Automatic Delineation of Reference Areas
Soil sampling by automatically identifying optimal reference areas reduces costs while maintaining model accuracy.
Exploratory Figures - Amazon, Brazil
Figure 1. Study Area Location and Sampling Design
Multi-scale overview of the Coari region, Amazonas, Brazil. The top panels display the regional context within Brazil and Amazonas State; the bottom panel details the manual reference area, training sites (represented by black points), and validation sites (denoted by green diamonds) over satellite imagery.
Figure 2. Environmental Covariates and Reference Area Overlay
Exploratory maps showing spatial variability of key environmental covariates (soil, geomorphology, geology, elevation, temperature, and precipitation). Training and validation points are overlaid on the manual reference area to illustrate environmental representativeness across gradients.
Figure 3. autoRA 50% Reference Area Coverage.
Map showing the automated Reference Area (autoRA) delineation at 50% coverage across the study region. Black circles represent training points and green diamonds represent validation points, illustrating the spatial distribution of sampling locations within and outside the autoRA coverage.
