A New Year 2000 Global Data set of Cropland and Pasture
Abstract
Human activities have dramatically altered terrestrial ecosystems through the appropriation and conversion of natural ecosystems to cropland and pasture. To quantify the environmental consequences of these changes in land use, it is necessary to map the extent of cropland and pasture. We have developed a new data set for the Americas using a combination of year 2000 national and subnational agricultural census data, a global satellite-derived percent herbaceous cover data set, and two satellite-derived land cover classifications. The agricultural census data, which is of finer detail than previously used to create global cropland and pasture maps, is used to train the satellite land cover classifications. We also present a new technique for merging land cover data from two different satellite-based land cover classification data sets. By utilizing the agreement and disagreement between the Boston University's MODIS global land cover product and the GLC 2000 global land cover data set, we are able to better predict the spatial pattern of the agricultural land than by using either data set alone, particularly in the tropics. We present a new 5 min (~10km) resolution cropland and pasture dataset for the Americas that is of greater accuracy than previously developed. Additionally, a status report outlining current efforts to apply this methodology to the rest of the globe is included.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2005
- Bibcode:
- 2005AGUFM.B51D0254E
- Keywords:
-
- 0402 Agricultural systems;
- 0434 Data sets;
- 1632 Land cover change;
- 1640 Remote sensing (1855)