Evaluation of ISLSCP Initiative II satellite‐based land cover data sets and assessment of progress in land cover data for global modeling

Abstract

[1] As an important component of the International Satellite Land Surface Climatology Project (ISLSCP) Initiative II data collection, eight state-of-the-art land cover/use data sets have been compiled and made consistent with the ISLSCP Initiative II land/water mask in support of global modeling efforts. These data sets contain new and improved global data sets at coarse resolutions (1/4, 1/2 and 1°) describing historical, recent and present land cover conditions and are a testament to the tremendous progress made in this area over the past decade. In addition to the historical data, data describing the subcell heterogeneity in land cover are also provided, both in terms of subcell proportions of land cover classes and vegetation continuous fields such as % tree, grass and bare cover. Here we present the various ISLSCPII land cover data sets and compare the principal satellite-derived data sets and the effect of their respective aggregation methods. We find that despite some notable disagreements among similar classes, the satellite-based data sets agree remarkably well over large portions of the Earth's surface (over 50% for all resolutions). We also find that the methods of aggregation, whether done by a strictly dominant type, or using more information on subcell tree cover, can have an important impact on the final output and need to be considered by the user. Finally, by integrating the vegetation continuous fields data into our analyses we are able to show that the principal differences in terms of discrete land cover classes are in fact transition zones between similar classes. Citation: Brown de Colstoun, E. C., R. S. DeFries, and J. R. G. Townshend (2006), Evaluation of ISLSCP Initiative II satellite-based land cover data sets and assessment of progress in land cover data for global modeling,

A New Characterization of the Land Surface Heterogeneity over Africa for Use in Land Surface Models

“…Using the normalized difference vegetation index (Tucker 1979) derived from the Advanced Very High Resolution Radiometer (AVHRR) instrument on board the National Oceanic and Atmospheric Administration (NOAA), gradually researchers developed quantitative relations to map biophysical variables (Gutman et al 1995;Sellers et al 1996b;Los et al 2000;Strugnell and Lucht 2001). It still represents today a challenge to the remote sensing community (De Colstoun et al 2006). More recently, reflectance data from the new generation of sensor systems with improved spectral directional or higher spatial resolution like the Polarization and Directionality of the Earth's Reflectance (POLDER), the Medium Resolution Imaging Spectrometer (MERIS), the Moderate Resolution Imaging Spectroradiometer (MODIS), the VEGETATION sensor, and the Spinning Enhanced Visible and Infrared Imager (SEVIRI) make them relevant for the mapping of land surface variables (Roujean and Lacaze 2002;Bacour et al 2006;Baret et al 2007;Houldcroft et al 2009;Trigo et al 2010).…”

ISLSCP Initiative II global data sets: Surface boundary conditions and atmospheric forcings for land‐atmosphere studies

“…They note that again the areas of disagreement are between similar classes such as the various forest types, open and closed shrublands, etc., and not between large core classes. While areal proportions of cover types showed the same trends as Hansen and Reed [2000], Brown de Colstoun et al [2006] found that the principal areas of disagreement were specifically related to the IGBP‐DIScover Mixed Forest class and the UMD Woodlands/Wooded Grasslands classes, particularly in Africa and boreal forest areas. Using the per‐class proportions for each cell as well as the UMD continuous fields data, they show that this disagreement is amplified because of the discrete nature of the classes and is in fact not as large when considering the percent tree cover of each of the classes.…”

“…This section provides a brief overview of the various Initiative II land cover data sets and provides insight into and compares their individual characteristics. For more in depth comparisons of the satellite‐derived data sets see Brown de Colstoun et al [2006] and Hansen and Reed [2000].…”

“… Hansen and Reed [2000] and Brown de Colstoun et al [2006] have compared the UMD and IGBP‐DIScover classifications, which are derived from the common 1992–1993 AVHRR data set. Hansen and Reed [2000] found that for broad classes such as forest/woodland, grass/shrubs, crops, etc, the per pixel agreement at 1 km resolution was 74%, decreasing to 48% when all common classes listed in Table 3 were included.…”

“…They also found that the overall agreement between IGBP‐DIScover and UMD was much greater (∼80% to ∼68%) at 1/2° resolution than the agreement of two well‐known nonsatellite land cover maps [ Olson et al , 1983; Matthews , 1985] which have been extensively used in the past for modeling studies. On the other hand, Brown de Colstoun et al [2006] analyzed the effects of the aggregation methods on the agreement of the two data sets and used the new data layers available in Initiative II to assess the areas of disagreement. They noted that when using a strictly dominant class criteria to label a pixel, the agreement between the two data sets increased with coarser resolution from 48% at 1 km to ∼52% at 1°.…”