Effects of land use and infiltration behaviour on soil conservation strategies

Soil erosion is a global problem because of its environmental consequences, including sedimentation and pollution in many areas of the world. Detachment of soil particles is mainly caused by rainsplash and the erosive force of overland flow. The main biophysical factor influencing the quantity of overland flow is the infiltration rate. Prediction of overland flow in catchments depends to a large extent on the characteristics of the infiltration process. Infiltration during a runoff-generating rainfall event is regulated by the hydraulic properties of the various soil layers, these are, the unsaturated or saturated conductivity and soil water retention characteristics, and the antecedent soil moisture conditions. This thesis deals with (i) the effects of spatial and temporal variability of soil physical properties, in particular the saturated hydraulic conductivity, on the generation and prevention of runoff in undulating loessial watersheds, and (ii) the application of an erosion model in the context of land-use planning and negotiation. Research for this thesis has been carried out in three small agricultural catchments in the southern loess area of the Netherlands (Limburg Province) and in an agricultural catchment on the Loess Plateau in China (Shaanxi Province).

Measurement of soil physical properties of soil horizons and land use units within the catchments was carried out, and the tendency of these soil layers to generate runoff was evaluated. Effects of heterogeneous saturated conductivity values on model outcome were evaluated using infiltration characteristics of both fully crusted and partially cracked crusted surfaces and were further examined by using a Simple Random Sampling approach to statistically identify different land use units. In loess areas, only differences in hydraulic properties of the surface layer caused differences in calculated discharge and soil loss, using the hydrological and soil erosion model LISEM. The China study showed that even if the soil type is the same, differences in land use and management may cause differences in hydraulic properties and, as a result of this, in calculated discharge and soil loss. If land treatment and soil type are the same, these land use units can be merged into the same soil physical unit, despite the fact that crops may differ. With this, a preliminary sampling scheme for comparable areas can be defined to reduce the number of samples needed to quantify K _s distributions.

To test the accuracy of the LISEM model on a small spatial scale, a single gully system was selected in the China catchment. It showed that measured and calculated hydrographs were in close agreement, but that the accuracy of the soil loss calculations decreased with smaller rainfall events. Alternative land use scenarios showed that reforestation of the gully floor significantly lowered water and sediment losses.

Effects of pre-defined land use scenarios on water and sediment losses were quantified for the China catchment, using the LISEM model and average and stochastic distributions of measured field K _s values. Use of stochastic K _s distributions and Monte Carlo analyses resulted in a range of model outcomes reflecting the effect of spatial heterogeneity upon simulated discharge and soil loss. In this way, probabilities of occurrence of the effects of alternative land-use strategies can be generated. Once a database has been compiled for a specific region, containing all necessary parameters of current and alternative land-use systems, on-the-spot simulations can be performed, allowing interactive negotiation with stakeholders. This database allows risk analyses to be performed for comparable areas to be conducted with relatively little additional effort.

The major challenge is to realise the on-the-spot use of the LISEM model to provide instant input into the negotiation process, which serves to stimulate the debate in defining conservation strategies.