Get reliable science

Rangeland degradation is a growing problem throughout the world. Differences in land use between developed and undeveloped countries obviate the need for assessment protocols suitable for detecting degradation resulting from both past and present practices. A statistically robust algorithm for comparing the trend and “status” of vegetation productivity to user defined reference conditions was developed. Vegetation productivity was estimated from 2000 to 2012 using annual maximum Normalized Difference Vegetation Index (NDVI) from the MODIS satellite platform.  Each pixel was compared with reference conditions derived from surrounding pixels on similar sites with nearly identical potential production. Trends in degradation were determined by comparison between the slopes of the linear trends in mean annual maximum NDVI at each pixel and reference conditions with a one-sample t-test.  In contrast, the status of degradation at each pixel was evaluated by comparing its mean annual response of NDVI between 2000 and 2012 to that of reference conditions over the same time period using a one-sample t-test. These procedures to evaluate trends and status of degradation were applied across northern and southern Great Plains of the United States. Trends in degradation were almost undetectable across the entire study area. In contrast the degradation status assessment revealed that 16% (7,330,625 ha) of the vegetation on the northern Great Plains and 9% (3,295,106 ha) of the Southern Great Plains were significantly different (degraded) (P ? 0.01) from reference conditions. The amount of annual net primary reduction lost resulting from these degraded lands relative to reference conditions was estimated at 1.99 Tg C yr-1, less than 1% of the total annual net primary production in the study area of 211 Tg C yr-1.