Effects of Climatic and Soil Data on Soil Drought Monitoring Based on Different Modelling Schemes

Abstract

Satisfactory requirements for the spatial resolution of climate and the influences of soil data in defining the starting points, endings, and the intensities of droughts have become matters of discussion in recent years. The overall inclusiveness of the modelling tools applied is also frequently discussed. In this light, five model setups (MSs) of the daily SoilClim water balance model were developed and tested for the Czech Republic (CR) in the 1961–2020 period. These included two versions of the SoilClim model, two sets of soil data, and two sets of climatic data at different spatial resolutions. MS1–MS4 were based on local, spatially-interpolated data from meteorological stations (500 × 500 m resolution), while MS5 was developed for global drought monitoring, based on the coarser ERA5-Land reanalysis (0.1° × 0.1°). During the 1961–2020 period, all the MSs indicated strong, statistically significant increases in the occurrence of 10th-percentile soil drought in the April–June season; however, trends remained largely non-significant for the remainder of the year. Variations among MS1–MS4 demonstrate that the range of soil property input data affects results to a lesser extent than different modelling schemes. The major simplification of the model grid in MS5 still led to an acceptable conformity of results, while the non-conformities disclosed may be explained by differences between meteorological inputs. Comparison with the Palmer Drought Severity Index (PDSI) confirmed that the SoilClim model depicts the variability of soil drought occurrence in greater detail, while PDSI tends to highlight the most severe events. The discussion arising out of the study centers around model uncertainties and the expression of soil drought episodes in different MSs.