Becsült talajhidrológiai paraméterek szimulációs vizsgálata a NAIK Erdészeti Tudományos Intézet két mintaterületénEvaluation of estimated soil hydraulic parameters at two forestry monitoring sites

The input data requirement of eco-hydrological models poses a growing challenge. Among the various input data the bottleneck seems to be the high-resolution quantification of soil hydraulic parameters. Recognising the situation, novel soil- and soil-hydrologic databases (e.g. MARTHA, DOSoReMi, Lucas, EU-SoilHydroGrids) were recently developed with different spatial characteristics (point or spatial datasets, various resolution). We aim to investigate at the soil profile level (i) the reliability of a currently implemented 3D soil hydraulic database of the Lake Balaton watershed (Hungary) and (ii) whether it improves the performance of hydrological calculations compared to some of the previously available data sets. Our analysis was carried out for two monitored sites (Szalafő – meadow, 2013-2014; Fiad – mixed oak-beech forest, 2016-2017) located in the catchment. For both sites meteorological and soil moisture time series data and information from field surveys and laboratory analysis were used to set up five soil profile models using Hydrus-1D. The model variants differed only in their following soil properties: soil layer depths, saturated hydraulic conductivity and water retention curve parameters. These properties were derived from: (i) calibration and validation (“presumably best” version), (ii) laboratory measurements of the sampled soil profile, (iii) estimations with European pedotransfer functions based on the measured easily available soil properties, (iv) estimations with European pedotransfer functions based on soil map information from the DOSoReMi.hu database and (v) the EU-SoilHydroGrids maps. We evaluated the model variants by the goodness-of-fit between the measured and simulated soil moisture time series (using NSME, RMSE, R²). The simulated water budgets were also compared for the five model variants. The calibrated soil parameters lead to the best model performance at both locations (NSME was 0.49 and 0.75 for Fiad and Szalafő). The EU-SoilHydroGrids maps provided second best agreement, while contrary to expectations, the simulation based directly on the measured soil-hydrologic values were the weakest. Major differences occurred in the simulated water budgets, which can be attributed to the different environmental conditions at the two sites. Water budget results are mostly in line with the model efficiency indicators: on-site measured soil hydraulic parameters led to different results from the others. Based on the presented results we can assume that the widely used soil profile description and sampling practice might not properly suit the input needs of hydraulic simulations and can increase the uncertainty of modelling. The performance of soil hydraulic parameters derived with pedotransfer functions is promising, we will continue the more detailed analysis of 3D soil hydraulic property maps.