An investigation on the stability of biochar-C in soils and its potential use to mitigate non-CO₂ greenhouse gases using near-infrared (NIR) spectroscopy : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Soil Science at Massey University, Manawatu, New Zealand
The global interest in using biochar for C sequestration-climate mitigation and soil improvement has driven rapid expansion in biochar research to understand its properties and application impacts. The potential for biochar application to increase soil organic carbon (SOC) stocks and its potential agronomic and additional environmental benefits, such as reducing soil nitrous oxide (N₂O) emission, are determined by its stability in the soil, which is dependent on its intrinsic properties and the soil conditions.
The inherent properties of biochar are highly influenced by feedstock type and pyrolysis temperature hence, the use of various types of feedstock and pyrolysis technologies leads to uncertainties in predicting the effect of biochar addition to soils. Previous research has established a general assumption that biochar stability is strongly influenced by the type of feedstock used and maximum pyrolysis temperature. Research is required, however, to produce practical and reliable techniques that can be used to verify the reported maximum pyrolysis temperature, regardless of the type of feedstock used and predict the likely stability of the biochar.
In addition to the pyrolysis process, the final particle size of the biochar and the method of incorporation into the soil may also influence the ability of the biochar to moderate soil properties and function. With previous research, there is general lack of attention given to these potentially influential parameters when assessing the impact of biochar application to soil.
Therefore this thesis evaluates (i) the use of near-infrared (NIR) spectroscopy technique for predicting the maximum pyrolysis temperature of biochar as it is a well-known, non-destructive, and rapid technique for analyzing organic material; (ii) the effect of biochar application, with special attention to biochar particle size and depth of placement, on N₂O soil emission and SOC stocks; and (iii) the integrated use of NIR spectroscopy for SOC measurement.
In the first study, we hypothesized that NIR spectroscopy can be used to predict the maximum pyrolysis temperature achieved during biochar production. Eighty-two carbonized materials produced from various feedstock types and pyrolysis conditions with the reported pyrolysis temperature ranged between 220 to 800 °C, were scanned using NIR spectrometer and were used as the calibration set. The NIR calibration model was built by correlating the NIR spectral data with the reported pyrolysis temperature using partial least squares regression (PLSR). A separate sample set (n=20) was compiled using laboratory-produced biochar made from pine wood at pyrolysis temperature ranged from 325 to 723 °C. The calibration model validated using (i) leave-one-out cross-validation (LOO-CV) and (ii) the prediction set, yielded good accuracy (LOO-CV: r²=0.80 and RMSECV:48.8 °C; prediction: r²=0.82 and RMSEP: 57.7 °C). Results obtained in this study have shown that NIR spectroscopy can be used to predict the maximum pyrolysis temperature of biochar and has the potential to be used as a monitoring tool for biochar production.
In addition to the first study, the predictive ability of the NIR model was evaluated further. We hypothesized that the variation in feedstock types and pyrolysis processes may affect the predictive performance of the NIR model in predicting the maximum pyrolysis temperature of biochar. Therefore, three sample sets were generated from a total of 82 carbonized materials and its subsets (Set A: n=82; Set B: n=68; Set C: n= 48) and were used for developing three calibration models. The selection of samples for Set B and C was made by reducing the variability associated with production conditions and feedstock type i.e. Set B consist of samples produced by slow pyrolysis and using the same pyrolyzer unit, while for Set C (a subset of Set B), samples produced from “processed feedstocks” were excluded. A separate sample set (n=18) consists of samples produced from animal manure, crop residue, and woody materials were used as the prediction set. This biochar was produced using the slow pyrolysis technique in a laboratory or under relatively high production controlled conditions at temperature ranged from 250 to 550 °C. These calibration models were validated using (i) leave-one-out cross-validation (LOO-CV) and (ii) a prediction set, with the model based on set C gave the best prediction (R2: 0.941; RMSEP: 27.3 °C), followed by the model based on set A (R2: 0.896; RMSEP: 35.6 °C), and set B (R2: 0.928; RMSEP: 37.3 °C). These results indicate that feedstock types have a considerable effect on the performance of the NIR model while the effect of pyrolysis conditions was less pronounced. Thus, data variability from samples needs to be taken into account in developing the NIR calibration model for predicting the maximum pyrolysis temperature of biochar.
Before studying the effect of biochar on N₂O soil emission and SOC stocks, the maximum pyrolysis temperature of biochar to be used in the experiment was predicted using the NIR spectroscopy technique. The estimated pyrolysis temperature – after scanning the 3-year old pine wood biochar and using the NIR model developed – was 500 °C, while the reported temperature was 550 °C.
A controlled glasshouse study was conducted to investigate the effect of biochar particle size and the impact of soil inversion (through simulated mouldboard ploughing) on N₂O emissions from soils to which cattle urine was applied. We hypothesized that the application of biochar may (i) affect N₂O emissions through changes in soil physical properties, specifically soil aeration and water retention; and (ii) the effects of biochar addition on these properties may differ depending on their particle size (e.g., a larger particle size may increase soil aeration whereas a smaller particle size may clog pores), and their placement in soil (e.g., the incorporation of a large particle size-biochar at depth may promote water movement from the top layer and increase the overall drainage of the soil). Pine biochar (550 °C) with two different particle sizes (<2 mm and >4 mm) was mixed either into the top soil layer at the original 0–10 cm depth in the soil column or at 10–20 cm depth by inverting the top soil to simulate ploughing. Nitrous oxide emissions were monitored every two to three days, up to seven weeks during the summer trial, and measurements were repeated during the autumn trial. The use of large particle size biochar in the inverted soil had a significant impact on increasing the cumulative N₂O emissions in the autumn trial, possibly through changes in the water hydraulic conductivity of the soil column and increased water retention at the boundary between soil layers. Thus, the importance of the role of biochar particle size and the method of biochar placement on soil physical properties and the implications of these on N₂O emissions was highlighted. In the same glasshouse study, the effect of biochar particle size and depth of placement was further evaluated in relation to soil organic C. We hypothesized that (i) the large-particle size biochar may affect soil aeration and accelerate soil C decomposition rate with increased oxygen availability, and this effect is greater when biochar is incorporated at depth due to the more compacted soil at deeper layer with poorer aeration compared to the surface layer; and (ii) the NIR spectroscopy technique can be used to predict the SOC concentration and SOC stocks in biochar-amended soil.
Carbon stocks were estimated using NIR spectroscopy coupled with partial least-squares regression analysis (NIR/PLSR) and direct organic C measurements using an elemental analyzer. The NIR spectra of the soil were acquired by scanning intact soil cores using the NIR spectrometer. By the end of the glasshouse trial (327 days), the large-particle size biochar applied at depth had induced significant soil C loss (9.20 Mg C ha⁻¹ (P < 0.05), possibly through the combination of enhanced soil aeration, and the interrupted C supply from new plant inputs at that depth due to soil inversion. This C loss did not occur in the treatment with the small-particle size biochar. Near-infrared (NIR) spectroscopy was able to predict the SOC concentration, however, the prediction accuracy may be negatively affected by an increasing biochar particle size and soil inversion, thus may affect the subsequent SOC estimates.
The information obtained in this thesis will inform the future use of biochar and contribute to the knowledge of possible factors affecting soil N₂O emission from biochar-amended soil, the mineralization of native SOC, and the changes in SOC stocks over time, particularly in the pastoral soils of New Zealand. Also, based on this study, the use of NIR spectroscopy technique may potentially be integrated as part of the methodology for SOC estimation.