Multi-scale hydraulic pedotransfer functions for Hungarian soils

Water and nutrient balance are among the main concerns about the sustainability of our soils. Numerous computer models have been developed to simulate soil water and solute transport and plant growth. However, use of these models has often been limited by lack of accurate input parameters. Often, the limiting input parameters are water retention and hydraulic conductivity. For many applications, estimation of soil hydraulic characteristics with pedotransfer functions (PTFs) may offer an alternative to costly and troublesome field or laboratory measurements. Many environmental problems are not restricted to national boundaries and therefore solutions require international co-operation. Soil particle-size distribution (PSD) is among the key predictors to most soil hydraulic PTFs. However, despite a number of recognised international standards, those data are rarely compatible across national frontiers, which hinders the establishment of international databases. The performance of four different interpolation procedures was evaluated to achieve compatibility of PSD data. Based on the number and distribution of measured points on the cumulative PSD curve, a general rule was formulated as when to fit a spline function or use a novel 'similarity procedure' to estimate missing values. The 'similarity procedure' uses an external source of soil information from which soils are selected with PSD that match the distribution of the soil under investigation. Fitting a non-parametric spline function to the PSD data showed similar accuracy and is independent of any external data sources, however, it is more sensitive to low data density on the PSD curve. International soil hydraulic databases can be valuable alternatives to smaller (e.g. national) databases when seeking solutions to soil water management related problems with international significance. Focusing on soil hydraulic data, the international UNsaturated SOil hydraulic DAtabase (UNSODA) was developed. The format, structure and operation of the redesigned second version of the database (UNSODA V2.0) were discussed. The 'similarity procedure' for the interpolation of PSD was tested for it's validity for soils of different geographical areas using data of UNSODA V2.0. While applying the procedure, a large European data set was used to help interpolate PSD of non-European soils. It was hypothesized that the procedure would not perform equally well for soils of other geographical areas, however, the study rejected this hypothesis. It was recognised that many soil hydraulic data have also been collected in Hungary in the framework of various independent projects. However, only part of those were stored in a common database. The HUNSODA database is introduced, which serves as storage for soil physical and hydraulic data as well as basic soil information on 840 soil horizons. The structure, contents and basic operations of the database are discussed. Data were then used to develop class pedotransfer functions for Hungarian soils according to both the FAO and USDA soil texture classes. There is considerable overlap between average curves of different classes if many small classes are distinguished. A less detailed classification system yields more distinction between class functions. In most PTF comparisons it remains unclear what the main sources of the estimation errors are. National, continental and intercontinental scale data collections were used to derive PTFs to estimate soil water retention. The same methodology (neural network model) and the same sets of predictors were used to allow the source database to be the only variable that is changed. We evaluated the performance of 11 different PTFs developed from each of the data sets, in order to study the influence of different combinations of predictors. All PTFs were tested using independent Hungarian data. Water retention estimations showed an improving trend as the list of input variables increased. Using a small set of relevant (local) data - when available - is better than using a large but more general data set. Estimated water retention curves (WRCs) were then used to simulate soil moisture time series of seven Hungarian soils. Small differences were found among the PTFs derived from different scale data collections. PTF estimates were only marginally worse than estimates using laboratory measured WRCs. Differences between estimations using different scale data sets (or measured WRCs) were not the main source of error in the simulation model. Scenario studies can be used in planning and prevention to fine-tune expert knowledge, as those give a quantitative dimension to the outcome of implemented changes. PTFs can provide important soil physical data at relatively low costs and at no risk to the environment. Three scenario studies demonstrate how PTFs may help improve land management planning. Exploratory modeling was used to give estimates of (i) benefits and risks associated with irrigating a heterogeneous field, in terms of soil water balance components and the soil's ability to supply the vegetation with water, (ii) the effect of changes in physical properties of a soil on flux balance and plant water deficit, and (iii) the risk of leaching to deeper horizons in different soil types of Hungary, based on the simulation of 20 years. Using an exploratory research approach, quantitative answers were provided to "what if" type of questions, allowing the distinction of trends in potential problems and successes.