Predicting use of habitat patches by spawning horseshoe crabs (Limulus polyphemus) along a complex coastline with field surveys and geospatial analyses

Abstract

Abstract This research described coastal habitat characteristics at a landscape scale and followed a geospatial modelling approach to predict the probability of habitat use by the population of Atlantic horseshoe crabs (Limulus polyphemus) in Long Island Sound (LIS, USA). This approach was developed to aid in decision‐making regarding management of horseshoe crab spawning locations. Geospatial data layers were created within which beach slope, wave exposure, surface substrate type, and distance from offshore aggregations of crabs (i.e. hotspots) were summarized for beaches in the western, central, and eastern regions of the Connecticut coast of LIS. Spawning abundances derived from field surveys conducted in May and June of 2009 and 2010 were used with remotely‐sensed habitat characteristics to develop a resource selection function from a candidate model set based on polytomous logistic regression. A single best model (Akaike weight = 0.97) for predicting the probability of habitat use by spawning horseshoe crabs was selected using an information‐theoretic approach. The parameter estimates predicted a greater probability of habitat use with increasing slope, decreasing wave exposure, and decreasing distance from offshore hotspots. Small ‘pocket’ beaches surrounded by rocky headlands, marshes, and developed areas are the typical habitat available to support horseshoe crab spawning in LIS. As spawning densities in LIS are relatively small and suitable spawning habitat is fragmented, geospatial methods are especially useful for identifying high‐use sites and maximizing the conservation value of management actions. Copyright © 2014 John Wiley & Sons, Ltd.