Elevational distribution affects pedodiversity (a measure or indicator of soil diversity) by controlling factors like climate, vegetation, and water drainage. It plays a vital role as a constant variable in soil formation processes (e.g., mineralization, eluviation, and illuviation, etc.) thus promoting diverse soil types at different toposequence formations. However, the relationship between pedodiversity and elevation at various spatial scales remains poorly understood and obscure, especially for dryland regions. Here, we first derive a national-scale pedodiversity map of Botswana, explaining more than 50% of the variance. For this map, elevation is among the most important environmental covariates influencing soil diversity. We further examine how spatial scale indicators such as spatial extent (i.e., countrywide, and locally) and resolution (i.e., 90m, 900m, 9000m, and 90000m) systematically influence landscape pedodiversity. While viewing the data countrywide, the relationship between pedodiversity and elevation maintained a negative or inverse linear trend (i.e., meaning that as elevation increases, pedodiversity decreases), but when viewed locally—for the small district, it showed a positive or direct linear trend (i.e., both elevation and pedodiversity increase simultaneously). This can be explained by differences in elevation patterns together with complex and dynamic interactions between scale-dependent soil-forming factors like land use type which tend to dominate local scales. Significant differences in pedodiversity related to the spatial resolution of geodata inputs were noticeable, for example, the local coefficient of determination values ranged from 0.06 to 0.65 for fine to coarse spatial resolutions respectively. Together, our findings demonstrate that the relationships between pedogenesis factors (e.g. elevation) and pedodiversity are scale-dependent. Even a small change in spatial resolution can lead to significant variations in pedodiversity, especially in semi-arid areas. Therefore, taking this into account can reduce overly optimistic conclusions about the landscape patterns we observe.
Read full abstract