Data Quality Influences the Predicted Distribution and Habitat of Four Southern-Hemisphere Albatross Species

Few studies have assessed the influence of data quality on predicted probability occurrence and preferred habitat marine predators. We compared results from four species distribution models (SDMs) for southern-hemisphere albatross species, Buller’s ( Thalassarche bulleri ), Campbell T. impavida grey-headed chrysostoma white-capped steadi based datasets differing quality, ranging no location to twice-daily locations individual birds collected by geolocation devices. Two relative environmental suitability (RES) were fit using minimum maximum absolute values each variable (1) monthly 50% kernel density contours background data, (2) primary literature or expert opinion. Additionally, two boosted regression tree (BRT) opportunistic sightings bird-borne electronic tags. Using model-specific threshold values, was quantified model. Model variables included distance land, bathymetry, sea surface temperature, chlorophyll- a concentration. Results both RES BRT model with those assess occupancy habitat. For all outperformed models. offer predictive advantage over being able identify relevant variables, incorporate interactions, provide spatially explicit estimates uncertainty. resulted in larger, less refined areas species. Our study highlights importance predicting albatrosses emphasises need consider pros cons associated different levels when SDMs inform management decisions. Furthermore, we examine overlap SDM fishing effort. discuss wide-scale distributions pelagic seabirds how these impacts could result protection measures.