High resolution habitat suitability modelling for an endemic restricted-range Hawaiian insect (Nysius wekiuicola, Hemiptera: Lygaeidae)

Determining potentially suitable habitat is critical for effective species conservation and management, but can be challenging in remote or sensitive areas. An approach that combines non-intrusive spatial data collection techniques and supporting field data can lead to a better understanding of landscape-scale species distributions. Here we present two habitat suitability models, at 1 and 10 m resolutions, for the endemic wēkiu bug Nysius wekiuicola, a poorly-understood resident scavenging arthropod species present on the summit of Maunakea in Hawai‘i. Our models reveal that the wēkiu bug, restricted almost entirely to portions of cinder cones above 3500 m elevation, has a high degree of habitat specificity and represents a classically rare species. Across the 55 km² study area, 850 ha of potentially suitable habitat were identified at the 10 percentile training threshold, with the core area located at the true summit. Our results show that elevation and surficial mineralogy were the strongest predictors of suitable habitat, with lesser contributions from aspect and slope. Climatic variables also likely influence wēkiu bug distribution patterns, but were not included in our models due to the coarseness of available climate data and high correlation between variables. Relatively minor differences between the two models, in terms of identifying the locations and amount of suitable wēkiu bug habitat, and a higher measure of performance for the 10 m resolution model, suggest that coarser resolution input variables may characterize suitable habitat more efficiently than very fine 1 m resolution data. The suitability models generated as a result of this study will be directly incorporated into conservation management and restoration goals, and can easily be adapted for other arthropod species, leading to a more holistic understanding of metacommunity dynamics at the Maunakea summit.