Detection of relative differences in phenology of forest species using Landsat and MODIS |
| |
Authors: | Bernard N Isaacson Shawn P Serbin Philip A Townsend |
| |
Institution: | (1) Department of Forest and Wildlife Ecology, University of Wisconsin—Madison, 226 Russell Laboratories, 1630 Linden Drive, Madison, WI 53706, USA |
| |
Abstract: | Landsat imagery is routinely used to characterize stand-level forest communities, but low temporal resolution makes pixel-wise
characterization of phenology difficult. This limitation can be overcome by using multi-year imagery, but organizing Landsat
scenes by calendar date ignores phenological gradients across the landscape as well as inter-annual differences in both scene-
and pixel-wise phenology. We demonstrate how a spatially generalizable, phenologically-informed approach for re-ordering Landsat
pixels can be used to characterize spatial variations in autumn senescence in several forest tree species. Using end-of-season
estimates derived from MODIS phenology data, we determined the “days left in season” (DLiS) across Landsat images to produce
a synthesized phenological trajectory of the normalized difference infrared index (NDII). We used ground-based species composition
data in conjunction with the NDII trajectories to model autumn senescence by species. Absolute phenology differed by one and
a half to 3 weeks between northern and southern Wisconsin, USA, but we show that the relative timing of phenology for individual
species differs across regions by only 1–3 days when considering senescence with respect to the local end of the season. The
progression of species senescence was consistent in lowland stands, starting with green and black ash, followed by silver
maple, yellow birch, red maple, and tamarack. The image analyses suggest that senescence progressed more rapidly in southern
than northern Wisconsin, starting earlier but taking about ten more days in the north. Our results support the use of MODIS
phenological data with multi-year Landsat imagery to detect species with unique phenologies and identify how these vary across
the landscape. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|