**4. Data**

An Autonomous Underwater Vehicle recorded the bathymetry in the Svacenický reservoir during field measurements carried out by the Slovak Academy of Science (SAS) in 2015 (22 September), 2016 (6 October) and 2017 (2 October). The EcoMapper is ideal for hydrographic spatial environmental monitoring in coastal and shallow-water applications. The speed of the AUV was 3.7 km/h and the sampling time interval was set at 1 s per sample. The total number of collected sample points is 2017 in 2015 and 2016, and 9211 sample points in 2017. The water level height was 4.45 m in 2015 and 2016, and 4.35 m in 2017. Figure 3 presents the results from the Digital Elevation Model (DEM) of the reservoir bottom with the 1 × 1 m spatial resolution. Post processing and data analysis were accomplished using Esri's ArcGIS software (Jack Dangermond, Redlands, CA, United States) (ArcMap 10.7). The DEM of the Svacenický Creek reservoir was created by geostatistical analyst tool through the Topo to Raster. Topo to Raster provides the functionality of incorporating other types of geographic features, which can assist in the creation of a DEM. Finally, the input DEM of the Svacenický Creek catchment (in grid cell size 10 × 10 m) is provided by the Esprit spol. s.r.o. (cartographic company).

In comparison with other physically-based models like EUROSEM (The European Soil Erosion Model) [32], WEPP (The Water Erosion Prediction Project) [33], and LISEM (The Limburg Soil Erosion Model) [34], the EROSION-3D model requires fewer soil input parameters (Table 2). Some of the input soil data were acquired during field measurements, and the laboratory processing of eleven soil samples followed (Figure 2a, Table 3). Also, the initial soil moisture (one of the most unstable and changeable parameters) was determined based on the field measurements in the study area. The importance of this model input parameter is due to its correlation with the previous precipitation index (Figure 4). The previous precipitation index was estimated as the 5-days sum of antecedent precipitation amount before each simulated event. The functions are derived based on rainfall experiments with known rainfall intensities and durations. The rainfall experiments took place at experimental plots in Slovakia (the Myjava region) and the Czech Republic (the Plzen region) in cooperation with the Czech Technical University in Prague and with the Research Institute for Soil and Water Conservation. A graphical representation of the functions is shown in Figure 4 (A—data from the Czech Republic, B—data from the Slovak Republic). The soil input parameters from the Parameter Catalogue were estimated for the specific crop and its growth phase in different months within the year.


**Table 2.** The input soil parameters required by the EROSION-3D model.

The rainfall data covers the time period investigated between September 2015 and October 2017. The rainfall events were used in the model calculations and the rainfall series consist of effective erosive events which occurred within the periods selected. Each rainfall event requires its own soil data set, whose parameters account for the current soil conditions and stages of crop growth of that date. The model runs were done for 95 rainfall events with specific rainfall intensities higher than 2 mm (for the model calculations, intensive rainfall events involved in soil erosion processes were selected). The summarized characteristics of the rainfall events selected are shown in Table 4. Figure 5 presents the occurrence of the selected rainfall events during the time period investigated.


**Table 3.** The soil characteristics in the study area.

**Figure 4.** The functions (dependence) between the initial soil moisture and previous precipitation index determined according to the field measurements in the Slovak and Czech Republics.

**Table 4.** Characteristics (minimum, maximum and mean value (x)) of the selected 95 rainfall events between September 2015 and October 2017.


According to the documentation of the local farmers, 49% of the arable land (from September 2015 to October 2016) was covered by wheat, which was followed by corn (19%), barley (14%), rye (13%), and lucerne (5%). In the next time period (from October 2016 to October 2017), corn covered 54% of the arable land, which was followed by wheat (31%), barley (10%), and lucerne (5%). Figure 6 shows the spatial distribution of the crops in the study area from September 2015 to October 2016.

**Figure 5.** Total rainfall amounts and date of occurrence of selected rainfall events during the two time periods: (**a**) September 2015–October 2016; (**b**) October 2016–October 2017.

**Figure 6.** Annual crop distribution in the study area: (**a**) 2015–2016; (**b**) 2016–2017.
