Predicting Soil Properties Using Topographic and Climatic Variables
Öz
Aim of study: The present study aimed to model soil physical and chemical properties through multiple linear and regression tree techniques.
Area of study: The study area is located between 41,07 – 41,33 N latitude and 41,74 – 42,27 E longitude in Artvin, which is in the Colchis part of the Black Sea Region of Turkey.
Material and methods: The multiple linear regression and regression tree models were used to predict soil properties using topographic and climatic features as independent variables. Besides, the relationships between soil properties and independent variables were determined by Pearson correlation.
Main results: The study results revealed that model accuracy by regression tree generally was higher than those of multiple linear regression. Up to 56% and 59% of the variance in soil properties was accounted for by multiple linear regression and regression tree, respectively. The easting, northing, elevation, and minimum temperature parameters were key drivers of both models. Increasing soil depth significantly increased the pH and reduced the organic carbon, total nitrogen, and carbon/nitrogen ratio.
Highlights: Topographic and climatic variables accounted for Up to 59% and 56% of the variance in soil properties such as texture, pH, organic carbon, total nitrogen, and carbon/nitrogen ratio by regression tree and multiple linear regression techniques. The most influential factors on soil properties were the minimum temperature, latitude, actual
evapotranspiration, mean temperature, distance to the ridge, and radiation index.
Anahtar Kelimeler
Forest, Modeling, Multiple Linear Regression, Regression Tree.
Kaynakça
- Aartsma, P. (2016). Climate, A Driving Factor Behind Soil Formation In Proglacial Areas In The European Alps? Master thesis. Master Earth and Environment Soil Geography and Landscape Group, Wageningen University.
- Adhikari, K. & Hartemink, A. E. (2016). Linking soils to ecosystem services—A global review. Geoderma, 262, 101-111.
- Aertsen, W., Kint, V., Muys, B. & Van Orshoven, J. (2012). Effects of scale and scaling in predictive modelling of forest site productivity. Environmental Modelling & Software, 31,19-27.
- Bishop, T. F. & Minasny, B. (2006). Digital soil-terrain modeling: the predictive potential and uncertainty. Environmental Soil-landscape Modeling, 185-213.
- Bolton, D. K., White, J. C., Wulder, M. A., Coops, N. C., Hermosilla, T. & Yuan, X. (2018). Updating stand-level forest inventories using airborne laser scanning and Landsat time series data. International Journal of Applied Earth Observation and Geoinformation, 66, 174-183.
- Brinkman, R. (1990). Resilience Against Climate Change? Pages 51-60, Developments in soil science. Elsevier.
- Buol, S. W., Southard, R. J., Graham, R. C. & McDaniel, P. A. (2011). Soil genesis and classification. John Wiley & Sons.
- Chai, T. & Draxler, R. R. (2014). Root mean square error (RMSE) or mean absolute error (MAE)?–Arguments against avoiding RMSE in the literature. Geoscientific model development, 7, 1247-1250.
- Demircan, M., Alan, I. & Sensoy, S. (2011). Increasing resolution of temperature maps by using Geographic Information Systems (GIS) and topography information. 27-29 in 5th Atmospheric Science Symposium.
- Dick, W. & Gregorich, E. (2004). Developing and maintaining soil organic matter levels. Managing soil quality: Challenges in modern agriculture, 103-120.