MODELING MOOSE DENSITY USING REMOTELY SENSED HABITAT VARIABLES

Abstract

Models for moose density were developed using subsets of remotely sensed habitat variables in north-central Alaska. Macro-habitat factors explained from 60 to 70% of the variation in November moose densities using a regression model. Use of logistic regression allowed correct classification of moose sample units into 3 moose density categories, based solely on habitat characteristics. Fire was less important to the model than anticipated, whereas river riparian zones were more important than expected. Fire was not the major determinant associated with high moose density in this interior Alaska study area. Models based on habitat alone may be useful for predicting moose density classes for some management purposes.