Po'ouli landscape bioinformatics models predict energetics, behavior, diets, and distribution on Maui

This study uses a spatially explicit microclimate/biophysical approach to examine the potential distribution of the Po'ouli on Maui to find either new habitats to search for existence or refine search efforts in previously occupied areas. We used specific physiological and behavioral ecology bird data, and Po'ouli morphological and spectral data obtained from museum specimens to address ecological and conservation-related questions about the Po'ouli that are otherwise very difficult to quantify. Laboratory and field tested microclimate and biophysical-behavioral animal computer models were integrated with remote sensing technologies. To show that the generic microclimate and endotherm models can predict metabolic and water loss requirements of Hawaiian Honeycreepers, we used the 2 species with known physiological properties, the Hawaiian Amakihi, Hemignathus virens, and the Hawaiian Anianiau, Hemignathus parvus. Predictions were within experimental measurement error of the laboratory measurements. Then using field rather than laboratory conditions as input data, we predict the field distribution of the Amakihi on Maui as the first spatial test of the models applied to birds. Results are consistent with Amakihi field distribution data. Fossils show that the Po'ouli once lived on Maui at low elevations in dry/mesic habitats on a likely diet of native tree snails and insects. The arrival of lethal mosquito-borne avian malaria in Hawaii exterminated low elevation Po'ouli forcing a population shift to mountain rainforests and possibly a snail diet instead of insects. To explore the maximum consequences of such a diet shift we assumed exclusive diets of snails versus insects at both low and high elevations. Snail diets require ∼4 times higher foraging rates than do insect diets, making a predominantly snail diet an unlikely prospect for the Po'ouli. Landscape scale simulations suggest that a snail diet would force a Po'ouli distribution inconsistent with observations. A predominantly insect diet is consistent with distribution observations. We show that as local environmental conditions change across the landscape in space and diurnal/seasonal time it is possible to quantify animal physiological and behavioral consequences of those variations in their local environment. This enables quantification of the requisite spatial and temporal distribution and amount or availability of resources that may affect species' potential for survival, growth, reproduction and distribution.