi ng in the columns becomes l ess compl ete and affects the s lope of the breakthrough curves . Sol ute mov i ng through macropores i s respons ib l e for an ear l y b reakthrough , wh i l e s low solute movement through m i cropores w i th in the aggregates i s respon s ib l e for a l ong ta i l sect ion in the breakthrough curves . Non -uni form f low becomes more evident when react ive solutes ? are used . Dav idson and McDouga l ( 1 973 ) suggested that in med i a w i th comp l ex pore goemetry , the observed t a i l i ng of the breakthrough curves cou l d be due to both s l ow d i ffus ion of solutes i nto the mi cropores , and to k inet i c adsorpt i on processes . Transport mode l s inc luding the effect s of pore geometry wi l l b e d i scus sed i n the fol l owing sect ion . 1 1 . 1 . 4 MOD I F I ED SOLUTE TRANSPORT MODELS Recent advances in predict i ng solute transport in soi l s i nc l ude efforts to incorporate the e ffects of pore geometry on f low ve l ocity distribut ion i n the transport mode l s . Some of the mode l s are : a . Convect i ve -D i spers ive Mode l w i th Lateral Di ffusion I n col umns of saturated soi l aggregates , soi l pores can be part i ? tioned into macropores , wh ich are pores between aggregates and micro? pores which are pores inside aggregates . Convect ive vi scous f low i s assumed t o occur on l y i n the macropores or mobi l e regions , whi l e di ffusion occurs ins ide the aggregates or immobi l e reg ions . The convect ive-d i spersive equat ion was modi fied to inc l ude l atera l d i ffus i on by Coats and Smi th ( 1964) and van Genuchten and Wierenga ( 1 976) . The equat ion for non-react ive solute transport may be wri t ten : + ? V 8 m m ac m az and the exchange between mobi l e and immob i l e regions given by 8 . l iD a c . liD ----at a ( C - C . ) m 1m ( 1 . 1 4) ( 1 . 1 5 ) I n these equat ions 8 and 8 . are the fract ions of soi l volume f i l l ed m 1m with mob i l e and stagnant l i quid respect ive l y , C and C . are the concen -m 1m trat ions of solute in the mobi l e and immob i l e regions respect ive l y> v i s the average pore water ve l oc i ty i n the mob i l e l iqui d , and a i s a m d i ffusional trans fer coeffic ient (T - 1 ) . Pas s ioura ( 1 9 7 1 ) derived and sol ved ana l yt ica l l y an equation s imi l ar to equation ( 1 . 1 4) . Van Genuchten and Wierenga ( 1 976a) inc l uded the e ffect of - adsorp? t ion-desorption in equat ion ( 1 . 1 4) and ( 1 . 1 5 ) . Ana lyt i ca l solut ions of equation ( 1 . 1 4) and ( 1 . 1 5 ) with l inear adsorption (N 1 ) are given by van Genuchten and Wi erenga ( 1 976a ) , whi l e for N ? 1 a numerica l solut ion i s given by van Genuchten and Wierenga ( 19 76b) . 1 2 . The d iffus iona l transfer coeffic i ent (a) has a l arge inf luence on the shape of the breakthrough curves . For non-react ive solutes , Rao et a l . ( 1 980) observed the diffus i on into spherical aggregates and found the a val ues were dependent upon aggregate s i ze , vol umetric water content in the aggregate ( 8 . ) , mol ecu l ar d i ffus ion in the J.m aggregate , the frac t i on of wat er content in the mobi l e reg ion ( 8 ) , m and di ffusion t ime . For react i ve solutes , a was found to depend on f l ow ve l oc i ty (van Genuchten and Wierenga , 1 977 ) , and i nput concentration ( Rao et a l . , 1979 ) . Thus treat ing a as a constant , as i s usual , i s not very sat i sfactory . b . Vi scous F low with Latera l D i ffus ion A mode l was deve l oped by Scatter ( 1 978) based on the assumpt ion of viscous solut ion f low down vert ica l cyl indr ica l channe l s or p l anar cracks , w ith s imu l taneous mol ecu lar di ffus ion of solute into the surrounding soi l . The theory and re l ated computations are g iven i n Appendix C . He assumed the movement occurred through channe l s of the s ame s i ze and uni form l y d i stributed across the cross section of soi l .. The resul t ing breakthrough curves for s trong l y adsorbed solute d id not differ s i gn i fican t l y from for the non -react ive solute breakthrough curves , when channe l d iameter and p l anar crack width were greater tha.n 0 . 2 and 0 . 1 mm respective l y . 1 . 5 GENERAL OBJECTI VES The work described in the fol l owing chapters aimed at i nvest igating the characteri s t i c s of chl oride , bromide and phosphorus movement . i n soi l w i th d ifferent pore geometries under both l aboratory and f ie l d condi tions . I t a l so aimed at assess ing the usefu l ness of some of the theory out l ined above to describe the observed solute movement , whi ch occurred mos t l y under preferent ia l f low condi tions . CHAPTER 2 ACCOUNTI NG FOR ADSORPTION DUR ING PHOSPHORUS MOVEMENT IN SOI L 1 4 . 2 . 1 I NTRODUCTION Phosphorus movement in soi l i s usua l l y qui te l imited , as i t i s retarded b y soi l -phosphorus interaction . The interact ion phenomenon i s general l y described by an adsorption -desorpt ion i sotherm , whi ch i s usua l ly non- l inear and t ime dependent ( Fox and Kamprath , 1970 ; Whi te and Tay l or , 1 9 7 7 ; Ryden et a l . , 1 9 77 ) . These i sotherms are usua l l y determined by the "batch method " , wh ich involves shaking suspens ions of soi l in various phosphorus solut ions and measuring the amounts of phosphorus coming into or out of solut ion . The adsorpt ion occurs rap i d l y in the first few hours , and then continues to occur s l ow l y over a l ong period of t ime . Adsorption and desorption processes of phosphorus are not s ingu lar , and some of the adsorpt ion react ions are i rreversib l e , due to precipitat ion and chemisorpt ion wi th h i gh adsorpt ion energy (Ryden et a l . , 1 9 7 3 ; Barrow and Shaw , 1975 ; Whi te and Tay l or , 1 9 77 ) . Common adsorpt ion -desorption mode l s for phosphorus have been described in Chapter 1 . Phosphorus transport through soi l s has often been stud i ed us ing misci b l e d i sp l acement techniques (e . g . Cho et a l . , 1970 ; Se l im et a l . , 1974 ; de Camargo et a l . , 1979 ) . The resu l t ing breakthrough curves are not as wou l d be expected from convent iona l convec t i ve-d i spers ive theory , as the curves are usua l l y asymmetri ca l w i th s ign i fi cant "tai l ing" . Many attempts have been made to quan t i fy the inf luences of adsorpt ion? desorption processes on breakthrough curves ( e . g . Dav idson and Chang , 1 9 7 2 ; Davidson and McDougal , 1 9 7 3 ; Kay and E l rick , 1 967 ; van Genuchten et a l . , 1 974 ; Manse l l et a l . , 1 9 77 ) . As phosphorus adsorpt i on i s t ime dependent , usi ng equi l ibrium adsorpt ion i sotherms to account for adsorpt ion during mi scib l e d i sp lace? ment in soi l often resu l t s in the predicted curves d iverging from the experimenta l data ( Se l im et a l . , 1974; Manse l l et a l . , 1 9 77 ) . The equi l ibrium adsorpt ion i sotherm can on l y g ive the upper bound of the amount of adsorpt ion of reac t i ve solute in soi l (Davidson and McDouga l , 1 9 73 ; van Genuchten et a l . , 1 974), and when used wi th the convect ive? di spers ive equat ion i s unab l e to predi ct correct l y the shape of the breakthrough curves in packed soi l columns ( Davi dson and Chang , 1 972 ; Manse l l et a l . , 1 977 ) . 1 5 . For reac t i ve solutes , the i nterac t ion between pore water ve loc i ty and adsorpt i on has been stud i ed by L i nstrom and Boesma ( 1 97 1 ) , Se l i m e t a l . ( 1 9 74 ) , Manse l l et a l . ( 1 977 ) , and d e Camargo e t a l . ( 1979) . They found more adsorpt ion occurred when pore -water ve loc i ty was decreased . Th i s suggests that the use of adsorpt ion i sotherms obtained using equi l j brat i on t i mes s imi lar to the contact t i mes for phosphorus in the soi l col umn mi ght be useful i n pred ic t i ng the break through curves . Th i s i dea i s inves t i gated in t h i s chapter . Adsorpt i on i sotherm determinat i ons usua l ly a im to measure equi l i br ium adsorpt ion , therefore the soi l - phosphorus soluti on suspens ion i s shaken unt i l the adsorpt i on rate i s neg l i g i b l y smal l . Li t t l e atten t ion has been pa id to the equi l i brat i on methods and t i mes adopted i n adsorpt i on i sotherm determinat i ons , carried out for use i n mode l s of adsorpt ion during m i s c i b l e d i s p l acement . For examp l e , Manse l l e t a l . ( 1 977 ) used adsorpt i on i sotherms determined after shaki ng 1 : 5 soi l to phosphorus soluti on suspens ion for 7 days to predi c t phosphorus movement in short sand columns in whi ch the d i sp l acement occurred in less than an hour . S imi l ar ly , de Camargo et a l . ( 1979) used i sotherms obtained from shaking a 1 : 20 soi l to soluti on suspens i on intermi t tent ly for 24 hours to predi c t phosphorus movement i n columns o f < 2 and < 1 mm soi l aggregate i n whi ch the d i s p l acement occurred w i t h i n about 1 2 hours . As s tated earl i er , the d i str ibuti on and adsorpt ion of phosphorus and other react i ve solutes are affected by soi l pore geometry or soi l structure . The movement of solutes into stagnant pores , or mi cropores w i th in soi l aggregates and so i l s t ructural un i t s , i s cons ide red to occur by molecu lar d i ffus i on a lone , wh i l e in the macropores between soi l aggre? gates convec t ion and hydrodynamic d i spersi on , i nduced by v i scous f low , are the maj or transport mechan i sms ( Pass i oura , 1 9 7 1 ; Skopp and Warrick , 1 9 74 ; van Genuchten and W i e renga , 1 976) . The adsorpt ion on soi l ? agg regate surfaces , or onto the wa l l s of mac ropores , wou l d reach equ i l i brium faster than on soi l surfaces w i th in the aggregates . Therefore , even non-equi l i brium adsorpt ion i sotherms determined using convent i onal methods , may be of l imi ted use for describing adsorpt i on during m i s c i b l e d i sp l acement . 1 6 . 2 . 2 OBJECTI VES The experiment s described i n th i s chapter aimed to compare the adsorpt ion i sotherms obtained from the convent ional batch method and from l eaching col umns . Then to determine the feas i b i l i ty of using i sotherms obta ined at appropri ate equi l ibrium t imes and concentrat ions for pred i c t ing phosphorus movement through co lumns of soi l aggregates . 2 . 3 MATER IALS AND METHODS 2 . 3 . 1 Convent ional Batch Method Adsorpt ion I sotherms Phosphorus adsorpt ion i sotherms were obtained us ing a 1 : 40 soi l to so lut i on rat i o . One gram of < 2 mm seived, a ir-dried soi l ( Tokomaru s i l t l oam , A hori zon) was shaken gen t l y in an end-over- end shaker w i th 39 m l of 100 ?g/ml ch loride as potas s ium . ch loride so lut i on ( KC l ) . One ml of the appropriate concentrat i on of phosphorus so lut i on as potass ium dihydrogen phosphate ( KH 2 P04 ) was added to the soi l suspens i on at vari ous t imes between 6 and 29 hours after commencing to shake the samp le s . Th i s method , proposed by Ryden et a l . ( 1977 ) , kept the shak ing t ime constant whi l e a l l owing the t i me for equi l i brat i on to vary from 1 to 24 hours . Pre- shak ing was needed because some aggregates were not broken down before 6 hours shak i ng . After 30 hours shak ing , the soi 1 suspens i on was centri fuged , fi l tered and a phosphorus determinat i on was made on the c l ear supernatant us ing the method of Murphy and Ri l ey ( 1962 ) . I n add i t i on , phosphorus adsorpt ion by vari ous soi l aggregate s i ze fract i ons ( < 0 . 2 , 0 . 2 - 0 . 5 , 0 . 5 - 1 . 0 , and < 2 mm) was determi ned at 1 0 hour equi l i brat i on t ime i n 28 . 5 ?g/m l phosphorus and 1 00 ?g/ml ch l or i de solut i on . As the aggregates were broken down during shak ing, tryis a l l owed any di fferences i n adsorpt i on of di fferent aggregate s i zes , concei vab ly due to d i fferent c l ay or organ i c matter contents , t o be assessed . In a l l cases potassium ch lor ide so lut ion was added to obtain the same total i on i c s trength as in the l eaching method descrjbed be low ( Sect i on 2 . 3 . 2 ) , and in the soi l aggregate co lumns ( Sect i on 2 . 3 . 3 ) . Three rep l i cate determinat i ons were made at each concentrat ion l eve l . 1 7 . 2 . 3 . 2 Con t i nuous - Leach j ng Method I sotherms A m i s c ib l e d i sp l acement techn ique was a l so used to determine the phosphorus adsorpt i on i sotherm . A sma l l p l as t i c tube 18 mm in ins ide d iameter and 40 mm l ong was packed w i th 5 g of < 2 mm seived , a i r-dried soi l , and was set -up vert i ca l l y . The soi l in the column was i ni ti a l ly saturated with d i st i l l ed water . Phosphorus solut i on was then percolated upwards through the soi l column at 0 . 5 ml /min . The three concentrat i ons used were 1 , S , and 1 0 wg/ml phosphorus as potass ium d i hyd rogen phosphate . The effluent was col lected at regu lar t ime interva l s by an automat i c fraction col l ector . One hund red wg/ml of ch lor i de as potassium ch l oride was mi xed w ith the phosphorus solut i on to act as a marker . The experiment was cont inued unt i l the concentration of the eff luent was the same as the i nfluent concentrat ion , w i thin the l im its of measurement . The re la t ive concentration of the eff luent solut i on ( C /C . ) was e 1 p lotted against the vol ume of the e ffluent , and the area between the chloride and phosphorus breakthrough curves was obtained by integrat i on using S impson ' s rul e . Th i s area was cons idered to represent the amount of phosphorus adsorpt ion . 2 . 3 . 3 Phosphorus and Ch loride Movement i n Columns of Soi l Aggregates Se ived , ai r dried soil aggregates , 0 . 5 - l mm i n e ffect i ve d i ameter , were uni forml y packed into 30 mm ins i de diameter , 90 mm l ong perspex tubing . These soi l columns wer e set-up vert i ca l l y , w i th the input sol ut ion entering the l ower end , and the effluent solution col l ected from the top of the column through the out - flow tube l ocated on the s i.de of the column ( F i g . 2 . 1 ) . "Dead space" at the base of the soi l column was minimi zed by us ing a s ing le l ayer of nyl on mesh ( 60 ?m) instead of a porous p l ate at the l ower end of the soi l column . At the top , the effluent soluti on was col l ected through the sma l l out flow-tube , minimi z ing "dead space" there a l so . The volume of effluent in the out fl ow- tube was neg l i gi b l e i n comparison w i th soi l pore volume ( l ess than 0 . 04%) . Eff luen t solu t ion out Frac t ion collec tor t I n f l uent solut ion in 1 8 . Fil ter paper Soi l aggrega tes Nylon mesh 4fiC--Out le t for f lush ing Fig . 2 . 1 Schema tic d iagram o f experimen tal se t-up for atudying phosphorus and ch l o r ido movement in soil -aggrega te columns . 19 . The soi l column was saturated s l ow l y from below w i th a so lut ion contain ing 293 ?g/m l potass ium n i trate ( KN0 3 ) and was l eached w ith the same solut i on for at l east 14 hours to ensure the soi l in the co lumn was comp l ete ly saturated . Potas s i um n itrate was u sed to equa l i se the dens i ty of the d i sp l aced solut i on and the d i sp l ac i ng soluti ons , so that there was no grav ity segregat i on effect during the d isp l acement (Rose and Pass ioura , 197 1b ; Starr and Par l ang e , 1976) . The d i sp lac ing solut ion , contain ing 20 ?g/m l phosphorus as potass ium d ihydrogen phosphate and 100 wg/m l ch loride as potas s j um ch l or ide , was app l i ed to the column after the previ ous so lut i on i n the chamber underneath the ny lon mesh had been fl ushed out . The flux dens i ty (q ) of the d i s p laci ng solut i on was maintained by a per i s ta l t i c pump at 9 x 10- 6 or 4 . 7 x 10- 5 m s - 1 ? The h igher f lux dens i ty was approx imate ly ha l f of the saturated hydrau l i c conduc t i v i ty o f the aggregates . Dup l i cate experiments were conducted at each flux densi ty . Effluent a l i quots were taken at appropriate t ime? i nterva l s and anal ysed for ch lor ide and phosphorus concentrat ion . A c h l o r ide s e l ect i ve- ion e l cc t roJl: was used to measure ch l oride i n the solut i on . 2 . 4 RESULTS AND D I SCUSS ION 2 . 4 . 1 Conventional Batch -Method Adsorpt ion I sotherms Phosphorus adsorpt ion data obtained us ing the convent i ona l batch method at 3 and 10 hour equi l ibrat i on t imes are shown in Fig . 2 . 2 . The adsorption i sotherms are non- l inear and described fair ly we l l by the Freund l i ch equat ion , equat i on ( 1 . 7 ) . The Freund l ich constants ( k and N) for the five equi l i brat ion t imes used are g iven i n Tab l e 2 . 1 . A l so g iven are the corre l at i on coeff i c i ents ( r) . Phosphorus adsorpt ion was strong l y t ime-dependent , as can be seen from F i g . 2 . 3 . The curves in thi s figure were obtained us ing a power- curve fi t t i ng procedure . The data indi cate a l arge part of the adsorpt i on occurred during ?he first hour , perhaps a lmost instantaneous ly . The rate of adsorpt ion decreased marked l y as the equ i l ibrat i on t ime l engthened . ,........ 00 ...._ 00 ;::l. .._, A ? ? 0 [J) [J) ? p::: 0 ::c p.. [J) 0 ::c p.. 200 160 1 20 k= l 5 . 4 80 40 / / / / / / / / / 10 / / k=6 . 5 ml /g -- ? ? _, " _, - - ,.,... ;/ " - - - - - - 1 -- / .., ,., -? - - :;-<- - -1 / / / 20 . 2 ? 5 EQUILIBRIUM CONCENTRATION 20 ( ).lg/ml) Fi g . 2 . 2 Adsorpt ion da ta an d the f i t ted Freundl ich isotherms , a t two equilibrat ion t imes ; 3 hours (- - -) and 10 hours (-) , ob t ained f rom the batch me thod . Arrows indicate the change in phospho rus concen trat ion in the so lut ion due to adsorp t ion . The l inear iso therms also shown were forced through the Freundl ich isotherms at 7 and 20 ).lg/ml . They were used to ob ta in solution dis tribut ion coef fic ien t s ( k ) fo r pred i c t ing phosphorus movement in the so i l co lumns . The tail and head o f the arrows indicate ini tial phosphorus concen tration and the concentration af ter adsorp t ion , respectively . 2 1 . Tab l e 2 . 1 Freund l i ch constants for batch method i sotherms . Equ i l i brat i on T ime k N r (hr) (ml/ g) 1 1 7 . 3 0 . 589 0 . 9 78 3 28 . 0 0 . 53 1 0 . 979 6 28 . 9 0 . 568 0 . 983 1 0 3 1 . 9 0 . 6 1 5 0 . 996 24 52 . 8 0 . 504 0 . 992 The amount of phosphorus adsorbed by d i fferent aggregate s i ze fract ions was checked , and the resu l t s are presented i n Tab l e 2 . 2 . ! The fract ions , 0 . 2 -0 . 5 and 0 . 5 - 1 mm adsorbed s igni fi cant ly more phosphorus after 1 0 hour equ i l i brat ion than the other two fract i ons ( d i fferences s i gn i fi cant at the 95% confidence l im i t us ing a t - test ) . The soi l aggregates were obtai ned by grind i ng and se iv ing the soi l samp les , and they must therefore have been bonded strong l y enough to res i st the force invo lved . The agg regate cement ing mater i a l s usua l l y observed i n soi l ?nc l ude c l ay m i neral s , col loi da l ox i des o f i ron and manganese and col loidal organic matter ( Foth , 1 9 78 ) . These cementing mater i a l s are presented in th i s soi l , as shown i n Tab l es A . l , A . 2 and A . 3 in Append ix A , and are known to be strong phosphorus adsorbers (Hsu , 1 964 ; John , 1 9 7 2 ) . The fract ion < 0 . 2 mm , which con s i s ted mos t l y of s i ng l e grain part i c l es , adsorbed the l east phosphorus , wh i l e the frac t i on < 2 mm , wh ich cons i sted main ly of f ine materia l , a l so adsorbed l ess than two aggregate fract ions . Tab le 2 . 2 Aggregate fract ion (mm) <0 . 2 0 . 2 -0 . 5 0 . 5 - 1 . 0 < 2 . 0 Phosphorus adsorpt ion data for d i fferent soi l aggregate? s i ze fract ions after 1 0 hr equi l ibrat i on . Solut ion phosphorus concen - Adsorbed phosphorus trat i on after equi l ibrati on (wg/m l ) (wg/g soi l ) 23 . 4 206 22 . 2 25 2 22 . 4 245 23 . 0 222 ,....._ 00 --... 00 ;:I. '-' A w gj 0 Cfl Cfl ? 0::: 0 ::r: P-< Cfl 0 ::r: p... 1 80 1 6 0 ? ? c . = 1 5 iJg /ml l I ? 1 4 0 / 0 1 2 0 t c . = 1 0 ?? g /ml 0 l lOO 8 0 I I /- ? c . = 5 \.lg /ml l 6 0 4 0 I If 0 I> c . 2 \.lg /ml - 2 0 I ? l 0 0 2 . . 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 TIME (hour ) Fig . 2 . 3 Phosphorus adsorption rates for dif ferent init ial solution concentrat ions ( C . ) . The curves l shown were f i t ted to the experimental data using a power curve f i t ting procedure . N N 23 . 2 . 4 . 2 Continuous - Leaching Method I sotherms Phosphorus and ch lor ide breakthrough curves for cont inuous l each i ng experiments are shown in F i g . 2 . 4 . On l y one rep l i cate i s shown , however amounts of phosphorus adsorpt ion were obtained from the average of the 4 rep l i cates . The ch l oride curves i nd i cate the d i s p l acement by the inf luent solut i on was compl eted i n l es s than one hour . Adsorpt ion of phosphorus occurred rap i d l y i n the f irs t few hours , then the adsorpt ion rate became s l ow after the re l at i ve concen? trat i on of the effluent reached 0 . 8 . The t ime requi red unt i l adsorpt ion approached equi l ibrium depended on the influent concentrat i on . I t took 7 . 5 ? 1 . 4 hours or approx imate l y 1 2 pore volumes when the influent con? centrat i on containe d 10 ?g/m l phosphorus , whi l e i t took 25 ? 10 . 4 hours or approx imate l y 36 pore volumes when the i nf luent concentrat i on was 1 ?g/ml . The actua l max imum adsorpt i on cou l d not be obtained u s i ng thi s exper imental techn ique , due to the very s low adsorpt ion occurring at l ong t imes , when the re l at ive concent rat ion was near uni ty . The l ong ta i l ing observed was probab l y influenced by both s l ow k i net i c adsorpt i on? desorpt i on reac t ions and s low d i ffus ion i nto soi l aggregates . As the d i ffus ion o? phosphorus i s s l ow i n compari son to the f low ve l oc i ty , the effluent concentrat ion measured might not be the same as the soi l solut i on concentrat i on w i th in the aggregates . I n contrast , us ing the batch method the solut j on concentrat i on i s uni form , due to the shaking and breakdown of the soi l aggregates . Couchat et a l . ( 1980) have pointed out that the l each i ng method for measuri ng adsorpt ion g ives erroneous resu l t s i f the reac t ion of solute and soi l i s very s low . The adsorpt ion i sotherms found after 6 and 10 hours of l eaching are presented in F i g . 2 . 5 . A l so shown for compari son are the resu l ts obtained from the batch method at the same t imes . The amount of adsorpt i on obtained from the l eaching method was l ower than from the batch method at both t imes , however the resu l ts from the two methods seem to d iverge l ess at 6 hours t han at 10 hours . The d i fferences between the resu l ts obtained by the two methods may be attFibuted to : - 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 0 1 . 0 0 . 8 0 . 6 Q) u 0 . 4 0 . 2 0 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 ,- - - -- --C?l?----------.,........-p;;------? I I I If I 0 I I I I I I I I I I I I 0 C l 600 1200 1 800 2400 p 600 1 2 00 1800 2400 ( a ) (b ) {- - - -- C l --- ....... - --------=-------- 1 p I ? ( c ) 24 . 3000 3000 OY---?---+--?----?--*---?---+--?----?? 0 200 400 600 800 1 000 EFFLUENT VOLUME (ml ) Fig . 2 . 4 Phosphorus and chloride breakthrough curves for continuous leaching exper iments . The three l evels o f phosphorus concen? tration are : ( a ) 1 , (b ) 5 , and ( c ) 10 ?g/ml . 160 140 ,....._ 120 .-I s --... CO ;::1. 1 00 ? i:LI 1'0 0::: 80 0 u: CJ'l ? 0::: 60 0 ;:r: p... CJ'l 0 40 ;:r: p... 20 0 0 Fig . 2 . 5 ? I fO I / / I 2 / / / / 4 / / /0 ? / / / 6 / 0/ / / / / . / Leaching Batch 8 10 Equilibration t ime ? .?hr ) 6 10 ? 0 12 1 4 SOLUTI ON CONCENTRATION ( ?g/ml ) 25 . 16 Comparison of adsorp tion iso therms ob tained from batch and and c ont inuous l eaching methods at 6 and 10 hour e quil ibrat ion t ime . For the batch method the curves shown are the Freund l ich iso therms . 26 . a . So i l structure di fference s So i l sampl es in the l eaching columns consisted of some aggregates , in wh ich adsorp t i on rate was l imi ted by s l ow d i ffus i on . Soi l aggregates i n the batch method had been broken down by shaki ng the soi l suspens ion , so that adsorpt ion woul d occur most l y d i rect ly onto the primary part i c les . The adsorpt ion rate for the l each i ng method would depend on both the kinet i c adsorpt ion reac t i ons and the d i ffus ion rate into the soi l aggregates , whi l e for the batch method the adsorpt ion rate wou ld depend on l y on the k i net i c react i on rate . D i ffus i on of phosphorus into soi l aggregates has been stud i ed by Gunary ( 1 964 ) and Evan and Syers ( 1 97 1 ) , u s i ng radi ographi c methods . The rate of d i ffu s i on observed was very s l ow , but varied from so i l to soi l . Tab le 2 . 3 shows cal cu l ated t imes for ch l or ide and phosphorus di ffus ion into spher ica l so i l - aggregates w i th constant concentrat i on at the surface . Detai l s re lat i ng to the ca lcu l at i ons are g i ven in Appendix B . I t takes on ly 3 m i nutes for the average ch l or ide concentration i n s ide the largest aggregates ( 2 mm ) t o reach 80% of the concentrat i on at infi n i te t ime , whi l e i t t akes as l ong as 7 hours for phosphorus at 1 0 ?g/m l surface so lut i on concentrat i on . The solut i on d i s tribut i on coeffi c i ents ( k ) obtained from the l i near Freund l i ch i sotherms were used i n these ca l cu l at i ons . Tab l e 2 . 3 Ca lcu l ated t imes for ch l or ide and phosphorus to d i ffuse into spheri ca l soi l aggregates , as suming a l inear adsorpt i on i sotherm for phosphorus ( see Appendix B for detai l s ) . Time for 80% rep l acement of so i l solut i on (min) Aggregate d i ameter (mm) 0 . 5 0 . 75 1 . 0 2 . 0 Ch l or ide 0 . 2 0 . 5 0 . 8 3 . 2 PhosEhorus Concentrati on k (?g/m l ) (ml /g) 1 . 0 36 . 0 69 1 37 274 1 096 5 . 0 1 7 . 6 34 67 1 34 536 7 . 0 1 5 . 4 30 60 1 20 479 1 0 . 0 1 3 . 2 25 so 1 00 420 - 27 . b . Phosphorus addit ion methods Phosphorus was app l i ed gradua l l y t o the so i l a t a constant rate i n the leaching method , wh i l e in the batch method , as ad sorption occurred phosphorus was removed from so lution and not rep l aced . Thus in the leaching method , the solut ion concentrat ion increased wh i l e adsorpt ion was tak i ng p l ace , but the concentrat ion decreased in the batch method ( see Fig . 2 . 3 ) . As phosphorus adsorpt ion and desorpt ion are hysteret i c , this wou ld l ead to more adsorption re l at ive ' to the fina l so lut i on concentrat ion in the batch method . A l so Barrow and Shaw ( 1 975 ) found that adsorpt ion of phosphorus was s l i ght l y l ess when smal l port ions of phosphorus were repeated l y added to soi l than when the total amount was added at once . Thi s was attributed to b l ocking of some of the adsorpt ion s i tes by the previous add i t ions . 2 . 4 . 3 Phosphorus and Chl oride Movement in Co l umns of So i l Aggregates Breakthrough data for phosphorus and chl oride through columns of 0 . 5 - 1 mm soi l aggregates are shown in Figs 2 . 6 and 2 . 7 for 9 . 2 x 1 0- 6 and 4 . 7 x 1 0- 5 m s - 1 f lux dens i ty respect ive l y . The physica l data for each co l umn are g iven i n Tab l e 2 . 4 . Ch loride breakthrough appeared ear l ier than expected at the l ower f lux dens i ty , w i th C /C . reaching 0 . 5 after e 1 approximate l y 0 . 8 pore vo l umes . Ear l y so lute breakthrough has been vari ous l y attributed to an ion exc l us i on from the negat ive l y charged c l ay surfaces ( Bi ggar and Nie l sen , 1 96 7 ; Krupp et a l . , 1 97 2 ; Thomas and Swoboda , 1970 ) , incomp l ete mixing in the stagnant pores ( Bi ggar and Nie l sen , 1 96 2 ) , and preferent ia l f low between the soi l aggregates ( B iggar and Nie l sen , 1 962 ; van Genuchten and Wi erenga , 1977 ; Green et a l . , 1 972 ) . I f anion exc l us ion had occurred , the area under the breakthrough curves wou l d be l ess than one pore vo l ume . The area under the breakthrough curves was obta ined by integrat i on using S impson ' s rul e . I n fact , the areas obtained for the l ower f lux dens i ty exper iment (Core 1 and Core 2 ) were 0 . 98 ? and 0 . 92 , suggest ing an ion exc lusion was not a s igni fi cant factor . . A l so i f i t was important , it woul d be expected to affect the resu l t s at both f lux dens i t ies . 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 ?rl 0 u 0 - (]) 1 . 0 u 0 . 8 0 . 6 0 . 4 0 . 2 0 0 ( a) CHLORIDE (b ) PHOSPHORUS ? COLUMN 1 -0- COLUMN 2 ? 6 / / . / . "' / / ? I I t:y I 1 . 0 R= 1 0 . 1 ... 1 0 PORE VOLUME .. .. ... -A- COLUMN 1 - ??COLUMN 2 -- --.... ?' ,. R= 1 9 . 7,, ' ;, ,' , ?' ,' , , , , , , , , , , , 28 . 2 . 0 20 Fig . 2 . 6 Dupl icate experimental and calculated b reakthrough curves -6 - 1 for (a ) chloride and ( b ) phosphorus a t 9 x 10 m s flux density . 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 ?M 0 u - Q) u 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 0 --- ------------------- 2 9 . (a) CHLORIDE 2 ? ... COLUMN 3 A COLUMN 4 1 . 0 2 . 0 - - ., - - (b ) PHOSPHORUS ., ., , D ? / , ? 0 ? , , , , , , , , , , " , , , , , , , R= l4 . 1 , , , ? COLUMN 3 , " , , o COLUMN 4 , , , , , , , , , - - - 1 0 20 PORE VOLUMES Fig . 2 . 7 Dup l icate experimental and calc ulated breakth rou )?, l l <;:?u rves . -5 - 1 for ( a ) chloride and (b ) phosphorus a t 4 . 7 x 1 0 m s flux density . 30 . When t he f l ux dens i ty was i nc reased app rox imate ly 5 t imes , the resu l t i ng breakthrough curves of ch l ori de sh i fted to the ri ght and bec ame somewha t ! la t t e r , w i t h t he e f f l uent concen t rat i on reach i ng C /C . = 0 . 5 at 0 . 95 pore vo l umes (1-"i g . 2 . 7a ) . The areas under the e L curves for the fas ter f l ux dens i t y exper i ment were 1 . 03 and 1 . 05 pore vol umes (Core 3 and Core 4 , respect i ve l y ) . The s l ight dev i at i ons from one pore vol ume 1n the area measured at both f lux dens i t ies were probab l y due t o expe ri menta l error . Phosphorus appeared in the effluent much l ater than ch loride . I t appea red aft er 5 pore vo l umes , or 9 hours after app l i cat ion t o the so i l col umns at the l owe r f l ux dens i t y , and a fter 2 pore vo l umes or 0 . 7 hours in the lt i ghe r f l ux Jens i t y co l umns . The concentrat i on in the e ff l uent a l so rose re l at i ve l y more qu i ck l y from the h i gher f l ux dens i ty co l umns , C /C . reach ing 0 . 5 aft er 8 . 5 pore vol umes , compared to 1 1 pore e 1 vo l umes for the l ower f l ux dens i t y colum?s . The differences in pos i t i on and shape of phosphorus break through curves at the di fferent f lux dens i t i e s ind i c ate more adsorpt i on occurred when res i dence t ime increased . 'fh i s behav i our i s 1 n agreement w i th prev ious stud ies reported by Se l im et a l . ( 1 9 74 ) and de Camargo et a l . ( 1 979 ) for phosphorus movement i n so i l columns . The s l ope of the phosphorus breakthrough curves decreased marked ly after C /C . reached 0 . 75 . The t a i l i ng of the breakthrough curves was e 1 probab l y i nf luenced by an interac t ion between s l ow k inet i c adsorpt ion and d i ffus i on i n the comp l ex soi l pore geometry i n the aggregate soi l col umns ( Davi dson and McDouga l , 1 9 7 1 ; Skopp and Warrick , 1 9 74 ; van Genuchten and W i erenga , 1 9 7 6 ) . 2 . 4 . 4 Mode l l i ng of Phosphorus Movement in Soi l Col umns No att empt was made to s imu l ate i n detai l the experimental break? th rough data; on l y a s imp l e , approximate mode l was used . The movement of phosphorus through the soi l aggregate column was described us ing equat ion ( 1 . 9 ) , for whi ch an anal yt i cal solut i on is avai l ab l e ( see Chapter 1 ) . Use of t h i s equat ion meant phosphorus adsorpt i on was assumed to fol l ow a }j near Freund l i ch i sotherm . I ' i I I I I 3 1 . Tab l e 2 . 4 Phys i ca l data for co lumn experiments us ing 0 . 5 - 1 mm aggregat e s . V pb Expe r i men t l l 0 - 5 - 3 l k g m ? ) m s - 1 ) I Col umn 1 1 . 35 8 7 1 Co lumn 2 1 . 34 855 Co l umn 3 6 . 86 877 Column 4 6 . 6 1 860 B Brenner number Co l umn l eng th ( nun ) -8 8 I 90 ' ' I I I I 88 I 90 EC l f) ( ] 0 - 8 m 2 s - 1 ) 0 . 68 1 4 . 5 8 0 . 685 3 . 7 2 0 . 687 23 . 3 0 . 7 1 3 2 9 . 5 T ime for one pore vo l ume (hr) 1 . 8 1 1 . 86 0 . 36 0 . 38 V /V c 0 0 . 80 0 . 86 0 . 95 0 . 95 v t he f l u i d a v e rage ve l oc i ty g i ven by the flux den s i ty ( q ) d i v i ded by the wat er f i l led poros i ty ( 8 ) F. C l d i spersion coeff i ci ent for ch l oride B 32 . 5 37 . 8 2 7 . 3 2 1 . 2 V volume of the effluent when re l at ive concentrat ion i s 0 . 5 , e ( L 3 ) V pore volume ( L 3 ) 0 pb bu l k dens ity (M L- 3 ) 32 . As phosphorus adsorption is very dependent on t ime and the ini tial solution concentrati on , the l inear sol ut ion d i stribut i on coeff i c i ents ( k ) used i n the mode l were obtained from the batch method adsorption isotherms so that the equi l ibration t imes and solut ion concentrat ions were s imi l ar to the contact t imes and effluent concentrat ions in the so i l columns . Two solut ion concentrat ions were used to ca lcu l ated k values . One was the influent concentration ( 2 0 ?g/m l ) , and the other ( 7 ?g/m l ) was approximate l y ha l f the fina l concentrat i on of phosphorus in the eff luent . Equi l ibration t imes were estimated as ha l f the t ime requi red for C /C . to reach 0 . 5 i n the soi l columns . Thus k va lues for e 1 the s l ow f l ux dens i ty column ( Column 1 and 2 ) were obtained from the adsorption isotherm at 10 hours , whil e for the fast f l ux dens i ty columns (Col umn 3 and 4 ) 3 hour i sotherm data were used . For compari son , i sotherm data after 6 hours equi l ibrat i on t i me , w i th 20 ?g/ml solut ion concentrat ion , ( R = 1 0 . 1 ) were a l so used in the mode l . The corresponding retardat i on factors (R ) are g i ven in Tab l e 2 . 5 . Tab l e 2 . 5 Retardat i on factor values (R ) at 7 ( R7 ) and 20 (R20 ) ?g/ml solution concentration , obtained from batch method i sotherms , for use in mode l l ing movement through soi l columns . Experiment Equi l ibrat i on time R7 R20 I simul ated (hr ) Column 1 1 0 1 9 . 7 1 2 . 5 6 1 0 . 1 Column 2 1 0 1 9 . 1 1 2 . 2 6 9 . 9 Column 3 3 1 4 . 1 8 . 3 6 1 0 : 1 Column 4 3 1 3 . 3 7 . 9 6 9 . 5 33 . The experimental ch l oride breakthrough data were used to determine d i spers i on coeffic i ents ( EC 1 ) , us i ng equat ion ( 1 . 1 ) and the method of ana l ys i s proposed by Rose and Pass ioura ( 197 1a) . An adj ustment was made to account for C /C . reaching 0 . 5 before one pore vo l ume . Thi s e 1 was accomp l i shed by rep l ac i ng the pore volume (V ) by an e ffective pore 0 vo lume (V ) in the ca lculations . The parameter V i s the vo lume of the e e e ffluent corresponding to a re l at ive concentrat i on of 0 . 5 . Thi s method of adj ustment was suggested by Rose and Pas s i oura ( 19 7 1a) and has been used by Casse l et a l . ( 19 75 ) , Cagauan et a l . ( 1 962) , and Se l im et a l . ( 1974 ) . The resu l t i ng Brenner numbers ( B = vd/E ) and E va lues are g i ven in Tab l e 2 . 4 . The pred i cted curves for ch l oride agree we l l w i th the experimenta l data , as shown in F i gs . 2 . 6a and 2 . 7a . The shape o f the predi cted phosphorus breakthrough curves does not agree so we l l w i th the experimental dat a , as might be expected ( F i gs . 2 . 6b and 2 . 7b ) . In a l l o f the predi cted curves , phosphorus appears i n the e ffluent l ater than was found experimental l y , and then the pred icted concentrati on r i ses more steep l y than the experimental data . I n F i g . 2 . 6b , the curve using k at 6 hour equ i l i brati on t ime and 2 0 ?g/ml s o lut i on concentration i sotherm (R = 1 0 . 1 ) g i ves the best approximati on of the experimenta l breakthrough dat a , whi l e the curve using k at 1 0 hour and 7 ?g/ml concentrat i on ( R = 19 . 7 ) appears much l ater . However , k at 1 0 hour and 20 ?g/ml concentrat i on ( R = 1 2 . 5 ) i s a l so reasonab l y c lose to the experimental dat a . For the fas t flux dens i ty co l umns , k at 3 hour and 2 0 ?g/ml (R = 8 . 3) g ives the best predict i on ( F i g . 2 . 7b ) . The resu l t s i nd i cate the importance of equi l i ? bration t ime when adsorpt i on i sotherms are used for pred i c ting the movement of react ive so lutes in so i l . A prob l em ari ses i f the s imp l e mode l used here i s to be used pred i c t ive l y . The phosphorus adsorpt i on parameters were obtained i ndependent l y of the mi scib l e d i sp lacement experiment ( d i ffering from most other work i n thi s regard) , but the appropri ate equi l ibrati on t imes for these parameters were inferred from the phosphorus break? through data . However , equi l ibrat i on t imes can a l so be est imated , w i thout aRy breakthrough informat i on , us i ng a s imp l e i terat ive procedure . 34 . Knowing the mass of soi l in the column and the rate at which phosphorus w i l l be added , and us i ng ini t i a l l y say the 24 hour adsorp? t ion value at the solution concentration to be app l i ed ( in this case 243 ?g P/g soi l ) , the t ime needed for enough phosphorus to enter the co lumn to sat i s fy the adsorpt ion demand and rep l ace the soi l solut ion i s ca lcu lated . Ha l f of thi s t i me i s then used as an est imate of the average t ime phcsphorus wi l l have to equi l ibrate in the column during mi sc ib l e d i sp l acement . Th i s equi l ibration t ime i s used to find a new adsorption value from the batch method dat a , and the who le calcu l at ion i s repeated . After 3 i terat ions , the equi l ibrati on t ime est imates for the co lumn experiment were found as 1 3 and 1 . 5 hours for the l ow and h igh flux dens i ty co lumns respective l y . These t imes result in R va lues ( for 20 ?g/m l solution concentrat ion) of 13 and 7 . These va lues are very s imi l ar to the values inferred from the breakthrough data of 1 2 and 8 (Tab le 2 . 5 ) and wou ld resu l t in s imi l ar predi cted breakthrough curves to the ones in F igs . 2 . 6b and 2 . 7b . 2 . 5 GENERAL D I SCUSS ION The results show the usefulness of us ing l inear adsorpt ion i sotherms , obtained at appropriate equi l ibrat ion t imes , for predicting when phosphorus wi l l appear in the effluent from soi l columns . The calculat i ons invo lved can be done s imp ly using a pocket cal cu lator . The same approach could be used to describe phosphorus movement under fi e l d condit ions , when the f low through the soi l i s uni form enough for convent ional convective-d i spers ive theory to app l y and for a d i spers ion coeffic ient to be found . However , as wi l l be shown in the fo l l owing chap ters , i f preferentia l f low is dominant , convective? d i spersive theory is of no use , and in fact adsorpt ion has l it t le effect on phosphorus movement . The various cond i t ions determining whether flow tends to be un i form or p re fe rentia l are d i scussed l ater in the thes i s . 35 . 2 . 6 CONCLUSIONS 1 . More phosphorus adsorpt i on occurred us ing the convent iona l "batch" method than the l eaching column method at comparab le equi l i ? brat ion t imes , part icular ly at l onger equi l ibrat ion t imes . 2 . Ch l or i de breakthrough curves , used as an indi cator of water movement in both l eaching co l umns and soi l aggregate co lumns , ind icated the d i sp l acement in the so i l co l umns occurred re l at ive ly uni form . The curves for the so i l aggregate co l umns were described we l l by the convect ive-di spers ive equati on . 3 . Phosphorus breakthrough curves were asymmetrical with "tai l ing" . The shape of the breakthrough curves could not be descr ibed adequate l y us ing convect ive-di spers ive theory wi th a s imp l e l inear- adsorpt ion i sotherm . However , th i s theory , us ing d i stribut i on coe ffic i ents (k ) determined independent l y at appropriate so lut ion concentrat ions and equi l ibrat ion t imes , predi cted fair ly we l l the pos i t i on of the breakthrough curves , and so the t ime de lay between phosphorus and ch loride breakthrough . 4 . Phosphorus adsorpt ion , and the resu l t ing shape of the break ? through curves , were influenced by the contact t ime of phosphorus in the soi l columns . Increas ing flux density decreased the amount of phosphorus adsorpt ion . 5 . Tai l ing of phosphorus breakthrough curves was probab l y due to s l ow d iffus ion of phosphorus into the aggregates , coup led w i th kinetic sorpt ion reactions . CHAPTER 3 AN ION MOVEMENT THROUGH ART I F ICIAL SO I L CHANNELS AND PLANAR CRACKS 3 . 1 INTRODUCTION 37 . Several workers have sugges ted that water and solute movement in natural soi l s often occurs preferent ia l l y through l arge soi l pores such as worm channe l s (Wi l l 1 ams and Al lman , 1969 ; Eh l ers , 1 97 3 ; Bouma et a l . , 1977a) , fis sures or p l anar cracks (Ri tchie e t al . , 1 972 ; B lake et al . , 1973 ) , worm channe l s and cracks (Wi ld , 1 97 2 ; Bouma and Dekker , 1 97 8 ; Omot i and Wi ld , 1979 ) , and root channe l s (Wi ld , 1972 ; Wi l l iams and Al lman , 1 969) . Such preferent ia l f low has been usua l ly thought to occur in re l at ive l y l arge channe l s . Wi l l i ams and Al lman ( 1969) refer to channe l s 2 - 1 0 mm in d i ameter whi l e Bouma and Dekker ( 1 978 ) and Omoti and Wi ld ( 1 979) suggest cracks greater than 2 mm wide are resposib l e . Such l arge pores wou l d on l y be e ffective in conduct ing water when the soi l is e ffective l y saturated . Jongerious ( 1 957 ) and Brewer ( 1 964) indicated that worm channe l s are on l y effective i n saturated soi l . However , most cracks are c losed by swe l l ing once the soi l is saturated . . F l ow in the smal l er pores becomes s igni ficant in the absence of l arge pores , or when the l arge pores are empty . Omot i and W i ld ( 19 79 ) observed p l anar cracks about 0 . 05 - 0 . 1 mm wide conducting water i n a weak ly s tructured l oamy sand in whi ch no cracks were apparent to the unai ded eye . Bouma et al . ( 1 977b ) , using dye tracer and morphometric techniques , observed cores of medium , subangu l ar b l ocky , structured soi l s . The conduct ing pores observed included mos t l y channe l s and vughs 1 0 . 1 - 1 mm in d iameter , 0 . 1 - 1 mm wide cracks , and a few pores greater than 1 mm in diameter . Anderson and Bouma ( 1977a) , us ing ch l oride as a tracer , found preferentia l f low in saturated soi l cores s imi l ar in structure to those studi ed by Bouma et a l . ( 1 977b) . I n a subsequent paper , ?nderson and Bouma ( 1 97 7b) used a thin crust at the soi l surface to obtain a matric potentia l of - 250 mm under the crust and unsaturated f low in the t op part of the cores . Presumab ly free water f lowed out the base of the cores , the ch loride breakthrough curves from these cores showed much less evi dence of preferent ia l f low than the saturated cores . re l at ive l y l arge voids , usua l l y irregul ar and not normal ly interconnected with other voids of comparab l e s i ze ( Brewer , 1 964 ) . 38 . At a matric potentia l of - 250 mm , cyc l indrical pores greater than 0 . 1 2 mm in d iameter and cracks greater than 0 . 06 mm wide wou ld be air- fi l l ed . Thus their work indicated that such pores l arger than this were respons ib le for most of the preferent ia l f low . However , chl oride s t i l l appeared in the eff luent somewhat ear l i er than expected even under part i a l l y unsaturated cond i t ions , probab l y due part ly to saturated f low near the bottom of the cores and poss ib ly a l so anion exc lusion . Comparab le di fferences between saturated and unsaturated break? through curves were also obtained by E lrick and French ( 1966) and Kisse l et a l . ( 1973 ) in natura l und i sturbed soi l cores , and by Bouma and Anderson ( 1 977 ) in art i fic ia l soi l columns containing vertical channe l s 5 mm in diameter . However , the porous p l ates or crusts , wh ich were used in a l l the unsaturated f low experiments , wou ld reduce the effects of any preferential solute movement through the soi l . Thus the breakthrough curves obtained , for examp le by Anderson and . Bouma ( 1977b ) , wou ld be influenced to some extent by the hydrodynamic d i spers i on in the crust on the top of the cores . Whi l e many works have attributed the ear l y appearance of app l ied so lutes in the effluent to preferent i a l f low through l arger soi l pores , no experimental work indicating the minimum s i ze of pore al l owing preferential f low appears to have been done . A theoretical mode l deve loped by Scot ter ( 1 978) , us ing a s imp l i fied pore geometry , suggest s that vert ical channels 0 . 3 mm in d iameter and cracks 0 . 1 mm wide are the minimum s i ze of conducting pores for both reactive and non-react ive so lute movement to occur preferentia l l y . Unfortunat e l y , due to the regu l ar shape of channe l and crack as sumed in the theory and the i rregul ar pore shape in rea l soi l , the resul t s of the model predictions cannot be quantitat ive ly compared with experiments using natural soi l . 3 . 2 OBJECTIVE The experiments described in this chapter aimed to observe the movement of sorbed and non- sorbed anions through soi l columns containing art i fic ia l vertical channe l s and p l anar cracks of regu l ar ? shape . 39 . 3 . 3 MATERIALS AND METHODS The soi l was taken from the Ah2 hori zon ( approximate ly 150 mm depth) o f Tokomaru s i l t loam (Appendix A) . Soi l samp les were air-dried and passed through a 2 mm s ieve . Soi 1 cyl inders containing an art i fic ia l channe l , approximate l y 47 mm i n d i ameter and 50 mm l ong , were made by casting a so i l s lurry around a l ength o f nyl on l ine 0 . 3 mm in d iameter . The soi l s lurry was made by mixing air- dried soi l w i th 330 ?g/ml ca lc ium ni trate solut ion (Ca (N0 3 ) 2 ) . After pouring , the cast s were air-dried and then oven-dr i ed at 105C . The presence of calcium ions and oven- drying , s tabi l i sed the soi l structure . The cast was then s l ow l y rewet to a sma l l pos i t i ve pressure potentia l and the nylon l ine was then removed . The s ide and most o f the bottom of some cast s was coated w i th paraffin wax . Other casts were coated al l over except for a 5 mm uncoated annulus around the entry and exit of the channe l . Soi l casts containing a s ing l e p l anar s l i t were prepared in a s imi l ar manner us ing brass shim 0 . 1 5 mm thick and? 1 2 mm wide , and the s ide and most of the bottom of these casts was a l so coated w i th wax . Each soi l cast was saturated s l ow l y from be low with a so lut ion containing 1 64 ?g/ml calcium ni trate , and 1 00 ?g/ml sodium az ide (NaN 3 ) to control microbial growth . I t was then l eached with the same so lut ion unt i l the flux was approximate l y constant . Next the Mariotte supp l y was removed , and as soon as free solut ion had disappeared from the soi l surface , the d isp lac ing solution , containing 1 0 ?g/ml phosphorus as potass ium dihydrogen phosphate ( KH2 PO? ) , 350 ?g/ml chl oride as potass ium chl oride (NC l ) , 1 15 ?g/ml cal cium n itrat e , and 1 00 ?g/m l sodium azide , was app l i ed and effluent a l iquots t aken a t suitab l e _ t ime interva l s . The cal c ium nitrate concentrat ion in the d i sp lac ing soluti on was chosen so that the d i sp l aced and d i sp l acing solution had the same dens i t y . The anal ys i s method of Murphy and Ri l ey ( 1962) was used for phosphorus , and t i trati on w i th s i lver n itrate ( Bower and Wi lcox , 1965) or a speci fic ion e l ectrode , for ch l oride . ----- ?----------------------- 40 . 3 . 4 COMPUTATIONS Breakthrough curves for ch loride and phosphorus were computed assuming v i scous f low of so lut ion down verti ca l cyl indrical channe l s and p l anar cracks , w i th s imu l taneous mo l ecu lar d i ffusion of solutes into the surrounding soi l . Deta i l s of the theory are g iven by Scatter ( 1978 ) and are presented in Appendi x C (Section C . l ) with a typ i ca l CSMP programme ( Section C . 3 ) . The retardat i on factor (R) for phosphorus was arbi trari ly se l ected by fit ting the ca lcu lated curves from the mode l to the experimental data for phosphorus movement through packed aggregate columns containing the same soi l materia l s ( Kanchanasut et al . , 1 978 ) . The channe l s i ze in the aggregate soi l co lumns was estimated from the retentivity curves . The R values used were 69 for phosphorus and zero for chl oride . As some s l ight swe l l ing and shrinkage was unavoidab le during pre? parat ion of the cast s , the actua l effect i ve d i ameters of the channe l s and widths of the s l i t s were ca lcu lated from the f low rates , us ing equat i ons ( 1 ) and (6) in Appendix C . l , respective l y . 3 . 5 RESULTS AND D I SCUSS ION Phys ica l data for the soi l casts are g iven in Tab l e 3 . 1 . The break? through data for ch l oride and phosphorus in the soi l cast s containing channe l s are shown i n F i g . 3 . 1 . B l ock ing the channe l s at the conc lus ion of the experiment typi cal l y reduced the f low to 1 % of i t s previous value , indi cating nearly a l l o f the f low was through the channe l rather than uni formly through the soi l . The movement of a rhodamine B dye-water mi xture subsequent ly app l i ed to the cast s a l so supported th is conclusi on , w i th on l y the channe l wal l or p l anar crack wal l , and l ess than 1 mm thickness o f the surround ing soi l matrix being affected by dye , as shown in F i g . 3 . 2 . 1 . 0 1- 0 . 8 0 . 6 'T'i u - ?Ill u 0 . 4 0 . 2 0 I ? 6. 0 I I I I I 11 I I 0 I J .... ... ? 6. 6. 0 0 ," //'' / J T l T I J 1 "' ... M - 6. 6. 6. 0 0 ? _.2--o--"-1>- - - - - - -- - ? .JJ - - - - - - - -"- -- 6. - --- (a ) 10 TIME (min) 1 . 0 ... 0 . 8 TIME 0 (min) 2 4 I I ...-?':in:- - ::J ? .... ? 0 (b ) 0 . 6{. { 20 Fig . 3 . 1 Breakthrough data for soil casts containing a s ingle channel . (a ) Breakthrough data for soil casts with exposed soil surface ; Cast I , chloride ( ? ) and phosphorus ( 0 ) ; Cast I I , chloride ( .A ) and phosphorus ( 6. ) . (b ) Breakthrough data for soil cast with almost completely wax-eoated surface ; Cast Ill , chloride ( ? ) and phosphorus ( o ) . Also shown are predicted breakthrough curves for Cast I and I l l , chloride (---- ) and phosphorus (- --) assuming R=69 for phosphorus . ? ...... F i g . 3 . 2 42 . Movement of Rhodamine B dye s o l u t i on i n the soi l casts containing art i fi c i a l ( a) vert i ca l channe l , and ( b ) p l anar crack . 4 3 . Tab l e 3 . 1 Phys i ca l data for so i l cas t s containing a channe l or crack . pt, (kg m- 3 ) Channel I 1 5 50 I I 1 5 30 I l l 1 590 Crack IV 1 350 V 1 3 1 0 Pb bu lk dens i ty K 8 ( 1 0 - 5 m s - 1 ) 0 . 38 0 . 67 0 . 38 1 . 1 8 0 . 35 1 . 04 0 . 46 3 . 00 0 . 46 2 . 5 0 K hydraul i c conduct iv i ty D i ameter Time for or width one pore k (mm) v0 lume (m1 g- 1 ) (min . ) 0 . 4 7 35 1 6 . 9 0 . 5 3 2 2 1 7 . 1 0 . 5 1 2 3 1 5 . 2 0 . 1 7 1 0 20 . 7 0 . 1 6 1 1 24 . 4 k adsorp t i on d i s tribut i on coeffic i ent of phosphorus 8 vo l umetri c water content . Pressure potent i a l I to drain I (mm) I - 63 - ss - 58 - 86 - 92 F i g . 3 . 1b shows the breakthrough data for ch l or i de and phosphorus in the soi l cast comp l ete l y wax - coated , except for j us t around the channe l . The exi t concentrat i ons of both ch l ori de and phosphorus ? rose s teep l y w i th re l at i ve concentrat ion (C /C . ) reaching uni ty in l es s than e 1 2 minutes . Thi s was i n good agreement w i th breakthrough curves computed us ing the s i mp l e theory in Appendi x C . 1 . In F i g . 3 . 1 a are shown the data for two casts w i th s l i ght l y d i fferent channe l s i zes and w i thout wax coat ing on the surface . The ch l or i de concentrati on d i d not r i s e as rap i d l y as i n F i g . 3 . 1b , a l though the re l at i ve concentra t i on of both casts s t i l l reached 0 . 5 in l es s than a minut e , after on ly a sma l l fract i on of a pore vo l ume had perco l ated , 44 . indi cat ing preferent ia l movement down the channe l . However , as near l y a l l the f l ow was down a s ing l e channe l , reference t o the number o f pore vo lumes has l i t t l e s i gn i fi cance . The phosphorus concentrat i on a l so rose rap i d l y , indi cat ing preferent ia l movement , and in 10 min . reached a re l at i ve concentrat i on of approximate l y 0 . 84 . A correcti on d i scussed in Appendix C was inc luded in the computer programme t o account for d i ffusion across the surface from the ponded so lut ion as we l l as rad ia l d i ffus i on out of t h e channe l . The d i fference between F i gs . 3 . 1 a and b shows the effec t of oi ffus ion across the surface of the cast s . Breakthrough data for ch l oride and phosphorus i n the dup l i cate casts containing p l anar cracks are shown in F i g . 3 . 3 . The computed theoret i ca l curves take approximate account of d i ffus i on across the surface of the cas t s , in the manner al ready referred to . Again , pronounced preferen t i a l movement of ch l oride and phosphorus occurred and the theoret ica l and experimental data are in reasonab l e agreement . The breakthrough data for the dup l i cate cas t s w i th s l i gh t l y d i fferent s i zed channe l s or crack s indicate the sen s i t iv i ty of f l ow to the s i ze of conduct ing pore . Scot ter ( 1978 ) has shown the predi cted breakthrough curves for d i fferent channe l and crack s i zes in F i g s . 2 , 3 , and 4 in Appendix C . The s hape of the curve i s s t rongl y dependent on pore s i ze . The val ue of R used gave phosphorus breakthrough curves in reasonab l e agreement with the experimenta l dat a . The values o f the adsor?i on di s tr ibut i on coeffi c i ent ( k ) for each cas t were c a l cu l ated ( R = ll k , see Sect i on 1 . 2 . 3) , i n order to compare them with the va lues obtained from the adsorption i sotherm experiments described in Chapter 2 . The k val ues obtained are g iven in Tab l e 3 . 1 . They are a l l somewhat h i gher than the measured i sotherm v? lues at 6 and 1 0 hour equi l ibrat ion t imes , whi ch were 1 0 . 7 and 1 3 . 2 m l g - 1 at 1 0 ?g/ml phosphorus concentrat i on , respect ive ly ( F i g . 2 . 5 ) . k values sma l l er than the measured i sotherm values m ight be expected , due to the t ime dependence of phosphorus adsorpt i on , and the phosphorus movement through the channe l s and cracks occurring within a few minutes after app l i cati on . However , soi l s in the casts had been mixed w i th ca l c i um i ons during preparat i on of the . ,...; u - (l) u 1 . 0 +-- 0 . 8 1- o . 6 r 0 . 4 t- 0 . 2 I 1 ?------ . . . . A -, - - - - - - ---- ----- - - - - - - - - - -- ?- -- - - - - - -- - - ? / 0 0 /o o o c. ll \l o g o A c. g I A 6 I f0 I : A I ?::,. I 6 - .... @ @ I 0 1 0 0 2 0 TIME (min ) Fig . 3 . 3 Breakthrough data for soil cast s containing a s ingle crack ; Cas t IV , chloride ( ? ) and phosphorus ( 0 ) ; and Cast V , chloride ( A ) and phos phorus ( A ) ? Also shown are predicted b re akthrough curves f or Cast IV , chloride (--- ) and phosphorus (-- --) assuming R= 6 9 for phosphorus . The c ircled symbols indicate t ime for one pore volume . +;. Vl - - - - - - - - - - -------------------- - ---- ----- - - 46 . cas t s , whi ch wou l d i nc rease the ads orp t i on capaci ty for phosphorus (John , 1 9 72 ) . The re l at i ve l y sma l l d i fference between the ch l oride and phosphorus breakthrough curves when preferen t i a l f low occurs ( F i g s . 3 . l and 3 . 3 ) indicates however that the actua l k v a lue s chosen are of l i t t l e i mportance . lbe resu l t s of th is expe r iment support the theory proposed by Scatter ( 1 9 78 ) . I t was found that so l ute adsorpt ion has l i t t l e e ffect o n preferen t i a l f l ow through channe l s a t l east 0 . 4 7 mm i n d iameter and cracks a t l tl'ast 0 . 1 7 mm wide . Whi l e the movement through c hanne l s or c racks c l oser t o the cr i t i c a l s i ze cou l d not be stud i ed , it is reasonab l e to as sume that the theory w i l l s t i l l be va l id for s uch f lows . Tab l e 3 . 1 , and Tab l e s 1 and 2 i n Appendi x C . l , show the pre s sure potent i a l s n eeded t o drain chann e l s and c racks of a range o f s i z es . The pre ssure potent i a l needed to drain t h e cri t i ca l s i zed pores i s - 200 mm , thus pre ferent i a l f l ow wou l d on l y h e e xpected t o occur in s o i l c l os e to o r a t s aturat i on , and not i n s o i l at " fi e ld capac i ty" o r dri er . Channe l s l arger than 0 . 2 mm in d i ameter and cracks w i der than 0 . 1 mm are probab l y qui te common i n natural so i l , a s many root s , p l anar crack s o r fi s sure s , and soi l organ i sms are o f the se magni tude s or l arger . I f such s o i l pores are water- fi l le d , and c ont inuous or interconnect ing w i th other l arge pore s , preferen t i a l f l ow wou l d be expec t ed . 3 . 6 CONCLUS I ONS 1 . Ch l or i de and phospho rus moved a lmos t i n s tantaneous l y t hrough cast s o i l c o l umns approximat e l y SO mm l ong con taining a s i ng l e channe l about O . S rrun in d i ameter , or c rack about 0 . 1 7 mm by 1 2 mm . The e ff luent concentrat i on of both an i ons reached SO% of the inf luent concentrat i on l e s s than a mj nute after the ir app l i cat i on to the surface . D ye patterns confirmed that a l mo s t a l l o f the s o l ut i on moved through these pores ra ther than uni form l y t hrough the s oi l . 47 . 2 . Experimenta l b reakthrough data for both ch l oride and phosphorus are in general agreement w i th mode l pred i c t i on s , a s sum ing a l l t he v i sc ous f l ow i s down the channe l or c rack , whi l e mo l ecu l ar d i ffus i on t ransports s o l ute i nto the so i l around the channe l or crack . The mode l imp l i e s that preferen t i a l f l ow wi l l o ccur on l y i f vert i ca l , cont inuous , wate r- fi l l e d channe l s at l east 0 . 2 mm i n d i ameter and/or cracks 0 . 1 mm wide are present in s oi l . CHAPTER 4 AN ION MOVEMENT I N SOI L CORES - - -- - - - ??------?------- --- - - 4 . 1 I NTRODUCTI ON 4 9 . M i s c i b l e d i sp l ac ement exper iment s c arr i ed out i n the l aboratory usua l l y invo lve u s ing soi l c o l umns of d i s t urbed and repacked s oi l s ( e . g . Dav i dson and Chang , 1 9 7 2 ; 1homas and Swoboda , 1 9 7 0 ; v an Genuchten e t al . , 1 9 7 4 ) . The pore geometry in such repacked soi l c o l umns i s very d i fferen t t o so i l under f i e ld cond i t i on s , and the extrapo l a t i on o f the re su l t s obta i ned to f i e l d s i tuat i ons may l ead t o l arge errors . D i fferences 1 n s o lu te movement in packed soi l c o lumns and re lat i ve l y und i s turbed soi l cores have been shown b y E l ri ck and F rench ( 1966) , Ki s s e l e t a l . ( 1 9 7 3 ) , McMahon and Thomas ( 19 74 ) and Cas s e l e t a l . ( 19 74 ) . The breakthrough curves for c h l oride obtained from the und i s ? t urbed soi l c ore s showed ch l oride appearing earl i e r , and t h e re l at i ve concen trat i on i n the e ff luent r i s i ng more rap i d l y than i n the repacked c o l umns , for wh i ch the c la s s i ca l "S- shaped" break through curves were foun d . The rap i d movement o f so lutes through und i s turbed so i l cores has been a t tr ibuted to the v i scous f l ow occurring predominan t l y through on l y a few re l a t i ve l y l arge con t i nuous pore s . Thi s has b e en confi rmed us ing dye- tracing t e chni ques b y R i tch i e e t a l . ( 19 7 2 ) , Omo t i and W i l d ( 19 79 ) and Anderson and Bouma ( 19 73 ) . I n format i on on so i l pore geometry and c onnec ting patt erns w i th in a so i l can b e i n ferred from the breakthrough curves . Anderson and Bouma ( 1 9 7 7a) found d i fferent breakthrough curve shapes for s o i l cores w i t h b l ocky s t ructure and those w i th pri smat i c s tructure . The rap i d movement o f ch l or i d e through the b lock - s tructured soi l indi cated more c on t inuit y of the c onduc t in g pores than in the pr i sma t i c s t ructured so i l . However , d i s t inct d i fferen ces b etween s imi l ar s t ructured s o i l s and rep l i cate cores were observed b y Anderson and Bouma ( 1 977a ) ?and Bouma and Wosten ( 1 9 79 ) , and even between rep l icates w i t h s imi l ar v ol umetri c water contents and hydraul i c conduc t i vi t i es by E l ri ck and F rench ( 1966 ) . The d i fferences were at tributed t o so i l h e t ero? gene i ty caus ed by crack s , root channe l s , w orm chann e l s and d i s t urbances caused by agricu l t ural prac t i ce s . 5 0 . To minim i z e the e ffec t s o f so i l heterogene i ty , R i t ch i e e t a l . ( 19 7 2 ) suggested t h e use o f l arge c y l indr ica l s o i l cores w i t h t h e d iameter l arger than the l ength , i n order t o reduce the e ffec t s o f d i s cont inui t i es i n conduc t ing pores c aused by the core wa l l s . A l though on l y sma l l d i fferences i n hydrau l i c conduc t iv i t y were obse rved i n cores o f di fferen t l engths, break? through curves for c h l or i de shown by Ki s s e l e t a l . ( 19 7 :) ) were s i gni fi ? cant l y di fferent . More pronow1ccd pre feren t i a l f l ow occurred through the short e r? cu re (b4 nun l ong) t han the l onger core ( 600 mm l ong) . Observed breakthrough curves for unsaturated so l ut e f l ow t hrough und i s turbe d soi l core s have been l e s s variab l e than for saturated core s . u?-t When s o i l i s un saturated l arge c onduct ing pores are dra i ned , and the movement o f water and so lutes i s through the sma l l er pore s , w i th a narrower s i ze range . The movement i s thus s l ower and more uni form . At a soi l pres sure po ten t i a l o f about - 300 mm ( - 30 mbar ) , E lr i ck and F rench ( 1966 ) found ch l oride movement o ccurred re l ative l y un i form l y and resul t s for rep l i cate cores were i n c lose agreemen t , i n c on t rast t o the saturated f l ow data referred to above . U s ing the same s o i l cores as in the previous l y described saturated f l ow experi men t s (Anderson and Bouma , 1 9 7 7a) , Anderson and Bouma ( 19 7 7b) obs erved the movement of ch loride became l e s s preferent i a l when unsaturated f l ow was i n duced by a surface gypsum cru s t 5 rrun th i ck . The corresponcling pres sure potent i a l s were about - 2 50 and - 35 0 mm for b l ocky and prismati c s tructured soi l respect i ve l y . A t these pres sure poten t i al s , chann e l s a t l east 0 . 1 2 mm i n d iameter and crack s at l ea s t 0 . 06 mm w i d e woul d b e draine d . The s e s i z es are s l i gh t l y l es s than the cri t i c a l s i ze s fo r pre feren t i a l fl ow proposed by Scat ter ( 1 9 7 8 ) . Howeve r , the resu l t i ng break through curve s s t i l l indi cated s i gni f icant preferen t i a l f l ow , probab l y due part l y to the experimen ta l method used , a s d i scussed i n Sect i on 3 . 1 . U s i ng a much h i gher pressure potent i a l o f - 5 0 mm ( at wh i ch channe l s greater than 0 . 5 mm i n d iameter woul d be drained) , Bouma and WBs t en ( 19 79 ) s t i l l observed a s i gn i fican t shi ft t o the r i gh t o f the ch l oride breakthrough curves for saturated und i s ? turbed core s . 51 . Conduc t ing pores in the unJ i s tu rbed Tokomaru s i l t l oam s o i l have been i n ve s t i gated us i ng dye t ra c ing t echniques . Corker ( 19 77) and 1>1cAu l i ffe l 1 9 7 8 ) app l i ed dye s o l u t i on to saturated s o i l cores a fter h yd rau l i c conduc t i v i t y measu remen t s cmd obs e rved worm channe l s , p l an t root s , and c ra c k s were the ma j or pa t hways for wat er movement . The hyd rau l i c conduc t i v i ty o f t he se so i l was great l y i n f l uenced by worm a c t i v i t i e s . McAu l i ffe ( 19 7 8 ) ob se rved changes in f l ow paths and hyd rau l i c conduc t i v i t y w i t h t i me due to worm act i v i t y . 4 . 2 OBJECT I VES The work d e s cr i bed in t h i s chapt e r a imed to s tudy the saturated and unsatura t ed movement o f an i on s through re lat i ve l y large und i sturbed s o i l core s . The e ffec t o f an i on so rpt i on on pre feren t i a l f l ow under s aturated cond i t i on s was i n ves t i ga t ed . Unsaturated f l ow was inve s t i gated to t e s t whether the cri t i ca l s i ze s of c y l indri ca l chann e l s and c racks for preferen t i a l f l ow p ropo sed by Scat t er ( 1978 ) were va l i d in natura l soi l core s . I mp roved t e c h n i ques for unsatura t ed m i s c ib l e d i sp l acement were a imed for . prov i d i ng a un i form matr i c poten t i a l throughout a fa i r l y l arge core , and avo i d in g the e ffec t s o f porous p la t e s a t e i ther end o f t h e core . 4 . 3 MATER IALS AND I>?ETHODS 4 . 3 . 1 Experimen t a l Set - up Mi s c i b l e d i sp l ac ement o f ani on s was i nve s t i gated using "undi st urbed" so i l c o re s und e r both s aturat ed and unsaturated f l ow cond i t i on s . The equipmen t used i s i l lu s trated i n F i g . 4 . 1 . I t was opera t ed s o that t here was a s teady s ta t e f lux w i t h i n the soi l and the grav ita t i ona l poten t i a l was t h e maj or dri v i ng force . Such cond i t i on s imp l y an hydrau l i c grad i en t c l os e t o un i t y and a near con stant pres sure poten t i al t hroughout the s o i l during s t eady f low . Cy l indr i cal cores were ob t a ined during w inter 1979 from th e top 1 50 mm o f Tokomaru s i l t l o am ( Hori zon A) i n the non - t i l e drained area growin g pasture in Dai ry Farm No . 4 , Mass ey Univers i t y , P?lmerston North . The a r?e ; i s adj acent t o the other experiment a l s i t e s descr ibed in Sec t i ons 5 . 3 and 6 . 3 . Cy l i ndr i ca l a lumin ium corers , 143 mm i n i ns i de F i 4 . 1 con t ro l l ed a i r p re s s u re t n perspex shee t a l um in i um corer s t a i n l es s s t ee l g l ass f u n ne l au tom a t i c f r a c t ion c o l l e c t o r .?. ? - -- man i fo ld ;? s o l u t i on i n t o wa te r manome t ers ???- t e n s i ome ters n y lon me s h f 60 J..(m t S 2 . C(' ncra l c>X jw r i.mc>n t :1 l s e t-up f o r mi s c i b l e d i s p l a ceme n t ::o t udy . 5 3 . d i ameter and 1 7 0 mm l ong w i th a beve l l e tl cutt ing edge , were forced into the s o i l t o the des i red depth . The wa l l thicknes s o f the cores was 2 . 5 mm , g iv ing an area rat i o 1 of 0 . 0 7 , and the corer cou ld then be defined a s a thin wa l l samp ler ( are a rat i o l es s than 0 . 2 , Loveday , 1 9 74 ) . So i l wat er c onten t when the cores were taken was approx imat e l y at fie l d capac i ty . To m i nim i se d i sturbance of s o i l s t ructure , excavat i on of the s o i l cores was carefu l l y made by remov i ng the so i l around each core to s l i ght l y deeper than the s o i l core depth and then break ing off the s o i l core . The excess s o i l was removed from the base o f the core by gen t l y chipping off the aggregates a l ong the natural fracture p l anes when the s o i l was p art l y dry , to prevent smeari ng and s o s ea l ing off water conduct ing pore s . A fine ny lon mesh w i th an e ffect ive p ore d i ameter o f 6 0 wm was u sed t o s ea l the bottom o f t he core s . I t was found more s at i s factory than a porous p l ate, as i t had a h igher permeab i l i ty and a con s i derab ly sma l l er p ore vo lume . The pore vo lume o f a fr i tted-g l a s s p l at e o f poro s i t y 3 and 4 . 6 mm t h i ck was found t o be 0 . 1 4 m l /cm 2 whi ch wou ld be approx imate ly 2% o f the pore vo lume o f the so i l core used in t h i s experimen t . The u se o f such a p l at e wou l d t end t o mask the actual s o lute breakthrough from the s o i l core , especi a l l y i f preferent i a l f l ow occur s . The n y l on mesh was p l aced acros s the co lumn base and s ea l ed t o the edge of the c ore u s ing s i l i cone rubber s ea l ant . Ten s iometers w i th water manomet er s w ere ins t a l l ed and seal ed i n t o the so i l core , 25 mm from both end s . F r i tted g l a s s fi l ter s t i c k s , 1 0 mm in d i ameter , were emp l oyed as the ten s i ometer s ensors . 4 . 3 . 2 Saturated F low iment Soi l c o l umns were i n i t i a l l y saturated w i t h a s o lu t i on containing 109 wg/ml po ta s s ium n i trate ( KN0 3 ) and 1 00 !Jg/ml sodium a z ide , and t hen l eached with the s ame s o l u t i on whi ch was supp l i ed by a per i s t a l t i c pump . A head of app rox i mate l y 1 0 mm was ma in t a i ned on the soi l surface . When t he f lux was con s t ant and the hydraul i c gradient was approx imate l y uni t y as i nd i cated b y t h e t en s i ometer s , t he inf luent s o l ut i on containing 1 0 0 ?g/ml ch l oride as pota s s ium ch l or i de , 2 0 wg/m l phosphorus as potas s ium d i hyd rogen phosphate and lOO ?g/m l s od ium a z i de was app l ied t o the s o i l as s oon as the p revi ou s s o l ut i on had d i sappeared from the' s o i l surface . 1 area rat i o area o f annu lus o f d i s p l aced so i l areaofthe sampkatthe- cut ing edge 5 4 . The eff luent s o lut i on was c o l l ected at appropriate t ime interval s us ing an automat i c fract ion c o l lector unt i l approx imat e l y 2 pore vo lwnes of the e ffluent were obtai ned . The e ffluent a l i quot s were ana lysed for ch l or ide and phosphorus u s ing a speci fi c i on e l ectrode and the Murphy and R i l ey method ( 1 96 2 ) , respec t i ve ly . Methy l ene b lue dye s o lut i on ( 0 . 1 % b y we i ght aqueous s o l u t i on ) was subsequent l y app l i ed to the so i l core and l eached for approximate l y 2 pore vo l ume s . The s aturated so i l cores were then we i ghed t o a l l ow the pore vo lume t o be determined l at er . 4 . 3 . 3 Unsaturated F l ow De saturat i on o f the s o i l core was then ach i eved by app l y i ng a pos i t ive air pressure to the c l o sed chamber above the s o i l surface and reducing the inf low rate . The air pres sure was maintained at 0 . 02 bar ( 20 0 ? 10 mm of water) and the corresponding pre s sure pot en t i a l s i n s ide the so i l cores as i nd i cated by the t en s i ometers were - 2 0 0 ? 1 0 mm o f water at both ends . When a new s teady s ta te had been reached , the inf luent s o l ut ion was changed t o a s o l u t i on c ontaining 5 0 0 ?g/ml of bromi de a s pota s s ium bromide ( KBr) , 1 00 ?g/ml sodium az ide and 1 % b y we i ght rhodamine B dye . Th i s s o l ut ion was i ntroduced i n t o the s o i l core s through a man i fo l d with S sma l l out l e t s whi c h al l owed i t to be fai r l y uni fo rm l y d i s t r i buted over the so i l surface . E ff luent a l iquo t s were co l l ected and anal ysed for bromi de unt i l 2 pore v o l umes were obta ined . No rhodamine B dye was v i s ib l e in any of the a l i quot s . At the conc lu s ion o f the experiment , the unsaturated s o i l cores w ere we i ghed and then oven-dried to con s tan t mas s a t l OSC . The saturated and unsaturated l iqui d - fi l l ed pore vo l umes were determined grav i me tr i ca l l y from the wet and dry mas s and the dens i ty of wat.er . 4 . 4 RESULTS AND D I SCUSSI ON 4 . 4 . 1 Saturated Fl ow E.x.periment Phy s i ca l data for the dup l i cate soi l cores are g iven in Tab l e 4 . 1 . Break th rough curves for ch l oride and phosphorus obtained from the saturated s o i l cores are p re s ented i n F i g . 4 . 2 . Both c h l oride and 1. ss . Tab l e 4 . 1 Phys ica l data for saturated and unsaturated f l ow . Saturated F l ow K T i me for ( 1 0 - 5 0 one pore s - 1 ) v o l ume m (hr ) Core I 2 . 44 0 . 562 0 . 90 Core I I I . 67 0 . 5 5 8 1 . 3 K hydrau l i c c onduct iv i ty e = vo l umetri c wat e r content E d i sper s i on coe ffi c i ent Unsaturated F l ow K Time for E ( 1 0 - 7 e one pore ( 1 0- 8 s - 1 ) vo l ume mz s - 1 ) m (hr) 1 . 03 0 . 535 2 15 1 . 80 1 . 56 0 . 5 3 1 1 37 0 . 5 8 56 . phosphorus appeared in the e ff luent a lmost i mmedi at e l y a fter app l i cati on t o the s o i l surface o f both co res , indi cat ing very pronounced preferen t i al movement o f s o l utes o ccurr i ng during the d i s p l acement p roces s . The breakthrough curves obtained are s imi l ar t o the one s obtained from sma l l "undi s turbed" s o i l cores ( 5 4 mm in d i amet e r and 6 0 nun l ong ) conta in ing a worm chann e l by Kanchanasut et a l . ( 19 7 8) and from the art i fi c i a l so i l core s c ontaining a channe l or s l i t described in Chapter 3 ( sect i on 3 . 5 ) u s ing the same soi l . I n F i g . 4 . 2 , breakthrough curve s for the dup l i cate cores are s l i ght l y di fferent ; C /C . o f ch l oride and phos -e 1 phoru s was 0 . 6 and 0 . 4 2 at 0 . 1 pore vo l ume respect ive l y for core I w i th an hyd raul i c conduct i v i ty o f 2 . 44 x 1 0 - 5 m s - 1 , whi l e the corresponding va lues were 0 . 4 and 0 . 1 4 for core I I w i th the s l i ght l y s lower hydrau l i c c onduc t iv i ty o f 1 . 6 7 x 1 0 - 5 m s - 1 ? The re l at ive l y l arge s i ze s o i l cores used i n th i s experiment provi de d a re lat ive l y narrow range o f mea sured hydrau l i c conduct iv i ty values . The mean and s t andard + ? s - 1 e rror for 5 c o re s was ( 2 . 4 2 _ 0 . 5 ) x 1 0 m s . F or sma l l e r , wax c oated cores 76 mm in d i ameter and l ength c o l l ected at the same s i t e , Corker ( 19 7 7 ) found l ower , but more variab l e conduct iv i ty va l ue s . H i s va l ue s were ( 1 . 5 -t 1 . 1 4 ) x 1 0- 5 m s - 1 for 1 0 core samp l e s obtained from 0 - 75 mm depth and ( 9 . 44 + 3 . 3) x 1 0 - 6 m s - 1 for S core samp le s obtained from 1 00 - 1 7 5 mm depth . The di fferences i 1 1 the measured hydrau l i c c onduct iv i ty values cou l d be due t o d i fference s in fie l d water s t atus and s o i l organ i sm act i vi t i es when c ore s amp l e s were t aken , rather than the d i fferent core vo lume s . i\ l so i n sma l l s o i l c ores , compres s i on during s ampl ing is mor e l i ke l y . Methy l ene b l ue dye , whi ch was emp loyed t o show the dominan t fl ow path s , reve a l ed that the movement of so lutes predominant l y occurred i n t he l arge pores c on s i s t ing o f worm chann e l s , d ecayed root channe l s , the space b etween roo t s and the soi l matrix , and fracture p l ane s . Cros s ? s e c t i on s o f one o f the so i l cores are shown in F i g . 4 . 3 . B lue- s tained worm channe l s are obv ious , part i cu l ar l y near the soi l surfac e . B l ue ? s ta ined fracture p l anes , root s and worm channe l s were obs erved i n verti ca l s ecti on s o f t h e s o i l core ( F i g . 4 . 4 ) . The evidence from F i g . 4 . 4 ind i cate s that the movement o ccurred in l arge interconnected pores o f a l l t ype s . Worm channe l s d id not nece s sar i l y have t o reach the surface to be e ffect ive i n conduct ing water and s o l ut e s , prov ided l . O 0 . 8 0 . 6 0 . 4 0 . 2 0 1 . 0 0 . 8 0 . 6 0 . 4 0 . 2 0 0 0 ( a ) CORE I ? ? ? ? I ? ? ? ? Chl o r ide? ? ? ? ? ? Phosphoru? A A A ? ? ? ? ? ? ? ? ? l . O (b ) CORr= I I ? ? ? ? ? ? Chloride ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 1 . 0 5 7 . 11- ?---?- . ...... .. - . 2 . 0 ... . . - ?-- ? ? ---- -. . . ... .-,...,. / / Phosphorus ? ? 2 . 0 T'ORF. VOLU!1ES F i g . 4 . 2 Breakthrough d a ta for ' undi sturbed ' soil cores : chlorid e ( ? ) and phosphorus ( ? ) for s aturated flow , and b romide ( ? ) for unsa turated f low at -200 mm pressure po tential . Calculated b reakthrough c urves for b romide are presented , assuming channel d iame ter o f 0 . 15 mm (-- ) and 0 . 1 mm (- -) . Duplicate cores ( a ) and ( b ) . ( b ) 5 8 . (a ) 20 mm be l ow the so i l surface 20 mm above the bottom end of the so i l core F i g . 4 . 3 Cross sec t i on s 1 4 3 mm 1 n d i ameter of a so i l core . The b lue co lour i nd i cates the dominant pathways for saturated f l ow and the p i nk co l our i nd i cates the dominant pathways for unsaturated f l ow at - 200 mm pres sure potent ia l . W = worm channe l , R = root channe l , C = p l anar crack . (b ) 59 . (a ) 20 mm b e l ow the so i l surface .,_.....;;;2...::.0_-11 mm 20 mm above the bottom end of the soi l core 20 F i g . 4 . 4 mm Ver t i ca l sec t i ons of a so i l core . The b l ue co l our i nd i cates the dominant pathways of saturated f l ow and the p i nk co l our i nd i cates the dominant pathways of unsaturated f l ow at - 200 mm pressure potent i a l . W = worm channe l , R = root channe l , C = p l anar crack . 60 . they were connected with other somewhat smal l er conduct ing channe l s , p l anar cracks and/or roots wh i ch d i d reach the surface . The s i ze o f worm channe l s vari ed ; the l argest s i ze observed near t h e so i l surface was 3 nun and near the bottom of the core was 2 mm . On ly a few worm channe l s near the bottom end o f the so i l core were b l ue - s ta ined . I f a s i ng l e v ert i c a l cy l i. IH.l r i ca l channe l of 1 mm in d i ameter went from the so i l surface to the bot tom of the soi l core , the hydrau l i c conduct i v i ty due to that channe l a l one ca l cu l ated from the l lagen - Po i seu i l l e equat i on (Append i x D ) wou ld have been 1 0 t imes greater than the va l ue measured , so there were obvi ous l y no c ontinuous vert ica l worm channe l s through the core s . Some conduct i ng channe l s were cut off by the core ? wal l , caus ing b l ue s ta ining on the s i de o f the soi l core , but there was no ind i cat ion of so l u t i on l eaking down between the so i l and the corer . Due to the very strong adsorpt i on of methy l ene b l ue by soi l , the co l our observed on l y shows the pathways o f h igh l y preferent ia l so lute movement . On l y those channe l s l arge enough for the v i scous f l ow of s o l ute in them to be greater than the tran s i ent d i ffus i on into the surround i ng soi l wi l l be stained to any depth . The pathways of phosphorus movement , whi ch i s a l so very s trong ly adsorbed , wou l d be expec ted to be simi l ar to methy l ene b l ue . The movement of non - adsorbed i ons such as ch l oride , wou l d be expected to oc cur through a l arger soi l vo l ume than that indicated by methyl ene b lue staining . 4 . 4 . '2 Unsat urated F l ow I :xpe r i mcnt Breakthrough data for brom i de from the unsaturated so i l cores at - 200 mm pres sure potent i a l are shown in F i g . 4 . 2 , a l ong w i th the s aturated f l ow data d i scussed above . Note that pore vol umes were determined from the l i qui d - fi l l ed pore vo l umes . The unsaturated hydrau l i c conduct i v i t i e s and vo l umetric water contents for both cores are g i ven in Tab l e 4 . 1 . The - 200 mm pres sure potent ia l in the soi l cores imp l i es that cy l i ndri c a l channe l s of minimum d iameter 0 . 1 5 mm , and p l anar crack s o f minimum width 0 . 0 7 mm , wou l d be air- fi l l ed and so ine ffect i ve in conduct i ng water or s o l utes . The app l i ed bromide s o l ut i on wou l d have to move through pores of sma l l er s i z e . Scatter ( 1 9 78 ) , as suming a very i dea l i zed soi l pore geometry , conc luded that so lute movement wou l d not be pre ferent i a l in so i l when the pressure 6 1 . potent i a l i s l e s s than - 2 00 mm ( - 0 . 02 bar) . The experimenta l data are i n fact s imi l ar t o the c l as s i cal "S - shaped"breakthrough curve s , o ften reported in the l i terature ( e . g . N i e l s en and B i ggar , 1 9 6 2 ) and a re qu i t e d i f fe rent to the s a t u rateJ breakthrough curves for the s ame s o i l core s . llowever , the b rom i de appeared in the e ff luent e ar l ier than expected from " c l a s s i c a l theory" , the re l at ive concentrat i on reach i ng 0 . :1 a f t e r 0 . 8 and 0 . 7 pore vo l ume had f l owed th rough the d up l i c a t e cores , rather than a ft e r one pore vo lume . The pos s ib l e rea son s fo r th i s arc d i s cu s s e d l a t e r . Rhodam i ne B dye was used to mark the preferenti a l p athway s i n t h e unsaturated s o i l cores . F i g . 4 . 3 shows the p i nk co l our i s concentrated i n certa in areas o f the s o i l c ore cros s - sec t i on , sugges t i ng s ome preferent i a l movement . l lowcvcr the break through data indi cate movement o f the bromide s o l ut i on t ook p l ace over a much w ider area than that s howing the p ink c o l our of rhodamine B dye , w h i ch i s a very s t rong l y ad sorbed s ub s t ance . The method o f s o l ut ion app l i cati on on the natural s o i l s urface , us i ng a man i fo l d of S point source s , may have been a fac tor induc ing the non-uni formity o f p i nk co l our near the s o i l surface . However , both F i g . 4 . 3a and F i g . 4 . 4a show the p i nk c o l our around the b l ue -wa l l ed l arger channe l s , sugge s t ing the movement of rhodamine B dye occurred to some extent in the same pathways as methy l ene b l ue . Such f l ow may perhaps be exp l ained by the e ffect o f short pore c on s t r i ct ions or pore necks near the surfac e , whi ch a l l ow s ome of the l arger pores to remai n water- fi l l ed and s o s ti l l ab l e t o conduct l iquid i n t h e uns aturated s oi l . To i nve s t i gate thi s pos s ib i l i ty , the equat i on was s o l ved for grav i ty- induced f l ow o f l iqui d through a vert i c a l cy l i ndr i ca l channe l w i th a short cons tri c t i on a t t h e t op . For examp l e , i f a con s tr i c t i on 0 . 1 5 mm i n d i ameter and 2 mm long i s connected. t o - a - 1 a 3 mm i n d i ameter channe l 1 4 8 nun l ong , a f l ow rate of 9 . 1 7 x 1 0 m s w i l l resu l t . Thi s i s 7 5 t imes greater than the f l ow conducted by a uni fo rm channe l 0 . 1 5 mm i n d i ameter , h owever both channe l s wou l d dra in at a pres sure poten t i a l o f - 200 mm . ( De ta i l s are g iven i n Appendi x D ) . Th i s t yp e o f pore geometry i s probab l y one o f the factors a l l owing s ome pre ferent i a l s o l ut e movement t o occar in un saturated s o j l , 0 . 0 0 1 . 0 P ORE VOLUME . 4 . 5 Unsaturated data for de ( ? , o ) and the (- - -, -.-) obta i ned from a convect ive - d i spers i ve mode l . 2 . 0 i c ted curv e s (J\ N 6 3 . Other po s s i b i l i t ie s t hat might cause uns aturated f l ow t o o c cur 1n the adj acent pathways to saturated f l ow , are associ a t i ons of l arge and sma l l pore s . P l an t root s have been observed t o grow and extend preferent i a l l y into the p l aces where s o i l i s l es s c ompact , such as i n the o l d worm channe l s , and p l anar cracks ( Scot t - Rus s e l l , 1 9 7 7 ) , t hus l arger and sma l l e r pre feren t i a l channe l s may tend t o o ccur together . An i on e x c lu s i on might be another p o s s ib l e reason for ear l y break? through . B romide exc l us i on has been observed by Smi th and Davi d ( 1 9 7 4 ) i n s o i l c o l umns wi th h i gh cat i on exchange capa c i t y . The so i l propert i e s o f Tokomaru s i l t l o am g iven in Appendi x A do n o t suggest t h e s oi l woul d cause much anion exc l u s i on however . 4 . 4 . 3 Computa t i on s 4 . 4 . 3 . 1 Convect ive - d i spersive mode l Unsaturated f l ow o f bromi de i n the und i sturbed so i l cores was f i tted t o the convect ive- d i spers i ve equat i on ( 1 . 1 ) u s ing the method proposed by Rose and Pa s s i oura ( 1 97 1 a) . Thi s equat i on as sumes un i form f l ow oc curring through the so i l and that the average p ore -water ve l o c i ty represen t s t h e fl ow ve l oc i ty i n the porous med ium . An adj ustment was made t o ac count for d isp l ac ement to t h e l e ft o f the breakthrough curve s , a s d iscussed ear l i er i n S e c t i o n 2 . 4 . 4 . The resu l t i n g pred icted curve s shown i n F i g . 4 . 5 are i n approximate agreement w i th the experimenta l data . The Brenner numbers obtained from s o l ving equat ion ( 7 ) g iven by Rose and Pas s i oura ( 19 7 1 ) were 1 . 6 and 7 . 4 for core I and I I respe c tive l y . However , thei r procedure i s s tr i c t l y val id on l y for Brenner numbers ran ging from 1 6 - 64 0 . The dev iat ion from a strai ght l i ne when the re l at i ve concentrat ion of the e ff l uent was p lo t ted against l n ( V/V ) on probab i l i t y paper was indicated b y corre l a -o t i on coeffi c i ents o f 0 . 9 7 3 and 0 . 99 7 for core I and core I I respec t i ve l y . Such dev iat ion s , part i cu l ar ly for core I , may indi cate n on -uni form f l ow occurring in the s o i l core . The exper imenta l resul t s d i s cussed e ar l i er in Sect i on 4 . 4 . 2 suggested some preferent i a l f l ow and/dr a n i on exclus i on o ccurred duri ng d i s p l acement t hrough t he s o i l core s . 64 . The ca l c u l ated d i sper s i on coeffi c i ents for both cores are g iven in Tab l e 4 . 1 . The s e values , however , are s omewhat in error , for the reasons d i s cussed above . 4 . 4 . 3 . 2 V i s cous f l ow w i th l atera l d i ffusi on mode l The c omputat i on method deve l oped by Scotter ( 19 7 8 ) for pred ict ing the movement of so lute through un i form vert i c a l c y l indri c a l chann e l s was a l so emp l oyed . H i s theory and the CSMP c omputer programme use d , are g iven in detai l i n Append i x C . The two channe l d i ameters a s sumed were 0 . 1 5 and 0 . 1 mm, wh i ch wou l d be drained at - 200 and - 30 0 mm pressurE? potent i a l respe ct ive l y . The mo lecu lar d i ffus i on coeffi c i en t o f brom ide in s o il was takf n as 7 . 5 x 1 0- 1 0 m2 s - 1 ( Robinson and Stokes , 1 9 5 9 ) a s suming a tortuos i ty- transmi s s i on factor o f 0 . 4 ( Scotter , 1 9 7 8 ) . The number o f channe l s , and the spac i n g between them needed t o g ive the measured uns aturated hydraul i c conduct iv i ty va lue s , are g iven i n Tab l e 4 . 2 . A uni form channe l spac ing i s as sumed . The computed breakthrough curves are shown toge ther w i th the experimen t a l data in F i g . 4 . 2 . Tab l e 4 . 2 Channe l s i ze s a s sumed and the spac ing between them . Hydrau l i c Channe l Spacing between Number o f conduct iv i ty d i ameter channe l s ch ann e ls (m s - 1 ) (mm) (mm) /m2 1 . 56 X 1 0 - 7 0 . 1 7 . 0 6465 0 . 1 5 ? 1 5 . 8 1 2 7 7 1 . 0 3 X 1 0- 7 0 . 1 8 . 6 4 2 7 1 0 . 1 5 1 9 . 3 844 The actua l geometry of the pore s respon s i b l e for most o f the f l ow i n the uns aturated so i l i s o f course much more comp l ex t han that a s sumed in the s i mp l e mode l . Thus c l ose agreement between the measured and the c a l cu l at ed break through curve s c annot be expected . However , the computed curves do show how s en s i t ive d i spers i on is t o sma l l d i fferences in the s i ze of the conduct ing pores i n s o i l w i th the s ame hydrau1 i c conduct i vi t y . Thi s makes the d i fferences between the re su l t s for the dup l i cate core s more unders t andab l e . 65 . For approx imate ly the f i rst pore vo lume , the computed curves enve lope the exper imental curves , but after that the experimental re lat i ve concentrat i on va lues tend to be h i gher than the computed ones . Thi s cou l d be due in part at l east to the non-un i form channe l d i ameter referred to ear l i er and the random , or even c l umped , d i stribut i on of the l arger conducting channe l s in the soi l , rather than the regu l ar shaped channe l s and un i form d i s t r i but ion assumed in the computat i ons . The resu l t s o f a random or c l umped d i s tri but i on wou ld be for certain sect i ons o f the cores to more quick l y approach concentration equi l i brium w ith the perco lat ing so lut i on , caus ing h i gher re l at i ve concentrat i on va lues in the e ff luent at intermed iate t imes than wou ld be the case w i th a un i form spac ing . 4 . 5 GENERAL D I SCUSS ION The experimental set-up used in th i s s tudy a l l owed saturated and unsaturated misc i b l e d i sp l acement experiments to be conducted on the same core . Pub l i shed l aboratory stud ies of unsaturateb mi s c i b l e d i sp l acement have usual l y i nvo l ved p lacing d i sturbed soi l , or i n a few cases a soi l core , between two porous p lates . The prob l ems o f contact between the so i l and the p l ates and di spersi on w i t h i n the p l ates then ari s e . Here the use of grav i ty- induced f low , w i th no porous materi a l on top of the core , and ny l on mesh at the base , v irtual l y e l iminated these prob l ems . A l so the experimenta l set-up and the samp l ing procedure , minimi zed d i sturbance of pore cont inuity at both ends o f the core . The precaut i ons taken would seem to be necessary i f rea l i s t i c breakthrough data are to be obtained , part i cu l ar l y i f any preferent ia l f l ow occurs in so i l . 4 . 6 CONCLUS I ONS 1 . Under saturated cond i t i on s , both ch loride and phosphorus moved preferen t ia l l y through 2 . 4 l i tre soi l cores from the A hori zon of a Fragiaqua l f under permanent pas ture . 2 . Methy l ene b lue dye , wh ich was used to mark the pathways o f saturated f low , i ndicated the movement predominant l y occurred in the l arger soi l pores cons i st ing of worm channe l s , decayed root channe l s , the spaces between roots and the so i l matr i x , and natura l fracture 66 . p l ane s . Ne tworks o f i nt erconnec ted pores o f d i fferent t yp e s s eemed to p rovi de the fl ow paths , rathe r than f l ow down a s i n g l e chann e l . 3 . Under un saturated c ond i t i on s , when - 0 . 02 bar ( - 200 mm o f water) p res sure potent i al was main ta ined w i th i n the s o i l cores during m i s c i b l e d i sp l acement , b romi de movement was more un i fo rm , the break through curves t end i n g t oward the c l as s i c a l "S- shape" . However , the f l ow was s t i l l t o s ome ex t ent pre fe rent i a l . 4 . 1be pathways o f pre ferent i a l movement in the unsaturated s o i l were v i s ua l l y i nd i cated by rhodam ine B dye . The rhodamine B dye s t a i n i n g around t h e l arge b l ue -wa l l ed channe l s i n d i cated t h e prefe rent i a l movement t ended to occur a l ong the s ame pathways as saturated f l ow . I t i s l i ke l y short con s tr i c t i ons near the top o f l arger channe l s keeps s ome channe l s from dra i n i ng , and s o they rema i n a s p referen t i a l conduct ing pathways i n un saturated so i l . 5 . 1be co l our observed f rom bot l1 dyes on l y shows the preferen t i a l pathways o f water and s o lu te movement , due to t h e s t rong adsorp t i on o f dye used b y t h e s o i l . 6 . Computed breakthrough curves for so i l containing un i form l y spaced , verti c a l , cy l i ndri ca l channe l s 0 . 1 and 0 . 1 5 mm i n d i ameter enve l oped the e xperi menta l un saturated m i sc i b l e d i s p l a cement data up t o near ly one pore v o l ume , and showed how sen s i t i ve the breakthrough curve s are to sma l l changes in channe l s i ze w i th in t h i s s i ze range . C l li\!'TER 5 WAl'l : R 1\N i i 1\N l ON MOVUvi i : NT TO ?10LF DFA I NS 6 8 . 5 . 1 I NTRODUCTI ON Mo l e - t i l e drainage systems are w ide l y used t o remove excess wat er from the root zone i n New Zea l and ( Hudson e t a l . , 1 96 2 ; Bow l er , 1 9 8 0 ) . I n the poor l y drained Tokomaru s i l t l oam , where a compact s i l t l o am fragipan ex i s t s be l ow 780 mm depth , mo l e - t i l e drains are o ft en i n s tal led i n the c l ay l oam B hor i zon t o a l l eviate water l ogging prob l ems . Mo l e- t i l e drainage systems have o ft en been found to remain e ffect ive over a fai r l y l ong peri od of t ime . Under favourab l e cond i t i on s , they may remain operat ive for 1 0 to 25 years under pasture ( Bowl er , 1 980 ) . Mo l e drainage can great l y enhance the hydrau l i c conduct iv i ty o f the B hori zon , as Scott er et a l . ( 1 9 9a) showed . They reported peak drainage f l ow s equivalent t o 25 mm/day water movement t o t h e s oi l , whi l e i n s i tu saturated hydrau l i c conduct i vi t y measuremen? i n the B hori z on at ad j acent undrained s i tes ranged from 0 . 02 t o 0 . 6 mm/day . Unl ik e the f l ow net o f drainage water t o the drain p i pe i n a s and t ank or i n homogeneous soi l , i n whi ch the s tream l ines are regu l ar and perpend i cu l ar to equipoten t i a l l ines ( K irkham and Powers , 1 9 7 2 ) , the f l ow of water to mo l e drains is be l i eved t o occur mos t l y a l ong the cracks c aused by the pas s age of the mo l e p l ough ( Hudson e t a l . , 1 96 2 ; Bow l er , 1 9 8 0 ) . Bow ler ( 1 980 ) has de scribed the act ion o f the mo l e p l ough in the s oi l . When the p l ough i s be ing pu l led through the s oi l , i t s movement is accompan i ed by heaving , which causes cracking extending to about 0 . 3 m each s ide o f the p l ough and may reach the surface i n pasture l an d . The cracks created b y t h e p l ough occur l arge l y a l ong t h e natural p l anes o f weakne ss . I n the summer fo l l ow ing the pu l l ing , the cracks widen due to shrinkage , providing gaps for p l an t roo t s to penetrate deeper into the so i l profi l e . Scotter et a l . ( 1 979a ) observed a lucerne (Medi cago sat iva L . ) root , and the other fine roots , growi ng in the mo l e -drains and where the s l i t l e ft by the mo l e p l ough h ad been 1 8 months ear l ier in Tokomaru s i l t l o am . As such roo t s shrink and decay , the remaining cav i t i e s cou l d provi de pathways for rap id water movement i nto the mo l e drains . Scatt er et a l . ( 1 979a ) a l so refer bri ef l y t o dye movement stud i es i nd i cat ing preferent i a l f l ow through pathways d i rect l y above the mo l e drains , often invo lv ing s tructural cracks , and worm and root channe l s . The d i sturbance caused b y the mo l e p l ough appeared t o a l l ow roots , humus and worms to better penetrate the B hori zon above the mo l e . 69 . Mo l e- t i l e d i scharge as a percentage of rainfa l l i s greatest when the so i l above the drains is at fi e ld capaci ty or wetter prior to a rain? fa l l event . However , even i f the s o i l is i n it i a l l y qui te dry , the drains can s ti l l sometimes f l ow . Bow l er ( 1 980 ) pointed out that the accentuated s o i l cracki ng associ ated with mo l ing and summer dryi ng cou l d cause f l ow t o occur under a rai nfa l l inten s i ty o f 3 0 mm i n 8 hours , whi le an inten s i t y of 1 0 mm i n 1 0 - 1 2 hours wou l d c ause f l ow onl y when t h e soi l above the mo l e s had rewet to fi e l d capac i ty . S ign i ficant l osses of app l i ed fert i l i z er i n the mo l e- t i l e drainage from Tokomaru s i l t l o am s o i l have been reported by Turner et a l . ( 1 9 76 ) , Sharp ley and Syers ( 1 9 79a ) and G andar and Gregg ( 1 979 ) . No fixed pattern of n i trogen and phosphorus l eaching was observed by Turner e t a l . ( 1976) , but a seasonal trend i n the concentrat i on of the drainage water appeared , w i th the h i ghest concentrat ion i n the e ar l y drainage events , t hen rap i d l y decreas ing c oncentrat i on dur ing success ive f l ow events . The h i gh concen? trati ons observed in ear l y drainage even t s were probab l y due to the rapi d l eaching o f nutrients a lready present i n o r adj acent to the c onduct ing c hanne l s . However , the resu l t s quoted were obtained by interm i ttent s amp l ing of the d i s charge from 0 . 1 2 5 h a p l ot s . Nearby , Sharpl ey and Syers ( 1979a ) found mo l e - t i l e drains contributed 2 7% of the s tream f l ow i n a sma l l catchment . and 2 2 % o f the di s so l v ed i norgan i c phosphate i n the s tream f l ow . Pub l i shed research re l ated to mo l e- t i l e drainage has reported water and nutri ent y i e lds from areas of at l east 1 0 0 0 m 2 (Macgregor et a l . , 1 9 7 5 ; Rennes et a l . , 1 976 ; Turner et a l . , 1 97 6 ; Sharp l ey and Syers , 1 9 79 a ; Gandar and Gregg , 1 979 ) . However n o one appears t o h ave studied the behav iour of an i nd ividua l mo l e drain, and the s o i l it is i n , on a sma l l er s ca l e , or to have l ooked i n detai l at the actual f l ow paths i n the soi l from the surface to the mol e . To understand the l arge sca l e res? l t s , sma l l er s ca l e invest igat i on i s necessary and such an i nves t i gation i s described i n thi s Chapter . 7 0 . 5 . 2 O BJECTIVES The experiment described aimed t o s tudy the f l ow paths in a mol e? t i l e drained so i l profi l e , and the movement of s orbed and non - sorbed ani ons from the so i l surface to the drainage system . 5 . 3 MATER IALS AND METHODS The fi e l d experiment was conducted i n the area near t o where the undi sturbed soi l cores described in Chapter 4 were co l l ected ( Sec t i on 4 . 3 ) . Tokomaru s i l t l oam i s c l as s i fi ed as an Aeric Frag iaqua l f ( So i l Survey Staff , 1 9 7 5 ) and as a g l eyed ye l low - grey earth ( New Zea l and Soi l Bureau Staff, 1 96 3 ) . The so i l profi l e has been described in detai l by Po l lok ( 1 975 ) , and his descript i on is g iven in Appendi x A . At the s i t e , the so i l con s i s t s of a s i l t loam A hori zon extending to approximate l y 250 mm , a c l ay l oam B hori zon to 7 80 mm , then a v ery compact s i l t l oam fragipan C hor i zon ext ending to 1 1 0 0 mm and under lain by a somewhat l es s compact s i l t l o am (Scotter et a l . , 1 9 7 9a ) . The experimental area was growi ng ryegras s -whi t e c l over pasture and had b e en mo l e -t i l e drained approximate ly 6 years prior t o the commencement o f thi s s tudy . The mo l e drains were pu l l ed at 2 m spac i ng in the B hori zon at approximate ly 450 mm depth on a gradi ent of 1 % . The t i l e drains were insta l l ed perpendi cu l ar to the mo l es at approximate l y 7 5 0 mm depth on a minimum gradi ent of 0 . 4% . The general experimenta l s et -up for this s tudy i s i l lustrated i n F i g . 5 . 1 . A n i nf i l t rometer ring , 3 8 0 mm i n d i ameter , was insta l l e d t o approximate l y 1 00 mm depth above a mo l e l ine . The inf luent s o l u t i on was p onded in the ring , and the effluent , wh ich perco l ated through the soi l profi l e , was co l l ected from the mol e i n a p i t l ocated adj acent t o the i nfi l trometer ring . The experiment was carri ed duri ng l at e autwnn and winter when the s o i l was at ' fi e l d capaci t y ' . I t was conducted i n a s imi l ar manner t o the l aboratory experiment us ing saturated so i l co l umns . The influent s ol ut i on , containing l OO 1Jg/ml ch loride as potass i um ch loride and 40 1-!g/ml phosphorus as potas s ium dihydrogen phosphate , was ponded i n the i nfi l trometer ring a t a near l y constant head o f 20 - 25 ' mm above the so i l surface . The e ff l uent s o lu t i on was co l l ected manual l y at 0 . 5 or 7 1. P o n d e d I n f i l t r o m e t e r r i n g F i g . ') . l Mo l e . h. ch a n n e l ? ? j ? , ' ' -- - ? - - -? -"' ?. E x p e r i men t a l s e t - u p f o r t h e m i s c i b l e d i s p l a c emen t e x p e r i m e n t a b o v e t h e mo l e d r a in . The i n f i l t rome t e r r in g w a s 380 mm i n d i ame t e r a n d t h e mo l e c h anne l l o c a t e d a t 400 ?n d e p t h . 7 2 . 1 m i nute t i me i n t e rva l s unt i I approx i mate l y one pore vo l ume had p a s sed th rough the so i l above the mo l e . To p revent dra i nag e from s ource s other than that app l i ed t o the i 11 f i l t rornet e r r i ng , the experiment was carr i ed out on a f i n e day when no na tura l mo le t i l e d ra inage was occurr i ng . The so i l surround ing i n f i l t rome t c r r i ng was wet at the s tart of the exp e r iment to m i n i m i e lat e ra l f l ow . Wat e r wa s then ponded in the r i ng unt i l the i n f i l t rat i on rate and the mo l e d ra i n f l ow we re near ly e4ual . A t t h i s s t age , wat e r mov ement Wii S a s sumed to occur w i t h i n the s o i l cy l i nder be l ow the r i ng to the mo l e depth , from wh i ch the pore vo l ume was e s t imated . l?e ponded water was then rep l aced wi th t h e c h l or i d e and pho sphoru s s o l ut i on . At the end o f the experi ment , approx i mate l y one pore vo lume o f 0 . 1 % methy l ene b lue dye s o l ut i on was app l i ed t o the soi l . The s o i l was a l l owed to drai n for 2 day s , then the profi l e was e xpos ed and the dye movement pathways were observed . ;, _ ,1 i <.J:SU LTS i\ND D I SCUSS I ON Breakthrough data show i ng t lte ch l oride and phosphorus concentrat ion 1 n t he e ff luent from the mo l e channe l are pre s ented i n F i g . 5 . 2 . Both ch l or i d e and phosphoru s appeared in the mo l e e ff luent a lmo s t immedi at e ly a ft e r app l i cat i on to the so i l surface , i nd i cat i ng h i gh l y pronounced p re ferent i a l f l ow occ11rred i n the saturated s o i l above the mo l e . During Expe r iment l , wh i ch was conducted i n l a t e autumn ( 1 5 Apr i l , 1 9 7 9 ) , the s teady s t a t e i nf i l trat ion rate was 6 . 7 5 x 1 0 - 5 m s - 1 ? The e ff l uent t o i n f l uent s o l u t ion rat i o ( C / C . ) of both ani on s rose qui ck ly t o approx-e 1 i mate l y 0 . 8 a fter on l y 5 m i nutes eH 0 . 1 pore vo l umes o f the s o l ut i on had been app l i ed t o the s oi l , but subs equent l y conti nued t o r i se on l y very s l owl y . Both ch l ori de and pho spho rus appeared l ater i n abs o l ute t ime when the exper i ment was repeat ed at an adj acent s i t e at the end of w inter ( 2 8 August , 1 9 79 ) ( Experiment l l ) . 1?e i nf i l trab i l i ty observed ( 6 . 39 x 1 0 - 6 m s - 1 ) was an order of magni t ude l es s than i n Experiment I . Pre fe ren t i a l movement o f both an ions was however s t i l l very pronounced . C / C . reached 0 . 49 for ch l or i de and 0 . 2 7 for phosphorus w i t h i n 6 2 m inutes e 1 o r 0 . 1 pore v o l ume o f the app l i ca t i on o f the i on s t o the surface . I n both experimen t s , the concentra t i on o f phosphorus i ncreas ed more s l ow l y 1 . 0 0 . 8 0 . 6 - 11) 0 . 4 u 0 . 2 0 . 5 . 2 0 0 . 2 ? ? 0 . 4 6 6 0 . 6 P O RE V OLUMES 0 . 8 1 . 0 Breakthrough da ta f or s aturated soil pro f ile above the mole drain ; chlo?ide (?\ and phosphorus (&--6! for Experimen t I with inf i l t ra t ion rat e of 6 . 75 x 1 0-S m s- l , and chloride (a--.) and phosphorus (D--0) for r imen t 11 with inf i l t ra tion rate o f 6 . 6 7 x 1 0 -6 m s - 1 . The curves have b een visua l ly f i t te d to the data pc in t s . Times corresponding t o one pore volume were 5 1 minutes and 9 . 5 hours fo r :- iment I and I I , respe c t ively . C /C . i s the r a t io o f e f f luen t to influent c oncen tra t ion . e 1 "-] ? 74 . than ch l oride . The curves s hown are s imi l ar to those obtained from the soi l columns containing art i fi c ia l channel s or p l anar cracks (Sect ion 4 . 4 ) , the saturated so i l cores ( Sect ion 4 . 4 . 1 ) , and those pred i cted by the s imp l e preferent ia l f low theory proposed by Scatter ( 1 978 ) (Appendix C . l ) . I t was not pos s i b l e to quant i fy the s i ze and number of the conduct i ng channe l s from the breakthrough data . However , i t can be i nferred that re l at ive l y l arge cont inuous channe l s were respons ib le for most of the f low i n experiment I , whi l e smal ler or fewer l arge channe l s were effect ive during experiment I I , whi ch showed l e s s pronounced preferent ia l f low . The app l i cati on of methy l ene b lue to the soi l after experiment I I showed that the dye so lut ion moved i nto the soi l through l arge pores , such as gap s between roots and the surroundi ng soi l , decayed root channel s , worm channe l s , soi l cracks and the i nc ip ient fracture p l anes between structura l uni t s . Gras s root channe l s and worm channe l s appeared to be the maj or pathways for water i ntake into the soi l profi l e . Re l at ive l y fewer b lue - stained root s and worm channe l s were observed i n the subsurface hori zon than i n the surface hori zon . F i g . 5 . 3 shows some of the preferent i a l pathways observed i n the soi l profi l e . Some conduc? t i ng worm channe l s extended from the soi l surface near l y to mo l e depth . Gras s roots were very extensive part i cu lar ly near the so i l surface , and extended down to mo l e depth and beyond . Some root s were found penetrating the mo l e cav i ty i tse l f ( F i g . 5 . 4 ) , and a l s o i n o l d worm channe l s . The methyl ene b l ue -staining tended to be concentrated i n the area d i rect l y above the mo le where the soi l had been d i sturbed a t the t ime the mo l e was pul l ed ( F i gs . 5 . 4a and b ) . Thi s was a l so obs erved in an ear l i er p i lot s tudy described by Scatter e t a l . ( 1 979) . Stain ing of l arge fracture p l anes was a l so observed qui te commonl y . F i g . S . Sa shows such a b lue stained natural fracture p l ane with assoc iated p l ant roots and worm channe l s . The presence of preferent ia l root growth between soi l s tructura l uni t s woul d prov i de con t inuous channe l s for rapid water movement as d i s cussed ear l i er . Po l l ok ( 1 975 ) and Soi l Survey Staff ( 1975 ) observed extens ive root growth between the s tructural soi l uni t s i n the fragipan , and sugges ted a l s o that the cavi t i es l e ft by these (b ) Conduct ing worm channe l s and root channe l s . 5 0 mm 7 5 . (a) Preferen t i al pathways i n the so i l profi l e d i rect l y above the mo l e . mm F i g . 5 . 3 Preferen t ia l pathways observed i n the so i l prof i l e above the mo l e dra i n , as indi cated by methy l ene b lue s tain ing . W = worm channe l , R = root channe l . F i g . 5 . 4 76 . 20 mm Mo l e dra in at 400 mm profi l e depth i n Tokomaru s i l t l oam so i l i n Da iry F arm No . 4 near Mas s ey Univers ity . Gras s root s penetrat ing i nto the m o l e were common l y obs erved . (b ) Cros s - sec t ion v i ew of conduc t i ng worm channe l s and the i nterconnected crack s . 20 .,_ ___ --tmm ?? ' ? . ?? ?- ? ? ? ?? ?' . ' ./' . ?j' ? ..... ? . f'.. ';, . ... ( -, ' ? '< : ?, I I ? ?I ' ' "" ? . . ' .. 7 7 . (a ) Verti ca l v i ew of conduct i ng channe l s i n a natural fracture p l ane . . "' 11 .. . 20 . - mm F ig . 5 . 5 Preferent i a l pathways observed i n the so i l profi l e . W = worm channe l , R = root channe l , C = p l anar crack . 78 . roots wou ld a l l ow some water to move through the fragipan . Preferentia l f low through a combination of conducting worm channe l s and cracks i s s hown in F i g . S . Sb . The conducting pores in the soi l profi l e were usua l ly a comb inat ion of grass root channe l s with worm channe l s ( F igs . 5 . 3a and b ) , worm channe l s with crack s ( Pi g . S . Sb ) or a combinat ion o f root channe l s , worm channe ls and cracks ( F i g . S . Sa ) . The s i ze and frequency of these channe ls presum? ab l y vary marked l y with season , due to root and worm ac t ivity and struc? tural cracking i nduced by drying (Corker , 1 9 7 7 ; Bow ler , 1980) . The l argest s i ze of the conducting worm channe l observed during thi s experi ? ment was approximate l y 3 mm in di ameter . The width of soi l cracks was not measured , but they were probab ly wider in autumn when experiment I was conducted , due to shrinkage during summer of c l ay mineral s and organic mat ter . The resul ts of thi s experiment wou l d? not be necessari l y app l y under natura l rai nfa l l . The hydrau l i c conductivities observed ( 243 and 23 mm/hour) were cons iderab l y h igher than common rainfa l l intens i t i es . Thus , w i th effective mo l e -t i l e drainage , surface ponding i s onl y observed at this s ite after sus tained , heavy rain . However , the resu l t s support and exp l ain the observation made by Macgregor and Gregg on the s ame s i t e , and reported b y Kanchanasut e t a l . ( 1978 ) . MacG regor found urea and col i forms in the mo l e - t i le drainage less than 2 hours after the spray appl i cat i on of dairy- shed waste at 10 mm hr- 1 in early spring , when the soi l was at fi e l d capac ity . Gregg found that when superphosphate was app l i ed to the soi l surface , the drainage d ischarge caused by 10 mm of rainfa l l the fo l l owing night showed a 10 fo ld increase ( from 0 . 1 t o 1 . 0 ?g/ml ) i n dissol ved i norganic phosphate over the preced ing drainage event . 5 . 5 PRACT ICAL IMPLICATIONS The resu l t s presented i n this chapter suggest a number of prac t i ca l imp l ications . F irst l y , i f any effluent i s app l ied to mo l e - ti l e drained so i l s under ponded cond i t i ons , virtual l y untreated effluent can appear in the drainage system within minutes . Whether or not the po l lutants in the effluent are capab le of absorption by, or reacti on with , the soi l matrix is of l i tt l e importance , as the effluent ' sees ' very l i t t l e o f the soi l i n i t s travel a l ong preferent ia l pathways . 79 . Second l y , the drainage from heavy rainfa l l inducing pondi ng may resu l t i n re l at i ve l y l i t t l e l eaching from the so i l matrix , as the preferenti a l f l ow pattern w i l l a l l ow for l i t t l e i nterac t i on w i th the soi l s o lut i on . l n genera l , drainage from l i ght or intermi t tent rain c ou l d be expected to l each more nutri ent s than the s ame amount of d rainage from heavy rai n . Thi rd l y , as b i o l og i ca l l y i nduced channe l s appear respon s i b l e for mos t of the flow to the mo l e drains , i t wou l d seem important to maintain good growth and a h i gh worm popu l at i on for mo l e s to remai n effective . F ewer roots and worm channe l s , as we l l as comp acti on due to whee l traffi c , cou l d be respons i b l e for any observed deter i orat ion i n s o i l permeab i l i t y and mo l e drain performance when a change i n l an d u s e from pasture t o croppi ng occurs . Converse l y , improved p asture growth , however i nduced , i s l ik e l y to enhance mo l e drain performance . 5 . 6 CONCLUS IONS ( l ) The observed breakthrough d ata for chl or i de and phosphorus indi cate that most o f the s o lute movement occurred preferent i a l l y through the saturated soi l p rofi l e to the mo l e drains . More pronounced prefer ? ent i a l f l ow was observed i n l ate autumn than i n l ate winter . ( 2 ) Dye tracer s tudies i nd i cated the pathways whi ch conducted water preferent i a l l y were mos t ly d irec t ly above the mo l e drain , c lo se t o where the mo l e b l ade had been ear l i e r . The e ffec t i ve c onduct i ng pathways i n c lude worm channe l s , root channe l s , and so i l crack s . Any parti cu l ar preferent i a l f l ow path from the so i l surface to the mo l e was usua l l y a c omb inati on of at l east two of these pathways . ( 3 ) Grass root s and worm channe l s were the most evi dent pathways near the so i l surface , whi l e soi l cracks were the mos t evident conduct ing pathways i n the B hori z on . Cl lAPTU? 6 BROfi H D1: LLACll 1 N(; UNDER F I E LD COND ITIONS 8 1 . 6 . 1 I NTRODUCT ION Resu l t s from Chapt er 5 i nd i cate that when water was pondcd on pasture , it moved pre feren t i a l l y through the so i l b e l ow and reached the mo l e drains in minutes , w i thout d i sp l acing mos t of the s o l ut i on i n the so i l matr ix . Therefore , l i t t l e l eaching o f surface app l i ed fert i l i z er from that so i l wou l d be expected , whi l e app l i ed e ffluent wou l d move s t ra i ght through . l lowever , w i th pas ture and an e ffec t ive mo l e drain system , ponding i s un l ik e l y under natura l rainfa l l cond i t i ons . I n contrast , loca l i sed ponding may o ccur under n atural rainfa l l on a cu l t ivated so i l , due to s l ower saturated hydrau l i c conduc t i vi ty i nduce d by repeated t i l l age , whi ch breaks down s o i l structure and b i opores ( such as worm channe l s and root chann e l s ) near the surface and a l so may i nduce a compacted l ayer or p l ough pan . Thi s chapter l ooks at leaching o f bromide under p onded cond i t i on s i n mol e -drained s o i l profi l e s under pas ture and crops , and under natural rainfa l l in so i l under pasture . F i g . 6 . 1 i l l ustrates vari ous p os s i b l e l eaching patterns for a s lug o f so l ut e in it i a l l y at the soi 1 surface . Curve ( b ) shows "pi s t on f low" when there i s on l y convect i on , with neg l i gi b l e d i spers i on . I f d ispers i on occurs , " c l as s i c a l " theory predi c t s a d i stribut i on shown as curve ( d ) on the F i gure , wh i ch is the s o lut i on o f equat i on ( 1 . 5 ) . However severa l stud i e s have shown that under natura l rai nfa l l the l eaching o f surface app l i ed so lutes can resu l t i n an asymmetri c a l vert i ca l d i s tribution o f s o l utes w i t h a " leading ta i l " ( O l sen e t a l . , 1 9 7 0 ; Bosewe l l and Anderson , 1 9 70 ; Cas s e l , 1 9 7 1 ; W i l d , 1 9 7 2 ; Ki sse l et a l . , 1 9 7 3 ; Burn s , 1 9 7 4 ; W i l d and Bab i k e r , 1 9 7 6 ; Cameron et a l . , 1 9 79 ) ( Fi g . 6 . 1 , curve ( c ) ) . However l eaching surface appl i ed s o lutes i n packed so i l columns o f sand o r d i sturbed s o i l does resul t i n a n ormal d i stributi on , as wou l d be expected from convect ive -d i spers ive th?ory ( B i ggar and N i e l sen , 1 96 2 ; Lev in , 1 964 ; Corey et a l . , 1 9 6 7 ; Evans and Levi n , 1 969 ; Ghuma et a l . , 1 9 7 5 ; Kirda et a l . , 1 9 7 3 ) ( F i g . 6 . 1 , curve ( d) ) . The peaks of concentrat ion in natural soi l profi l e s have o ften been observed at sha l l ower depths than expected . F or examp l e , Wi l d and Babiker ( 19 76 ) observed the peak o f n i t rate and ch l oride concentrat i on at approximate l y ha l f the expected depth , w i th ta i l ing to at l east 800 mm depth after 50 mm o f water was app l i ed ( F i g . 6 . 1 , curve ( c ) ) . () 200 (c) 600 HOO 8 2 . 1 \EL!\T I VE C: l l l ,O l\ I lll': CONCENTRi\T l ON 0 . 6 O . H 1. 0 ( il ) - - - - - - - - - - - ( b ) ( a ) I n it i a l d i s t ri b u t Lon , ( b ) P i s t on f l ow c l i s p l ac e rnen t , ( c ) Ob se r ve d l i l ? l d c: l d o r i de d i s p L: tc: erne n t und c: r i n t e rm i t t e n t i r r i ga t i on o f b >:Ll p:::: 0 ...-1 ? "'"' z H ;2 JULY AUGUST S EPTEHBER OCTOBER 300 ' S econd S ampl 200 100 (46 First Sampl ingl mm excess \. rain) ET ( 1 82 mm excess rain) o l ? , ' , , ' ' 1 , I 0 20 40 60 8 0 lOO 1 20 DAYS Fig . 6 . 2 Daily rain fall and evapo tran sp ira t ion ( ET ) data dur experimen t in 1 9 7 9 . CO '-.] 88 . fo r about 3 hours , w i th occas i ona l l y s t i rr ing . A brom i de se l ect ive e l ect rode (Orion Solid state bromide e l ectrode) and reference e l ectrode , connected to a meter ( Ori on Research M i croprocessor I onal yzer 9 0 1 ) were used to measure brom i de co1 1Centrat i on . The e l ec trode s were i mmersed i n to the c l ear supernatant above the s o i l suspen s i on . The reading was recorded when the number d i s p l ayed had s t ab i l i s ed , usua l l y after 1 m inutes . A s im i l ar procedure for brom i de extrac t i on and measurement i n s o i l was found sat i s factor i l y by Abda l l a and Lear ( 1 9 75 ) . 6 . 3 . 4 . 2 P l ant t i s sues To determine bromide i n p l ant t i s sues , p l ant s ampl es were oven? dr i ed at l OSC for 1 5 hours , then ground fine l y . One g ram of dr i ed p l an t s amp l e was m ixed w i th 50 ml d i s t i l l ed water conta in ing 2 % of I SA s o l ut i on in a p l as t i c contai ner w i th a t i gh t l y f i t t ed l id . The s amp le s were thorough l y shaken and then p l aced in an u l trason i c bath for 10 m inutes . The supernatant was then fi l t ered and brom i de determin ? at i on was made on the fi l t rate u s i ng the brom ide s e l ec t ive e l ec trode . The methods o f extrac t i on and measurement were , i n genera l , the s ame as those proposed by Abda l l a and Lear ( 1 9 75 ) . 6 . 3 . 5 1 Rainfa l l Rhodamine B dye was emp l oyed t o ma rk the pathways o f preferent i a l movement under natural rainfa l l cond i t ion s , e speci a l l y near the s o i l s urface . Two r ing s , 1 4 3 mm i n d i ameter , were pushed i n t o the s o i l t o 1 5 0 mm depth and then 5 mm o f 1 % rhodamine B dye i n aqueous s o l u t i on was ponded on the s o i l surface . La ter , after recei v i ng 65 mm of exce s s rainfa l l , t h e r i n g s and so i l enc l osed were removed and brought t o t h e l aboratory . The s o i l cores were secti oned ver t ic a l l y at f i e l d moi sture c ontent and the dye patt e rn observed . Th i s experiment was conduc ted during Novembe r and Decemb e r when the s o i l was s omewhat dri e r than fi e l d capac i t y nos :? o r the t ime . t?los t o f the dye movement probab l y occurred on two i nd i v i dual days when 2 7 mm and 58 mm of rain fe l l . However , pond ing probab l y d i d no t occur , due t o the re l at ive l y dry ant ecedent water conten t s . 89 . 6 . 3 . 6 So i l Hu l k Den s i t y S i x s o i l cores , S O mm l ong and SO mm i n d i ame t e r we re t ak e n from e a c h p ro f i l e d e p t h t o m e a s u r e bu l k d e n s i t y , us i n g t ec h n i q ues d e s c r i bed by Lov e day ( 1 9 74 ) . Bu l k den s i t i es a t t h e t wo l o cat i on s a rc shown 1 11 F i g . 6 . ? . 6 . 4 RESU LTS AND D I SCUSS I ON 6 . 4 . 1 Con t i nuous Pond i ng Expe r i ment As a l ready i nd i c a ted , t h i s ex p e r i men t was condu c t e d ; 1 t two l o c a t i on s , one unde r p e rmanent p a s t u re a n d t h e o t h e r under cu l t i v a t i o n . F i g . 6 . 4 s hows t h e r e s u l t s o f t h ree re1 1 l i ca te s ( a , b , and c ) o bt a i ned f rom t h e l oca t i on u n d e r pas t u re a n d F i g . 6 . S s h o w s t h e r e su l t s o f 2 rep l i c a t e s ( a and b ) from t h e l ocat i on unde r cu l t i va t i on . Dat a p r e s e n t e d i n t h e f i g u r e s a rc t he mean v a l ue s , t l 1 e i nd i v i d ?a l b rom i d e concent rat i on mea s u r e ? men t s a r e g i ven i n Appen d i x E . 6 . 4 . 1 . 1 P re - l each i n g measuremen t The b rom i de founJ i n samp l e s t ak e n a p p rox i m at e l y 3 hou r s a ft e r app l i c at i on t o t h e so i l surface had moved d e e p e r i nt o t h e s o i l p r o f i l e t ha n e x p e c t ed . The 2 mm o f b rom i d e s o l u t i on app l i ed wou l d h a v e been un i form l y d i s t r i bu t ed w i t h i n the top 4 mm i f t h e moveme nt had occurred a s p i s t on f l ow , a s s um i ng a s at u r a t e d vo l ume t r i c w a t e r c on t e n t o f O . S . Wh i l e t he concent rat i on o f brom i de observeJ was h i gh e s t i n t h e t o p 1 0 mm , a s i gn i f i c a n t amoun t o f brom i de was found be l ow t h i s d e pt h . However near l y a l l of the b rom i d e rema i ned in t he t o p 50 mm , a s s h own i n F i g . 6 . 4 a n d ? i g , 6 . 5 . Tab l e 6 . 1 , g i v i n g t h e recovery p e r c e n t a g e s o f a pp l i ed brom i de , i n d i c a t e s on l y sma l l amoun t s o f b rom i d e b e i ng recovered b e l ow 50 m d e p t h . l lowevc r , re l a t i v e l y l ow recovery pe rc e n t a g e s ( av e rag i ng 46% ) were o b s erved a t b o t h l oc a t i on s . The p o s s i b l e rea s o n s fo r t h i s w i l l b e d i s c u s sed l a t e r i n Se c t i on 6 . 4 . 5 . 0 . 0 0 . 1 0 . 2 r:s p_, w 0 1-W c-l H f.'-< 0 . 3 0 ? p_, 0 . 4 0 . 5 BULK DEN S l TY !! ( ) () HOO __,._ Und e r p a s t u r e -3 (kg m ; 1 200 ? \ \ \ I I ? 90 . l 600 -?- Under c u l t i v a t i o n . 6 . 3 Soil bulk density distribut ion in the soil p ro f i l e s u n d e r p a s t u r e a n d c u l t iv a t i o n ( m e an s a n d s t an d a r d d ev i a t i o n s o f 6 r e p l i c a t e s a t e a c h d e p th ) . 0 l OO .-' I I I I 200 \ .J I .J HROMIDE CONCENTRAT LON ( Pg ml- l ) 200 4 00 600 ( a ) (b ) ( c ) P r e - l e ac h i n g Po s t - l e a c h in g 91 . B r o m i d e d i s t r i b u t i o n i n t h e soil pro?ilc b e f o ? e a n d a f t e r o o n d i n g 50 mm o f wa t e r on the soil surface in c o n t i n u o u s p a s t u r e g row i n g a r e a . The c o n c e n t r a t i o n o f b r om i d e b e f o r e po n d i n g w a s t h e a v e r a ge o f 5 c o r e s a mp l e s , a n d a f t e r p o n d i n g w a s t h e av e r a g e o f 1 3 c o r e s a mp l e s . ? r:L w p w c-l H ? 0 p::; r:L 9 2 , BROMIDE CONCENTRATION ( )J g /ml ) 200 400 600 0 r .J ? - - - - - - - - - - - - - J I _ _ _ _ _ _ ..J ( a ) l OO I r j ) I L I ? l I 1 2 00 i 300 I ) I ,J I I I I 0 l OO ? ' : i r ?--?r_.! _ _ _ _ _ J . ; _ _ r_ _ _ _ jj r - -? I ? 0 . 06 mm d i ameter) 8 :::: b !).. "' c ? :-.. c "' !).. . 6 . 6 VOU'METRH cu;\ ( m m 0 04 2 0 . !. 0 . 6 o . o s 0 . 10 0 . 1 5 ? j . I ; r' r \ ? ?" o . zo ' ' ' ' 0 . 2 5 8 r' r' t c I " ' ' ' I = 8 mrr. h r 1 - 1 I ' mrr; h r- l I = 4 mr.: h r ' 2 . 0 . 30 0 . 2 0 . 1. 0 . 6 0 . 2 0 . 4 0 . 6 0 I 0 . O 'J 0 . 1 0 t tt 1 .? __ Pos t - 1 Pre-leaching 0 . 15 1 0 . 20 0 . 25 d ? t e - 1 - 1 0 . 30 1 l ? 2 mm hr ? I ? 2 . 5 mm hr Volume t r i c wa t e r c on ten t dis t ribu t ion in the s o i l p ro f i l e s b e f ore and a f te r and poro s i ty . I in f i l tra t i on r a t e h r ) i n the f ir s t hour a f te r pa s t ure ( a , b , and c ) an d un d er c ul t iva t i on ( d and g . Under (.0 0'? 9 7 . drain ing a t " fi e l d capac i t y " , whi l e the s o i l be low i s unsaturated . In fac t , a perched water t ab l e was obs erved after pond ing i n one o f the rep l icat e s . Th i s supports the a rgument for the thro t t l ing e ffec t of the c ompacted layer propos ed a bove . Surpri s i ng l y , the infi l t ra t i on ra tes during the fir s t hour of p onding into the cu l t i vated and non - cu l t i vated s o i l were fair l y s imi l ar . Thi s was perhaps due t o a few very l arge pores or crack s , induced by the cul t i va t i on , compen sat i ng fo r t he l owe r vo l ume o f macropore s in the cu l t i vated soi l . The 5 0 mm o f ponded water soaked i n i n l e s s than 20 hours on two of the pas t ure p l ot s , but took 2 days to s oak into the t h ird pasture p l ot and the two cu l t i vated p l ot s . The b romide concentrat ion remain ing i n the t op 20 mm o f s o i l was more than 240 ?g/ml i n the cu l t i vated area , whi l e appoximate ly 1 5 0 ?g/ml was found i n the pasture area . The finding of un l e ached b romide near the soi l s urface i nd i cate s pre feren t i a l movement o ccurred dur i ng l eaching . Leaching o f b romi de was l e s s i n the c u l t i vated are a , where the crop row spacing was 1 5 0 mm , than in the p as ture area where past ure p l ants were more un i fo rm l y d i s t r i buted on the so i l surface . The c onduc t i ng pores as sociated w i t h p l an t root s in t he cul ti vated area woul d have been w ide l y spaced and a l s o c l umped w i th in the rows . B romide which was l ocated between the crop rows wou l d t end t o move re l at ive l y s low l y down the profi l e and wou l d move even more s l ow l y i f t h i s e xposed s o i l had deteriorated in s tructure due t o cu l t i vat ion and rainfa l l impac t . However the roo t - re l ated conduct ing pores i n the pasture area wou l d be more c l o s e l y spaced and un i form l y d i s tr ibuted , which wou l d a l l ow b romide t o be l eached mor e e ffect i ve l y . 6 . 4 . 2 Natura l Rainfa l l Experiment s The l eaching o f bromide unde r natural rain fa l l was under t rans i ent f l ow cond i t i ons . F i g . 6 . 7 shows the mean va l ues o f bromide concentrati on found i n the s o i l profi l e . () l OO 200 300 4 00 500 600 HROl'll DE CONCENTRAT ION ( JJg/ml ) l OO 2 00 300 4 0 0 Prc- l cach ing Fi r s t samp l i ng Second samp l 98 . 500 . 6 . 7 B romide d i s t r ibu t ion in the soil pro f i l e unde r natural ra in fa l l i n the con t i n uous pas t ure g rowing a rea . 99 . ' 6 . 4 . 2 . 1 P re - l each i ng measurements The i n i t i a l d i stribut ion of app l i ed bromide was measured on the day aft e r app l i cat i on when the s o i l water content was at ' ' fi e l d capac i ty" . The maximum concentrat ion was in t he top 1 0 mm segment , the concentrat i on then decreased sharp l y in the succeed ing depths . The average concentrat i on was 4 2 ?g/ml in the 40 -50 mm depth segment , s uggcs t i. ng t he movement of some b rom i de be l ow thi s depth . Th i s was p robab l y one reason for the l ow recovery percentage of app l ied bromi de i n the top 50 mm depth . Some unusua l l y h i gh bromide concentrat i ons were obse rved at 40-50 mm depth ( 2 1 8 , 240 , and 4 1 7 ?g/ml ) , and variat ion o f brom i de concent rat i on j n t he rep l i cate samp les was qui te con s i d e rab l e . Th i s w i l l be d i s cus sed later in Sec t i on 6 . 4 . 3 and 6 . 4 . 5 . 6 . 4 . 2 . 2 Pos t - l each i ng measurements D i s tribut i on of bromide concentrat io?n with depth was measured after 46 mm and 182 mm of excess rai n fa l l over evapotransp irat ion had i n fi l trated into the so i l profi l e . a . F i rst s ampl i ng Dai ly rai n fa l l records before the first samp l i ng i ndicate rainfal l was l i ght and re l a t ive ly uni formly d i s tributed over t hat peri od ( F i g . 6 . 2 ) . The max imum dai l y rainfal l during thi s period was 1 l mm , p robab l y not enough to cause saturat i on or surface run -o ff . Leach i ng o f brom i. de thus p robab l y occurred unde r unsat urated cond i t i ons mos t of the t i me . The movement under these condi t ions woul d be more uni form than under pending cond i t i on s , as was shown i n the l aboratory experiment descri bed i n Chapte r 4 . The soi l moi sture profi l e at th i s samp l i ng t ime i s shown i n F i g . 6 . 8 . I t i nd i cates the water contents near the soi l surface ( 0 to 50 mm depth ) were s l i ght l y l ower than under the ponded condi t i on s ( F i g . 6 . 6 ) wh i l e they were s imi l ar be l ow 50 mm . l lowever the water conten t s duri ng l eaching under natura l ra i n fa l l wou ld have varied from t ime to t ime , accord ing to rainfa l l inten s i ty and d i s tribut i on . MASSEY UNIVEIWIJ l:lll.RAJU 0 l OO /""' E 200 E ;:r: E:-< p... ? 0 JOO w .-l H ? 0 0:: p... ..-4 400 H 0 (/) 500 600 VOLUMETRIC WATER CONTENT (m 3 m - 3 ) 0 . 2 0 . 4 0 . 6 1 00 . STANDARD DEVIATION +0 . 05 0 -0 . 05 ,- - - .1 r - .J ,- -A ,J r--J ,- J ,-I ,J ,.. J I I I ,.. J I I - r : _ _ J 1 : 1 : I ; , : .?.: Firs t samp l ing Second samp l ing Soil porosi ty Fig . 6 . 8 Vo l ume t ric wa ter con ten t d i s t ribut ion in the soi l pro f i le under natu ral rainfal l cond i t ions a t both sampl ing t imes ( a f t er 4 6 mm and 1 8 2 mm excess rainfal l ) , and soil poros i t y . Means and s tandard deviat ions ?o f 4 0 samples . The s.tandard devia t ions re fer to the f i rs t sampling . The s i t e was under ' . pas t ure . 1 0 1 . The resu l t s however suggest p referential f low did occur to some extent , as ind i cated by the h i gh concentrat ion of bromide remaining in the top segment ( F i g . 6 . 7 ) . The average concentrat ion of bromide in the top 20 mm was 7 1 ?g/ml , wh i ch was the maximum throughout the profi le . The average concentrat i on in the succeeding depth dropped to 38 ?g/m l , a lmost ha l f of the above . Then the concentration increased s l ight l y be low thi s depth and a second peak of bromide concentrat ion ( 4 7 ?g/ml ) appeared at about 1 30 mm depth . Such doub l e peak s of concen? trati on were a l so observed by Cameron et a l . ( 1979) for n i trogen and ch l or i de l each ing in bare c l ay l oam so i l , and by Saffigna et a l . ( 19 7 7 ) i n a sandy loam profi l e under potatoes . The 32 ?g/ml average concentra? t ion measured at 250- 300 mm depth suggest s some bromide was l eached be low 300 mm depth . The appearance of a second concentrati on peak , a l though a very flat one , suggests more uni form than preferent ia l l each ing , presumab ly due to unsaturated f l ow condi t ions . I f the so i l water content i s approx imate l y constant in t ime and spac e , s imp l e l eaching mode l s predict the depth of the peak to be i / 8 where i i s the equiva l ent depth o f water moving through the so i l and 8 is the vo lumetr i c water content (Gardner , 1965) . As suming water for evapotranspirat ion i s removed from at or near the so i l surface , i may be found as the excess of rainfal l over evapo? transp i rat i on for the period in quest ion . Dayananda et a l . ( 1980) treat more comp l i cated cases . Tak i ng 8 as 0 . 42 ( the average va lue over the samp l ing depth , F ig . 6 . 8 ) , and finding i from F i g . 6 . 2 , the pred icted depth of the concentration peak at the t ime of the first samp l ing i s 1 10 mm, s l i ght l y sha l l ower than the observed depth of 1 30 mm . Bromide remaining in the top 50 mm of soi l after l eaching by natural rainfa l l was approximate ly ha l f that remaining after leaching by a comparab l e amount of cont inuous l y ponded water , as indi cated in Tab le 6 . 1 . Hi gher efficiency of l eaching by natura l l i ght rainfal l rather than continuous ponding was a l so found by Wi l son et a l . ( 19 6 1 ) . 1 0 2 . b . Second s ampl i ng There were a few heavy rain s t orms between the f irst and s econd s amp l ing , as i ndi cated by the ra infa l l data i n F i g . 6 . 2 . The heavy ra i n fal l s o f 33 and 20 mm day - 1 were probab l y l arge enough to cause mo l e d i s charge and l o ca l i s ed s u r face p onding . Leaching of bromide wou l d then be part l y under saturated condi t i on s , during which much o f the f l ow wou l d b e pre ferent i a l l y through the l arger channe l s . Some o f the bromide apparent l y moved b e l ow 6 00 mm depth , and presumab l y s ome '' : 1 ;; car r i ed znvay w i th the mo l e drain s at 400 mm depth . The s o i l water c ontent be l ow 300 mm dept l1 was near s aturat i on ( Fi g . 6 . 8 ) , indi cat ing the mo l e s were suscep t ib l e to drainage at that t ime . The tota l bromide recovered to the samp l ing depth was 4 1 % (Tab l e 6 . 1 ) . The first 4 6 mm o f exce s s rainfa l l was much more e ffec t i ve i n l eaching bromide out o f t h e surface s oi l than t h e n ex t 1 36 mm, as can be s een from F i g . 6 . 7 . The l i ghter rain fa l l events b efore the f irst s amp l ing p robab l y caused unsaturated and more uni form f l ow through the s o i l . A l so the bromide remaining i n the t op 50 mm of s o i l a fter the fi r s t s amp l i ng wa s probab l y l ocated w i thin so i l aggregates and i n p l aces r emoved from the f l ow pathways , and so was l e s s p rone to l eaching . After 1 8 2 mm o f exce s s rai n fa l l , the s econd p eak o f bromide concen? trat i on had moved to approxi mate ly 450 mm depth and f la t tened out a l most comp l et e l y . The s imp l e l ea ch i ng theory described above predi cts a depth o f 4 30 mm for the concentrati on peak at the t ime o f the s e con d s amp l i ng . Due to the presence o f the mo l e d rains at thi s depth , and the poor so i l s tructure in the B hor i zon , the agreement be tween the obse rved and c a l cu l ated depths i s probab l y l arge l y fortui t i ous . 6 . 4 . 3 Var iab i l i ty i n Bromi de D i s t ribut ion [ n th i s experiment , some unusua l l y hi gh bromide c oncerl t rations were found i n s everal s amp l e s from every depth , as s hown i n Appendi x E for the con t inuous ponding experiment , and in the h i sto grams showing the frequency d i s tribut i on of bromi de in the s o i l profi l e for the experiment under natur a l rain fa l l i n F i g . 6 . 9a . Presumab l y thes e s amp l es i n c l uded preferen t i a l pathway s . Some cores tended to have h i gh bromide concen? trat i on at a l l depth s . However other c or e s had h i gh va lues on l y at certain 40 20 (a) Pre-leaching 0- 100 lll!ll o I ?_I 1 1,,u 1 1 t:2 cn 8oo ? 40 ? ? ,,,_,'"'"'"' ffi 30-40 mm ? 20 "' "" z 8 o D-0= ? = r !; lOO 200 300 4 00 "' 40 ? \ eo" - l m > iog 8 0-50 mm 20 ' ? 0 ? d] (h, , D cp ? 1 0 20 30 40 50 60 70 80 40 20 Pos t-leaching 400-500 mm O l 1 1 1 111 1p ! D 10 20 30 40 c ( \lg/m1 ) 50 (b) l nfh1:'"' 0 - 1 00 nnn 2?0 2 . 5 3 . 0 0 1 . 0 Pre - l eaching 30-40 nnn 2 . 0 3 . 0 l I I Pos t -l each ing 0-50 mm I f==1 I I I 1 I I I I 1 1 . 0 1 . 5 2 . 0 l Pos t-leaching ' Jill:: I 0 . 5 1 . 0 1 . 5 ln C ( \l g/m1 ) Fig . 6 . 9 His tograms showing f requency dis t rib ut ions o f bromide concen t ra t ion ( C ) in so i l be fore and a f t er leach ing by 182 mm of excess rain fal J . ( a ) Fre quenc y dist ribu t ion o f C , (b ) Frequency dist ribu t ion o f ln C . f.-' 0 (.N l n C (l I 2 3 4 5 6 2 ,-- ( a ) Pre-leach ing r=0 . 987 0-20 rrrrn 0 0 ? 00 Mean = 4 1 3 ,ug/ml 00 Mode = 346 ,ug/ml 01 Med ian = 389 ?g/ml #0 0 ? or!' o ? # 0 8 l - 1 1-- D 0 ? ? 0 ? D . D r=0 . 9 75 ? 0 ? 7 8 0 1 2 ? ( c ) ?re-leach ing . 30-40 mm ? 0 0 0 0 rf'o I I ? ? ? c I I ln C 3 4 5 t '' : . . 0 . ? ? ? ' .r' ,, ,. . ? Nean = ? 5 .ug /m l Mode = 5 . 4 .ug/rnl Med ian = 2 2 . 2 ?g/ml Ul -2 ? I X I X 0 0 2 0 1- - 1 ? -2 0 100 200 300 400 500 600 c l n C 2 3 4 5 6 (b ) Pos t- l eaching ? 0 0-50 mm r=0 . 987 ? 0 ? 0 . 0 0 ;'? f?t " ! 0 Mean = 3 0 . 5 J.Jg/rnl Mode = 2 2 . 3 .ug/ml Med ian = 2 7 . 4 ,ug/ml D ? a ? e ? r=O . 9 2 3 1 0 20 30 . 40 50 60 c 700 800 0 lOO 200 300 c ln C 7 8 I 2 3 4 ( d ) Pos t-leaching r.= 0 . 985 . 0 400-500 mm ? ? 0 c l' l l . ?1 Mean ? 1 4 . 3 ?g/ml Mode ? 1 1 . 9 )Jg/ml D J Median = 1 3 . 4 )Jg/ml ?a 0 ? 0 ? 70 8 0 0 20 30 c Fig . 6 . 10 Fra c t i le diagrams showing the re l a t ionsh ip o f probab i l i t y un i t s ( ( x-? ) / s ) and b romi de concent ration in soil ( C or ln C ) , where x=C ( o ) or ln C ( ? ) , x = mean values , s = s tandard devi a t ion , and r = corre l a t ion c oe f f i c ien t . excess rain fa l l . Po s t l ea ching data shown were a f ter 1 82 mm ,.... 0 ..,. 105 . depths , suggesting non-vert ica l preferen t ia l pathways had been inte rcepted . The frequency d i stribut ions of bromide under natural rainfa l l were asymmetri c , as seen from F ig . 6 . 9a . The frequency d i stribut ions i n r: i g . 6. 9a were trans formed to l ogari thmic d i s tribut i ons and these are shown in F i g . 6 . 9b . The h i s t og ra 11 1s , each of wh i c h i nc ludes a t l east 40 measu rement s , i nd i cate b rom i de concen t rat ion va r i ab j l i t y in the soi l prof i l e was more l og ? no rma l l y t han norma l l y d i s t r i butcd . F ract i l e di agrams , wh i ch have been used by Bi ggar and N i e l sen ( 1976 ) as an i nd i cat or of norma l or l og- norma l d i s tribut i on , we re a l so constructed . The deta i 1 s for cons truc t i on of fract i l e di agrams , fo l l owing B i ggar and Nie l sen ' s procedure , are given in Append ix F . Examp l e s o f fract j l e di agrams arc shown i n F i g . 6 . 1 0 , w i th the corresponding l inear corre l at i on coeffi c i ent va lues ( r) . The probabi l i ty uni t s ( (x - x) / s ) were more l inear ly re l ated to ln c than C , as indicated by the r va lues , a l so sugges t i ng a log-norma l d i stribut ion . The resul t s of fract i l e di agram construct ion for observat i ons under natural rainfa l l l eaching , both pre - l eaching and after 1 8 2 mm exces s ra i nfa l l , g i ven i n Tab l e F . l , i ndicate the r va lues for the re la t ionships o f probab i l i t y uni t s and ln C are cons i stent l y greater than the re lat ionship w ith C . The mean , mode , and med ian value? obtained from fract i l e d iagrams and shown i n F i g . 6 . 1 0 and Tab l e F . l , are di fferen t . Mean values are h i gher than moda l and med i an va lues i n a l l cases . Data for the cont inuous ponding experiment are not presented as frequency d i s tributions , as not enough rep l i cates were taken to a l l ow the data to be ana l ysed thi s way . The i nd i cator of the vari ab i l i ty i n bromide concentrat i on at any depth in the soi l profi l e i s gi ven by the exponen t i a l of the standard deviat i on of l n C (wri t ten as exp s L) ' and the coeffic i ent of var iat ion (C . V . ) , where C . V . = 1 00 s L/ x L , and x L i s the mean of l n C . The s i xteenth and e ighty- fourth percent i l es of the l og-normal d i s tribution can be e s t imated as exp (xL - s L) and exp ( x L + s L) respect ive l y ( Rokosk i , 1 9 72 ) . The C . V . values for the ponded water experiments were ca l cul ated as suming log-norma l di stribut ions , and the values are shown i n Tab l e E . l and E . 2 in Appendix E . These C . V . va lues for the ponding exper iment s are much greater than for the expe riment under nat ura l rai n fal l 1 06 . c ond i ti on s (Append i x F ) , probab l y due to more preferen t i a l f l ow occurring under ponded condi t i on s , and al so p erhaps l es s red i s tribut i on t ime b eing a l l owed b efore t ak ing pre - l e aching samp l e s ( 3 h ours compared to 24 hours i n the natural ra i n fa l l experiment) . The C . V . values tend i ncrease w i th so i l depth , except under cu l t i vat i on , where the c . v . greate s t at the depth of the compacted l ayer ( Tab l e E . 2 ) . I n the natural rainfa l l experiment , the C . V . v a lue s for the first s amp l ing (Tab l e F . 2 ) are sma l l er than for the s e cond samp l ing ( Tab l e F . l ) . i s The sma l l er C . V . va l ues for the fi rst samp l ing are probab l y due t o the l i ght rain fa l l prior to samp l ing , whi l e preferenti a l f low induced by heavy rain wou l d cause larger variab i l i t i e s i n the s e cond samp l ing measurement s . D i s tr ibuti on s o f saturated hydrau l i c conduc t i vi ty i n the fie l d have b e en obs erved to b e l og-normal by Mason et a l . ( 19 5 7 ) , Rokoski ( 19 72 ) , N i e l sen et a l . ( 1 9 7 3 ) , and Bak er ( 19 78 ) . Thus l eaching , whi ch i s re l ated t o f l ow ve l o c i ty , m i gh t be expected t o resu lt i n l og-normal d i s tribut i on s o f s o l ut e concentrat i on i n the s o i l profi l e . B iggar and N i e l sen ( 19 76 ) and Van der Po l et a l . ( 19 79 ) , present ing fi e l d l eaching data i n t erms o f f l ow ve l o c i t i es and d i spers i on coeffi c i ent s , found these l eaching paramet ers were l og-norma l l y d i s tributed . Large varia? b i l i t y in hydrau l i c c onduct iv i ty data has been reported within a s o i l s eri e s . However the data have usua l l y been obtained from a fai r l y l arge are a , e . g . N i e l sen et a l . ( 19 7 3 ) measured t h e s o i l hydrau l i c conduct iv ity i n twenty 6 . 5 m2 p l ot s d i stributed over 1 50 h a . The variab i l i ty i n bromide concentrat ion reported here i s on a much smal l e r s ca l e , over 40 s amp l es being taken from a n area as sma l l as 0 . 1 6 m2 . Thi s rai s e s interest ing que s t i on s as t o the s ca le o f var i ab i l i ty i n to the fie l d . Perhap s the l arges t vari ab i l i t y i n s o i l occurs over d i stances re l ated to parameters such as p l ant spac ing , s tructure uni t s i ze , and t h e spacing between l arge b i opores . 6 . 4 . 4 Leaching F l ow Pathways under Natura l Rain fa l l P l an t i ntercept i on and s t em f l ow , s o i l micro-re l i e f or micro? depres s i on s caused by animal pugging and l arge pores open to the s o i l surface , cause non-un i form eritry o f rain into the s o i l surface . 1 0 7 . A subs id iary experiment was conducted to observe the l eaching pathways under natural rai n fa l l in the fie l d , as descr ibed i n Sect ion 6 . 3 . 3 . Rhodamine B dye was used as a tracer to mark the pathways . F i g . 6 . 1 1 shows the movement of dye tended to be concentrated in the area a round and d i rect l y unde r the p l ant s tems . The dye stain seems to be as soc i ated w i th root s ( F i g . 6 . 1 1a and c ) , worm channe l s ( F i g . 6 . 1 1b and c ) , and cracks ( F i g . 6 . 1 l c) . The accumu lat i on of dye near the base of p l ants was p robab l y caused by stem- flow , which had been i n i t iated by the co l l ect ion of rain by the aer ia l parts of a p l ant . The ? stem- fl ow wo u l d de l i ve r i t to th e so i l at the base o f the p l ant , where i t in fj l t rate s i n an amount cons iderab l y greater than i n adj acent areas . Thus the intercept ion of rain by fol i age prevented water reaching the soi l surface at some spots beneath the p lants , but tended to funne l water stra i ght down the stem i nto the so i l . The reported examp l e s of s tem f low have usua l l y referred to trees , l arge shrubs or l arge - l eaved p l ants ( G l over and Gwynne , 1962 ; Saffi gna et a l . 1 9 76) , however G l over et a l . ( 1962) found that i t was an important phenomenon in gras s l and . The intercepted water might a l s o drip from the l eaf t ips d irect l y to the soi l , caus ing point sources of water i nfi ltration . Soi l mi c ro - re l i e f or sma l l depre s s i ons may cause loca l i sed surface ponding . Cameron et a l . ( 19 79 ) have d i scus sed the effects of soi l m icro-re l i e f on non -uni form l eaching of ch l oride and n i trate into so i l under natura l rainfa l l . Surface run off into sma l l depress i ons meant more i nfi l trat i on , and so so l utes moved deeper there than they did i n the surrounding soi l . Much variat ion in mi cro-re l i e f was observed in the experimenta l area , as we l l as many sma l l depress ions caused by sheep pugg ing . Soi l micro -depres s i ons , in as soci at i on w i th open . channe l s to the soi l surface , wou ld cause non-uni form f low onl y i f enough rain occurred to cause ponding i n the ho l l ows . However, i f the who l e surface was ponded , these depress i on areas mi ght conduct water at a s l ower rate than the surround ing soi l , due to l ower saturated hydrau l ic conduct iv i ty of the compacted soi l . The non-uni formity of i ncom i ng water due to p l ants and micro -re l i ef has often been neg l ected i n l each i ng stud i es . Most exper iments have been conducted in bare soi l w i th a d i s turbed and l eve l le? soi l surface , such as report ed by Wi l d and Bab i ker ( 19 72 ) , Wet s e l aar ( 1962) , Mi l l er et a l . ( 1965 ) , B i ggar and N i e l sen ( 19 76 ) , Van De Pa l et a l . ( 19 77 ) . 1 08 . ( a) (b ) F i g . 6 . 1 1 F i g . 6 . 1 1 ( c ) Photographs i nd icat ing s tem- f l ow respon s i b l e for non uni formi ty o f water intake at the so i l surface under natura l rai nfa l l . W = worm channe l , R = root , C Crack 1 09 . 6 . 4 . 5 Pos s i b l e Factors Tend ing to Cause Percentages 1 1 0 . Low Bromide Recovery A re l at ive l y low recovery percentage of bromide , ranging from 40 to 86% , was obtained (Tab l e 6 . 1 ) . Some pos s ib l e reasons for t i t i s are d i scussed be l ow : - a . Retent ion o f bromide i n soi l Near ly comp l ete recovery of app l i ed bromi de from soi l was reported by Abda l l a and Lear ( 1 975 ) , us ing water extract ion and a bromide s e l ect ive ion e l ectrode for analys i s . To check the ir resu l t s under the cond i t i ons in th i s s tudy , the subs id iary labora? t6ry experiment described in Sect ion 6 . 3 . 2 was carried out . I t i nves ? t i gated short- term (18 hour equi l ibrat ion) bromide retent ion in so i l under s i mi l ar cond i t i ons i n the f i e l d exper i ment . The resu l t s we re 94% and 99% recovery from 0-20 mm and 1 00- 1 20 mm core depths , respect ive l y . The h igher retent ion i n the top l ayer was probab ly due t o a higher organi c mat ter content and so more anion adsorpt ion , and/or p l ant uptake . Retent ion of bromide in so i l in thi s experiment was therefore not a maj or factor . b . Retent ion of bromide by plants Bromide uptake by p l ants was measured 99 days after i t s app l i cati on to the soi l surface . During th is period 1 1 3 mm of evapotranspirat ion had occurred . The amoun? of bromide measured from the two sub-p l ot s we re 1 3 . 5% and 1 1 . 5% of the app l i ed amount . The rate of p l ant uptake was probab l y re l at ed to the transp ir? at ion rate and the amount of bromide avai l ab l e i n soi l . The amount of bromide uptake by p l ant s at the t ime of the f irst samp l ing was probab l y much l es s , a s on l y 2 7 mm of evapotransp irati on had occurred b y th?m . c . Non-un i form bromide app l i cat ion Bromide so lut ion was app l i ed on the soi l surface us i n g ;1 watering can at the start of the experiment under natural rainfa l l , and using an atomi zer in the ponded water experiments . Severa l pract ice runs were made w i th the watering can prior to the actua l app l i cat ion , in order to produce as uni form a d i stribut ion as poss ib le . However , some non-uni form i ty of app l i cat ion i n both exper iments was unavo idab le . 1 1 1 . d . Retent i on of brom i de on p l ant s The concentrat i on of brom i de solution used was re l at ively high and the amount app l ied re l at ively l ow , so that the drop l et s of s o l ut i on remaining on the p l ants wou ld contain a s i gn i fi cant amount of bromide whi ch wou l d not be recovered in the pre - l each ing samp l i ng . e . So i l sampl i ng method The so i l corer used in this experiment was re l at ive ly smal l in comparison w i th the p l ant s i ze and spacing . For ease of samp l ing , the so i l samp les tended not to inc lude p l ant crowns and the soi l d i rect l y under them . Thus , g iven the di scuss ion in Sec t i on 6 . 4 . 4 concern ing i n t ercept ion and stem f low , an unconcious b i a s probab l y occurred in the so i l samp l i ng , and this was probab l y the ma i n reason fo r the l ow percentage recovery of bromide . 6 . 4 . 6 Computat ions Leac l1 i ng of so lute from a th i n surface l ayer i nto the s o i l profi l e beneath , under saturated cond i t i ons , was mode l l ed numerical ly using CSMP . The computer programme used was s im i l ar to that used for mode l l ing preferent i a l so lute movement by Scatter ( 1 978 ) ( see Appendix C . l ) The preferent i a 1 f l ow was ideal i zed as occurri ng in uni form l y s raced , vert ica l , cy l i ndri ca l channe l s . V i scous l ami nar f l ow , fo l l owing the Hagen - Po i s eu l l e equat i on , was as sumed . An examp le of the computer programme for bromide leach i ng is g iven in Appendix C . 4 . Brom i de was assumed i n i t i a l l y to be present on ly i n the top 1 0 mm of the so i l profi l e . Cont i nuous l each ing for 20 hours was s imu l ated , and in that t ime 50 mm of l each i ng water had passed through the soi l , as suming a saturated hydrau l i c conductivity of 6 . 94 x 1 0- 7m s - 1 ( i . e . the measured fi e l d infi l t rat i on rate of 2 . 5 mm hr- 1 ) . During the l each i ng process , brom i de in the so i l matrix in the top 10 mm of -soi l d i ffu sed towards the channe l s , and was then trans l ocated down i nto the profi l e with the vi scous f l ow in the channe l s . Be low 1 0 mm depth , bromide from the channe ls then d i ffused back into the surrounding so i l matr ix , due to the reversa l i n concentrat ion gradient . I 1 1 2 . l n the s imulat ion of so i l cont a i n i ng l arge channe l s , the viscous fl ow in the channe l s was much fa ster than the rad ia l d i ffus i on i n to or out of the surround i ng s o i l . Thus re l at i ve ly l i t t l e bromide wou ld be l eached from the top 1 0 mm so i l , but the sma l l amount l eached wou l d penetrate qu ite deep l y into the so i l pro fi le . The converse wou ld ho l d for smal ler channe l s . The so i l prof i le model l ed was 300 mm deep and was d iv ided into 6 l ayers of th i cknes s 1 0 , 1 0 , 20 , 40 , 80 and 1 4 0 mm , the thi ckest l ayer be i ng the deepes t . The concentra t i on of bromide i n each soi l layer , and i n the effluent so lut i on at 300 mm depth , were computed as a funct i on of t ime . Three channe l s i zes were cons i dered , w i th d i fferent spacings between them so a s to g ive the same hydrau l i c conduct iv i ty . The channe l s i zes and re l ated data are g i ven i n Tab l e 6 . 2 . The soi l poros i ty as sumed was 0 . 5 2 , the average value perta in ing i n the f ie ld . Tab l e 6 . 2 Channe l s i ze and re lated data .as suming hydrau l i c conduct iv i ty of 6 . 94 x 1 0- 7 m s - 1 D i ameter of Maximum Pressure potent ia l channe l s d i ffus i on rad ius t o drain (mm) (mm) (mm) 0 . 1 0 2 . 8 1 - 300 0 . 1 5 6 . 3 2 - 200 0 . 20 1 3 . 2 8 - 1 50 -- The resu l t s of the computat ions are shown in F igs . 6 . 1 2 and 6 . 1 3 . The movement o f so lute through 0 . 2 mm diameter channel s was found t o be preferent i a l by Scat ter ( 1 9 7 8 ) , thus the shape of the concentrat i on d i st r i but i on curve appeari ng i n F i g . 6 . 1 2 for the soi l conta i ning 0 . 2 mm channe l s i s the resu l t of l each i ng under preferent ia l f low . A h i gh concentrat i on of bromide remains in the top 1 0 mm soi l , and a neg l ig i b l y smal l concentrat ion , too l ow t o appear on the graph , occurs in the profi le be low . Most of the bromide that i s leached from the t op l ayer moves be low 300 mm depth . F i g . 6 . 1 ? shows the re lat ive concentrat i on (C/C0 ) 0 0 I ,....:, lj. ? ::..'! I i L ?? ? I I I ; ........ 1 0 0 ? .....__ ::r:: E--< p... w I I A _ .. .J w I ....:1 . H ? 0 0::: 2 0 0 t I I p... . I 3 0 0 Fig . 6 . 1 2 c I c 0 0 . 2 0 . 4 0 . 6 0 . 8 CHk?EL DIAMETER (mm) 0 . 2 0 . . . . . . . . . . . . 0 . 1 5 - ?? - ? ?- 0 . 1 0 Computed bromide concentration d i s tribution af ter leaching in so i l containing uni formly spaced , vertical , c y l indrical channels . Ini t ially the soil solu t ion concentrat ion (C ) in the top 1 0 mm of soil was C 0 1 . 0 ...... ...... lN 0 . 0 1 4 --- - 0 . 0 1 2 I - - - - - 0 . 0 1 0 ?0 . 2 mm -- -- - .,- - / "" "" -"' / / 0 . 008 1 "" / 0 / u ., -''h-o . 1 5 mm ......_ / / u 0 . 006 / / / / / / 0 . 004 .L r' I I I 0 . 002 J. I I I I I I . 0 5 1 0 1 5 20 THIE (hour ) ..... Fig . 6 . 1 3 Computed re l a t ive concen t ra t ion o f b romide l eached f rom top l 0 mm to b e l m,? 300 mm dep th ...... ? in soil con t a in ing ver t ical cyl indrical channe l s , 0 . 2 and 0 . 1 5 mm in diame ter . 1 1 5 . o f bromide in the e ffluent increased qui ck l y t o 0 . 0 1 1 6 after 1 0 minutes , and then dropped s low l y to 0 . 009 after 20 hours . The total amount of bromide l eached be low 300 mm depth was 1 2 % of the amount i n i t ia l l y present in the top soi l l ayer . I n the fie l d , continuous channe l s 0 . 2 mm in d i ameter or greater wou ld const itute a "short circui t" to the mol e- t i l e drainage sys tem , but would cause re l at ive ly l i t t l e leaching . The concentrat ion d i s t r ibut ion o f bromi de in the so i l containing 0 . 1 5 mm diameter channe l s i nd i cates re l at ive ly less preferent i a l f low occurring during l eaching ( F i g . 6 . 1 2 ) . The concentrat ion of bromide in the effluent increased stead i l y with t ime and reached 0 . 0 1 3 C/C 0 after 2 0 hours ( Fig . 6 . 1 3 ) . S l ight l y l ess bromide was l eached i n this case , 9% of the in it ia l amount moving be low 300 depth . No bromide moved below 300 mm depth in the soi l w i th 0 . 1 mm channe l s , and the concentration di stribut ion curve was effectively symmetrical with a d ist inct peak at approxin1ate l y 120 ?mm depth ( Fig . 6 . 1 2 ) , which i s s imi l ar to the c l ass ica l d i s tribution curve ( curve (d ) in Fig . 6 . 1 ) . The re lat ive ly large segment thi ckness used be l ow 200 mm depth in the numerical ana lys i s caused the "bumpiness" in the distribut i on curve . The ponded water l eaching data i n F igs . 6 . 4 and 6 . 5 , w i th the h igh bromide concentrat ion remaining at the surface , and the absence of a second peak l ower in the soi l profi l e , resemb l e more the curve for s o i l contain ing 0 . 1 5 mm diameter pores in Fig . 6 . 1 2 suggest ing pores approx? imate l y this s i ze were responsib le for much of the f l ow . The natural rainfa l l l eaching data in Fig . 6 . 7 shows l ower concentrations at the surface , and a second concentrat ion peak in the soi l profi l e , resemb l ing more the curves i n F ig . 6 . 1 2 for a s o i l w i th 0 . 1 mm pores . Thi s suggest s the l arger pores were drained and the pressure potent i a l was l ess than -200 mm for much of the t ime in most of the so i l profi l e under nitural rainfal l . 1 16 . 6 . 5 CONCLUSIONS 1 . Some bromide was leached to300 mm profi l e depth by both 50 mm of ponded water and 46 mm of ex cess ra infa l l in th i s experiment . Al so p refe rent i a l movement i s i nd i c a t ed by the high concentrat i on rema in ing un l eached near the soi l surface . About 3 1 to 62% of app l i ed bromide remai ned in the top 50 mm uf sui l under ponded l eaching cond i t i ons , whi l e 1 5% remained under natura l rai nfal l cond i t i ons . 2 . Bromide was l eached more effective l y by natura l rainfa l l than by cont inuous pend ing , when the same amount of l eaching water was used . 3 . Large variabi l ity 1 n bromide concentrat ion in the rep l icate core samp l e s was observed . Unusua l l y high concentrat i ons of bromide were found in some samp l es at a l l depths , which caused the frequency d i stribut ion to be l og-norma l rather than norma l . 4 . Intercept ion and stem- f low , combined wi th the soi l samp l ing techn ique used whi ch inadvertent l y avoided samp l ing p l ant crowns and the so i l d irec t l y under them , probab l y resu lted i n some b ias in the s amp l ing . The re l at i ve ly l ow recovery percentages of the app l i ed bromide (40 to 86% ) , part i cu l ar l y in the pre- l eaching samp l ing , were probab l y l arge l y due to th i s . 5 . P l ant uptake of bromide was probab l y sma l l during the experiment involving l each ing wi th ponded water . But approximate l y 10% of the app l ied bromide was taken up by p l ants by the end of the experiment under natura l rainfal l cond i t i on s . 6 . Doub le peaks in bromide concentrat i on were observed i n the soi l profi l es under ponded water condit i ons i n soi l which had been cul tivated , and under natura l rainfa l l condi t i ons in the pasture area . The second peak was probab l y the resu l t of unsaturated , and so non-preferent ia l f low . In the cu l t ivated area th is was probab l y brought about by the compacted l ayer thrott l ing the flux i r tto the soi l beneath . Under natural rainfa l l cond j t i ons , unsaturated f low p robab l y prevai led most o f the t ime , due to the rainfa l l intens i ty being l ess than the saturated hydraul i c conduct iv ity . 1 1 7 . 7 . N o t ran sport mode l s eems t o be ab l e to describe quant itat ive l y b rom ide movement i n th i s exper iment , part i cu l ar l y the effects of prefer? ent i a l f l ow . The resu l t s of the .computat i on s for bromi de l each i ng i n an idea l i zed soi l , contai ning un i form l y spaced , vert i ca l , cyl indri ca l chann e l s , indi cated that brom i de wou ld be l eached more effect ive ly from so i l where cont L nuous channe l s 0 . 1 5 mm or l arger were e ither absent , or a i r- fi l l ed and so i neffect i ve . Such channe l s i f un i form , wou l d d ra i n at - 200 mm pres sure potent i a l . 8 . A pract i ca l i mp l i c at i on o f the re su l t s 1 s that s o l ub l e fert i l i zers app l i ed to pasture or crops may not be as prone to l eaching as is often thought . Intercept ion , stern f low , and preferent i a l f low in th i s soi l may comb ine to l eave a s i gn i fi cant fract i on of the app l i ed ferti l i zer near the surface for a cons i derab ly period . On the otJ?.er hand , prefer? ent ial f l ow wi l l l each some of the app l i ed fert i l i zer more quick l y b e l ow the root zone than wou ld more un i form movement in the soi l . APPEND I X A SO f L P fWF f L I : f l i :SCH I PT f ( )NS /\Nn PI I Y .'-; l C/\ L AN i l C I IEM f CA I . D/\T/\ 1 19 . l 'ah i l' A . I l ' ru f 1 i l' l lesL? r i p t i on of Tokoma ru s , i t J .oam i n Ua i ry farm No . 4 , Massey Un i vers i t y , l 'a l mc r s t on North ( l'o l J oJ.. , 1 9 7 S ) . ? --? .--- ---------,------- - - ------------------- , l iP ! ' I .'.U I I IJes?r ip t i on . - ? ? - ?-- ---- --+------- ------------------- Ah l 0 t o l U + I S i i t l oam ; da rJ.. grey i sh brown ( 10 YR 4 / 2 ) , w i t h s l i !(ht yel lowish? r l ' d ( ? ? Y H ? 1 / t ? , ? . ; t , ) 1uut t J i n? '-! ruunU ? r?u? :-. rou t ? ; mulh.:ratc ! y dPvt : l opC'd f i r l l ' : t nd m<'d i urn c rumh s t ruct ure' ; frl ah l c.? ; con s i derab l e humus ; nume rous , f i ne grass roots ; s l ight ly ir.?perfect interna l d ra i n : tg<' ; mo i ?, t : fa i r l y d i s t i nct anc.J even boundary , ----- -?----- ------+---- --- ---------------------------------------------? Ah? lU + I to 20 + 1 S i l t l oam ; da rJ.. grey i sh brown ( 10 YR 4 / 2 ) , w i th some ye l l owish? n?J l? Y l l ?/h , ??/<> ) mot t l i ng around grass root s ; moc.Jera t c l y deve l oped , m<'d i um c rumb , cast granu l ar and nutty s t ructure ; fn ab l e ; mode r:? t e amount of humus ; numerous , fine grass root s ; somL' <':J r t hwo r11" ; some i ronstone conc ret i ons ; s l i?:h t l y imperfect i n t t? rn ; r l d r; r i l l :q?, l ' ; mo i s t ; rather i nd i s t i nc t , s l i ?? h t l y I IOCVf'n bounc.Jary . AB 11;: Cxg . . - - - - . - - - - . - - - . --- 20 + I to 2b + 2 ---- -------? - . . . . . . - ?----------------- ?-- ?-------- -- S i i t l oam ; g rey i sh brown ( 2 . 5 Y 5 / 2 ] , moderate ly set t l ed ye l l ow i sh - rec.J [ S YM 4 / 6 ) , dark reddish brown ( 2 , 5 YR 3 / 4 ) and dusJ.. y red ( 1 0 M 3/ 4 ) in a ret icu late pattern ; moderate l y to s t rong l y devc l opl'J , med i um nut ty structure ; fr i ab l e - fi rm ; some humus ; moc.Je ra t e number of grass roots ; some earthworms ; some i rons t one corH: rct i ons ea 6 mm in d iame ter ; impcrfL?c t i nternal d r: t i l l : t )?l' ; mo i q ; r:t t hcr i nd i s t i nc t , s l i ght l y unev l y?onu l s t nu.: turc ( pcn t a unJ l l t ' )l ; l \ t d l l !ru r ; r r . r 1 1 d I r? ;q H ? :'. t H? o l r mrna r a f' t c.? r Hn?wt: r ? , I ?H tt1 ) w i t h t he.? pol ygons vary 1 11g from I S to 40 cm i n width and separated by l arge, III.J J i t l y VL' J ' l i 4,: .J I I y U J' i c l l tcJ , SU! l - f i ! l cJ c racks L-4 Clll W lli c , both pC'ds : tnd c r?cJ.. ? hC' i n)? cont i nuous w i t h th C' hori zon hC'nC':J t h ; th i c k c l ay s k i ns ; so i l w i t h in peds compact and ext reme l y hard when dry , moc.Jc ra t c l y s t i d, )' :.111c.l p las t i c when wet and d i spers i b l e in water ( frag i pan) ; v i rtual l y no humus wi thin peds , but a l i t t l e in the v i c i n i t y of decay i ng roots w i th in the c rack s ; some old bush roots and a few more recent l i v i ng roots w i th i n cracks , but v i rtua l ly none w i th i n ped s ; no fauna seen ; numerous pi nhead Fe/Mn concre? t ions w i th i n peds ; i mpeded i nternal drainage , the ma i n avenue for water movemen t bei ng down the soi l -f i l led cracks ; moi s t ; d i ffuse ( impercept i b ly merg i ng] boundary , Cwg l 1 1 ? ? :; tu 1 ? 7 ? I S i i t J ua m ; ? u l ull l ., as for hori zon above , but w i th l i ght o l i vl> grey (S Y h / 2 ] now a s sum i ng predominance over strong b rown ( 7 . 5 YR 5/6) in t he mott l i ng pa t tern w i th in ped s ; structure , cracks and c l ay s k i n s as in t he hori zon above ; not iceab l e change i n cons i s tency , w i th t h e soi l hecom i ng l e s s hard and compact and los ing t h e pro? pert i e s of a frag ipan ; l i t t l e i f any humus ; no roots ; no fauna seen ; p i nheaJ concret ions become darker i n co lour ( b l ack) and more mangan i ferOL JS ,?ompa red w i th the hori zon above ; i mperfect i nternal dra i nage w i th the ma in avenue for water movement be ing down the soi l - f i l l ec.J c ra c k ,, be tween pcds ; moi s t ; J i s t i lll: t , even boundary . L_ ______ ?-------------?--------- --------------------------------? CONTlNUEU Tab 1 t' A . 1 ( .:ont i nued) ---- I IO I' l ZUI I llej.l t h ( .:m) t :wg2 1 4 7 . 1 t o 1 !>7 - . 1 - 1 20 . Descript ion Si J t l oam ; l l g i l l o I i ve g rey ( 5 y 6/2 ) to o"! i ve ( 5 y 5/3) , mod er- a t e l y IIIO t t l l?li ?t run? brown ( 7 . 5 YR 5/6 , 5/8 ) in ret i cu l ate pattern ; 111oderatc l y <.levc loped , very coarse polygona l struc ture ( pent a and hex a co lwnnar an <.I t rapezo.:o lumnar after Brewer , 1964 ) . with the soi l - fi l l e<.l c racks between s tructura l uni t s tending to be finer and more wide l y spaced co111pared w i t h the Cxg and Cwg l hori zons above ; fi rm ; no humuS 1 roots or fauna ; munerous dark grey and b l ack re/Mn p i nhca<.l concretions ; i mperfect interna l drainagv ; mo i s t ; over ! i es terrace grave ls at depth . 1 2 1 . Tab l e A . 2 Some Phys ical and Chemical Properties o f Tokomaru Si l t Loam ( Po l lok ) 1975 ) . Tab le A . 2 . 1 Part i c l e S i ze Ana ly s i s Hori zon Samp l e Symbol Depth Lab . No . Depth (cm) ( cm) Ah 1 0 -8 494 1 0 - 8 Ah2 8 - 20 4942 1 0- 1 8 ABg 20- 38 4943 23-33 Btg 38 - 76 4944 5 1 -64 1Cxg 76 - 145 4945 9 7- 1 2 7 1Cwg 1 145 - 2 2 1 4946 168- 198 2C 2 2 1 - 234 4947 2 2 1 - 229 1Cwg2 234- 250+ 4948 2 37- 248 Equivalent Spherica l Diameter (mm) 0 . 006 0 . 0 2 0 . 06 0 . 2 2 . 0 <0 . 00 2 t o to to to to 0 . 002 0 . 006 0 . 0 2 0 . 06 0 . 2 Cy fSi mSi cseSi fS cseS 23 . 0 7 . 6 1 7 . 9 4 3 . 0 8 . 0 . 0 . 5 2 2 . 0 7 . 1 1 8 . 5 4 3 . 7 8 . 2 0 . 5 23 . 8 7 . 4 1 8 . 0 4 1 . 2 8 . 7 0 . 9 30 . 2 7 . 8 1 5 . 5 39 . 3 6 . 9 0 . 3 1 8 . 4 4 . 8 16 . 3 46 . 6 1 2 . 7 1 . 2 1 8 . 0 6 . 4 1 5 . 0 45 . 9 1 3 . 1 1 . 6 8 . 9 6 . 2 1 7 . 8 26 . 5 22 . 0 1 8 . 6 16 . 3 6 . 3 16 . 7 50 . 3 9 . 5 0 . 9 1 2 2 . Tab l e A 2 . 2 Chemical Ana lys is pH Cat i on Exchange Propert ies Organi c C and Soi l Total N Horizon meq/1 00 g soi l in in BS c N water nKC 1 % % % C/N Na K Ca Mg TEB CEC Ah1 5 . 0 4 . 2 0 . 1 5 0 . 2 1 3 . 1 5 2 . 4 0 5 . 9 20 . 2 29 3 . 32 0 . 3 1 1 1 Ah2 4 . 9 4 . 1 0 . 05 0 . 1 8 3 . 1 5 2 . 1 2 5 . 5 1 8 . 0 3 1 2 . 08 0 . 28 7 ABg 5 . 3 4 . 4 0 . 1 6 0 . 1 2 2 . 55 2 . 58 5 . 4 1 3 . 8 39 1 . 98 0 . 1 4 1 4 Btg 5 . 0 3 . 6 0 . 49 0 . 3 1 1 . 35 3 . 28 5 . 4 1 6 . 0 34 0 . 83 0 . 1 5 6 1Cxg 5 . 1 3 . 4 0 . 99 0 . 1 5 2 . 00 3 . 90 7 . 0 1 2 . 9 54 0 . 1 2 0 . 04 3 1Cwg1 ? 6 . 1 4 . 1 1 . 30 0 . 1 9 3 . 1 5 3 . 90 8 . 5 1 1 . 9 7 1 0 . 09 0 . 02 5 2C 6 . 3 4 . 6 0 . 76 0 . 08 1 . 80 2 . 70 5 . 3 5 . 4 98 0 . 03 0 . 01 3 1Cwg2 6 . 4 4 . 1 1 . 1 7 0 . 20 3 . 33 3 . 82 8 . 5 1 2 . 0 7 1 0 . 09 0 . 02 5 ? N ? ? t;l::) n n n n rt :( :( >< (IQ (IQ (IQ (IQ N ? ? ? ? VI C]\ 00 00 00 0 VI 1.0 0 +:>. N + + + + + + + + + + + + + + + + + I + + + + + + + + + + I + + + I + + + + I + + + + I I I I I ... + I I I I + I + o--3 o--3 o--3 I 'i 'i 'i + . I I I I o--3 'i o--3 + o--3 o--3 o--3 'i 'i ? 'i ?v ?-v I I I I I + + + + + + + ? + ? ?v ?-v + o--3 o--3 + 'i 'i + + )> ? ? t;l:l (IQ N N N VI N VI 00 0 0 + + + + + + + + + + + + I + I + I + + + + + + + I I I + + + + + + + + + I I I I o--3 I 'i o--3 o--3 I 'i 'i ?v I I I + + + + :{ ? ?-v ?v o--3 + + 'i ?v Soi l Hori zon % C l ay M i ca/ I l l i te C l ay-verm i cu l i te Heat - co l l ap s ib l e 1 4 ? C l ay C l ay-Ch lorite Interstrat i fi ed C l ay Montmor i l l onite Kao l in i te Hal l oys i t e A l l ophane Fe ldspar Quart z FeO (OH) -- o--3 ? ,_. CD ;t> N VI n ,_. Ill '< ::s:: ... . ::I CD 'i Ill ,_. 0 (IQ '< ? N VI APPENDIX B DI FFUSION OF SOLUTE I NTO SPHER ICAL SO I L AGGREGATES 1 25 . Cons i dering soi l crumbs to be spherica l , the d i ffusion of non-adsorbed so lute into a soi l crumb can be described by (Crank , 1 956) , ac at + 2 r a c ) a r ( B . 1 ) where C = solute concentrat ion in soi l solution (M L- 3 ) t t ime for di ffus ion (T) D = di ffusion coeffi cient of solute in soi l ( L2 T-1) s r radial d i stance from centre of sphere ( L) . As suming adsorption of react ive solute by soi l i s instantaneous and l inear l y related with solut ion concentrat ion , the d i ffus ion of sorbed so lutes into the sphere can be expressed as , where ac at R D s 1 + R 2 r ac ay- ) a retardat ion factor whi ch i s equiva l ent to ( B . 2 ) pb soi l bu lk density (M L - 3 ) , 8 i s volumetric water content ( L 3 L- 3 ) , and k i s the solut i on d i stribution coeffici ent obtained from the l inear adsorption i sotherm ( L 3 M-1) (Eq . 1 . 7 ) . For unsteady s tate d i ffus i on , when the surface concentration i s maintained constant and the ini t ia l concentration in the sphere of radius a is uniform , then the solution for the total amount of d i ffus ing substance entering or leaving the sphere (Mt ) can be expressed as a fraction of the correspond ing quantity after infinite t ime (M00) by the re lation , 00 1 ? = 1 n 2 ( B . 3 ) where D i s D for non-adsorbed so l utes ?1 n d i s D I ( 1 + R) for adsorbed s s so lutes . Figure B . l shows the re l ationship between Mt/M00 and (Dt/a 2 ) ? given by Crank ( 1 956) . 1 . 0 r---+---?--?---+--?----???--, 0 . 8 0 . 6 0 . 4 0 . 2 0 Fig . B . 1 0 . 2 0 . 4 2 I ( Dt / a ) >:l 0 . 6 0 . 8 2 !.::: Relat ionship be tween M /M and ( Dt / a ) 2 , as shown t 00 by Crank ( 1956 ) . 1 26 . APPEND I X C THEORY OF PREFERENTIAL SOLUTE MOVEMENT THROUGH LARGER SOI L VO IDS AND TYP ICAL COMPUTOR PROGRAMMES 1 2 8 . C . 1 "Preferentia l solute movement through l arger soi l voids . I Some computat ions using s imp le theory" . by D . R . Scotter reprinted from Austra l i an Journa l Soi l Research . 16 : 2 5 7 -67 Preferential Solute ?Movement through Larger Soil Voids. I Some Computations Using Simple Theory D. R. ScoiiC!r Oepartmcnt of Soil Scicm:c, Masscy Unavcrsity, Palmerston North, New Zealand. Abslrut?l Aust. J. Soil Res., 1978, 16, 257-67 Viswus solution flow down verti.:al .:ylindri?.:al ?.:hannds and planar cracks, with simultaneous mole.:ular ditl'usion of the solute into the surrounding soil, was modelled. Chloride and phosphate were .:hoscn as representative of nun-?orbcd and ad,urbcd ions respectively. Jn channels at least 0 ? 2 mm in diameter, and cracks at least 0 ? 1 mm wide, almost instantaneous preferential movement of both chloride and phosphate was pred icted . Lill lo.: or no preferential movement was predicted in smaller channels or .:ra.:ks. For example phosphate was predicted to move to a depth of 200 mm within 10 min in saturated soil containing 0 ? 2 mm diameter continuous channels. However, it would take phosphate over 2 months to reach the ?ame depth in simillll soil with the same hydraulic wnd1 1ct ivity , but containing only 0 ? 05 mm diameter channels. Channels and cracks permitting preferential 'olute movement would be a;oiution-filii:J uuly at pre?sure potcntials above -0 ? 2 m, so such movement ?:an only occur in near saturated soil. Although highly ideal ized soil-void geometries were assumed , the results have a number of practical implica? tions related to the movement of nutrients and pollutants in field soils. Introduction Growing concern for groundwater and stream pollution, and continued interest in salt and fertilizer leaching, arc the main reasons for the present strong interest in the movement of solutes in soil. Theory dcsl:ribing miscible fluid displacement in uniform soils exists (Nielsen and Biggar 1 962 ; Kirkham and Powers 1972), and experimental results for packed columns of sand or soil aggregates are in general accord with the theory (Biggar and N ielsen 1 967). Hydrodynamic dispersion occurs at the interface between the displacing and displaced solution, but i n general non? adsorbed solute applied at one end of a ?aturated column breaks through after the application of approximately one pore volume of solution. Adsorbed solutes appear i n the effluent even later, after a number of pore volumes have been displaced. Field experiments have often shown quite different behaviour, however, with some of the applied solute moving through the soil much faster than expected. Kolenbrander ( 1 970) found surface applied ni trate becaml! more dispersed in cracking-clay soils than in more uniform soils, while Kissel et al. ( 1 973), using a solution of chloride and ft uorescein dye, showed preferential movement down the cracks in a swelling? clay soil. Even in a weakly structured loamy sand, field experiments by Wild and Babiker ( 1 976) showed asymmetric leaching patterns for both chloride and nitrate. They attributed this to preferential solute movement down larger channels, such as earthworm channels typically 5 mm in diameter, and possible smaller channels. 1 29 . 1 30 . 251! D. R. Scotter Comparisons between chloride breakthrough curves for 'undisturbed' cores and repacked columns for a ?ilt loam ( Llrick and French 1 966), and a swelling-clay ( Kissel et al. 1 973), show the chloride appearing earlier from the cores in both cases. The authors attributed this to the pre?cnce of continuous channels and cracks in the und isturbed cores and their absence in the repacked columns. Bouma et al. ( 1 976) found very pronounced preferential ch loride movement in saturated cores of Dutch 'knik' clay soils, with the solute appearing in the effluent after the application of only a thousandth of a pore volume of solution. They attributed this to the bul k of the flow being through a relatively few large, continuous soil pores. Other data, showing preferential chloride movement in saturated cores of silty clay loam soil with sub? angular blocky structure, are given hy Anderson and Bouma ( 1 977). The same authors (Anderson and Bouma 1 973) earlier applied a water-Rhodamine D dye mixture to soil cores which were then dried, impregnated with plastic, sectioned, and polished. They present diagrams showing dye movement down continuous channels and planar voids, but not through the bulk of the soi l . While preferential sol ute movement through relatively large continuous voids is wel l documented, no attempts to de?crihe the phenomenon theoretically appear to have been made. The exi?ting miscihle di?placemcnt theory already referred to treats the soi l as a continuum on a 'macro!>copic' level (Raats and Klute 1 96H), and does not describe fluid and solute behaviour at a 'microscopic' level i n and adjacent to indi? vidual voids. An exception is the theory of Passioura ( 197 1 ) for hydrodynamic dispersion in aggregated media, which is developed from a microscopic viewpoint for viscous flow in the intra-aggregate pore space, coupled with molecular diffusion within the aggregates. I t is predicted that slight preferential movement may occur in aggregated media under certain conditions (see Fig. 2 of Passioura 1971) . However, t he size of the intra-aggregate voids is not considered explicitly, and neither is the effect of solute adsorption. The work described here is also microscopic in viewpoint, but involves a different approach. The effect of isolated relatively large channels or cracks of k nown size on solute movement, with and without solute adsorption, is investigated. Particular attention is addressed to three questions. Firstly, how large does a continuous channel or crack have to be for sol ute!> to move through i t preferential ly ? Secondly, at what water potential do such pores drain, and so cease to conduct solutes ? Thiruly, as adsorbed and non-ad?orbed ion? arc known to move differentially in packed soil columns, to what degree to strongly adsorbed solutes move preferentially through larger continuous voids ? Theory Brewer ( 1 964) suggests the larger i nterconnected soil voids fal l i nto two broad . classes, channels and planes. Two ideal ized pore geometries wil l be considered here, vertical cyl indrical channels, and vertical planar voids. Cylindrical Channels The Hagen-Poiseuil le equation for viscous flow through a hollow cylinder, applied to gravity induced flow in a vertical channel, gives (Childs 1 969) q ? (rcpcga4)/(HI'J), ( l ) Solute Movement through Soil Voids. I 259 whcrc tf i? the flow rate ( m 3 s - 1 ) . ,,, i s t he fl u i d densi t y ( kg m - 3), g is the acceleration d ue to g ra v i t y (m ? - 2). a i s t he rad i u s ol' the t u be ( m ), and 11 the viscosity ( Pa s). The ?ol u te tl u x i n to such a tu be is qC; ( m o l ? 1 ) where C; is the concentration of the app l ied sol u t ion (mol m - 3) . T h e s i m ple model, used for ft ow t h ro u gh a s a t urated soil containing channels, con?idcrs v i ?cous flow down a n u m ber ol' u n i for m l y ?paced c y l i n d rical c h a n ne l s o f a certa i n d i a m e ter, with negligi b le v i scous flow i n the rest of the soil. The fourth power relat ionsh ip between pore rad i u s and flow means that, while the presence of a few larger channels may make o n l y a very m i n o r contribution to the total soil porosity, nearly all t he flow i s t h ro ugh t he?e channel? ( Ho u ma and Anderson 1 973). In t he soil surro u n d i ng t h e cy l i n J rical c h a n nels, assuming radial symmetry, molec u l a r d i ffus ion of t h e sol u te may he lbcribed by (Gardner 1 965) j ? - - DJJCjdr, (2) w here j i? t he solute tl u x cro?sing u n i t cro!>s-?cc t i o n a l area per unit time (mol m - 2 s - 1 ), D. is t h e effective solute diffusivity in the soi l ( m 2 s - 1 ) ,/is the soil porosity (m3 m - J), C is t he concentration of the soil sol u t ion ( mo l m - 3), and r is radial distance (m). I t i s assu med that the channels contribute negl igi b ly to f For transient diffusion of a solute fol lowing a l inear adsorption isotherm Olution and air as 7 ? 3 ? 10 - l N m 1 ? I n the CSMP73 programmes, the ?oi l volume around each channel or inside each sl it was div ided into four layers along the .:-axis ofO ? 02, 0 ? 04, 0 ? 06 and 0 ? 08 m thick? nes?. Each layer wa!> divided in to u?ual ly eight, or sometimes four, geometrically spaced annu l i . Each annulus was I ? 2 I ? X times as thick as the one preceding it, with the thinnest annulus adjacent to the channel or slit. The method used to solve equation ( 3 ) was very s imi lar to that descri bed by de Wit and van Keulen ( 1 972). The rectangular in tegration method was used, with finite ditlerence methods to compute the flux in and out of each annu lus. The thickness of the first annulus was chosen to minimize computer time, whi le avoiding serious computational errors at short times. This could be checked by comparing the computed flux into the . first annu lus in the top layer with the analytical solution. The integration time increment was chosen so that C/C; in the first annulus in the top layer was less than 0 ? 2 after the first i teration, ensuring numerical stabil ity. To a_ccount for adsorption the real Table J . Channel si;tcs assumed a nd related data Diameter of Maximum Pressure Fraction of soil channels dill'u?ion potential volume occupied radius to drain by channels (mm) (mm) (m) 0 ? 05 0 ? 4 -0 ?60 0 ?0036 0 ? 1 1 ? 6 -0 ? 30 0 ?0009 0 ? 2 6 ? 6 -0 ? 1 5 0 ?0002 0 ? 4 26 -0 ?07 0 ?00006 volume of each annu lus was mult ip l ied by ( I + R) to give an effective volume. I f, as sometimes happened for narrow channels or slits at short times, the computed flux into the first annulus was greater than the solute flow through the channel or sl i t at that depth, then the flux was put equal to CLqfL for channels or CLpf2L for sl its. where L is the thickness of the layer and CL the concentration of solution entering the layer. For sol ution of this k ind of transport problem in soil, CSMP was found convenient and inexpensive to use. Four channel sizes were considered, with different spacings between channels so as to give the same hydraulic conductivity. The diameters assumed, and related data, are given i n Table I . The channels constitute a negligible fraction of the total pore . volume, in all cases less than 0 ? 8 %. The Reynolds number for the largest channel considered is 1 60, so equation ( I ) would apply in all cases (Bird et al. 1960). Computed breakthrough curves for chloride and phosphate are shown in Fig. 2. C" is the sol ution concentration at 200 mm depth. For soil with 0 ? 4 mm diameter channels both ions broke through elfectively instantaneously, with C./C; reaching 0 ? 8 or higher before 0 ?0 1 pore volumes or 6 min had passed. The spacing between the channels was such that even after 20 h there was no interference between the ditfusion shells around each channe l . Further computations, assuming a 2 m soil depth also containing 0 ?4 mm vertical channels and the same K, showed C./C; still reaching 0 ? 8 after 0 ? 0 1 pore vol umes or l h . Solute Movement through Soil Voids. I 263 The breakthrough curves for 200 mm ?oil depth contatntng 0 ? 2 mm diameter c.: h a n neb abo show mar ked prefere n t i a l movement of both ions down the channels. A fter 20 h or two pore vol umes there was some interference between adjacent dif? fu s i o n shel ls fo r ch loride, b u t n o t for pho?phate. For soil containing 0 ? I mm diameter channels the breakthrough curve for chloride shows l i ttle preferential movement and almost 'classical' behaviour, resem- 1 1 l ? tJ ' O?X? v 1 1 ?6 --- ' v? I 1 1 ? ?1 0?:! 11 Pore volumes 0??4. Cl and P ' 11 ? 1 . ("1 0?05, Cl 1 0 Time ( h ) 1 5 Fig. 2 . Breakthrough curves for chloride (--) and phosphate (- - - -) moving through 200 mm of soil depth containing cylindrical channels. The numbers on the curves are channel diameters (mm). c.tc, is the ratio of effluent to influent concentration. bling the S-shaped breakthrough curves symmetrical about one pore volume found for sand or disturbed soil ( Biggar and N ielsen 1 967). The breakthrough curve for phosphate i n this case is shown in Fig. 3. Note that the horizontal scale has been expanded by a factor of 200, and phosphate is not predicted in the effluent unt i l four pore vol umes or 2 days have passed, and then the effluent concentration rises only very slowly. For soi l containing 0 ? 05 mm diameter channels very steep breakthrough 0?4. 0?2 0 100 Pore volumes 200 0? 1 Channel ,.. " 0?02 Sl?t ' ? U?US Channel 50 1 00 Time (days) 400 I SO Fig. 3. Breakthrough curves for phosphate in soil containing channels (- - - -) or slits (--). The numbers on the curves are either the channel diameter or slit width (mm). curves, suggest ive of 'piston flow' (Nielsen and Biggar 1 962) are predicted for both ions ( Figs 2 and 3). Similar steep breakthrough curves have been found for chloride in uniformly sized glass beads ( Biggar and N ielsen 1 967). Phosphate is not predicted in the effluent unti l 1 94 pore volumes or 80 days have passed . h>u r planar void widths were considered, again spaced so as to give the same hydraul ic conductivity of 10 mm/h. The slit widths assumed and related data' are 1 35 . 1 36 . 264 D. R. Scotter given in Table 2. The slits oecupy only a small part of the total porosity, less than 2% in all cases. The computed breakthrough curves are shown in Figs 3 and 4. Fig. 4 ?how? that with 0 ? 2 11 1 1 1 1 ? l it width both chloride and phosphate appear almost instantaneously in the effl uent. With 0 ? 1 mm slit width, preferential move? ment still occurs with both ions, but i? k?? pronounced for phosphate. With 0 ? 05 mm sl i t width, chloride stil l appears wi th in a few minutes, but phosphate (shown in both Figs 3 and 4) is not predicted in the etlluent until 7 h have passed. Table 2. Planar void widths assumed and related data Width of Outer radius of Pressure Fraction of soil slit soil annulus potential volume occupied enclosed by sI i t to drain by slits (mm) (mm) (m) 0 ?02 2 ? 6 -0 ? 74 0 ?0085 0 ?05 4 1 - 0 ? 30 0 ?0014 0 ? 1 320 -O ? I S 0 ?0003 0 ? 2 2600 -0 ?07 0 ?0001 With 0 ? 02 mm width slits, the data in Fig. 4 show a steep breakthrough curve for chloride which appears after one pore volume, indicating negligible preferential movement. I n Fig. 3 the curve for phosphate movement in the same soil shows a similar shape, but with no phosphate appeari ng in the effluent unti l 1 76 pore volumes or 73 days have passed . Fig. 4. Breakthrough curves for chloride (--) or phosphate (- - - -) moving through 200 mm depth of soil containing vertical planar slits. The numbers on the curves are slit widths (mm). The computations descri bed Jo not take account of di ffusion across the soil surface from a ponded . solution before it enters the larger voids. Fig. 5 shows the solutions to equations ( 1 0) and ( 1 2) for the values of K, f, D. and R used in the computations, and gives an indication of the importance of such diffusion. The curves for both chloride and phosphate indicate that when pronounced preferential movement does occur, the etrect or ?urface diffusion may be the same order of magnitude as the radial di trusion out from the larger voids in the soi l . However, in these cases both diffusion processes e ffect only relatively slight changes i n the con- centration of the percolating solution. ? Solute Movement through Soil Voids. I 265 Discussion The two soil-void geometries considered are highly idealized, so care must be taken in interpreting the significance of these results to solute movement in field soils. Whi le only continuous vertical channels and slits of uniform width have been con? sidered, in real soil a short constriction in an otherwise uniform channel or slit may very significantly affect its solution conducting capacity. Some of the results of the analysis were unexpected, in particular the strongly wntrasting behaviour predicted for solute movement through 0 ? 2 mm and 0 ?05 mm diameter channels. Phosphate could move through 200 mm long 0 ? 2 mm diameter channels in soil within a few minutes, but would not emerge from 0 ? 05 mm channels until over 2 months after i t was applied. The minimum channel diameter for prefer? ential flow of both non-sorbed and strongly adsorbed solutes is approximately 0 ? 2 mm. The minimum crack width for preferential flow i s approximately 0 ? 1 mm. v ..._ v" Por? volumes u l , ... r--=========='f'=====--_ -_ - - _-_-_-_-_...,_ - IHI 11?6 11?4 O?l 0 1 0 IS Tome ( h) Fig. 5. Estimated concentration ?.:hangc due to diffusion through soil surface from pondcd solution before it enters larger voids, for chloride (--) and phosphate (- - - -). The less pronounced influence of sorption on preferen tial movement in channels as compared to planar voids is a result of the different diffusion geometries in the soil adjacent to the voids. For radial diffusion from a channel, the soil close to the channel provides a larger part of the 'diffusion resistance' than for planar voids, but contains a relatively small soil volume per unit radial distance. Thus the diffusive flux there quickly approaches a steady state, and so is less dependent on sorption. I n the root-zone horizons of many soils it is probable there are some channels greater than 0 ? 2 mm diameter resulting from the activity of roots and soil fauna, and so some _ preferential solute movement is l ikely. Seminal and first lateral roots of wheat have been measured as 0 ? 3-0 ? 45 mm in diameter (Russell 1973), and presumably when decayed may leave channels approaching the same dimensions. Structural cracks are also common, particularly in swelling soils. Although cracks often close partly or completely when the soil is at or near saturation, the associated swelling takes some time, and meanwhile preferential water and solute movement can take place. 137 . 1 38 . 266 D. R. Scottcr The data presented here for ?oi l contammg continuous larger voids show that preferential solute movement in such soil may be much more pronounced than indicated by Passioura's theory for aggregated media. His maximum dispersion occurred in the l imiting case of zero Brenner number and gave Cc/C1 equal to only 0 ? 2 after 0 ? 2 of a pore volume. Practical Implications The results have several practical implications. Channels larger than 0 ? 2 mm in diameter and planar voids wider than 0 ? I mm will be air-filled if the pressure potential in the soil is less than -0 ? 1 5 m, so preferential solute movement can occur only in soi l considerably wetter than the usual 'field capacity'. If it does occur, both non-sorbed and strongly adsorbed ions will move preferentially. When potential ground water pollutants are applied by sprinkler irrigation, and a long solute residence time in the topsoil is desired to allow biodegradation or plant uptake, the application rate or frequency should be such that the soil stays slightly unsaturated, and the pressure potential is maintained below -0 ? 1 5 m. Preferential movement would not then occur. A nother implication is that movement of sol utes, particularly strongly adsorbed solutes, through columns of disturbed and repacked soil, is l ikely to be very dif? ferent from movement through undisturbed soil i n the field. The soil chemical properties may be relatively unchanged, but in many situations the pore geometry may be more important than the sorption capacity in determining the movement of nutrients and pollutants. Coarser-textured soils with high saturated hydraulic con? ductivity are often considered unsuitable for effluent disposal, but finer-textured soils with a lower conductivity, but containing cracks or channels, may i n fact allow more groundwater pollution. While the movement of solutes into the soil has been considered here, the results are of course just as relevant to the leaching of solutes out of soils. When water move? ment occurs preferentially, the concentration of the percolating water will be very much lower than the concentration i n the bulk of the soil, and little leaching will take place. References Andersoo, J. L., and Bowna, J. (1973). Relationships between saturated hydraulic conductivity and morphometric data of an argillic horizon. Soil Sci. Soc. Am. Proc. 37, 408-13. Andersoo, J. L., and Bouma, J. ( 1 977). Water movement through pedal soils : I. Saturated flow ? ? ? J. Soil Sci. Soc. Am. 41, 413-18. Bird, R . B., Stewart, W . E., and Lightfoot, E. N. (1 960). 'Transport Phenomena.' (John Wiley: New York.) Biggar, J. W., and Nielsen, D. R. (1 967). M iscible displacement and leaching phenomenon. In 'Irrigation of Agricultural Lands'. Agronomy 1 1 , 254-74. (Eds. R. M. Hagan, H. R. Haise and T. W. Edminster.) (Am. Soc. Agron. : Madison, Wise.) Bouma, J., and Andersoo, J. L. ( 1 973). Relationships between soil structure characteristics and hydraulic conductivity. In 'Field Soil Water Regime'. (Eds. R. R. Bruce et al.) (Soil Sci. Soc. Am.: Mad ison, Wise.) BoWIIII, J., Dekker, L. W., and Verlinden, H. L. ( 1 976). Drainage and vertical hydraulic conductivity of some Dutch 'knik' clay soils. Agric. Water Manage. l, 67-78. Brewer, R. ( 1964). 'Fabric and Mineral Analysis of Soils.' (John Wiley : New York.) Carlslaw, H. S., and Jaeger, J. C. (1 959). 'Conduction of Heat in Solids.' 2nd Edn. (Oxford Univ. Press : London.) Solute Movement through Soil Voids. I 267 Childs, E. C. ( 1 969). 'An Introduction to the Physical Basis of Soil Water Phenomena.' (John Wiley : London.) Gardner, W. R. ( 1 965). Movement of nitrogen in soil . In 'Soil Nitrogen'. Agronomy 10, pp. 550--72. (Eds. W. V. Bartholomew and F. E. Clark.) (Am. Soc. Agron. : Madison, Wise.) Kirkham, D., and Powers, W. L. ( 1 972}. 'Advanced Soil Physics.' (John Wiley : New York.) Kissel, D. E., Ritchie, J. T., and Burnett, E. ( 1 973). Chloride movement in undisturbed swelling clay soil. Soil Sci. Soc. Am. Proc. 37, 2 1 -4. Kolenbrander, G. J. ( 1 970). Calculation of parameters for the evaluation of the leaching of salts under field o.:onditions, i l lustrated by nitrate. Plant Soil 32, 439-53. Nielsen, D. R., and Biggar, J. W. ( 1 962). Miscible displacement : lll. Theoretical considerations. Soil Sci. Soc. Am. Proc. 26, 2 1 6-2 1 . Olsen, S. R., Kemper, W. D., and van Schaik, J. C. ( 1 965). Self-diffusion coefficients of phosphorus in soil measured by transient and steady-state methods. Soil Sci. Soc. Am. Proc. 29, 1 54-8 . Olsen, S. R., and Watanabe, F. S. ( 1 970). Diffusive supply of phosphorus in relation to soil textural variations. Soil Sci. 1 10, 3 1 8-27. Passioura, J. B. ( 1 97 1 ). Hydrodynamic dispersion in aggregated media : 1 . Theory. Soil Sci. 111, 339-44. Porter, L. K., Kemper, W. D., Jackson, R. D., and Stewart, B. A. ( 1 960). Chloride diffusion in soils as influenced by moisture content. Soil. Sci. Soc. Am. Proc. 24, 460--3. Raats, P. A. C., and Klute, A. ( 1 968). Transport in soils : The balan<;e of mass. Soil. Sci. Soc. Am. Proc. 32, 1 6 1 -6. ? Robinson, R. A., and Stokes, R. H. ( 1 959). 'Electrolyte Solutions.' 2nd Edn. (Butterwortbs: London.) Russell, E. W. ( 1 973). 'Soil Conditions and Plant Growth.' l Oth Edn. (Longman : London.) Taylor, G. I. ( 1 953). Dispersion of soluble matter in solvent flowing slowing through a tube. Proc. R. Soc. (London) 219A, 1 86-203. Wild, A., and Babi.ker, I. A. ( 1 976). The asymmetric leaching pattern of nitrate and chloride in a loamy sand under field conditions. J. Soil. Sci. 27, 460--6. de Wit, G. T., and van Keuleo, H. ( 1 972). 'Simulation of Transport Processes in Soils.' (Pudoc: Wageningen.) Manuscript received 5 December 1 977 1 39 . 1 40 . C . 2 L ist of Symbo l s in CSMP Programmes . A , B , C , D , E , and F = BB = c CSA , CSB , . . . = cso = D = DELT D I FC = D I FD D I ST F = Suffix indicat ing l ayer from the soi l surface downwards , with A nearest the surface ( 1 + R) , where R i s the retardation factor described in text re l at i ve concentrat ion of so l utes in soi l solution re lat ive concentrat ion of so lute in macropores concentrati on imposed at the soi l surface mo l ecular di ffusion coeffic ient of solutes in soi l ( L2 r- 1 ) ( for programme in Sect ion C . 2 . 2 on ly ) computation t ime interval (T) mo l ecu lar di ffus ion coeffi c ient of so lute in soi l ( L2 T- 1 ) ( for programme in Sect ion C . 2 . 3 ? on 1 y ) distance between the centers of two succeeding annul i ( L) ( see F i g . C . l and C . 2) d i stance from the centre of the channel to the midd le of any annulus ( L) ( see F i g . C . l and C . 2 ) f lux of solute between two adjacent annuli (M L- 2 T- 1 ) F INTIM = FLOW = LA , LB , . . . N = OUTDEL = P I 2 = POR RAD RITCOM = s SFLUX = TCOM THCKNS = TOPA VA , VB , . . . = 1 4 1 . final t ime for computation (T) vi scous flow rate of so lution in the cy lin? drica l channel ( L 3 T- 1 ) thi ckness of column l ayer ( L ) number of annul i around the channe l t ime interval when output i s g iven (T) 2TI water fi l led poros i ty radius of cy l indrical channe l ( L ) a factor giving a geometric increase i n the thi ckness of adj acent annul i net flux of so lute into or out o f an annu lus (M L- 1 T- 1 ) f l ux of so lute across the soi l surface (M L - 2 T- 1 ) thi ckness of annulus ( L) ( see F i g . C . l and C . 2) maximum di ffusion radius from the channe l ( L) cros s - sect iona l area o f the so i l surface ( L2 ) re lat ive amount of so lute i n each soi l l ayer (M L- 1 ) c . 3 . 142 . Prograrrune for miscib l e d isp l acement of so lutes in a soi l w ith uni form vertica l cy l indr i cal channe l s . ..- - - -,. -- ..... ...... .,"' ....... , / ' / ' 1 4 3 . t'" ..... --- .... , ', I /.,. CSA ' \ I I / ? \ \ I I I ? I LA I I I CSB I I - - --?--- -(-- -?- - - - ?F71=-l )- -!- ? - - f' - - - - i- - LB I F ( l+1 ) I F{ I) I ' : I I I I ( i I csc I I .., ?- ?--L - ---r ----; - - - - - - i -- -- , -- + ,-- I I LC I I I I I I I I I I I CSD I --- -? --- ---? - - - -, - - - ? ? __ i _ _ _ _ i _ _ I I I I I LD I I TCOM ( I ) I I I I I I I I _l ( 1+ 1 ) I I ' I I I I I lH FD(l) I I DIST{ l ) I I I I I RAD I I I f-i I I I I I I ?t.,_--??------?--( I-1 ) CSE Fig. C . l Geometry of the system and symbols used in the programme for misc ible displacement of solutes in a soil col umn wi th un iformly spaced vertical cyl indrical channels . Arrows indicate d irection o f flow . M ;. ? .. :, f \ l r. I \. r J. ? . I I \ ------ ------?-- -- ------- - . - - ?-? ??- - ?- ---- ------- . -------? -? ? svl T II A : ( t ) 0 ? P II I : ( Q ) " ? J If ? " C I l " ? 9 1' ? : ( 1) ) '0. ( \I I. : ( C ) W ? " " ? " ( tr ) ll ? r ?t ? = C r ) I' A Z \1 ? 1: ( 7 ) W A l l' ' z f t ) "?? ll ?-? ?-?- l l? " S II ? r n n <: , ,., , , ? , J ? I ? P n ? c / n ? ? I'I ) 1?9 J ? I " L ? ? c r. l'l ? ' " l ,?9 1 NI ? 9 0 ? C ( 0 ? ' " l 1 ? ? J ? l a <; ? ? ( h r. ? ? II ) 1 ? 9 J ? J ? tr ? ? r r n r ? n ) 1 ? 9 t "J l a ( 0 ? C "nc: ? l' rl)I!> J N T a Zil ? C T II ? ' " l l ? 9 l lli i ? I II A ( ? J ? t ll ) 1 ? !> ) ? , ? ? 1 ? ( / J ? ? ? ) 1 ? ' ) "J z J 1 ? c n J ( ? " 1 1 N' J ? J = Q 1 ? r ? J ? ? ? l 1 ? ? J "J i a?; 1 ? ( tr J ? t n ) 1 N ? I ? l ? ? , ? r r J ? ? ? J 1 )1? J ? T ., f 1 ? ( ? J ? tn ) ,?? l? ?li? C T J ? ? II ) 1 ? ? J ? I ? 1 1 ? r o ? ? ? ? , , ? , J ? l ? 9P \ r / P r t ? ) 1 ? !> J ? l " L R ? r n P ? ? n ) l )l? l N J c Q R ? r c P ? ? ? ) 1 N' l ? I c c; q ? r ? P ? ? II ) 1 H , J H ! r lr q , r r P ? ? r l 1 N ? ! "J l ? r R ? l ?P? I II J l? j ? J c l ? ? ( T P ? t n ) 1 N ? ) ? : t ll ? r ? v ? ? n ) 1 ?? ? ? ? r 9 V ft l / ? ? ? n ) 1 H ? J ? c J ? ? r ? v ? ? " J l)I? J ? T c Q ? ? r c ? ? ? r , , ? ? ? ? I ? ? V ' C ? V ? ' " l 1 H? J ? I c tr V ? r r v ? , ,., , , ? ? l ? J ? f Y ? C ? V ? t n y,?? l N ? ? Y ? C T Y ? ? ?'? l 1 ?? 1 N I ? t ? ? HH'!<; t o t n n o t n '- I T n on n l J T O on n t:n n '! l T 0 -'71 ft ., , , , t J T n " J t O 1\l)lO O Q T n ) Q ! II ll 9 t n f:.I) T II "9 t n t' Q t n 1. 9 t n T Q T I'I 11 9 t n '- c; t o oc; T n }l;T 0 oc;,n c c; t n lr <; t n r c; ? n :O c; t n t ?; t n n c t n " " tr\ o tr t n J tr t n tl tr t l') ?c; ., ? n ., ., ; n f tr T 0. ., , , , " 1 lr T I'\ ,., , , llr.t n ( II ) J ? II O J 0 ( ! 11 ??r b , J a bO:l - -? -H T II ( O ) S ? II O ? ? ( T O C J l S ? l O S ?; ( T O ( Q l S ? Q O? " ? t n r C ) <; a <;O<: . - ?tT II , .. , ? . .. c c: 7-r t n C ? ) S a ( C ? to n C Z ) S z Z C ? n r t n C t l <: a t ?<: - - ' 11 1. T O r t + T H? C T H?: r n <: Il l o z t n H ? t ? l - 1 Q G / ? T n C T ) Y ? I ( { ) J . ( t . T l 1 l ? ( J l j ? l O ;> T n " ' 7 " 1 r t on C 7 t o "HII! l t ? 0 , II C lr ?. T O ' l'l a l S j r ;> t n ,., , , ? r 1 l ? " ? , = c t l J J C ;> z t o Q ? ? J C ? J ? c 1 ? 1 l l 1 ? t n ? n ? J t " ? ? r t ? J- 0 <: 1 ., 1 ? ' ,., ., , , r t l Y. o C r t ) ? . r <: j l ? r t l ? ? ? ? n ( ! ) 1 r A t C J \ II ? ? ( J l j 7. ! ? . t n "J ' ! :r i ? 1 00 l t T II r o 1 <: : o j <: ? t T n r ; l <: : J j ? C t t n C ? l ? : Q J S lr t T n r ? ' ' " c; ' ? r t t n - l ? l <: :r ? J <: ,, , , r r \ <: :r r 1 <: t t t n ( 7. ) <: : /. J <: , , , , ( t l <: a t ? <: ? n y n n + T l .? - f l ) b ( J ) C: l T o n t n ? , , ., , t t o n J n t n ( T l Y o i ( J ) 1 . ( 1 - l l J ' z C ! l J li T Q n t n ? ? Z = I n t o o c n t n ?ft? T l N 0 1 Q C ll ll t n ? r : n ? ' r n t n ? 1 , ? 11 1 J o 1<; 1 a ( ! l ? c; c ., n T n Q C ? !; C: ? !: <; c 0<: 1 ) q t n t n .,.,,j , ? 1 ? C t l J . J S J : O? j ?" T n 146 . - - - ? ? ----------------.---- -- - - -- - ? -? -------- - - ---- - - - - ------- ---?-???-?? --- -- ------- 1 4 7 . '- ... "' I N " I 0 I ? I . ., >- ,.... v. c:: 0 ...... :;.. .., 0 .::. :::> ;) ,_. IF. I > V. I ::> ? : <> ? ? I "' :::> :::> ol\ I 1\ er 0 0 i I .., I "' I "' ,., I 0 ::> I I I .1\ I "" ...... N 0 .;) I I I ? I ? "' :::> I 0 I I E ? o t ? 5 0 6 ? iJ l 1 o 0 o 8 5 ? ? 0 t 9 ? ! v 3 1 ? ? o ;: 6 ? t 3 ? o <: ?A 2 Il l C ::. C (. (. q C li 3 ? ? -? . 6Y 9 1 E ? O j cp c 1 ? o . o o ? O o O U C :i G ?"O+I I DL O ! H lJ S J u '; r f>O I?il l rHi ., ,.HR T l Hl . T l H[ ? 7 a 2 0 1 5( 0 2 C A 1 ? 2 . 9 u OIIE ? oL C S u ? 2 o H7 3 ? ? 0 2 C B q t 6 ? 3 :> 6H ?n C C J ? 6 o 5 o 1 3 E ? 0 2 C 0 2 ? 6 t 9 1 9 0 [ ? Q 2 C E 1 ? 6 o 9 2 5 o E ? O ? ? e ? r ? l ? ? t t? E ? o:.:: c 4 ? I o 1 \1 8 6 E ? o ? 1 ? 2 0 0 1 ( 0 3 C A l ? 5 ? 1 2 1 9E ?u 3 - - - - - ? ? --css -. - -- ? 1 -5 ;.? ( ? 0 3 C B 4 ? 1 ? 8 o25E ? 0 2 C C J ? ? o 3 7 3 6 E ?o;: C02 ? ? 2 o 9 E ? o 2 - - -- ---- e?r ? - -, ? eHoE?o:.:: 8? ? t 7 U 6 3 E ? 0 2 4 ? 2 ? 1 9 0 6 [ ? 0:.:: l A 2 (. lj l c s c C C '? (. I.J 3 C. l 2 (. f 1 C !I C. t A 2 C B I C :i C t c 4 (. I.J ) fH C :i G ? .. ? ? ? ? ? a ? ? & ? ? ? ? ? .. ? "' ? ? ? ? ? ? ? ? ? ? ? ? ? 5 o 6 5511i. ?O ! 0 3 1 ? 6 0 3 1 l ? O I C 6 2 l ? '1 4 5 3 t. ? O ! c c 1 3 ? ? 4 '1 g l ? 0 2 C S .; 1 ? 7 3 t. ? 0 2 C D 4 O a O O C ? 3 O t O O -e r 2 o . o o j ? 7 9p L ?O l ? ? 3 o 4 4 5 L ? O 8 2 1 ? 4 86 3 l ? O l c c 1 ? ? 2 6 7 ? l - 0 2 ? s ., t O O l l ? 0 2 [i 4 - ? t o 4 t L ?o' lf l 0 . 0 0' 2 o . o o s . 3 9 2 7 t. ? o 2 C A 3 3 . ? 4 9 8 ( ? 02 C B 2 3 o 5 6 3h ? o 2 c c 1 9 ? 9 4 0 9 t. ?0 2 C S u ? 2 6 5 0 [ ? 0 2 C [i 4 6 o 0 6 1 3 l ? 02 c ( 3 2 o 0 5 1 5t. ?0 3 Cf 2 o . oo 9 t 52 3 1 [ ?0 3 C A 3 ? . 6 t 09 l ? 0 3 tB 2 6 o 5 6 t1 6 l ? 0 3 c c 1 ? t 5 7 4 1 t. ?O ? C S ? o 6H 9 t. ? O C D 1 ? l! 6 7 5 [ ?0 2 ""'C[ 3 t 3A ? l l ?o 2 c r 2 o . o ? ? ? ? ? ? - ? ? ? ? ? ? ? ? ? ? ? ? ? . ? ? ? ? ? ? ? R E C T l N TE.?oR ii T l lJ N 7 . 50 39(. ? 0 ! C A 4 & 1 o 1 5 !'J[? 0 1 C b 3 a 1 ? 3 5 ? 2 [ - 0 ! t C 2 ? 9 . b b 4 7 [ - 0 2 C i.J 1 a 9 . 2 3 8 4 ? 0 3 C H ? o . o u C E ll ? O o OU C f 3 ? 4 . 8 6 t g? ? o l C A 4 & l o 2 6 6 ? o C b 3 ? l ? ? 4 9 J l ? O t c c. 2 ? 1 . t 0 ? 6 [ ? 0 t ? li l ? 4 t 2 U ; n ? 0 2 S E. ? 1 ? 1 2 6 Jl - 0 3 Cf 4 . . . o u c 3 ? R EC T 1 N TE. ?.o F ? T l L? h 6 t 'l 1 2 b [ ? 0 2 C A ll ? 5 . t J i o E ? C z C c J ? 4 o b 3 4 o t. ? 0 2 C C 2 ? ? ? 6 2 1: n ? o2 C v l ? o 4 6 2 5 E. ? o z CH ? 5 ? 55 9 7 l - 0 2 C L 4 ? ? ? 41 HH ? o ? C f 3 ? 1 ? 2 2 0 7 ? ?02 C A ? ? 1 ? ? 2 9 1 [ ? 0 2 Cb 3 ? l o 4 9 8u [ ? 0 2 C C 2 a 1 .11 9 4 0 [ ?0 2 C U I ? 3 ? 8 7 2 4 ? ? 0 2 CH ? 5 ? ?282[ - 0 2 C E ? - 2 ? 365[ ? 0 2 e r l ? Cl o 50 u oL ?IJ! 7 o 4 1 0 5L ? v l a . 4 1 ? 2 L ? u z 5 o l:l 4 7 9 t. ? v t U t U O v o v u \l o V V ) o ? 0 9?L ? v t 1 t 1 4 7 [ ? U t 1 o U 4 1 8 t. "; v l () o 4 'i C. 2 l ? v <: 2 o tl 3 :> b l ? v 2 l t 0 9 .tt ) l ? U j O t U G 7 o o 3 ? o t. ? '-' < :> t 9 1.. l iE. ? iJ .{ 5 t Y 1 o 2 t. ? v 2 ? ? J 2 :> 8 t. ? u ? t Y C l 8 t. ? u .:: ? ? j 3 o )L ? v <: :1 ? 5 ?\ll - v lt 1 e 38 J 5t. ?0 2 l t 1 1 ?4 [ ? U 2 2 ? 0 ? o t L ? ? 2 ? ? o 3 o ? t: - 0 2 4 o 2 9 o J t. ? v 2 ) o 6 7 02L?v:.:: 2 t l9?6l ?02 ? V1 0\ APPEND IC U DERIVATION OF EQUATION FOR F LOW THROUGH NON-UN IFORM CAP I LLARY .TUBES 1 58 . The gravity induced , v iscous flow through a vertical cyl indrical , cap i l l ary tube of length L and radius r can be described by the Hagen - Poiseu i l l e equat ion (Chi l d s , 1 969 ) , Q where Q pf g n ? and ? 1 2 = = Tir4 pf g (? J - ? 2 ) Bn L f l ow rate (m 3 s - 1 ) f lui d density (kg m - 3 ) acce lerat ion due to gravity (m s - 2 ) v i scosity of fluid (Pa s ) (D . 1 ) pressure potential at inflow and out flow respective l y (m) . When v i scous f low occurs through a tube made by j oining together two d i fferent d iameter cap i l l ary tubes A and B ( as shown in the figure be low) , at steady state the rate of fluid flow in tube A i s equal to the f l ow rate in tube B , thus rnflow ?A T LA A ?AB t LB B l 1/JB 1 out flow or nr? pf g (?A - ?AB) 8 nLA Rearrangement of equat ion (0.2) g ives = + 159 . CD . 2) + (D . 3 ) Thus , once the potential at the j oin has been found using equat ion ( 0 . 3) , equat ion ( 0 . 1 ) may be used to find Q. APPEND I X E BROMIDE CONCENTRATI ON MEASUREMEN1S FOR I ND IV IDUAL REPL ICATES UNDER CONTINUOUS POND ING COND ITIONS I Tab I e E . 1 Bromide concentrat ion measurements for i nd iv idua l rep l i cates under continuous ponding cond i t ion i n the pasture area . A l og -normal dist ribution was assumed . PLOT ( a J Pr(' leach i ng Bromide COIH.:cntra t i on (JJg/m l ) - exp (xL) exp ( sL) IJepth X C . V . ( 111111 ) Core I I I I l l I V V ( JJg/m l ) % t-- 0- 1 0 - 549 504 764 396 553 538 1 . 3 4 . 3 I l l ?? ? 2 1 1 1 1 ??? 711 H!i 1 00 84 2 . 0 1 5 . 4 L0- 30 20 8 1 20 1\6 55 52 43 2 . I 1 9 . 3 3 0 - 4 0 t:\ 62 4 87 1 2 35 21 3 . 6 42 . 3 4U-SU 1 4 1 3 2 5 1 5 1 0 8 2 . 4 4 3 . 0 50-75 2 1 2 2 8 3 6 4 2 . 3 56 . 8 75- 1 00 2 3 2 4 4 3 3 1 . 6 44 . 6 Post - Leaching Bromi de ( lJg/m l ) - exp (xL) exp ( sL) Depth concentrat i on X c . v . (mm) Core I ! I I l l I V V V I V I I V I I I I X X X I X I I X I I I (lJg/m l ) % 0- 20 1 1 9 96 1 89 1 49 1 30 1 24 1 6 1 1 1 7 223 1 55 160 108 166 154 150 1 . 3 4 . 5 20-40 90 96 148 7 1 46 38 48 40 1 02 97 1 20 47 90 79 7 3 1 . 6 1 0 . 6 40-60 59 68 1 2 2 47 4 5 36 56 27 1 1 4 4 1 89 35 33 59 so 1 . 6 1 2 . 0 60-RO S6 90 67 I ll 1 02 4S n ?0 1 611 3 1 1 07 57 1 8 66 56 2 . 0 1 7 . 2 80- 100 29 97 24 3 1 84 26 46 33 70 30 63 6? 1 3 5 3 38 2 . 8 2 8 . 4 I OU- 1 20 23 ss 8 2 1 6 1 10 1 9 25 2 1 24 I S 62 5 33 1 9 ? . I 38 . 5 1 20- 1 40 1 3 1 3 4 5 1 46 22 33 23 1 2 1 1 4 ? 40 3 25 1 4 3 . 0 42 . 1 140- l bO 1 2 !l 2 9 1 43 3 1 42 26 9 1 2 5 28 3 25 1 3 3 . 2 44 . 8 1 60- 1 80 1 2 1 I 7 8 1 1 0 32 64 1 5 8 9 1 5 1 3 3 24 14 2 . 6 36 . 0 180- 200 7 7 1 0 8 85 36 1 1 7 7 8 8 1 3 1 7 3 25 1 2 2 . 9 42 . 6 200-225 7 7 1 2 5 25 34 96 7 5 8 1 5 24 4 19 12 2 . s? 37 . 3 225- 250 1 9 ?7 6 4 1 3 27 50 9 5 5 1 2 1 8 4 14 10 2 . 2 34 . 2 250- 300 6 27 6 3 14 1 7 35 5 6 7 1 2 1 2 3 1 2 2 1 . 3 35 . 3 PLOT ( b ) Pre- leach ing Bromide concentrat i on ( lJg/m l ) - exp (xL) exp ( sL) Depth X c . v . (mm) Core 1 1 1 I l l I V V ( lJg/m l ) % 0 - 1 0 602 882 452 74 1 884 7 1 2 691 1 . 3 4 . 4 1 0 - 20 3 1 6 1 1 1 67 102 72 33 6 . 8 54 . 3 20-30 44 - 1 108 7 40 14 7. 2 74 . 3 30-40 1 2 3 1 3 6 5 5 2 . 7 85 . 1 40-50 2 0 0 0 0 0 X X X 50-75 0 0 0 0 0 0 X X X 75- 1 00 0 0 0 0 0 0 X X X Post - l each i ng Dep th Bromide concentrat i on (lJg/m l ) X exp (xl.) exp (s l. ) c . v . (mm) Core I ! I I l l IV V V I V I I V I I I I X X X I X I I X I I I (lJg/m l ) % 0-20 l3b 1 64 95 89 1 33 1 1 9 1 1 3 1 1 0 103 1 59 1 4 9 1 5 1 1 92 1 32 1 29 1 . 3 4 . 8 20-40 97 84 48 56 - 73 94 99 69 7 1 1 25 1 27 1 1 7 88 85 1 . 4 7 . 0 40-60 92 57 25 4 1 8 7 69 79 88 6 1 56 104 96 69 7 1 67 1 . 5 9 . 4 60-80 64 - 69 28 50 85 58 72 82 64 86 75 92 53 68 70 1 . 4 7 . 6 80- 100 4 1 70 33 4 1 8 3 48 67 1 02 62 1 1 3 80 14 3 35 71 64 1 . 6 1 1 . 2 100- 1 20 1 8 80 27 5 1 85 42 80 70 59 104 74 1 1 7 1 8 63 6 1 1 . 9 1 5 . 6 1 20 - 1 4 0 8 5 1 J(t 53 71 36 58 7 1 ss 29 5 1 8 8 7 43 35 2 . 4 24 . 6 140- 1 60 4 45 5 42 74 1 3 68 52 46 28 67 9 1 7 42 29 3 . 0 32 . 5 1 60 - 1 80 I 47 6 35 72 6 56 27 37 1 8 63 93 9 36 22 3 . 7 42 . 4 1 80- 200 2 40 1 2 29 6 1 3 42 20 29 1 3 81 108 5 34 20 3 . 5 4 1 . 7 200-225 2 40 1 3 1 7 49 3 45 9 30 1 5 63 84 4 29 1 8 3 . 4 42 . 7 225- 250 4 . 38 1 2 1 6 3 1 6 27 10 20 1 3 67 55 5 23 15 J.. 4 32 . 0 250- 300 5 24 4 1 3 2 1 4 1 8 9 1 2 1 6 6 1 42 6 18 1 9 2 . 7 33 . 9 CONTINUED . . . 16 l . 1 62 . Tab 1 e E . 1 ( con t i nucd) I'LOT [ c ) l ' r l? - l t: at:h i ng llepth Brom i uo concen t rat i on l lJ&/m l ) X exp (xL) exp ( s L ) c . v . (nun) Core I 1 1 I l l IV V ( 1Jg/m 1 ) \ 0- 1 0 442 5<>5 4!! 1 74 1 4<>5 539 529 1 . 2 3 . 3 10-20 248 96 65 3<> 9 91 54 3 . 4 30 . 7 20-30 54 52 42 23 3 35 24 3 . 2 36 . 5 30-40 26 1 1 1 3 6 2 4 . 5 191 . 8 40-50 2 1 2 0 I 1 5 1 6 . 8 520 . 9 50-75 0 1 0 0 0 0 + + + 75- 100 0 0 0 0 0 0 + + + Post - l ca<.: h i ng Depth Bromide concentrat ion ( lJg/m l ) ?; exp (xL) exp (sL) C . V . (nun) Core I I I I I I IV V V I V I I V I I I I X X X I X I I X I I I (1Jg/m 1 ) \ 0-20 208 1 84 207 152 172 154 2 1 0 109 145 20 1 184 1 8 1 1 4 7 1 7 3 1 7 1 1 . 2 3 . 7 20-40 86 36 35 50 88 57 99 7 48 52 40 107 22 56 49 2 . 1 19 . 1 40-60 1 1 8 27 1 5 22 38 1 1 4 38 4 32 1 8 1 8 91 33 44 3 1 2 . 5 26 . 9 60-80 68 28 10 3 1 4 84 22 1 5 3 30 53 36 28 1 5 3 . 7 48 . 3 80- 100 19 7 3 2 1 2 4 1 3 1 2 9 1 1' 53. 32 1 3 7 3 . 4 62 . 5 100- 1 20 24 3 3 0 9 0 3 1 0 39 1 54 25 1 2 2 10 . 7 27 . 3 1 20 - 1 40 3 3 I 4 0 25 0 0 0 0 48 0 4 7 2 4 1 4 + + + 1 40- l hO 34 () <> 0 2(> 0 0 () u 32 0 40 24 1 2 + + + 160- 180 35 0 6 0 29 0 0 0 0 34 0 32 26 1 2 + + + 1 80- 200 28 0 8 0 29 0 0 0 0 47 0 25 24 1 2 + + + 200-22S 2Y () 1 2 () J :l u () ( I () 4 :1 () 23 14 l U + + + 225- 250 2 1 0 14 0 23 0 0 0 0 4 1 0 - 6 9 + + + 250- 300 1 8 0 1 1 0 - 0 0 0 0 32 0 - 3 6 + + + + As In 0 i s not defined , these parameters cannot be computed . Symbo l s used in the Table are defined in sect ion 6 . 4 . 3 . I I I - r;?b l c E . :! Bromide concentrat i on measurements for indi v i dua l rep l i cates under cont inuous ponding cond i t ions in the cu l t i vat ed ;?rea . A J og -norma l d is tr ibut ion was assumed . ?\.!!.. Pre - leach ing ll rom iJ<' ...:un.._?.._? n t rat i on ( IJg/m l ) - cxp (xL) cxp ( sL) lk"p t h X c . v . l nun ) Cu re l 1 1 I l l I V V ( IJg/m l ) % t l - I l l i.>OO 744 600 6!!2 71l? 683 679 1 . 1 1 . 9 l l l - 20 (J J 3 2 1 J 3 .\1:1 1 8 1 5 2 . 0 26 . 2 2i l - 30 3 20 1 3 2 1 4 1 0 7 2 . 9 54 . 6 ."i0 - 40 4 3 I I ? 2 4 2 . 5 64 . 4 ? I l l - :.o 2 3 0 I ::, 2 2 2 . 8 2 50 . 9 5 0 - 7 5 2 2 2 2 2 2 2 1 . 1 1 7 . 2 75 - 1 00 3 3 2 I 2 2 2 1 . 6 86 . 7 l ' ost - l cach i ng --- llcp t 11 Brom i de concentra t i on (IJg/m l ) X exp (xL) exp ( sL) c . v . (mm) Co re I 1 1 I l l I V V V I V I I V I 1 1 I X X X I X I I XI I ! ( IJg/m l ) % U - 20 297 302 287 308 2 1 7 205 22<> 244 300 2 34" 268 2 2 1 286 26 1 259 1 . 2 2 . 6 20-40 2 10 262 1 89 292 1 8 9 14 5 294 276 408 2 1 7 288 16 1 258 245 1 55 1 . 3 5 . 2 40- b O 4 1 66 4 0 69 1 2 7 4 3 1 56 1 37 1 26 1 04 1 19 24 86 88 75 1 . 8 1 3 . 6 b il - l:lO 4 5 4 7 2 8 6 2 3 29 71\ 39 1 2 1 1 1 1 9 1 0 3 . 2 49 _ 6 8 0 - 1 00 3 ?? 1 2 3 1 2 5 5 3 4 3, 1 3 6 3 2 . 8 9 9 . 3 ? 100- 1 20 3 I 3 1 4 3 1 2 2 1 9 3 2 1 2 3 5 3 2 . 5 80 . 0 1 2il - 1 40 2 4 1 9 4 I J 1 1 0 6 54 2 7 8 4 3 . 3 95 . 0 1 40- 1 60 3 3 1 2 4 4 2 1 1 2 5 1 7 94 1 20 1 5 5 4 . 7 95 . 8 160- 1 80 4 1 2 6 3 1 1 1 1 6 1 37 1 8 5 2 1 38 1 7 8 4 . 2 68 . 1 1 80- 200 1 1 26 16 4 1 1 3 1 1 7 2 1\ 9 2 4 4 8 3 3 1 2 2 2 3 . 5 1 99 . 7 2 00 - 2 2 5 22 3 1 24 44 1 0 3 2 3 2 46 3 4 1 1 6 "19 2 1 I S 2 . 7 36 . 9 225 - 250 1 1 ; 1 4 1 8 40 9 3 1 3 1 2 4 2 ? 30 26 8 1 4 1 8 1 5 2 . I 26 . 8 2 5 0 - 300 1 3 I 9 1 2 I 2 7 6 3 1 4 3 2 2 1 9 24 4 1 5 1 3 2 3 2 1 . 3 1 02 . 9 I PLOT ( b ) Pre - l each i ng IJcpth llrom i de concent rat ion (IJg/ml J X cxp ( xL) cxp ( sL) C . V . (mm) Cor<' I 1 1 I l l I V V ( pg/m 1 ) % 0 - Hl 38h (,R.1 795 709 (>RI\ 6 5 2 389 1 . 3 4 . 4 1 0- 20 9 1 6 647 1 04 20 1 5 9 4 5 5 . 7 4 5 . 6 2U- 3U 2 1 1 349 !! I 74 t! 1 I . 4 1 1 4 . 8 3 0 - 4 0 I 9 60 6 I 1 5 ? 7 . 1 1 40 . 1 40- 50 11 9 1 2 8 0 6 2 1 2 . 0 5 98 . 5 5 0 - 7 5 0 I 2 1 0 I ? + + 7S - J llll 0 I I 0 1 1 ( ) . ? . Post - l each i ng llepth Bromide concent rat i on (IJg/m l ) X exp (xL) exp ( sL) c . v . (mm) Core I I ! ! I T IV V V I V I I V I I I I X X X I X I I X I I I ( 1Jg/m1) % 0 - 2 0 1 69 1 9 3 3 4 2 2 3 1 335 269 1 94 257 232 !67 1 90 220 1 39 226 2 1 2 1 . 3 4 . 9 ::!0- 40 b7 64 200 l i S 232 1 44 l OS 1 90 ! 6 1 90 92 25 1 so 1 35 1 1 7 1 . 68 1 0 . 9 40-60 3 2 29 40 65 53 33 38 45 39 3 1 9 1 1 5 26 4 3 1 0 1 . 8 1 6 . i t.0 - 80 24 7 3 1 34 2 9 I 1 8 6 " 7 I 29 33 1 7 1 0 3 . 6 54 . 7 8 0 - 1 00 8 1 1 1 7 2 1 83 I J I 1 24 I 7 4 7 1 7 6 5 . 2 R9 . 7 1 00- 1 20 K 5 s I 1 91l 1 I 0 0 (J I 0 3 1 2 0 3 5 . 3 1 36 . 8 1 20- 1 4 0 2 7 2 5 I 238 2 I 0 4 1 0 0 8 2 2 2 1 0 . 4 4 86 . 8 1 40- 1 60 45 0 2 2 I 1 69 1 1 () I I 0 I 1 4 3 23 2 1 2 . 5 333 . 1 160 - 1 80 22 0 70 0 1 24 20 0 22 1 2 33 0 33 2 5 3 1 7 . 4 235 . 9 I B0 - 200 1 8 1 0 69 0 1 1 0 ss 0 59 0 25 79 0 2 1 34 6 1 9 . 2 8 1 60 . 3 2110 - 2 25 7 1 5 2 66 1 7 80 74 1 8 8 5 6 5 2 6 1 1 0 60 so 4 1 2 . 6 25 . 8 2 2 5 - 250 64 4 70 7 7 54 6 3 S I 8 8 3 h 4 9 33 5 1 4 so 38 2 . 76 28 . 0 250- 300 2 1 2 39 79 so 7 2 fifi 63 59 1 9 3 7 59 0 4 2 2 7 4 . 1 4 2 . 9 ? As I n 0 i s not def i ned , these parameters cannot be computed . Symbo l s used i n the Tab le are defined i n sect i on 6 . 4 . 3 . 1 6 3 . APPEND I X F FRACTI LE D IAGRAM CONSTRUCT ION 1 6 5 . Deta i l s concern i ng the const ruc t i on of a fract i l e J i ag ram have been g i ven by Hald ( 1 952 ) and B i ggar and N1 e l sen ( 1 976) . The diagram i s haseJ on t he cumu l a t i ve d i s t r i but i on funct i on P (x ) fo r a no rma l d i s t r i bu t i on : P (x ) ( x -x) / s f -oo ( F . 1 ) where i n thi s case random var i ab l e x i s the concentrat i on of bromide i n the so i 1 ( C , or l n C ) x the me::tn va lue of C or ln C , and s the standard , deviat ion . The values of P (x) a re obtained by rank ing the observed values of x in i ncreas ing order of magni tude . Pre- leaching data at 30-40 mm depth are g iven here as an examp le . P ( x) i s equa l to i/n , where n = 48 , the tota l number of the samp les , and i = 1 , 2 , 3 . . . , For the l argest observed value of x , P (x ) = 1 and for any value of x 4 7 , 48 . l e s s than the smal lest observed value , P ( x ) 0 . By us ing these values of P ( x ) , the corresponding values of ( x-x) / s obtained from the tabul ar va lues of Eq . ( F . 1 ) (Mood and Grayb i l l , 1 963) were p l otted versus C and ln C . The l inear i ty of the resu l t i ng data was determined us i ng l inear regression ana lys is . I n this examp l e , the probab i l ity uni t s were more l inearly re l ated to ln C (r = 0 . 993 ) than C ( r = 0 . 808) as shown i n F ig. o . 1 0C , thus bromide concentrat i on variab i l i ty in the soi l after app l i cat ion to the so i l surface was approximate ly log-normal ly distributed . For a normal d is tributi on , the mode (the most frequent ly observed va lue) , the median ( the value with an equal number of values above and be l ow) , and the mean are al l identical . For a l og-normal d i stributi on , the mode , median and mean are obtained from : mode exp ( xL 5 2 ) L 5 . 4 ]..lg/ml median = exp (xL) 2 2 . 2 )..lg/ml mean exp (xL t 0 . Ss? ) 45 . 0 )..lg/m l where xL and sL for ln C were 3 . 1 and 1 . 2 ]..lg/ml , respective l y . The resu l ts o f the analys i s for other l eaching data under natural rainfal l conditions are g iven in Tab l e F . 1 and F . 2 . Tab l e F . 1 Depth (Imn) n Pre- l eaching 0 - 10 48 1 0 - 20 4 8 20-30 48 30-40 48 40-50 48 1 66 . Resu l t s from fract i l e d iagrams for experiment under natural rainfal l . - I Mode I exp (xL) I Medi an - lexp ( s L) X c . v . r 1 r2 % (]Jg/ml ) 0 . 975 0 . 98 7 4 1 3 346 389 380 1 . 4 1 5 . 8 0 . 9 1 7 0 . 979 14 1 9 3 1 23 1 1 7 1 . 70 1 1 . 0 0 . 8 1 7 0 . 980 65 24 47 4 1 2 . 25 2 1 . 0 0 . 808 0 . 993 45 5 22 24 3 . 28 38 . 3 0 . 745 0 . 980 4 1 3 1 6 14 3 . 9 1 49 . 0 Post - leaching (after 1 8 2 JJUn excess rainfal l ) 0 -50 40 0 . 92 3 0 . 986 30 . 5 2 2 . 3 2 7 . 4 24 . 6 1 . 5 8 1 3 . 8 50 - 100 40 0 . 878 0 . 978 1 3 . 1 1 1 . 3 1 2 . 5 1 2 . 0 1 . 37 1 2 . 4 100- 150 40 0 . 9 1 9 0 . 968 1 1 . 4 1 0 . 4 1 1 . 0 1 0 . 4 1 . 30 1 1 . 0 1 50 - 200 40 0 . 889 0 . 978 1 2 . 4 1 0 . 7 1 1 . 8 1 1 . 1 1 . 36 1 2 . 5 200 - 250 4 0 0 . 969 0 . 992 1 1 . 9 1 0 . 9 1 1 . 5 1 1 . 4 1 . 2 8 1 0 . 0 250 - 300 40 0 . 964 0 . 998 1 2 . 6 1 1 . 0 1 2 . 0 1 1 . 6 1 . 36 1 2 . 2 300-400 ? 40 0 . 955 0 . 979 14 . 3 1 1 . 6 1 3 . 3 1 2 . 3 1 . 46 14 . 6 400-500 40 0 . 934 0 . 985 14 . 3 1 1 . 9 1 3 . 4 1 2 . 8 1 . 4 2 1 3 . 6 500-600 40 0 . 957 0 . 984 1 1 . 1 8 . 9 1 0 . 3 9 . 8 1 . 46 16 . 4 r 1 and r2 = corre lation coefficients for (x-x) /s vs concentrati on C and ln C , respect ive ly , n = number of samples . Median = median . obtained from experimental data . Other symbo l s used in the tab l e are defined in secti on 6 . 4 . 3 . 1 67 . Tab l e F . 2 Data for bromide concentrat ion d i stribut ion after infi l trat ion of 46 mm excess rainfal l over evapotranspirati on . Four core samp l es obtained from 0 . 32 x 0 . 5 m subp lots were bu lked t ogether i n each repl i cate . Depth Range - exp ( s L) exp (x L) c . v . n* X (mm) % ( l-!g/m l ) 0 - 20 1 0 37- 105 7 1 1 . 38 68 7 . 6 20-40 10 2 7 -49 38 1 . 16 38 4 . 0 40-60 10 26-60 37 1 . 26 36 6 . 5 60-80 10 22 -59 38 1 . 36 37 8 . 6 80- 1 00 10 2 1 - 73 44 1 . 45 4 1 1 0 . 0 1 00 - 1 2 0 10 2 2 - 70 48 1 . 45 44 9 . 8 1 20- 140 1 0 2 2 -63 47 1 . 4 1 so 8 . 8 1 40 - 160 10 2 5 -67 48 1 . 36 47 8 . 0 160- 1 80 1 0 28-67 49 1 . 35 49 7 . 7 180- 200 10 2 8 -6 1 45 1 . 3 3 43 7 . 6 200 - 250 1 0 24- 70 40 1 . 44 38 1 0 . 0 1 50 -300 1 0 14-60 32 1 . 64 29 14 . 7 * n = number of samp les . Other symbo l s are defined in sect i on 6 . 4 . 3 . REFERENCES 1 69 . Abdal l a , N . A . , and B . Lear . 1 975 . Determinat ion of inorgani c bromide in soi l and p lant t i ssues w i th a bromide sel ective- i on e l ectrode . Commun . in Soi l Sc ience and P l ant Ana lys i s . 6 : 489-494 . Al l i son , F . E . 1 966 . The fate of n itrogen app l i ed t o soi l s . Advan . Agron . 1 8 : 2 1 9 - 258 . Anderson , J . L . , and J . Bouma . 1973 . Relat ionships between saturated hydrau l i c conductivity and morphometric data of an Agri l l i c hori zon . Soi l Sci . Soc . Am . Proc . 37 : 408- 4 1 2 . Anderson , J . L . , and J . Bouma . 1977a . Water movement through pedal soi l s . I . Saturated f low . Soi l Sc i . Soc . Am . J . 4 1 : 4 1 3-4 1 8 . Anderson , J . L . , and J . Bouma . 1 977b . Water movement through pedal soi l s . I I . Unsaturated f low . So i l Sci . Soc . Am . J . 4 1 : 4 19-423 . Ay lmore , L . A . G . , and M . Karim . 1968 . Leaching of fert i l i zer ions in soi l co lumns . Int . Congr . Soi l . Sci . Trans . 9th (Ade l aide , Aust . ) , 1 : 143- 1 5 3 . Baker , F . G . 1 9 78 . Variabi l ity of hydrau l i c conductivity wi thin and between nine Wiscons in soi l series . Water Resour . Res . 1 4 : 103- 108 . Ba lasubramanian , V . , Y . Kanehi ro , P . S . C . Rao , and R . E . Green . 197 3 . F i e l d s tudy of so lute movement in a highly aggregated oxi so l w i th intermi ttent fl ooding . I . Ni trate . J . Environ . Qua l . 2 : 359- 362 . Barrow , .N . J . , and T . C . Shaw . 1975 . The s l ow reactions between soi l and anions : 2 . Effect of t ime and temperature on the decrease in phosphate concentration in the . so i l solut ion . Soi l Sci . 1 19 : 167- 1 7 7 . Barrow , N . J . 1978 . The descript ion of phosphate adsorption curves . J . Soi l Sci . 29 : 447-462 . B iggar , J . W . , and D . R . Nie l sen . 1 962 . Miscible d isp lacement ; I I . Behaviour of tracers . Soi l Sci . Soc . Am . Proc . 26 : 1 2 5 - 1 2 8 . 1 70 . Bi ggar , J . W . , and D . R . N ie l sen . 1 963 . Miscible d isp lacement : V . Exchange processes . Soi l Sc i . Soc . Am . Proc . 2 7 : 6 23-627 . Bi ggar J . W . , and D . R . N i e l sen . 1 967 . Miscible d isp lacement and l eaching phenomenon . p . 254- 2 74 . In R . M . Hagan et a l . ( eds . ) , Agronomy No . 1 1 , Am . Soc . Agron . , Mad i son , Wisconsin . B i ggar , J . W . , and D . R . Ni e l sen . 1 9 76 . Spat ia l variabi l i ty of the l each ing character i s t i cs of a fi e l d soi l . Water Resour . Res . 1 2 : 78-84 . B l ack , C . A . 1968 . So i l -p l ant re l at ionships . John Wi l ey , New York . S lake , G . , E . Sch l icht ing , and U . Zirnrnerman . 1 97 3 . Water recharge in a soi l with shrinkage crack s . Soi l Sci . Soc . Am . Proc . 37 : 669-672 . Bo l ton , E . F . , J . W . Ay l esworth , F . R . Hare . 1970 . Nutrient losses through t i l e drains under three cropping systems and two fert i l i ty l eve l s on Brookston c lay l oam . Can . J . Soi l Sci . 50 : 2 75- 279 . Boswe l l , F . C . , and O . E . Anderson . 1970 . Nitrogen movement compari sons in cropped versus fa l l owed soi l s . Agron . J . 62 : 499-503 . Bouma , J . , and J . L . Anderson . 1 9 7 3 . Re lat ionships between soi l structure characterist ics and hydrau l i c conductiv ity . F i e ld soi l water regime . Soi l Sci . Soc . Am. Spec . Pub l . No . 5 . Am. Soc . of Agronomy . Mad i son , Wiscons in . p . 7 7 - 105 . Bouma , J . , and J . L . Anderson . 19 7 7 . Water and chloride movement through soi l co lumns simulat ing peda l soi l s . Soi l Sc i . Soc . Am. J . . 4 1 : 766- 770 . Bouma , J . , L . W . Dekker , and H . L . Ver l inden . 1977a . Drainage and verti ca l hydraul ic conduct ivity of some Dutch "knik" c l ay soi l s . Agricul tural Water Management . Vo l . 1 : 6 7 - 78 . 1 7 1 . Bouma , J . , A Jongerious , 0 . Boesma , A . Jager , and D . Schoonderbeek . 19 77b . The funct ion of d i fferent types of macropores during saturated f low through four swe l l ing soi l hori zons . Soi l Sci . Soc . Am. J . 4 1 : 945 -950 . Bouma , J . , and L . W . Dekker . 1 978 . A case study on infi l tration into dry c lay soi l . I . Morpho logica l observati on . Geoderma . 20 : 2 7-40 . Bouma , J . , L . W . Dekker , and J . H . M . Wosten . 1 978 . A case study on infi l trat ion into dry c l ay soi l . I I . Physical measurement . Geoderma . 20 : 4 1 -5 1 . Bouma , J . , and J . H . M . Wosten . 1979 . F l ow patterns during extended saturated flow in two , undi sturbed swe l l ing c l ay soi l s with di fferent macrostructures . Soi l Sci . Soc . Am . J . 4 3 : 1 6 - 2 2 . Bowden , J . W . , S . Nagaraj ah , N . J . Barrow , A . M . Posner , and J . P . Quirk . 1 980 . Describing the adsorption of phosphate , c itrate , and selenite on a variab le charge mineral surface . Aust . J . Soi l Res . 1 8 : 49-60 . Bower , C . A . , and L . V . W i l cox . 1965 . Solub le sa lts . In Methods of Soi l Anal ys i s . Agronomy . No . 9 . , p . 9 33-95 1 , Am . Soc . Agron . , Madi son , Wiscons in . Bowler , D . G . 1 980 . The drainage of wet soi l s . Hodder and Stoughton , Sydney . pp . 25? . Bres l er , E . , and R . J . Hanks . 1 969 . Numerical method for estimating simu l taneous flow of water and sa l t in unsaturated soi l s . Soi l Sci . Soc . Am . Proc . 33 : 8 2 7 -832 . Brewer , R . 1 964 . Fabric and mineral ana lys i s of soi l ? . John Wi l ey and Sons . I nc . , New York . 470 pp . Burns , I . G . 1 974 . A mode l for predict ing the redi stributi on of sa lts app l i ed to ?al lowed soi l s after excess rainfal l or evapotranspi rati on . J . Soi l Sci . 25 : 165- 1 7 8 . 1 72 . Burwe l l , R . E . , D . R . T immons , and R . F . Ha l t . 1975 . Nutrient transport in surface run -off as j n f l uenced by so i l cover and seasona l periods . Soi l Sci . Soc . Am . Proc . 39 : 5 23-528 . Cagauan , B . G . , Jr . , L . S . Lau , R . E . Green , and G . Uehara . 1962 . So lute d i spersi on in two Haw i i an soi l s under saturated f low . Trans . J . Meet . Comm . V I . , Int . Soc . Soi l Sci . p . 1 85 - 1 94 . Ca lvert , D . V . 1975 . N i trate , phosphate , and potass ium movement into drainage l ines under three so i l management systems . J . Environ . Qua l . 4 : 1 83- 1 86 . Cameron , D . R . , and A . K lute . 1 9 77 . Convect ive-di spersive so lute transport with a combined equi l ibrium and kinetic adsorption mode l . Water Resour . Res . 1 3 : 1 83- 188 . Cameron , D . R . , C . G . Kowalenko , and K . C . !?arson . 1 978 . Ni trogen and ch l oride d i stribut ion and ba lance in a c lay loam soi l . Can . J . Soi l . Sci . 58 : 7 7 - 88 . Cameron , D . R . , C . G . Kowalenko , and C . A . Campbe l l . 1979 . Factors affecting n itrate n itrogen and chl oride l eaching variabi l i ty in a fie l d p lot . Soi l Sci . Soc . Am. Proc . 4 3 : 455-460 . Casse l , D . K . 1 97 1 . Water and so lute movement in Svea l oam for two water management regimes . Soi l Sci . Soc . Am . Proc . 35 : 859 - 866 . Casse l , D . K . , T . H . Krueger , F . W . Schroer , and E . B . Norum . 1974 . Solute movement through d i sturbed and undisturbed soi l cores . Soi l Sci . Soc . Am . Proc . 38 : 36-40 . Casse l , D . K . , M . Th . van Genuchten , and P . J . W ierenga . 1975 . Predicting an ion movement in d i sturbed and undi sturbed soi l s . Soi l Sci . Soc . Am. J . 39 : 1 0 1 5 - 1 0 1 9 . Cho , C . M . , J . Strong , and G . J . Racz . 1970 . Convect ive transport of orthophosphate ( P - 3 1 and P-32 ) in several Mani toba soi l s . Can . J . Soi l Sci . 50 : 303- 3 1 5 . 1 73 . Chi lds , E . C . 1969 . The phys ical bas i s of soi l water phenomena . John Wi l ey , New York . pp 493 . Coat , K . R . , and B . O . Smith . 1964 . Dead-end pore volume and d i spers ion in porous media . Soc . Pet . Eng . J . 4 : 73-84 . Corker , R . B . 1977 . Saturated hydrau l i c conductivity measurement on the Tokomaru s i l t l oam . D i p . Agri . Sci . Thes i s , Mas sey Univers ity , New Zealand . Corey , J . C . , D . R . N i e l sen , and D . K irkham . 1967 . Mi scib l e d i sp lacement of nitrate through so i l columns . I bid . 3 1 : 49 7-50 1 . Couchat , Ph . , F . Bris saud , and J . P . Gayraud . 1980 . A study of strontium- 90 movement in a sandy so i l . Soi l Sci . Soc . Am . J . 44 : 7- 1 3 . Crank , J . 1956 . The mathematics of di ffus ion . C larendon Press , Oxford , Eng l and . Davidson , J . M . , C . E . Re ick , and P . W . Sant l eman . 1968 . Influence of water flux and porous material on the movement of se lected herbicides . Soi l Sci . Soc . Am. Proc . 32 : 629-633 . Davidson , J . M . , and R . K . Chang . 1 9 7 2 . Transport of p ic loram in re l at ion to soi l phys ical condi t ions and pore-water veloci t y . Soi l Sci . Soc . Am . Proc . 36 : 2 5 7 - 26 1 . Davidson , J . M . , and J . R . McDougal . 1 973 . Experimental and predicted movement of three herbicides in a water saturated soi l . J . Environ . Qual . 2 : 428-433 . Day , P . R . 1 956 . Di spers ion of a moving salt -water boundary advancing through saturated sand . Trans . Am. Geophys . Union . 37 : 595 -60 1 . Dayananda , P . W . A . , F . P . W . Winteringham , C . W . Rose , and J . Y . Parlange . 1980 . Leaching of a sorbed solute : A mode l for peak concentrati on di sp l acement . I rrigation Sci . 1 : 169 - 1 75 . 1 74 . de Camargo , O . A . , J . W . B i ggar , and D . R . N i e l sen . 1 9 79 . Transport of inorganic phosphorus in an A lfi sol . Soi l Sci . Soc . Am . J . 4 3 : 884- 890 . Duxbury , J . M . , and J . H . Peverly . 1978 . N i trogen and phosphorus losses from organic soi l s . J . Environ . Qua l . 7 : 566-5 70 . Eh lers , W . 1973 . Observat i on on earthworm channe l s and infi l trat i on on t i l led and unt i l l ed loess soi l . Soi l Sci . Soc . Am . Proc . 37 : 242- 249 . E l rick , D . E . , and L . K . French . 1966 . M i sc ib le d isp lacement patterns on d i sturbed and undi sturbed soi l cores . So i l Sc i . Soc . Am . Proc . 30 : 1 5 3- 156 . Evans , G . , and I . Levin . 1969 . The di stribut i on of salt by infi l trat ion of water into dry soi l and sand . Aust . J . So i l Res . 7 : 2 1 - 2 7 . Evans , T . D . , and J . K . Syers . 1 97 1 . An app l i cat i on o f autoradiography to study the spherical di stribut i on of 3 2 P - l abe l l ed orthophosphate added to soi l crumb . Soi l Sci . Soc . Am . Proc . 35 : 906 -909 . see below Fox . R . L . , E . J . Kamprath . 1 970 . Phosphate sorpt i on i sotherm for eva luating the phosphate requirement of soi l s . Soi l Sci . Soc . Am . Proc . 32 : 902-907 . Gandar , P . W . , and P . E . H . Gregg . 1 979 . Des i gning a l ong-term nitrogen ba l ance experiment . Proceeding of the Agronomy Soc iety of New Zealand . 9 : 29-34 . Gardner , W . R . 1965 . Movement of ni trogen in soi l . p . 550- 5 7 2 . I n Bartho lomew , W . V . , and F . E . C l ark . ( eds . ) , Soi l Ni trogen . Pub l . Am . Soc . Agron . , Inc . , Mad i son , Wi scons in . Gast , R . G . , W . W . Ne l son , and J . M . MacGregor . 1974 . N itrate and ch loride accumu lat i on and di stribut ion in fert i l i zed t i le-drained soi l . J . Environ . Qual . 4 : 406- 4 1 2 . Foth , D . H . 1 9 7 8 . Fundamentals of soil science . 6 th Edition . John Wiley and Sons . New York . pp . 436 . 1 75 . Ghuman , B . S . , S . M . Verma , and S . S . Prihar . 1975 . Effect of appl i cation rate , init ia l soi l wetnes s , and red i s tribut i on t ime on sa lt d i sp l ace? ment by water . Soi l Sci . Soc . Am . Proc . 39 : 7 - 1 0 . G l over , J . , and M . D . Gwynne . 1962 . Li ght rainfa l l and p l ant survival in East Africa . I . Mai ze . J . Eco l . 50 : 1 1 1 - 1 1 8 . G l over , P . E . , J . G lover , and M . D . Gwynne . 1962 . Li ght rainfal l and p l ant survival and E . Afri ca . I I . Dry grass land vegetat ion . J . Eco l . 5 0 : 1 99 -206 . Green , R . E . , P . S . C . Rao , and J . C . Corey . 1 972 . So lute transport in aggregated soi l s : Tracer zone shape in re lation to pore-ve locity di s tribut ion and adsorption . Symposium on Fundamental of Transport Phenomenon in Porous Medi a . Proc . 2nd 2 : 732 - 75 2 . Gunary , D . 1 964 . Phosphate di ffus ion i n intact soi l . 8th I ntern . Congres s of Soi l Science , Bucharest , Roman i a . 5 73 -5 7 8 . Hal l , J . K . , and N . L . Hartwig . 1 978 . Atraz ine mob i l ity in two soi l s under convent ional t i l l age . J . Environ . Qual . 7 : 6 3-68 . Hudson , A . W . , H . G . Hopewe l l , and D . G . Bow ler , and M . W . Cross . 1 96 2 . The drainage of farm l ands . 2nd Ed . Massey Co l l ege , Palmerston North . Humphreys , F . R . , and W . L . Pri chett . 1 9 7 1 . Phosphorus adsorption and movement in some sandy forest so i l s . Soi l Sc i . Soc . Am . Pro? . 35 : 495-500 . John , M . K . 1972 . Factors affect ing the adsorpt ion of tagged phosphorus by soi l s . Communicat ion in Soi l Sci . and P lant Analys i s . 3 : 1 97 - 205 . Johnston , W . R . , F . I t t ihadieh , R . M . Daum . and A . F . P i l l sburry . 1 965 . N itrogen and phosphorus in t i l e drainage effluent . Soi l Sci . Soc . Am . Proc . 29 : 287- 289 . Jongerius , A . 1957 . Morphometric investi gations on soi l structure . Vers l . Landbouwkdi . Onder . 6 3 : 1 2 . 1 76 . Kanchanasut , P . , D . R . Scatter , and R . W . Ti l l man . 1 9 78 . Preferent ia l so lute movement through l arger soi l voids . I I . Experiments w ith saturated so i l . Aust . J . Soi l Res . 1 6 : 269- 276 . Kay , B . O . , and D . E . E l rick . 1 967 . Adsorpt i on and movement of l indane in soi l s . Soi l Sci . 1 04 : 3 1 4 -3 2 2 . Kirda , C . , D . R . N i e l sen , and J . W . Bi ggar . 1 9 7 3 . of ch l oride and water during infi l trat ion . Proc . 3 7 : 339- 345 . Simultaneous transport Soi l Sci . Soc . Am . Ki rkham , D . , and W . L . Powers . 1 97 2 . Advanced soi l phys i cs . Wi l ey? Intersc ience , New York . pp . 5 34 . K i sse l , D . E . , J . T . Ri tchie , and E . Burnett . 1 9 7 3 . Chl oride movement in undi sturbed swe l l ing c l ay soi l . Soi l Sci . Soc . Am . Proc . 37 : 2 1 - 24 . K i sse l , D . E . , J . T . Ritchie , and Earl Burnett . 1974 . N i trate and ch loride l eaching 1n a swe l l i ng c l ay soi l . J . Environ . Qua l . 3 : 4 0 1 - 404 . Krupp , H . K . , and D . E . E l rick . 1 969 . Dens i ty effects in miscib l e d i sp l acement experiment . So i l Sc i . 1 0 7 : 37 2 - 38 0 . Krupp , H . K . , J . W . B iggar , and D . R . N i e l sen . 1 .9 7 2 . Re l ative flow rates of sa l t and water in so i l . Soi l Sci . Soc . Am . Proc . 36 : 4 1 2 - 4 1 7 . LaF l eur , K . S . , W . R . McCask i l l , and D . S . Adams . 1 974 . Movement of Prometryne through Congaru soi l into ground water . J . Environ ? Qua l . 4 : 1 3 2 - 1 33 . LaF l eu r , K . S . , G . A . Woj cck , and W . R . McCask i l l . 1 972 . Movement of toxaphene and fluometron through Dunbar soi l to underlying ground water . J . Envi ron . Qua l . 2 : 5 15 - 520 . Lapidus , L . , and N . R . Amundson . 1 95 2 . Mathemat ics of adsorption in beds . V I . The effect of l ongi tudinal di ffus i on in ion exchange chromatograph ic columns . . J . Phys . Chem . 56 : 984-988 . 1 7 7 . Lawes , J . B . , J . H . Gi lbe rt , and R . Warington . 1 882 . On the amount and composit ion of the rain and drainage waters co l l ected at Rothamsted . Wi l l i am C l owes and Sons , LTD . London . Levin , I . 1 964 . Movement of added nitrates through soi l co lumns and und i sturbed so i l profi l e s . l?ans . 8th l nt . Congr . Soi l Sci . ( Bucharest ) , 4 : 1 0 1 1 - 1 022 . L i nd st rom , F . T . , L . Roc rsma . 1 97 1 . 'fhc theory of mas s t ransport of previous ly d i stributed chemica l s i n a water saturated sorb ing porous med ium . Soi l Sc i . 1 1 1 : 1 9 2 - 1 99 . Loveday , J . 1974 . Methods for so i l analysis of irrigated soi l s . Tech . Commun . No . 54 , Commonwea lth Agricul tura l Bureaux , Farnham Royal . MacGregor , A . N . , J . K . Syers , M . A . Turner , and D . G . Bow ler . 1975 . Land disposal of untreated dairy shed effluent at Massey Univers ity . New Zealand Soi l News . 2 3 : 38-40 . MacLean , A . J . 1 9 77 . Movement of ni trate n itrogen with di fferent cropp ing system in two soi l s . Can . J . Soi l Sci . 5 7 : 2 7 - 33 . Manse l l , R . S . , H . M . Se l im , P . Kanchanasut , J . M . Davidson , and J . G . A . r i s k e l l . 197 7 . Experimenta l and s imulated t ransport of phosphorus through sandy so i l s . Water Resour . Res . 1 3 : 1 89 - 194 . Ma son , D . D . , J . F . Lut z , and R . G . Peterson . 1 957 . Hydrau l ic conductivity as real ted to certain soi l propert ies in a number of great so i l groups , samp l ing errors invo lved . Soi l Sci . Soc . Am. Proc . 2 1 : 554- 560 . McAul i ffe , K . W . 1 978 . The effect of land d i sposal of dairy factory waste on soi l propert ies . Master Thes i s . Massey Univers ity , New Zealand . McMahon , M . A . , and G . W . Thomas . 1 9 74 . Ch loride and tri t i ated water f low in d isturbed and undi sturbed soi l cores . Soi l Sci . Soc . Am. J . 38 : 7 2 7 - 732 . 1 7 8 . Meek , B . O . , L . B . Grass , A . J . McKenz ie . 1 969 . App l ied n itrogen l osses in re l at ion to oxygen status in soi l s . Soi l Sci . Soc . Am. Proc . 33 : 575 - 578 . Mi l l er , R . J . , J . W . Bi ggar , and D . R . N ie l sen . 1965 . Ch l oride d i sp lacement in Panache c l ay l oam in re lat ion to water movement and d i stribut ion . Water Resour . Res . 1 : 6 3- 7 3 . Minsha l l , N . , M . S . N i cho l s , and S . A . Wi t ze l . 1 969 . P l ant nutrients in base fl ow of streams in Southern Wisconsin . Water Resour . Res . 5 : 706- 7 1 3 . Mood , A . M . , and F . A . Graybi l l . 1 96 3 . Introduction of theory of stat i st i cs . McGraw-Hi l l , New York . pp . 43 1 . Murphy , J . , and J . P . Ri l ey . 1 962 . A modi fied s ingle solut i on method for the determinat ion of phosphate in natural water . Ana l . Chim . Acta . 27 : 3 1 - 36 . New Zealand Soi l Bureau Staff . 1 968 . Soi l of New Zealand . Part I . Soi l Bureau Bul l . No . 26 ( 1 ) . ( DS IR : We l l ington) . N i e l sen , D . R . , and J . W . Biggar . 1 96 1 . M iscib l e d i sp l acement in soi l : I Experiment informat i on . Soi l Sci . Soc . Am . Proc . 25 : 1 - 5 . N i e l sen , D . R . , and J . W . B i ggar . 1 96 2 . M iscible d i sp l acement : I l l Theoret i ca l consi derat ions . Soi l Sci . Soc . Am . Proc . 26 : 2 16- 2 2 1 . N i e l sen , D . R . , J . W . Biggar , and K . T . Erh . 1973 . Spatia l variabi l i ty of fie ld measured soi l -water propert ies . Hi l gadi a . 42 : 2 15 - 259 . O l sen , R . J . , R . F . Hen s l er , O . J . Attoe , S . A . W i t ze l , and L . A . Peterson . 1970 . Ferti l i zer nitrogen and crop rotat ion in re l at ion to movement of ni trate nitrogen through soi l profi les . Soi l Sci . Soc . Am . Proc . 34 : 448 - 452 . Omoti , V . , and A . Wi ld . 1979 . Use of fluorescent dyes to marke the pathways of so lute movement through soi l s under l eaching condit i ons : 2 . F i e l d experiments . Soi l Sci . 1 2 8 : 98- 104 . 1 79 . Oster , J . D . , L . S . Wi l lardson , and G . J . Hoffman . 1972 . Sprink l ing and ponding techniques for rec laiming sal ine soi l s . Trans . ASAE . 1 5 : 1 1 15 - 1 1 1 7 . Passioura , J . B . 1 9 7 1 . Hydrodynamic di spersion in aggregated media . I . Theory . Soi l Sci . T I I : 339- 344 . Pol l ok , J . A . 1975 . A comparat ive study of certain New Zea land and German soi l s formed from loess . Ph . D . thes i s , Friedrich-Wi lhe lms? Universi tat , Bonn . Priest l ey , C . H . B . , and R . J . Tay l or . 1 9 7 2 . On the assessment surface heat flux and evaporat i on us ing large-scal e parameters . Month . Weather Rev . 1 00 : 8 1 -92 . Qui senberry , V . L . , and R . E . Phi l l ips . 1976 . Percolation of surface app l ied water in the fie ld . Soi l Sci . Soc . Am . J . 40 : 484-489 . Qui senberry , V . L . , and R . E . Phi l l ips . 1 978 . D i sp lacement of soi l water by s imul ated rainfal l . Soi l Sci . Soc . Am. J . 4 2 : 675 -679 . Rao , P . S . C . , R . E . Green , V . Balasubramanian , and Y . Kanehiro . 1974 . F i e l d study of so lute movement in a h igh l y aggregated oxi so l wi th intermittent f lood ing . J . Environ . Qua l . 3 : 1 9 7 - 20 1 . Rao , P . S . C . , J . M . Davidson , R . E . Jessup , and H . M . Se l im . 1979 . Eva luati on of conceptual mode l s for describing nonequi l ibrium adsorption-desorpt ion of pest i c i des during steady f low in soi l s . Soi l Sci . Soc . Am. J . 4 3 : 2 2 - 2 8 . Rao , P . S . C . , R . E . Jessup , .D . E . Ro l ston , J . M . Dav idson , and D . P . Ki lcrease . 1980 . Experimental and mathematical descript i on of nonadsorbed solute transfer by di ffus ion in spherical aggregates . Soi l Sci . Soc . Am . J . 44 : 684 -688 . Rennes , L . , R . W . Ti l lman , J . K . Syers and D . G . Bow ler . 1976 . Effects of mo l e drainage on surface run-off from a soi l under pasture . N . Z . J . Agric . Res . 20 : 45-59 . 1 80 . Ritchie , J . T . , D . E . Ki sse l , and E . Burnett . 1 972 . Water movement in undi sturbed swe l l ing c l ay soi l . Soi l Sci . Soc . Am. Proc . 36 : 874- 879 . Robinson , R . A . , and R . H . Stokes . 1959 . E l ectro l yte so lution . 2nd Edn . Butterworths , London . Rokowski , A . S . 1 972 . Watershed physics : Soi l variabi l ity cri teria . Water Resour . Res . 8 : 1 0 1 5 - 102 3 . Rose , D . A . , and J . B . Passioura . 197 l a . The ana lysis of experiments on hydrodynamic di spersion . Soi l Sci . 1 1 1 : 2 5 2- 257 . Rose , D . A . , and J . B . Pass ioura . l97 l b . Gravity seggregation during mi sc ib le di sp l acement experiments . Soi l Sci . 1 1 1 : 2 58- 265 . Ryden , J . C . , J , K . Syers , and R . F . Harri s : 1973 . Phosphorus in run-o ff and streams . Advan . Agron . 25 : 1 - 45 . Ryden , J . C . , J . R . McLaughl in , and J . K . Syers . 1 9 77 . Mechanisms of phosphate sorption by soi l s and hydrous ferri c oxide ge l . J . Soi l Sci . 28 : 72-92 . Saffigna , P . G . , C . B . Tanner , and D . R . Keeney . 1976 . Non-uni form infi l trat ion under potato canopies caused by interception , s temflow , and hi l l ing . Agron . J . 68 : 337- 342 . Saffi gna , P . G . , D . R . Keeney , and C . B . Tanner . 1 97 7 . Lys imeter and fie l d measurements o f chl oride and bromide l eaching in an uncultivated loamy sand . Soi l Sci . Soc . Am . J . 4 1 : 4 78- 482 . Scotter , D . R . 1978 . Preferent ia l so lute movement through l arger soi l voids . I . Some computat ions us ing simp l e theory . Aust . J . Soi l Res . 1 6 : 2 5 7- 267 . Scotter, D . R . , B . E . C l othier , and R . B . Corker . 1979a . Soi l water in a Fragiaqual f . Aust . J . So i l Res . 1 7 : 443- 453 . 1 8 1 . Scotter , D . R . , B . E . C l othier , and M . A . Turner . 19 79b . The soi l water balance in a Fragiaqual f and i t s effect on pasture growth in centra l New Zea land . Aust . J . Soi l Res . 1 7 : 455 -465 . Scott -Russe l , R . 197 7 . P lant root systems : The ir funct i on and interact ion wi th the so i l . McGraw-Hi l l . London . pp . 298 . Se l i m , H . M . , P . Kanchanasut , R . S . Manse l l , L . W . Z e lazny , and J . M . Dav idson . 1974 . Phosphorus and ch loride movement i n a spodoso l . So i l and Crop Sci . Soc . F la . 34 : 1 8 - 2 3 . Se l im , H . M . , J . M . Davidson , and R . S . Manse l l . 1976 . Evaluat i on of a two s i te adsorpt i on-desorpt ion mod e l for describing so lute transport in soi l . Proc . of Summer Computer Simu lei;t i on Conf . , Washington , D . C . Sharp ley , A . N . , and K . J . Syers . 1 979a . Phosphorus i nput into a s tream draining an agri cul tural watershed . I I Amounts contributed and re lat ive s i gni ficance of run -off types . Water , Air , and Soi l Polut ion . 1 1 : 4 1 7 -428 . Sharp ley , A . N . , and J . K . Syers . 19 79b . Loss o f n i trogen and phosphorus in t i l e drainage as influenced by area app l i cat i on and graz ing ani ma l s . N . Z . Journal of Agri cul tura l Research . 2 2 : 1 2 7- 1 3 1 . Shuford , W . J . , D . D . Fri tton , and D . E . Baker . 1978 . N itrate -nitrogen and ch loride movement through undisturbed fie l d soi l . J . Environ . Qua l . 6 : 255 - 259 . Skopp . J . , and A . W . Warrick . 1974 . A two-phases mode l for the mi sc ib l e displacement of react ive so lutes in soi l s . Soi l Sci . Soc . Am . 3? : 545 -550 . So i l Survey Staff . 1 975 . Soi l taxonomy . Handb . No . 436 . (U . S . D . A . : Washi ngton ) . Smi th , S . J . , and R . J . Dav i s . 1974 . Re lat ive movement of bromide and n i trate through soi l s . J . Environ . Qua l . 3 : 1 5 2 - 155 . Spencer , W . F . 195 7 . D i stribut i on and avai labi l i ty of phosphate added to a Lake land fine sand . Soi l Sci . Soc . Am . Proc . 2 1 : 1 4 1 - 144 . 1? .? Starr , J . L . , and J . Y . Par l ange , 1976 . So l ute di spersi on i n saturated soi l co lumns . Soi l Sci . 1 2 1 : 364- 372 . Syers , J . K . , M . G . Brownman , G . W . Smi th , and R . B . Corey ? 1 9 73 . Pho sphate sorption by soi l s eva luated by the Langmuir adsorption equat ion . So i l Sc i . Soc . Am . Proc . 37 : 358- 363 . Tal sma , T . 1 96 7 . Leaching of t i l e-drained sal ine soi l s . Aust . J . Soi l Res . 5 : 37- 46 . Taylor , N . H . 1956 . Di scussion on measurement of soi l moi s ture . Proceedings of the con ference on soi l moi sture . N . Z . Department of Scienti fi c and I ndustria l Research . pp . 1 7 4 . Tay lor , S . A . and G . L . Ascroft . 1 972 . Phys i ca l edapho l ogy : The phys i c s of i rrigated and nonirri gated soi l s . W . H . Freeman and Company , San Franc i s co . pp . 533 . Thomas , G . W . , and A . R . Swoboda . 1 970 . Anion exc lusion effects on chloride movement in soi l s . Soi l Sc i . 1 10 : 163- 1 66 . Thomas , G . W . , and R . E . Phi l l ips . 1979 . Consequences of water movement in macropores . J . Environ . Qua l . 8 : 149- 1 5 2 . Turner , M . A . , D . R . Scotter , D . G . Bow ler , and R . W . Ti l lman . 1976 . Water harvesting : The concept and use in a humid c l imate . Proceed ings of Soi l and P l ant Water Symposium . Palmerston North . p . 168- 1 74 . van De Pol , R . M . , P . J . Wierenga , and D . R . Ni e l sen . 1 97 7 . So lute movement in a fie l d soi l . Soi l Sci . Soc . Am. J . 4 1 : 1 0- 1 3 . ? van Genuchten , M . Th . , J . M . Davidson , and P . J . Wierenga . 1974 . An eval uation of k inetic equi l ibrium equat ions for the prediction of pest icide movement in porous media . Soi l Sci . Soc . Am. J . 38 : 29- 35 . van Genuchten , M . Th . , and P . J . Wierenga . 1 9 76a . Mas s transfer studies in sorbing porous med i a : I Ana lyt ical so lutions . Soi l Sci . Soc . Am . J . 40 : 4 73- 480 . 1 8 3 . van Genuchten , M . Th . , and P . J . Wi erenga . 19 76b . Numerical so l ution for convect ive di spers ion with intra-aggregate di ffus ion and non- l inear adsorpt i on . p . 275 - 292 . ? G . C . van Steenkiste ( ed . ) System Simulation in Water Resources North Ho l l and Pub l . Co . , Amsterdam . van Genuchten , M . Th . , and P . J . Wi erenga . 1 97 7 . Mass transfer studies in sorbing porous media : 1 1 . Experimental eva luation with tritium ( 3H20) . Soi l Sc i . Soc . Am . J . 4 1 : 2 7 2 - 278 . van Genuchten , M . Th . , P . J . Wi erenga , and G . A . O ' Connor . 1977 . Mas s trans fer studies in sorbing porous medi a : I l l . Experimenta l eva luation with 2 , 4 , 5 -T . So i l Sci . Soc . Am . J . 4 1 : 2 78- 285 . Warri ck , A . W . , J . W . Bi ggar , and D . R . N ie l sen . 1 97 1 . Simu ltaneous so lute and water transfer for an unsaturated soi l . Water Re sour . Res . 7 : 1 2 16- 1 2 25 . Wetse laar, R . 1 962 . N itrate dis tribution in trop ica l soi l s . I l l . Downward movement and accumu l ation of n itrate in the subso i l . P l ant and Sc i . XV I : 19- 3 1 . White , R . E . , and A . W . Tay l or . E ffect of pH on phosphate adsorption and i sotop i c exchange in acid soi l s at l ow and high additions o f sol ub l e phosphate . J . Soi l Sc i . 2 8 : 48-6 1 . Wi l d , A . 1 972 . Nitrate leaching under bare fal l ow at a s i te in northern Nigeria . J . Soi l Sci . 2 3 : 3 1 5 - 324 . Wi l d , A . , and I . A . Babiker . 1-9 76 . The assymetric l eaching pattern .of n i trate and chloride in a l oamy sand under fi e l d condit ion? . J . Soi l Sci . 2 7 : 460- 466 . Wi l l iams , R . E . , and D . W . A l l man . 1 969 . Factors affecting infi l tration and recharge in a l oess covered bas in . J . Hydrol . 8 : 265- 28 1 . Wi l son , L . G . , J . N . Luthi n , and J . W . Bi ggar . 196 1 . Drainage sal inity investigations of the Tul e l ake l ease l ands . Ca l . Agr . Expt . Sta . Bul l . 7 79 . 1 84 . Z irnmerman , U . , K . O . Munn i ch , and W . Roether . 1 96 7 . Downward movement of soi l moi sture traced by means of hydrogen i sotopes . Geophys i ca l Monograph . No . 1 1 . Amer-Geophys . Uni on . p . 2 8 - 36 .