Models combining multiple scales of inference capture hydrologic and climatic drivers of riparian tree distributions

Predicting species geographic distributions is key to managing invasive species, conserving biodiversity, and understanding species' environmental requirements. Species distribution models (SDMs) commonly focus on climatic predictors, but other factors can also be essential, particularly for with specialized habitats defined by hydrologic, topographic, or edaphic conditions (e.g., riparian, wetland, alpine, coastal, serpentine). Here, we demonstrate a novel approach capturing strong effects of both hydrologic predictors in SDMs riparian plants, merging analyses targeted at drivers within ecosystems across the western USA (3.8 × 106 km2). We developed presence-background from five algorithms three trees (Tamarix ramossisima/chinensis [saltcedar], Elaeagnus angustifolia [Russian olive], Ulmus pumila [Siberian elm]) native Populus spp. (cottonwoods). used separate background datasets develop different spatial scales inference: (1) spatially filtered random points represent available habitat study area (2) target-group Salix (willow) occurrences habitat. Random-background captured tree relative largely upland USA, whereas Salix-background context ecosystems. Combining predictions two backgrounds identified hydrologically suitable climatically regions, resulting fewer false “absences” than either alone, improving over previous SDMs, providing more complete information guide management decisions. Surprisingly, predicted U. pumila, newly recognized invader, was as extensive deltoides/fremontii, T. ramossisima/chinensis, E. angustifolia, most common complexes USA. Watersheds constituting 20% contained no occurrence records, indicating high risk future unrecognized invasions. ecosystem-specific may improve many habitats, method link localized features while broad-scale