Vegetation Disturbance on a Multi-year Scale in the Great Basin, U.S. as Identified by AVHRR

The effects of global environmental change on vegetation health and vigor are proposed to be manifested at the landscape scale. Recognizing affected ecosystems is challenging on such a broad scale, but remotely sensed data offer the opportunity to test for changes in vegetation. While this method has been applied to tropical and temperate systems, less effort has been directed toward arid and semi-arid regions (with the exception of the African Sahel). Here we present results from the Great Basin, US in which we use AVHRR 1-km, biweekly composite NDVI data acquired from 1989-1999 to characterize vegetation phenology and identify anomalous response patterns. To begin, we created a 1-year average phenology for each pixel using the 11-year NDVI dataset. We then subtracted the annual average from the full time series to create a residual difference NDVI. We fit a line to the residual for each pixel to determine the slope and root mean square (RMS) difference. The slope records the change in vegetation (in NDVI/yr) outside of the average phenology. RMS shows the variability in the residual over the 11-year period. The average slope for the entire Great Basin (700,000 sq. km) is 0.60 +/- 0.23 NDVI/yr. We assume that this amount of increase is due to systematic errors from the AVHRR instrument and the change from NOAA-11 to NOAA-14 satellites in 1995 (e.g. Gutman, 1999). Against this systematic increase, we can then target regions whose slopes exceed one standard deviation of the mean. Our initial analyses have focused on areas of anomalous NDVI increase. One particular area of interest is located in northwest Nevada between 39.5-41N, 117-119W. The increase is limited to the valley floors in an elevation range of 1200-1400 m and does not extend to mountaintops or high valley slopes. Valley floors are contiguous and cover an area of 2000 sq km. The average slope increase of these regions is 1.29 +/- 0.30 NDVI/yr, three standard deviations above the mean for the Great Basin as a whole. Attempts to correlate the anomalous NDVI increase to rainfall are inconclusive to date. The high rainfall from El Nino in 1998 undoubtedly affected vegetation abundance, however this event occurred throughout the Great Basin and cannot sufficiently explain local anomalies. The region displays a variety of increase patterns. Some areas show a continuous rise over the 11 years, others have one or two years of anomalous increase. In several instances the NDVI increase is much more prominent in summer months relative to the rest of the year. These results indicate that multiple processes are likely affecting the semi-arid ecosystems of the Great Basin. As yet it is unclear whether these changes are caused by geophysical and/or climatological disturbances, by invasion of non-indigenous species or by some other environmental factors. Future work will scale down this investigation to anomalous increases within the spatial extent of a Landsat scene. Gutman, G.G., J. Geophys. Res. (D6), 104, 1999