Spatiotemporal variability and simulation of tillaged loess microtopography in water erosion

Abstract

The objective of this study was to reveal the spatial variability of the microtopography in the Loess Plateau region of China during a simulated soil erosion process. Methods of artificial rainfall, laser scanning measurement, and geographic statistics were adopted to generate microdigital elevation model (M-DEM). Spatiotemporal variability characteristics under the condition of 1.5 mm min⁻¹ (0.059 in min⁻¹) rainfall intensity were studied. The results indicated that a probability division of relative elevation of microtopographic tillaged slope follows the normal distribution. It is valid to put microtopography elevation data into the geostatistics models. The fluctuations of different microtopographic tillaged slopes present a general trend in nearly all erosion phases: contour tillage > artificial digging > artificial backhoe > rake wear leveling. Tillage system is the critical factor for the spatial variability of M-DEMs. Spatiotemporal variability is affected by both structural properties and random attributes during the soil erosion process and comparatively more by the structural property. The contour tillage system should be the first choice in the agricultural tillage layout for soil and water conservation. According to the determined coefficient of the semivariogram function and the related coefficient r of the Kriging Cross validation, the natural tillage microtopography of loess slope can be imitated with a semivariogram function fitting curve model; in general, the Gaussian model and Spherical model can be widely adopted, except for the exponential model for artificial digging with a slope of 15°. Most importantly, the work has established a novel methodological framework for microtopography simulation of microscale (cm) area. Our study reveals the essence of microtopography of loess slope and lays a theoretical foundation and a support datum for further refining the boundaries of various erosion stages.