Figure 2

(A and B) Gini Importance of explanatory variables for iterative random forest models utilizing (left) Santa Cruz Wharf (SCW) data and (right) SCW + inland data. (C and D) The 8 most stable interactions recovered by iRF with the highest NSE utilizing (left) Santa Cruz Wharf (SCW) data and (right) SCW + inland data. Black triangles represent stability values while blue dots represent precision values.