Experimental results and temporal surrogate modeling of particulate organic carbon released during interrill erosion

Despite its temporal nature, the release of particulate organic carbon from soils during interrill erosion lacks high resolution description via experimental datasets and has not been analyzed and predicted with time series approaches that might be adopted research and application. Hence, current research approaches have produced limited success to describe the interrill process that temporally flushes particulate organic carbon off of landscapes and into waterways. Therefore, the goals of this work included three-fold innovative components and was to (i) perform rain simulation experiments and collect a comprehensive particulate organic carbon erosion dataset during erosion; (ii) utilize the dataset to describe particulate organic carbon release as a function of governing transport variables using a multivariate autoregressive model, which is argued as a versatile tool for description of temporal erosion and inclusion within predictive modeling structure; and (iii) to apply a laboratory-based surrogate approach to predict erosion release, in absence of rainfall simulation data, that is coupled with the temporal modeling methodology. Data from the rainfall simulation experiments was collected on agricultural field sites from soils of different textures, including silt and silt loam soils, and with a variety of tillage systems. The experimental datasets reflected both the initial non-equilibrium detachment process followed by near-equilibrium detachment later in the simulations associated with aggregate structure and fluid energy, with the more aggressive tillage systems exhibiting less time to reaching equilibrium. Thereafter runoff and fine sized soil particle release was utilized as predictive variables within autoregressive temporal modeling of the carbon detachment process. The applied variables predicted carbon release well (i.e. 95% of the carbon loss was predicted) and notably both the non-equilibrium and equilibrium phases of erosion was predicted. The surrogate approach was then applied by replacing the variable indicating soil release with a time series of fine soil released during laboratory slaking. Findings show that the temporal stability of the surrogate variables could describe the non-equilibrium and equilibrium detachment for the different soil types and tillage systems. The state space coupled with the surrogate approach proved to be an important and promissory tool to explain the particulate organic carbon release interrill erosion, even in absence of predictive data.