Spatial variability of soil structure and its impact on transport processes and some associated land qualities

This thesis treats the impact of soil spatial variability on spatial variability of simulated land qualities. A sequence of procedures that were done to determine this impact is described in chapters 2 and 3. The subchapters correspond to seven manuscripts that either have appeared in or have been submitted to peer-reviewed journals.In chapter 2 attention is paid to methods to inventory spatial variability of soil characteristics related to the structure of the soil. A method was developed to construct confidence intervals to point count results in case of spatial dependency of the point observations on a soil thin section. It was concluded, that confidence intervals obtained following the traditional method by assuming all observations independent, will be much narrower than those where spatial dependency structure is taken into account.Two other papers in chapter 2 describe a method to translate soil profile descriptions into soil physical input data for computer models that simulate solute flow. The concept of <em>functional layers is</em> introduced. A functional layer is a combination of soil layers showing comparable soil physical behaviour related to water flow. The functional layer approach was tested and accepted for examples of disturbed and thinly stratified soils by calculating functional properties of the layer under defined hydrological conditions. When functional layers are established, mapping the thickness, starting depth and type of functional layers provides spatial information about soil physical characteristics. In one paper in chapter 2 the number of necessary observations in this mapping procedure is optimized by application of geostatistical methods and a sequential sampling test.In chapter three the impact of variability of the structure of the soil on variability of crop yields and nitrate leaching is investigated. One paper describes a field scale empirical study where barley grain yield variability is correlated to variability of soil characteristics and simulated transpiration deficits. Simulation model inputs were obtained using the functional layer approach described in chapter 2. Regression functions based on simulated transpiration deficits only could explain 43% of the variance in yields, which suggested that variability of transpiration may be an important factor causing yield variability. This hypothesis was tested in a next paper in which remote sensing estimates of the leaf area index were used to obtain estimates of the potential transpiration with a high spatial accuracy. Incorporating space- and time series of the leaf area index into a crop growth model resulted in a prediction of yield variability that could explain 39% of measured variability. Variability of plant- available water, expressed by the actual transpiration, is an important factor causing yield variability.Two papers in chapter three describe how a combined solute flow and crop growth model was used to evaluate the spatial varying effect of fertilizing scenarios. 'Me spatial interpolation method <em>Disjunctive Kriging</em> was used to translate spatial variability of simulated nitrate leaching into maps of the probability that a threshold leaching concentration is exceeded. It was also investigated, whether the number of simulations could be minimized using <em>Disjunctive CoKriging</em> and available spatial information. It was concluded, that different soil units within one agricultural field showed a different leaching response and crop yield response to identical fertilizer treatments, and that yield variability will increase when fertilizer levels approach the level for maximal production.