Historical extension of operational NDVI products for livestock insurance in Kenya |
| |
| Institution: | 1. Research Ecologist, USDA-ARS Rangeland Resources Research Unit, Fort Collins, CO 80526, USA;2. Assistant Professor, Department of Wildland Resources and Ecology Center, Utah State University, Logan, UT 84322, USA;1. Department of Earth and Environment, Boston University, Boston, MA 02215, USA;2. Geospatial Sciences Center of Excellence, South Dakota State University, Brookings, SD 57007, USA;3. Department of Geography, South Dakota State University, Brookings, SD 57007, USA;4. Department of Geography, Environment, and Spatial Sciences, Michigan State University, East Lansing, MI 48824, USA;5. Center for Global Change and Earth Observations, Michigan State University, East Lansing, MI 48823, USA;6. Department of Forestry and Environmental Resources, North Carolina State University, Raleigh, NC 27695, USA;1. Department of Ecology, Evolution and Environmental Biology, Columbia University, New York, NY, USA;2. The International Research Institute, The Earth Institute, Columbia University, Palisades, NY, USA;3. The Nature Conservancy, Central Science Department, Arlington, VA, USA |
| |
| Abstract: | Droughts induce livestock losses that severely affect Kenyan pastoralists. Recent index insurance schemes have the potential of being a viable tool for insuring pastoralists against drought-related risk. Such schemes require as input a forage scarcity (or drought) index that can be reliably updated in near real-time, and that strongly relates to livestock mortality. Generally, a long record (>25 years) of the index is needed to correctly estimate mortality risk and calculate the related insurance premium. Data from current operational satellites used for large-scale vegetation monitoring span over a maximum of 15 years, a time period that is considered insufficient for accurate premium computation. This study examines how operational NDVI datasets compare to, and could be combined with the non-operational recently constructed 30-year GIMMS AVHRR record (1981–2011) to provide a near-real time drought index with a long term archive for the arid lands of Kenya. We compared six freely available, near-real time NDVI products: five from MODIS and one from SPOT-VEGETATION. Prior to comparison, all datasets were averaged in time for the two vegetative seasons in Kenya, and aggregated spatially at the administrative division level at which the insurance is offered. The feasibility of extending the resulting aggregated drought indices back in time was assessed using jackknifed R2 statistics (leave-one-year-out) for the overlapping period 2002–2011. We found that division-specific models were more effective than a global model for linking the division-level temporal variability of the index between NDVI products. Based on our results, good scope exists for historically extending the aggregated drought index, thus providing a longer operational record for insurance purposes. We showed that this extension may have large effects on the calculated insurance premium. Finally, we discuss several possible improvements to the drought index. |
| |
| Keywords: | NDVI AVHRR SPOT MODIS Index insurance Intercalibration |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
