Integrating remote sensing and local ecological knowledge to monitor rangeland dynamics

Abstract Rangelands are among the most extensive anthropogenic landscapes on earth, supporting nearly 500 million people. Disagreements over the extent and severity of rangeland degradation affect pastoralist livelihoods, especially when impacts of drought and over-grazing are confounded. While vegetation indices (such as NDVI, or Normalized Difference Vegetation Index) derived from remotely sensed imagery are often used to monitor rangelands, their strategic integration with local ecological knowledge (LEK) is under-appreciated. Here, we explore these complementary approaches in Kyrgyzstan’s pasture-rich province of Naryn, where disagreements regarding pasture degradation could greatly benefit from additional information. We examine a time series of MODIS satellite imagery (2000–2015) to characterize browning trends in vegetation as well as to distinguish between climate- and grazing-induced trends. We also compare and contrast measured trends with LEK perceptions of pasture degradation. To do so, we first examine statistical trends in NDVI as well as in NDVI residuals after de-trending with meteorological data. Second, we use participatory mapping to identify areas local pasture managers believe are overgrazed, a particularly useful approach in lieu of reliable historical stocking rates for livestock in this region. Lastly, we compare the strengths and weaknesses of LEK and remote sensing for landscape monitoring. Browning trends were widespread as declining trends in NDVI (and NDVI residuals) covered 24% (and 9%) of the landscape, respectively. Local managers’ perceptions of pasture degradation better reflected trends seen in NDVI than in climate-controlled NDVI residuals, suggesting patterns in the latter are less apparent to managers. Our approach demonstrated great potential for the integration of two inexpensive and effective methods of rangeland monitoring well-suited to the country’s needs. Despite limitations due to terrain, our approach was most successful within the semi-arid steppe where pasture degradation is believed to be most severe. In many parts of the world, sources of long-term spatially extensive data are rare or even non-existent. Thus, paired LEK and remote sensing can contribute to comprehensive and informative assessments of land degradation, especially where contentious management issues intersect with sparse data availability. LEK is a valuable source of complementary information to remote sensing and should be integrated more routinely and formally into landscape monitoring. To aid this endeavor, we synthesize advice for linking LEK and remote sensing across diverse landscape situations.