High resolution habitat suitability modeling for a narrow-range endemic alpine Hawaiian species

Abstract

Mapping potentially suitable habitat is critical for effective species conservation and management but can be challenging in remote areas exhibiting complex substrate heterogeneity. An approach that combines a diverse set of nonintrusive spatial data collection techniques with field validation can lead to a better understanding of landscapes and species distributions. Nysius wekiuicola, commonly known as the wēkiu bug, is the most studied arthropod species endemic to the Maunakea summit in Hawai‘i, yet details of its life history and geographic distribution remain poorly understood. The wēkiu bug, a species of concern, provides an excellent opportunity to employ nonintrusive spatial data collection techniques to answer previously elusive questions about habitat quality and composition. To predict the geographic distribution of N. wekiuicola, MaxEnt habitat suitability models were generated from fifteen years of species occurrence data and a variety of spatial datasets, including high resolution digital elevation models, surface mineralogy based on hyperspectral remote sensing, and climate variables. MaxEnt model results indicate that the variables with the highest influence (in terms of percent contribution) were elevation (78.2%), presence of nanocrystalline hematite surface minerals (13.7%), and minor contributions from aspect, slope, and other surface minerals. A limitation of this study is that many historic trapping sites were placed near roads and other accessible pre-determined locations instead of being systematically or randomly placed, meaning final model results may be biased and not entirely indicative of true wēkiu bug distribution. Although climate data is available, these climatic variables were auto-correlated and at too coarse of a spatial resolution to include in the final analysis. A trapping experiment based on surface mineralogy and geomorphic position affirmed that both elevation and surface mineralogy play significant roles in the spatial patterns of wēkiu bugs, but observed presence upslope on a cinder cone and absence downslope, even within the same predominant surface mineral, suggests that other habitat variables may be at play such as competition/predation. The models of wēkiu bug range and predicted suitable locations will be incorporated into management efforts and restoration goals for land managers of Maunakea. In addition, environmental data layers created in this initiative have now unlocked the ability to create suitability models for other species of interest on Maunakea.