Estimating Palila (Loxioides bailleui) Habitat Using Remote Sensing

Abstract

The remaining population of palila (Loxioides bailleui), which is tied closely to mamane (Sophora chrysophylla) and naio (Myoporum sandwicense) tree species, has been in long-term decline. This decline is caused by black rats (Ratus ratus), malaria, predation, and parasitism of moth species that palila consume, and the destruction of the mamane naio woodland by ungulate browsing and invasive flora. Using LiDAR and hyperspectral data collected in 2009, this study derived measures of the mamane and naio habitat. Using Bayesian information criterion (BIC) stepwise selection this study found that LiDAR ground return intensity, area of mamane trees and water band index (WBI) were the best explanatory variables of palila bird count data and that mamane and naio may be used at different scales. The browsing and area of mamane variables support existing knowledge; the canopy water content however suggests a problem for palila in wetter areas, contra to established expectation.