Spatial modelling of soil humus content using the drone-based measurements

Abstract

The aim of the study was to assess the potential of spatial modeling of soil humus content through measurements with unmanned aerial vehicle (UAV), using digital relief model combined with geostatistical and regression methods. Location and time of the study. The study area was in the Middle Urals (Perm municipal district of the Perm region) in the southern taiga zone of the Central Russian province. The key area of the surveyed territory is located on the educational and scientific experimental field of the Perm State Agro-Technological University (56°34-37'N and 57°92-93'E). Soil samples were taken with a drill from a depth of 0–20 cm in May 2023. Methods. The survey data was obtained using the drone DJI mini 2 UAVS from a height of 50 m. The orthophotoplan and digital elevation model of the key site were made using the Drone Deploy software. The object of research was the soil cover of the site, represented by sod-podzolic soils. The soil humus content was determined using potassium digestion method. The UAV data processing was performed using the QGIS 3.22 system and SAGA 9 software. Spatial modeling was carried out using the geostatistical method “Ordinary Kriging” employing the Geostatistical Analyst tool of the ArcGIS 10.8 geographic information system. Results. Humus content in in the 0-20 cm soil layer varied from 0.9 to 3.0%, following normal distribution, thus allowing using the normal labels method without data transformation. The U-shape indicated a strong trend direction from north to south and from west to east. The correlation analysis showed the association of humus content with the morphometric parameters, the correlation coefficient ranging from -0.48 to 0.75. A regression equation was constructed for predicting soil humus content by using the parameters with the maximal correlation coefficients such as height (ELEV), distance to watercourses (Channel Network Distance) and relative position of the slopes (Relative Slope Position), having the correlation coefficients of 0.75, 0.75 and 0.66, respectively. This equation was found to model spatial humus with a relative divergence from Kriging interpolation of -0.6…-1.1%. Conclusions. The use of highly detailed UAV images to establish the influence of geomorphological conditions on soil properties allows not only to determine their spatial distribution, establish relationships with surface elevations, distance to watercourses, but also to simulate property changes in properties due to anthropogenic intervention, which, among other things, can enhance water erosion.