| Summary: | Thesis (M.S.) University of Alaska Fairbanks, 2000 Twice-monthly AVHRR-derived NDVI were used to estimate growing season length across Alaska, north of the Alaska Range. An algorithm, based on the ratio of NDVI to annual maximum NDVI for each pixel, was used to represent percent of maximum greenness for each composite period. Greenup and senescence commenced when NDVI values rose above and fell below a selected percent of maximum greenness. Six different percent of maximum greenness threshholds, ranging from 25 to 50 percent, were evaluated. This algorithm eliminates complications of landscape-specific NDVI thresholds and year-to-year variability. The algorithm was tested against 1) air temperature data from 23 weather stations located in northern Alaska from 1991 to 1997, 2) observed greenup at two sites in Fairbanks, Alaska, from 1991 to 1997, and 3)phenology observations on the Seward Peninsula during the 1996-1997 growing seasons. Best results were obtained with NDVI values at 30% and 40% of maximum NDVI.
|
|---|