Epistemic uncertainties in the assessment of regional soil acidification
- Authors: Benke, Kurt , Robinson, Nathan , Norng, Sorn , Rees, David , O’Leary, Garry
- Date: 2022
- Type: Text , Journal article
- Relation: Environments - MDPI Vol. 9, no. 8 (2022), p.
- Full Text:
- Reviewed:
- Description: The increasing acidification of soil due to pollution and agricultural management practices is a growing problem worldwide, where food production is already under threat by climate change, more frequent droughts, and soil nutrient depletion. Soil acidification is quantified by pH measurements and is a primary metric for soil health. High soil acidity is a constraint on the production of grains and other crops because it decreases the bioavailability of important plant nutrients while increasing soil toxicity arising from an imbalance of essential soil elements. Field pH can be estimated by colour test kits which are very cost-effective and particularly suitable for developing countries where laboratory services are not available or fail to provide timely results. Because the pH test kit is based on visual colour matching between a colour card scale and a soil sample in solution, there are epistemic uncertainties, such as variability in expert opinion, differences in colour vision, measurement error, instrumentation, and changes in daylight spectral content. In this study, expert human observers were compared in experiments conducted using a standard pH test kit under a range of environmental conditions. A significant difference in uncertainty in colour discrimination was evident between male and female experts, whereas changes in daylight conditions had lower impact on the variance of pH estimates. In a group of subject matter experts, the male standard error (0.35 pH) was 57% higher on average over the range of pH values (pH = 4
Development of pedotransfer functions by machine learning for prediction of soil electrical conductivity and organic carbon content
- Authors: Benke, Kurt , Norng, Sorn , Robinson, Nathan , Chia, K. , Rees, David , Hopley, J.
- Date: 2020
- Type: Text , Journal article
- Relation: Geoderma Vol. 366, no. (2020), p.
- Full Text:
- Reviewed:
- Description: The pedotransfer function is a mathematical model used to convert direct soil measurements into known and unknown soil properties. It provides information for modelling and simulation in soil research, hydrology, environmental science and climate change impacts, including investigating the carbon cycle and the exchange of carbon between soils and the atmosphere to support carbon farming. In particular, the pedotransfer function can provide input parameters for landscape design, soil quality assessment and economic optimisation. The objective of the study was to investigate the feasibility of using a generalised pedotransfer function derived with a machine learning method to predict soil electrical conductivity (EC) and soil organic carbon content (OC) for different regional locations in the state of Victoria, Australia. This strategy supports a unified approach to the interpolation and population of a single regional soils database, in contrast to a range of pedotransfer functions derived from local databases with measurement sets that may have limited transferability. The pedotransfer function generation was based on a machine learning algorithm incorporating the Generalized Linear Mixed Model with interactions and nested terms, with Residual Maximum Likelihood estimation, and a predictor-frequency ranking system with step-wise reduction of predictors to evaluate the predictive errors in reduced models. The source of the data was the Victorian Soil Information System (VSIS), which is a database administered for soil information and mapping purposes. The database contains soil measurements and information from locations across Victoria and is a repository of historical data, including monitoring studies. In total, data from 93 projects were available for inputs to modelling and analysis, with 5158 samples used to derive predictors for EC and 1954 samples used to derive predictors for OC. Over 500 models were tested by systematically reducing the number of predictors from the full model. Five-fold cross-validation was used for estimation of model mean-squared prediction error (MSPE) and mean-absolute percentage error (MAPE). The results were statistically significant with only a gradual reduction in error for the top-ranked 50 models. The prediction errors (MSPE and MAPE) of the top ranked model for EC are 0.686 and 0.635, and 0.413 and 0.474 for OC respectively. The four most frequently occurring predictors both for EC and OC prediction across the full set of models were found to be soil depth, pH, particle size distribution and geomorphological mapping unit. The possible advantages and disadvantages of this approach were discussed with respect to other machine learning approaches. © 2020 Elsevier B.V.
Assessment of error sources in measurements of field pH : Effect of operator experience, test kit differences, and time-of-day
- Authors: Robinson, Nathan , Norng, Sorn , Rees, David , Benke, Kurt , Davey, Michelle
- Date: 2018
- Type: Text , Journal article
- Relation: Communications in Soil Science and Plant Analysis Vol. 49, no. 3 (2018), p. 269-285
- Full Text:
- Reviewed:
- Description: Various methods exist to measure soil pH, and while there is general agreement between the existing published laboratory and field-based methods, the latter are subject to uncertainties including test kit reliability, accuracy, precision, and environmental factors. The contribution of this study is to quantify three uncertainties that affect the conversion between field pH and laboratory pH measurements, namely operator experience, choice of test kit, and the time-of-day for measurement. Soil samples from western Victoria, representing the pH range 4.5–10.0, were used in a randomized complete block design with 10 assessors split into two groups representing experienced and inexperienced users. Statistical analysis of laboratory and field pH was based on using the Maximum Likelihood Functional Relationship (MLFR) to determine if there was any bias between the two methods. Significant differences were found between experienced and inexperienced users, and between test kits. © 2017 Taylor & Francis.
Improving the information content in soil pH maps: a case study
- Authors: Robinson, Nathan , Benke, Kurt , Norng, Sorn , Kitching, Matt , Crawford, Deborah
- Date: 2017
- Type: Text , Journal article
- Relation: European Journal of Soil Science Vol. 68, no. 5 (2017), p. 592-604
- Full Text: false
- Reviewed:
- Description: Uncertainties associated with legacy data contribute to the spatial uncertainty of predictions for soil properties such as pH. Examples of potential sources of error in maps of soil pH include temporal variation and changes in land use over time. Prediction of soil pH can be improved with a linear mixed model (LMM) to analyse factors that contribute to uncertainty. Probabilities from conditional simulations in combination with agronomic critical thresholds for acid-sensitive species can be used to identify areas that are likely, or very likely, to be below these critical thresholds for plant production. Because of rapid changes in farming systems and management practices, there is a need to be vigilant in monitoring changes in soil acidification. This is because soil acidification is an important factor in primary production and soil sustainability. In this research, legacy data from south-western Victoria (Australia) were used with model-based geostatistics to produce a map of soil pH that accommodates a variety of error sources, such as the time of sampling, seasonal variation, differences in analytical method, effects of changes in land use and variable soil sample depth in legacy data. Spatial covariates that are representative of soil-forming factors were used to improve predictions. To transform spatial prediction and estimates of error in soil pH into more informative and usable maps with more information content, simulations from the conditional distribution were used to compute the probability of a soil's pH being less than critical agronomic production thresholds at each of the prediction locations. These probabilities were mapped to reveal areas of potential risk. Highlights: Can maps of soil pH be improved by accounting for temporal variation and change in land use? First example of taking account of temporal variability in sampling for pH in spatial models. Key factors for uncertainty in spatial prediction include time of sampling and sample depth. Accuracy improved by accounting for additional sources of error combined with conditional simulations. © 2017 British Society of Soil Science