A bias-aware indicator framework for prioritizing surveys of data-deficient freshwater leeches in China

Sparse and spatially biased occurrence records complicate survey planning for freshwater invertebrates. We evaluated a bias-aware indicator framework linking taxon documentation with a fixed 40:40:20 ranking of environmental similarity to documented occurrence contexts, documentation gap and freshwater support. Contemporary accessibility was evaluated separately to characterize documentation associations and inform survey logistics. Habitat-scope and coordinate audits retained 198 records representing 44 freshwater leech taxa in China; 88.6% of taxa had fewer than ten records. Using 197 predictor-eligible records and 4907 freshwater-background points, random-record validation indicated repeatable environmental structure within the compiled dataset (median background-rank AUC = 0.753). Geographically blocked transfer was weak (AUC = 0.511) but exceeded a randomized-reference median of 0.293 ( p = 0.005). At 50 two-degree cells selected nationally within each fold, the reference recovered 55.3% of held-out records but none in two folds. Environmental similarity plus documentation gap recovered 57.9%; relative performance changed with budget. A 75:25 acquisition–exploration portfolio recovered 56.3% with a smaller environmental coverage gap (1.048 versus 1.364); a 50:50 portfolio retained 52.0%. An external retrospective cell-level adaptation using 158 filtered Procambarus clarkii records recovered 22/36 held-out record-containing cells versus a random median of six ( p = 0.001). Environmental transfer was heterogeneous, and the aggregate randomized-reference comparison was inconclusive (weighted AUC = 0.629; null median = 0.564; p = 0.0939). National mapping identified 985 freshwater-background points meeting both environmental-similarity and documentation-gap criteria. These comparisons expose budget-dependent indicator value and acquisition–coverage trade-offs, not validated field priorities or ecological learning. Standardized prospective surveys are required.