Fig. 1: Spatial distance distributions in global mapping studies.
From: Machine learning-based global maps of ecological variables and the challenge of assessing them
Spatial distribution (left; equal Earth projection) and distribution of nearest neighbor distances (right; sample-to-sample distance in pink, prediction-location-to-sample distance in blue) for three different publicly available datasets: cation exchange capacity in the soil from the WoSIS database23 as used for global soil mapping3 (A), specific leaf area from the Try database24 as used for the global mapping in Moreno-Martinez et al. (2018)25 (B), and the nematodes dataset compiled by Van den Hoogen et al. (2019)2 (C). For comparison, the fourth dataset is a simulated completely spatially random sample of the same size as the nematode dataset (D). Distance distributions were calculated and visualized using the R package “CAST”26.
