资源环境科技发展态势分析平台

Most uses of remotely sensed satellite data to characterize wildlife habitat have used metrics such as mean NDVI (Normalized Difference Vegetation Index) in a year or season. These simple metrics do not take advantage of the temporal patterns in NDVI within and across years and the spatial arrangement of cells with various temporal NDVI signatures. Here we use 13 years of data from MODIS (Moderate Resolution Imaging Spectroradiometer) to bin individual MODIS pixels (5.3 ha) into phenoclasses, where each phenoclass consists of pixels with a particular temporal profile of NDVI, regardless of spatial location. We present novel procedures that assign sites to phenoclusters, defined as particular composition of phenoclasses within a 1 km radius. We apply these procedures to Mexican spotted owl (Strix occidenralis lucida) nesting locations in the Sacramento Mountain range in south-central New Mexico. Phenoclasses at owl nest sites and phenoclusters around owl nest sites differed from those at and around points randomly placed in forest types that are known to support nesting owls. Stand exam data showed that the phenoclasses associated with owl nest sites are dominated by Douglas-fir (Pseudotsuga menziesii) and white fir (Abies concolor). The availability of phenoclusters and phenoclasses on Mescalero Apache tribal lands differed from those on adjacent National Forest lands within the Sacramento Mountain, consistent with different elevations and forest management practices. Nonetheless owls predominately used the same phenoclasses and phenoclusters in both land ownerships. MODIS phenoclasses and phenoclusters offer a useful means of remotely identifying forest conditions suitable for wildlife. Because the remote sensing data are freely available and regularly updated, they can be part of a cost effective approach to monitor and assess forested wildlife habitat over large temporal and spatial scales.