Boron availability in top- and sub-soils as affected by topography and climate

Abstract

Available boron (B) is essential to the normal growth of crops. Previous studies on available B have focused on topsoil; hence, information about available B variation and its relationships with environmental variables (topography, climate, vegetation, soil property and parent material) in subsoil is limited. The current study collected 132, 124, and 87 soil samples, respectively, from A, B, and C horizons of arable land in purple hilly areas of southwestern China. Classical statistics, semivariogram analysis, and boosted regression trees (BRT) were applied to investigate available B variation and its affecting factors in various horizons. Samples of each soil horizon were randomly divided into calibration (80%) and validation (20%) sets. Mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R2) were employed for evaluating model performance. Results indicated that available B content decreased with soil depth and showed the strongest spatial autocorrelation in the A horizon. Approximately 50%, 58%, and 51% of available B variability in the three horizons could be explained by the BRT models. Values of MAE varied between 0.038 (C horizon) and 0.053 mg kg−1 (A horizon), and RMSE changed between 0.048 (C horizon) and 0.069 mg kg−1 (A horizon). The relative importance of environmental variables to available B variability varied with soil horizons. Precipitation, flow path length, and topographical aspect were the most critical factors for the A, B, and C horizons, respectively. The importance of valley depth, elevation, and temperature enhanced, whereas precipitation and normalized difference vegetation index decreased in subsoil.