Model-based prediction of a vacant summer niche in a subarctic urbanscape: A multi-year open access data analysis of a ‘niche swap’ by short-billed Gulls

Abstract

Gulls belong to the seabird group, and are widely considered to be useful ecological indicators. Increasing urbanization throughout the Anthropocene has led to the rise of the urban landscape (‘urbanscape’) as a globally dominant habitat. Though historically less developed, similar urbanization patterns are emerging in the boreal forest, the world's largest forest ecoregion. Here we evaluate the ecological position of migratory Short-billed Gulls (Larus canus) in a subarctic urbanscape. We attempt to investigate a summer niche swap theory, in which gulls annually fill a vacant niche otherwise occupied by common ravens in winter. For that investigation we conducted seasonal plot surveys for Short-billed Gull presence from 2013 to 2016 in the Fairbanks, AK municipality. All data were made publicly available. We compiled them in a open source and ESRI geographic information system (GIS) platform and then added 68 open-access predictor layers, including socio-economic U.S. Census data. We trained, tested, and evaluated the performance of an ensemble of machine learning models, resulting in predictions of gull-abundance hotspots and coldspots, at 100-m resolution for inference. We find that Short-billed Gulls prefer the synergy of industrial areas near man-made water bodies, impervious surfaces, gravel pits, strip malls, transfer sites (garbage dumps) and some young forest vegetation. This study is a first-known attempt to utilize a blended ‘Big Data’ approach, in combination with traditional multi-year field-based data collection and alternative model assessments, in order to characterize an urban seabird niche. Our findings, and the digital infrastructure herein, provide an interdisciplinary baseline for potential applications in urban planning and monitoring the spread of disease reservoirs.