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. LANDSCAPE ECOLOGY OF THE RANGIPO DESERT A thesis presented in partial fulfilment of the requirements for the degree of ~ ·'2..,L.e_ Master of -Philo..:g}?fl-y in Soil Science at Massey University Andrew Mark Purves MA1r11r,y1, ur,1i1R1111I11111iii1R y 1990 1061750612 Massey University Library CHAPTER 1 GENERAL INTRODUCTION 1 . 1 Locat i on Of The Stud y Area In 1847, William Colenso, missionary, explorer, and botanist made his first crossing of the Rangipo Desert in foul weather: "At 3pm we crossed the sand desert called Te Onetapu (Rangipo)- a most desolate weird­ looking spot, about 2 miles wide where we crossed it; a fit place for Macbeth's witches or Faust ' s Bracken scene!" Colenso (1884) was describing a landscape clothed by tussock, dunefields, barren gravelfields and deeply scoured watercourses. The Rangipo Desert is situated on the exposed south-eastern portion of Mount Ruapehu's ring plain. Mount Ruapehu is New Zealand's largest andesitic massif and its summit is the highest peak in the North Island, reaching 2,797m. To the east of the Rangipo Desert uplifted ranges of indurated sandstone, argillite, and conglomerate make up the Kaimanawa Range (Grindley, 1960 ) . The field study was , in the main, confined to the west and south of the Whangaehu River, to the north of the Waihianoa Aqueduct, and to the east of Makahikatoa Stream (Fig. 1 .1 ) . The field area, therefore, covers about 40km 2 , most of which was gazetted to the New Zealand Army in 1943 (Birch, 1987). The altitude wi thin the study area gradually increases from 910 metres above sea level at Waihianoa Aqueduct, to 1200 metres above sea level at the north-eastern margin of the study area . The landscape of the Rangipo Desert is unique for its contrasts within a few kilometres. In one area eroding light yellow tephra, colonized by only scattered hardy plants, abuts against healthy enclaves of N. solandri; while, in another area only a kilometer away, dunes colonized by Phyllocladus and Halocarpus bidwillii are interpersed across a grey lag pavement covered by scattered boulders and even fewer plants. These contrasting features make the task of delineating the landscape into discrete regions simple. 1 Crater Lake RUAPEHU . .. . . . . ... . . . ~ . . . . . ~ Iv. ·• • .--. . . . ' <:Jt . •• • ~ /01') • • • • • • •• ~ Q/ •• • • ..... ~ /:, • •. • • V q,..f •• ''~ ~s~ ~\>'~ t>-1~ ~~ ~\)~ j\)"" \>'\)G ss 0 Dmu .. ~i Der 0 ~. .. . 0 . 0 0 Dmup 1 0 Den 5 . . . . . . () oo" ;;o· Dcrt • . . o. C).: Dmu • " 0 O· • . - Dmup " 0 " 0 tJ 0 50cm . 0 • "'0 Dmut .o • .\· 'i·): 0 . . • . . ".1 ·:~·.J:,.'l· Dmut 0 0 Dmup 0 9 2 Plate 2.7 Recent stream channel cut exposes 4 deposits typical of lithofacies assemblage 1. Lithofacies Dmu (base of the spade) is the common lithofacies found in the topographically lower, wide channels on the active fan. Fluvial re-working has occurred in much of the upper part of the sequence. Dmu = matrix supported, ungraded diamicton . Den = c las t supported , normal graded d iamicton. Der = c l ast supported, reverse graded diamicton . Dcu = clast s upported, ungraded diamicton. 53 Ocu Ocr Omu Lithofacies assemblage 2 (Plate 2.3) is found on the topographical highs of t he active fan and also on the inactive fan within wide channels. This assemblage suggests that large high- competence debri-flows have over­ topped these topograhical highs, or over-topped The Chute, and became unconfined . The result is a loss of velocity behind the flow head, and the stranding of boulders in a thin matrix. Rodolfo et al. (1989) found a similar type of facies at Mount Mayon where boulders up to Sm in intermediate diameter projected from surface deposits that were less than 2m thick in over-bank areas . Li thofacies assemblage 3 (Plate 2.8) is found on topographic highs ad jacent to the wide channels described above. The presence of these deposits show that the large high-competence debri-flows that over -topped The Chute, and became unconfined, were also large enough to deposit debris on some of these adjacent higher areas . However, the lithofacies suggest that these over-flows were small and dilute, thus indicating it wa s only the marg i ns of the large lahars that reached these areas . Boulders associated with these flows are assumed either eroded or buried under fluvial and more recent debris-flows in the wide semi-confined channels. Lithofacies assemblage 4 (Plate 2.9) is found in a relict channel on the s outhern margin of the inactive fan. The assemblege suggests that debris­ f lows in this area were dilute. The boundary between diluted debris-flow deposits and hyperconcentrated flow deposits is poorly understood (Pierson and Scott, 1985; Smith,1986; Costa, 1988), however, the absence o f internal stratification and the presence of outsized clasts with d iscernible orientation, indicates that lithofacies Dmup is formed from debris-flows, albeit diluted. Dilution has occurred because debris from t hese flows was stranded as the flows became very thin up stream in the semi-confined channels before reaching this area. This lithofacies was not observed at a similar distance from source along the Whangaehu River, indicating that the river channel was the major artery for large and competent lahar flows down the catchment . 54 Pla te 2.8 Photograph of section which represents l ithofacies assemblage 3 located on t opographic highs on the inactive f an. The Taupo Pumice ign imbrite i s also preserved i n these areas and can seen in the photograph between 1 .3m and the bottom of the photograph. The two over-bank deposits, typical of lithofacies assemblage 3, are located between 1 .1m amd 1 .3m. The lower deposit is a clast-supported diamicton whilst the upper deposi t is a matrix supported diamicton. 55 Plate 2.9 Photograph showing the three top debris-flow deposits belonging to lithofacies assemblage 4 . Dilution of the debris-flow deposits is evidenced by the small number of outsized clasts in a poorly sorted matrix. Fine fluvial deposits found by the pen, just below 1 .Orn, and at the pocket knife separ a t e each deposit. Note the erosive base of the top debris- flow deposit by the pen . Overlying the top depos i t are Makahikatoa Sands which extend to the surface of the pr ofile . 56 2.4 Conclusions The constructional landscape of the Rangipo Desert has been inundated by ten major lahars since the Taupo Pumice eruption of c. 1764 yrs B.P. Six of these events are shown to have over-topped The Chute, and evidence strongly suggests another 4 events also over-topped The Chute. Minimum volume estimates of the deposits show that debris-flow deposit 6 was the largest and deposited at least 8.9 x 10 6 m3 of debris, but quite possibly 12.0 x 10 6 m3 . This is between two and three times the volume of any other deposit. The 1975 lahar's total volume, calculated from a discharge hydrograph of the Whangaehu River, was 1 . 86 x 10 6 m3 (Page and Patterson, 19 ) , yet the volume of the resultant deposit is estimated at only 0.09 x 10 6 m3 . This gives a 1 : 20 ratio of the total volume to amount deposited in the study area for the 1975 lahar . This ratio has been applied to the other deposits, and the results indicate that the cumulative total volumes of major lahars which crossed the study area from~- 1764 yrs B. P. t o 160 yrs B.P. far exceeds earlier predictions for the Whangaehu catchment. The potential hazard of a future lahar causing damage in the Whangaehu catchment area is, therefore, higher than earlier predicted . Periodicity of major lahars to inudate the study area since the Taupo Pumice ignimbrite is 1 every 160 years . However, between 850 yrs B.P and 665 yrs B.P., the periodicity increased to 1 every 37 years . Since 500 yrs B.P., only one major lahar has inundated the Rangipo Desert . Debris­ flow deposit 6, the largest deposit recorded in the study area, is estimated to have been deposited by a lahar ~- 800 yrs B.P. It is therefore believed to correlate to the lahar deposit found further downstream dated at 756 ± 56 yrs B.P. (Campbell, 1973). Approximately 50- 70% of the total post-Taupo Pumice debris transported by the major lahars was deposited during the period between 850 yrs B.P. and 665 yrs B.P . This was a major period of aggradation on the constructional surface, both as lahar deposits and subsequent fluvial deposits. Before and after this period fluvial degradation was the dominant process. Today, floods and smaller lahars are cutting channels into the debris-flow deposits and the older aggradational fluvial deposits. 57 Most of the lahar deposits in the Rangipo Desert are debris-flow deposits. The lithofacies associated with these deposits have been grouped into 4 assemblages that were found in six regions of the study area. Lithofacies assemblage 1 shows that both large high-competence and small dilute flows were common on the topographical lows on the active fan. Lithofacies assemblage 2 suggests large high-competence flows inundated the topographical highs on the active fan and wide channels on the inactive fan. Where flow volumes were exceptionally large, small dilute over flows washed from the wide channels onto adjacent topographical highs to deposit lithofacies assemblage 3. As these flows continued down the wide channels, boulders became stranded in a thin matrix resulting in a dilution of the flow. These diluted flows then re­ channeled down slope to bulk (evidenced by lithofacies assemblage 4) and continue flowage throughout the entire Whangaehu catchment to the sea. 58 3.1 General Introduction CHAPTER 3 AEOLIAN DEPOSITS The aeolian sands in the Rangipo Desert have added diversity to both the physical landscape, in the form of dunes, and to the biotic landscape by providing additional unique vegetation associations. A relationship between the aeolian sands and the vegetation was discussed by Cockayne (1908) who recognised that the trapping of aeolian sand by plants such as Muehlenbeckia axillaris and Podocarpus nivalis were crucial for dune fixing to begin . Cockayne states: "A shrub whose seeds can germinate, producing a seedling that can tolerate the station, may arrest such a drift, and, provided it can increase by rooting near its growing-point, will build up smaller or larger mounds where other plants can settle" ...... "A slightly less unstable substratum or a more sheltered position and true desert dunes are formed." . Cockayne also recognised that curious vegetated mounds, which were still accumulating aeolian sand, represented the remnants of a once stable landscape. The presence of these mounds, referred to here as "pedestals", in the Rangipo Desert and on the slopes of eastern Tongariro led Topping (1974) to suggest that current erosion was largely due to Polynesian deforestation (chapter 1). In this chapter the post-Taupo Pumice aeolian deposits are formally designated as the Makahikatoa Sands. The depositional and weathering history of the Makahikatoa Sands have been characterised using stratigraphy and laboratory procedures. Laboratory procedures are also used to compare the physical and chemical fertility of the Makahikatoa Sands with that of an incipient soil found on the fluvial/laharic surfaces. The results in this chapter, together with the findings of chapter 2, provide new information for the initiation of the present day erosion found in the Rangipo Desert. 59 SECTION A: Depositional And Weathering History Of The Makahikatoa Sands 3.2 Methods 3.2.1 Introduction One of the major factors controlling the weathering of tephra or aeolian sand is the time weathering processes are able to operate effectively (i.e., time at land surface). Thus, if deposition rates of the tephra or aeolian sand increase, the opportunity for weathering and pedogenesis decreases (Gibbs,1971) . The depositional and weathering rates of the Makahikatoa Sands were studied by measuring and describing stratigraphic sections , and then analysing for carbon, allophane and phosphate retention from selected samples in the laboratory. Carbon provides a measure of the time period that has permitted accumulation of organic matter, whilst allophane provides a measure of the time period that has permitted formation o f short-range order clays from the primary minerals (Lowe, 1986). An indirect measure of the allophanic content of the soil is also provided by analysing the soil's phosphate retention (Saunders, 1965 ) . Particle size analyses of sand, silt and clay fractions were performed to elucidate the provenance of the Makahikatoa Sands, to provide information on the environment of deposition, and t o quantify rates of weathering . 3.2.2 Laboratory Procedures Sections were described in the field' and then selected samples were taken back for laboratory analysis . A representative sample was taken from each sample bag and the procedures described below were followed f or each analysis. Ratings given to any analyses in this study follow the criteria of Blakemore et al. (1987). ' Methods used to describe stratigraphic sections are given in section 2. 2. 1 . 60 Tam.ms Oxalate-Extractable Iron, Silica, And Aluminium 61 The acid oxalate shaking extraction method as described by Blakemore et al. (1987) was used. A 1+5 diluent was used for the Al and Si extracts, while a 1+5 and a 1+9 diluent was required for the Fe extracts. The extracts were analysed using an aa spectrophotometer. Pyrophosphate-Extractable Iron And Aluminium The method described by Blakemore et al. (1987) was used. Extracts were cleaned by centrifuging at 14000 rpm. for 50 minutes and the addition of 'superfloc'. The extracts were analysed using an aa/ae spectrophotometer. Allophane was estimated using the formula of Parfitt (1986). Carbon Gravimetric determination of the% organic carbon was measured us ing a Leco high induction furnace (after Nelson and Sommers, 1982.) Carbon (multiplied by 1 .7) provides an approximate measure of organic matter content (Blakemore and Millar, 1968) . Phosphate Retention This method followed that of Blakemore et al. (1987). Particle Size Analyses In this study chemical dissolution , using acid-oxalate extraction of short range order c lays and organic complexes (SROCO), was employed as a pre-treatment procedure for particle size analysis (after Alloway, 1989). Literature suggests that acid-oxalate selectively dissolves SROCO materials that are composed of allophane or ferrihydrite, and if conducted in the dark, dissolution of crystalline oxides and layer silicates is very limited (see Alloway p . 243, 1989). Organic matter was removed initially using hydrogen peroxide, but later this pre-treatment was abandoned because differences were found to be negligible between treated and untreated samples. The samples were air dried and gently disaggregated so as to pass through a 1mm sieve. Two representative 3 gram sub-samples were respectively placed into two 250ml centrifuge bottles, and each bottle was filled with 0.2M acid oxalate (pH 3.0-3.5). These were shaken end over end in the dark at 20°C overnight. The samples were then poured through a 63µm sieve, thus separating the dissolved SROCO material, and the lattice clays and silt, from the sand fraction. The captured SROCO material, plus lattice clays and silt were left to settle for 12 hours. The dissolved SROCO material was then decanted from the residual clay and silt. The residual clay and silt fraction is then transferred with water to centrifuge tubes, and is separated at 850 rpm for a length of time, depending on the temperature (hence the viscosity) of the liquid. The suspended clay is decanted off and the procedure repeated until the supernatent is relatively clear. The silt and clay fractions are dried and weighed. The sand fraction, and any silt that was not originally filtered, was wet sieved. SROCO content was calculated as the difference between the other fractions combined weight and the original weight . Therefore, Weight (or%) of sand+ unfiltered silt+ filtered silt+ lattice clays+ residual SROCO material= weight (or 100%) of sub-sample. Samples were duplicated or triplicated, where necessary. This method was reasonably fast, particularly since there was only a small percentage of lattice clays present. However, a major limitation is that in most of these samples the <63µm residue is too small for accurate pipette analysis. 62 3 . 3 Results And Discussion 3 . 3.1 Makahikatoa Sheet Sands And Dune Sands The Makahikatoa Sands found in the Rangipo Desert are described here as either sheet sands or dune sands . The distinction is based both on their extent and their geometry. The sheet sands typically occur over wider areas than the dune sands, although, they are found on small, eroding, pedestals in some localities . The physiography of the sheet-sand deposits is relatively flat (Plate 3.1), whereas the dune sands occur as discrete convex deposits (Plate 3 . 2 ) . A f ew of the larger dunes have developed bas ins as sands have built up at f aster rates around the periphery . Neither t he dune sands nor the shee t sands f ound in t he s tudy area exhibit any di agnostic bedding f eatur es. The dunes sands cannot be grouped and classified into specific dune types because of the absence of cross-bedding and no prefer red al i gnment di rections . The pr esence of intact Tufa Trig members in most dunes confirms Cockayne ' s ( 1908) observation t hat the dunes were stationary f eatur es that could increase i n s ize , providi ng pl ant s were present to ensur e stability. There are a number o f large dunes , upwar ds of 5 meters in height , but med i um sized dunes ( >1m in height) are more numerous , and a mult i tude of s ma l l dunes are pr esent. Sheet - s and deposits overlie the Ma ngatawa i Tephra and the Taupo Pumice i gnimbrite preserved in non-eroded areas . The dune sands usually overlie the laharic or fluvial deposits but also have formed occasionally on areas where the Mangatawai Tephra and the Taupo Pumice ignimbrite are preserved . The dune sands have also formed on areas where tephric coverbeds have been eroded in recent times . Both the sheet-sand deposits and large dunes are covered by N. solandri and Phyllocladus forest. Phyllocladus scrub occurs almost exclusively on medium sized dunes whilst shrubs are found on small dunes. Seral shrublands and tussocklands have colonised both the sheet sands and dune s ands as a response to historical fire . 63 Plate 3.1 Characteristic uniform terrain formed by sheet sand deposits which overlie the post-Taupo Pumice paleosol. Plate 3.2 In contrast to the sheet sands, the dune sands form a distinctive dune like geometry. In this photograph dunes colonised with Phyllocladus or tussock provide relief to the otherwise barren laharic surface . Mount Ngauruhoe is in the background. 64 3.3 .2 Description Of Sheet Sands Type Section For The Makahikatoa Sands The type section (section 13 ) for the Makahikatoa Sands is located on a large sheet sand pedestal (Plate 3.3), immediately adjacent to a widespread erosional surface at T20/398064 . The formation name, Makahikatoa Sands, is named after the nearby Makahikatoa Stream (Fig. 1 . 1 ) which runs into the Whangaehu River (T20/390030). A f eature of the Makahikatoa Sands observed in the type section (Fig. 3 . 1, Plate 3.4) . is a 70cm thick post - Taupe Pumice paleosol developed in Makahikatoa Sands which overl ies the Taupo Pumice ignimbrite, and is in t urn overlain by a further 3m of Makahikatoa Sands. The boundary between the post-Taupo Pumice paleosol and the overlying Makahikatoa Sands is def ined at an unnamed ash, found immediately below Tufa Trig member 4 at a depth of 3.52m. The bottom 10cm of the paleosol has developed in the Taupe Pumice ignimbrite, while the rest of the paleosol has developed in the Makahikatoa Sands; f or simplicity the paleosol is referred t o as being developed from Makahikatoa Sands. Morpho l ogical differences between the paleos ol and the overlying sands were distinctive. The paleosol was distinguished by its : 1) yellower hue, 2 ) greasier texture, 3) firmer consistence, 4) moderate block structure, a nd 5 ) presence of rhizomorphs. Cracking, due t o differential shrinking, wh i ch occurs in some paleosols (Topping 1974) , is not present in the pos t-Taupo Pumice paleosol (hereaf ter referred t o as pTPp). In addition, a llophane (Fig. 3.2) and carbon (Fig . 3 . 3) percentages, measured from samples taken from the type section, were higher in the pTPp than the overlying sands. However, the decline in allophane and carbon percentages was no t inversely proportional to the corresponding 5 fold increase (0 .66mm/yr t o 3.8mm/yr) in the accumulation rate of the overlying sands (Fig. 3 . 4). Allophane percentages declined from a high of over 8 percent in the pTPp to about 5 to 6 percent in the overlying sands. Phosphate 65 Plate 3.3 The remnant pedes tal of N. solandri seen on the left of the photograph is the locality for the type s ect ion (section 13) of the Makahikatoa Sands. Part of the eroding surface seen in this photograph provides a stark contrast to the N. solandri forest enclaves. Section 15 is located near the outer margin of the stand of N. solandri to the right edge of the photograph . 66 67 Figure 3.1 Type section for the Makahikatoa Sands (!20/398064) Informal members of the Tufa Trig formation, numbered T.1, 1a, 2, 4, 5, 14, 15, 15a, 16, and 17, have been correlated throughout the Rangipo Desert. Tufa Trig members 4, 5 and 14-17 are black lapilli or ash deposits, while members T. 1 and 2 are light brown and orange pumiceous lapilli deposits, and T.1a is a cream and orange fine ash deposit. The dashed lines represent Tufa Trig members which were not correlated to other stratigraphic sections. A single dashed line represents a very thin unnamed ash. The arrow represents a lahar or fluvial deposit that appears to have admixed with Makahikatoa Sands. For a detailed description of the type section (section 13) see appendix 1. Symbols Makahikatoa Sands Paleosol developed in Makahikatoa Sands Black lapill i and ash Light brown or orange pumiceous lapilli and ash Taupo Pumice ignimbrite D B socm 2•76m 0 0 0 0 . ,, . 3 •11 m 3•52m -1~1 :- 1~ • • 4 • •• 0 T.17 T.16 T.15a T.15 T .14 ~ T.5 T.4 T .2 68 Plate 3.4 Type Secti on For The Makahikatoa Sands The Taupo Pumice ignimbrite i s f ound at t he base of the section. The bot t om arr ow poi n t s t o Tufa Trig member 1 which is one of the pumiceous tephra interbedded in the post-Taupo Pumice paleosol. The second and third arrows from the bottom point to bands of oxidised (red) iron which sandwich a band of r educed (grey) iron . These redox changes in i ron indicate latera l movement of wa t er t hrough the upper part of the pa l eosol . The f ourth arrow from t he bottom poi nts to Tufa Trig member 4 , a c oarse tephra deposited ~ - 850 yrs B.P . The boundary between the pal eoso l and the overlying Makahikatoa Sands i s found 10cm below Tuf a Tr i g member 4. The upper arrows point to Tufa Trig members that are interbedded with the Makahikatoa Sands. 69 0 Sm % allophane 0 2 4 6 8 0 I I ' I T.18 50 100 150 .- T.17 200 r 250 c r J 300 r T.5 r 350 ~ l 400 1 .- ' I I .... (cm) Depth Figure 3.2 Percent allophane1 for section 13 (type section) 0 0 50 100 150 T.17 200 250 300 -T.5 350 400 (cm) Depth T.18 % carbon 2 3 Figure 3.3 Percent carbon for section 13 (type section) 1 General explanation of the graphs 10 4 Vertical lines between the dot s on t he graphs represent the thickness and depth of each sample taken for analyses. The sloped lines represent the interval between samples. On some graphs no lines are s hown due to only a few sampling intervals . To help compare the graphs with the described stratigraphic sections in the text or appendix 1, some Tufa Trig members are i ndicated . On some graphs a question mark after a Tufa Trig member i ndicates that member is tentatively ident ified, based on stratigraphic position . The raw data for the graphs i s presented in appendix 2. 70 Accumulation Rate (mm/yr) (cm) 0 2 4 6 8 0 I 50 100 ~ 150 ~ 200 250 -- ·· -· · · ·· · · · · · ns 300 T.4 350 >- 400 450 500 I I · · · Tp + T.4 +T. 15 - Tp + T.4 Fig . 3.4 Accumulation rates' for section 13 (type sec tion ) 'Explanat ion of the above graph 10 The solid line Tp + T.4 means that two accumulation rates have been calculated. The first rate is based on thickness of the Makahikatoa Sands between the Taupo Pumice ignimbrite (Tp) dated at 1764 yrs B.P . and Tufa Trig member 4 (T . 4) with an estimated date of 850 yrs B.P. ie. 63mm 7 (1764-850) = 0.66 mm/yr. The accumulation rate between T.4 and the surface is calculated from a profile thickness of 2890mm divided by 850 yrs. (ie 850 yrs B.P. to the present) . The same procedure is done with the dotted line except Tufa Trig member 15 (estimated dated of 200 yrs B.P.) is included in the calculation. Therefore the profile thickness between T.4 and T.15 is measured and divided by the age: 850 yrs B.P . - 200 yrs B. P .) . The same procedure is done between T.15 and the surface. Some graphs shown later may have another time plane present (ie,. another dated Tufa Trig member) so a further rate can be included in the profile calculation. Other time planes are described in the post-Taupo Pumice chronology (Fig 2.5; section 2.3.2). 71 retention trends (appendix 1b) paralleled allophane as expected (Saunders, 1965); they declined from over 80 percent in the pTPp to 60 percent in the overlying sands. Similarly carbon levels declined from over 2 percent to between 1 and 2 percent in the overlying sands. Carbon levels are nonetheless low in the pTPp which suggests that organic cycling was limited despite the slower accumulation rates measured for the pTPp (Fig. 3.4). In the overlying Makahikatoa Sands at a depth between 3.11m and 2.56m the deposit becomes less sorted and coarsens with granules and one 2.5cm pebble present. These granules and the pebble was unexpected given the relatively large amount of pre-weathered sand also found within the deposit. An explanation could be that this deposit represents the distal edge of a lahar deposit . This is possible given that the deposit immediately overlies Tufa Trig member 5, which is known to have been deposited at the same time as when the largest post-Taupo Pumice lahar crossed the study area (chapter 2). This locality is found within the inferred distribution of this lahar (Fig. 2.4, Map B) . The high content of weathered silt and sands is assumed to be the result of pre-weathered Makahikatoa Sands being incorporated into the deposit as the margin of the flow crossed the landscape. Alternatively, the deposit may be of fluvial origin. It is possible that coarse material located up slope from the type section was re-worked, and deposited at the type section, during a period of surface runoff. The coarse deposit is also noted for the increased presence of re-worked pumice derived from the underlying Taupo Pumice ignimbrite (hereafter referred to as re-worked Taupo Pumice). From the coarse deposit to the surface, the Makahikatoa Sands grade into a soft, friable, and well sorted medium to coarse sand. Re-worked Taupo Pumice remains conspicuous to the surface. 72 A decline in carbon and allophane 1 percentages (Figs. 3.2, 3.3) indicate the lahar or fluvial deposit is comparatively less weathered than the aeolian deposits, but is still decisively more weathered than the primary tephra deposits. Regardless of whether the primary tephra deposits were laid down just after the Taupo Pumice eruption or in the last few hundred years, amorphous clay contents are almost non-existent. Moreover, very low residual carbon percentages, indicative of limited organic matter buildup, suggest the raw nature of the primary deposits are inimical to plants roots. Particle size analyses show (Fig. 3.5) a decrease in the silt fraction from 35% in the pTPp to 15-20% in the overlying Makahikatoa Sands . The amount of silt being deposited has probably not decreased in real terms, and in fact, may have increased; however, the proportionately greater increase in the deposition of fine to very fine sand from the surrounding eroding surfaces results in drop in silt percentage. The medium to coarse sand fractions increased dramatically as expected in the lahar or fluvial deposit but have also declined in the upper 2 metres of the profile to levels be low that of the pTPp. In contrast, the fine sand and particularly the very fine sands fractions have increased in the upper two metres of the profile. This indicates that the modal grain size of the source material is predominantly a fine or very fine sand. The decline in the medium to coarse fraction being deposited may be due to the pedestal surfaces becoming increasingly isolated from the surroundi ng deflating surfaces, thus making it increasingly difficult for the larger grains to saltate (Pye, 1987 ) up onto the surface of the pedestal. 1 Percentage allophane was calculated using the method of Parfitt (1986) involving acid-oxalate Al and acid-oxalate Si ratios (appendix 2). The ratios are consistently between 1 .6 and 2 in most deposits in this profile, although some ratios reach up to 3 in other sections. These Al:Si ratios suggest that the short range order clays formed in the study area are dominantly proto-imogolite allophane, with some silicate tetrahedra absent in the proto-imogolite allophane when ratios of >2 are present (Parfitt, 1986). 73 (cm) 0 0 100 200 300 400 500 Figure 3.5 Particle size distribution for section 13 (type section) 10 . - . ···· · t:_, .. ! T:::i, r Silt 63-125)Jm 20 % \ 30 125-250,um 250-1mm 40 T.5 T.1 74 50 Reference Sections For The Makahikatoa Sands Reference stratigraphic sections for the sheet-sand deposits are located adjacent to the laharic/fluvial surfaces, south-east of the type section, both near and on the fault escarpment (Plate 2.2). These sections, unlike the type section, are characterised by fine grained aeolian deposits that are less than a meter thick (Fig. 3.6). Profile thicknesses vary little between these sections which is reflected in a uniform physiography found in the areas where these deposits occur (Plate 3.1). In addition, the Tufa Trig members decrease in thickness and number as the section sites occur further from Mount Ruapehu. Accumulation rates for the Makahikatoa Sands were generally consistent between the reference sections (Fig. 3.7), and were five to seven times slower than in the type section. However, despite the decline in the accumulation rates there was no field indication that the sands had weathered to any greater degree due to increased pedogenesis. This was subsequently verified when allophane (Fig. 3.8) percentages were found to be lower in the reference sections compared to the type section. An allophane percentage measured in the pTPp for section 14 was almost identical to the allophane percentages in the pTPp of the type section but measurements in the overlying sands varied between 2 and nearly 4 % in section 14 compared to 6-7% in the type section. Carbon percentages also declined above the pTPp in section 14 (Fig. 3.9) but unlike the type section percentages near the surface increased, reflecting an increase in organic cycling. The similarity between the particle size distribution and profile thicknesses between section 1 and a sample from section 15, which occurs on the fault escarpment (appendix 3b), indicates the very fine sand and silt fractions are being transported across the landscape and deposited on the fault escarpment at comparable rates. However, local site conditions appear critical for deposition of the coarser grain size. For example coarser textures occur in section 14, compared to section (Fig. 3.10), despite the latter's closer proximity to the extensive laharic/fluvial surfaces (Plate 2.2). This has been attributed to a number of large dunes near section 1 that are acting as a local trap for the larger, saltating sand grains. 75 76 Figure 3.6 Correlation columns showing the change in thickness of the Makahikatoa sheet-sand deposits between section 13 (the type section) and reference sections 1, 14 and 15. Informal members of the Tufa Trig formation, numbered T.1, 2, 4, 5, and 7a, were correlated between these sections, while T. 16 and 17 were only found in section 13. A Tufa Trig member which is question marked is a tentative identification. Tufa Trig member 4 is the datum line for the columns. Deposits 01 and 02 represent clast supported and matrix supported diamictons respectively. For detailed descriptions of the sections see appendix 1. Symbols Makahikatoa Sands Paleosol developed in Makahikatoa Sands Black lapilli and ash D B D Light brown or orange pumiceous lapilli and ash CJ Taupo Pumice ignimbrite T.4 13 --- T. 17 <1 • . a o o 0 • 0 . T. 16 T.15 T.14 T.5 T.2 50cm 77 1 14 15 · · .. . . .• T. 7a? T.5? . 0 .. • 0 o 0 lD:·:2~==::::::::::::::::=:::::::::P~~~ T.1 _______ p::x(==T.1? t-;----'-.:_____J D. 1 \ \ j / ) ) Fig.3. 7 Accumulation rates for reference sheet sand sections Section 1 Accumulation Rate (mm/yr) (cm) o 0.5 1 1.5 2 0 ,·----·-~-----.----------·1---1---~--· 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 Tp + T.4 TP-..T.41T.7a TP+ T.4+ T.7a+ T 15? (cm) 0 0 20 40 60 BO 100 120 140 160 Section 4 Accumulation Rate (mm/yr) 0.5 Charcoal T.7A T.4 TP t T 4 TP1T4tT7n1-Ch 1.5 TP 1- T 4 t Ch -----· ------------- 2 Section 14 Accumulation Rate (mmlyr) (cm) 0 0.5 1.5 2 0 r-----rr----,-----.----, 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 T.7A r---'" . '. .. · T.4 Tp + T.4 Tp+T.4+T.7a --i (X) 79 Section 1 % allophane 0 2 4 6 8 10 0 10 20 30 40 T.7a 50 T.5 60 (cm) Depth Section 14 % allophane 0 2 4 6 8 10 0 20 40 T.5 60 80 100 (cm) Depth Fig. 3.8 Percent allophane for reference sheet sand sections (cm) 0 10 20 30 40 50 60 (cm) 0 20 40 60 80 100 80 Section 1 % carbon 0 2 3 4 5 ' I ' i- I ' I C I T.7a r T.5 Section 14 % carbon 0 2 3 4 5 ~~-~---i-----~------~- i r I ~-7 a ,----------u ! T4 _j Fig. 3.9 Percent carbon for reference sheet sand sections (cm} 0 0 10 20 30 40 50 60 (cm) 0 0 10 20 30 40 50 60 10 I [ __ I I I ) / r 10 T.7A Section 1 % 20 30 T.15? T.7A Section i4 % 20 T.4 40 30 125-25Dµm 250-1rnm 40 Fig. 3.10 Particle size distribution for reference sheet sand sections 81 50 60 50 Temporal changes also occur in particle size distributions. In section 14 (Fig. 3.10) the medium and coarse grain size fractions are greater between Tufa Trig member 4 and 7a than those higher in the profile. As well as the coarsening of modal grain size, an increased accumulation rate (resulting in over-thickening), and comparative decrease of allophane and carbon percentages also occurred. These changes are attributed to increased local influxes of sand. It is not known whether this increase was due to different site factors during this time, or due to an increased amount of source material available for re-working. The latter is a possibility as a large volume material was deposited from lahars throughout the study area, between the times when Tufa Trig members 4 and 7a were deposited (chapter 2). 3.3.3 Discussion of Sheet Sands The ubiquitous presence of re-worked Taupo Pumice throughout the Makahikatoa Sands above the lahar or fluvial deposit (in the type section) suggests that the major source of the sands is from adjacent eroding surfaces. The re-worked Taupo Pumice indicates that its source became established when the Taupo Pumice ignirnbrite, deposited on the adjacent surfaces became exposed and began to deflate. Despite accumulating at a rate some 16 times faster than the reference sections, the Makahikatoa Sands in the type section are comparatively more weathered. This is a further indication that most of the aeolian sand deposited at the type section was derived from pre-weathered tephra found on adjacent eroding surfaces. Today hoar frosts are observed to fragment the exposed, partially cemented, pre-weathered tephra. Wind then either suspends or saltates the tephra particles for a short time (Pye, 1987) off the eroding surface (Plate 3.5), and onto the adjacent areas covered in vegetation. Re-working by water after tephra fragmentation is also often an intermediary process. 82 Plate 3.5 This photograph was taken on an erosional surface during a dust storm created by strong nor t h-west winds on 28th January, ·1988 . Dusts storms carrying dust to a height of 300 metres were witnessed. 83 It is therefore concluded from the depositional history of the section that the present eroding surfaces began to deflate during or just after the time when a coarse lahar or fluvial deposit disturbed the pre­ existing vegetation pattern. Where the coarse deposit is absent, the stratigraphy suggests adjacent surfaces began to erode c . 800 yrs B.P. Field and laboratory results indicate that the initial rise in the accumulation rates from£· 850 yrs B.P (at a depth of 3.52m) was probably due to an increase in source material caused by a newly formed lahar or fluvial deposit, or perhaps a period of extreme climatic conditions (ie. increased winds) . This is because low allophane (Fig . 3.2) and carbon (Fig. 3.3) percentages, and the lack of re -worked Taupo Pumice, indicate that the aeolian sediment deposited at this depth was not derived from adjacent deflating surfaces. The friable nature of Makahikatoa Sands in the upper 2 metres of most profiles, together with nearly 1 .5 metres thickness sandwiched between Tufa Trig members 16 and 17, indicate accumulation rates escalated in the upper part of the profile. This suggests either the source area has i ncreased or- erosion rates have increased (or both.). This is probably due t o gullies continuing to expand and undermine the remnant interfluves on the eroding surfaces; acceleration of this process due to increased c limatic instability may be possible . In contrast to the type section, the provenance for the Makahikatoa Sands f ound i n the reference sheet sand sections is the surrounding l aharic/fluvial surfaces. These sections, unlike the type section, are f ound in locations which are either comparatively distal from their s ource areas or are surrounded by dunes that trap much of the sediment that is being re-worked from the so