Land Use, Land-Use Change and Forestry

Other reports in this collection

2.4.4. Measurements Using Remote Sensing

Satellite remote sensing may be used to determine initial areas of different land-cover types, including forests; to determine the extent, rates of change, and locations of activities that result in forest cleaing and regrowth; to determine the extent, rates of change, and locations of some of the additional activities that may be agreed on; and to determine other natural and human-induced changes in terrestrial ecosystems. The utility of data from satellites raises the question of whether such data will be used as a primary source of data by countries reporting sources and sinks of carbon or whether they will be used for verification.

2.4.4.1. Determining Initial Conditions

Satellite data (NOAA-AVHRR, synthetic aperture radar) have been used to create global maps of land cover and forest cover (Belward et al., 1999; De Grandi et al., 1999; Rosenqvist et al., 2000). Existing maps generally have resolutions of 1 km, but future capabilities at 30-m resolution may be possible with data from Landsat-7 in conjunction with other satellites. National-level and global observation strategies have been proposed for monitoring changes in forest cover (Skole et al., 1997).

Several active research efforts in laboratories in different countries continue to explore the potential applications of these data sets and seek to update them with new information as new optical measurements become available from SPOT-Vegetation and MODIS (e.g., Running et al., 1994a,b; Townshend et al., 1994; Belward et al., 1999). One approach consists of direct parameterization of land-cover characteristics that are relevant to carbon studies, such as vegetation structure and percentage of woody vegetation (DeFries et al., 1995, 2000). Another option is the development of a multi-resolution approach to global forest mapping. Such an approach would use coarse-resolution satellite data to determine a sampling frame within which higher resolution data would be nested. All of these methods are feasible within a scientific research agenda, but they would require further development-including calibration and evaluation with ground truth-for routine operational use.



Other reports in this collection