Elsevier

Remote Sensing of Environment

Volume 94, Issue 1, 15 January 2005, Pages 123-132
Remote Sensing of Environment

A comparative analysis of the Global Land Cover 2000 and MODIS land cover data sets

https://doi.org/10.1016/j.rse.2004.09.005Get rights and content

Abstract

Accurate and up-to-date global land cover data sets are necessary for various global change research studies including climate change, biodiversity conservation, ecosystem assessment, and environmental modeling. In recent years, substantial advancement has been achieved in generating such data products. Yet, we are far from producing geospatially consistent high-quality data at an operational level. We compared the recently available Global Land Cover 2000 (GLC-2000) and MODerate resolution Imaging Spectrometer (MODIS) global land cover data to evaluate the similarities and differences in methodologies and results, and to identify areas of spatial agreement and disagreement. These two global land cover data sets were prepared using different data sources, classification systems, and methodologies, but using the same spatial resolution (i.e., 1 km) satellite data. Our analysis shows a general agreement at the class aggregate level except for savannas/shrublands, and wetlands. The disagreement, however, increases when comparing detailed land cover classes. Similarly, percent agreement between the two data sets was found to be highly variable among biomes. The identified areas of spatial agreement and disagreement will be useful for both data producers and users. Data producers may use the areas of spatial agreement for training area selection and pay special attention to areas of disagreement for further improvement in future land cover characterization and mapping. Users can conveniently use the findings in the areas of agreement, whereas users might need to verify the informaiton in the areas of disagreement with the help of secondary information. Learning from past experience and building on the existing infrastructure (e.g., regional networks), further research is necessary to (1) reduce ambiguity in land cover definitions, (2) increase availability of improved spatial, spectral, radiometric, and geometric resolution satellite data, and (3) develop advanced classification algorithms.

Introduction

Two new global land cover data sets, Global Land Cover 2000 (GLC-2000) and MODerate resolution Imaging Spectrometer (MODIS) global land cover (MODIS land cover) have recently became available. The Joint Research Center (JRC) of the European Commission (EC) implemented the GLC-2000 project in partnership with more than 30 partner institutions around the world, using Satellite Pour l'Observation de la Terre (SPOT) VEGETATION 1-km satellite data (Fritz et al., 2003). Boston University prepared the MODIS land cover data using MODIS 1-km satellite data on board the Terra satellite (Friedl et al., 2002).

Both data sets were prepared with the same fundamental goal: to improve our understanding of the extent and distribution of the major land cover types of the world. The information generated can be used for various applications including ecosystem and biodiversity assessments, climate change studies, and environmental modeling (Brown et al., 1999, Giri et al., 2003, Loveland & Belward, 1997, Loveland et al., 1999, Reed, 1997). The main objective of GLC-2000 was to prepare a harmonized land cover database of the world for the year 2000, primarily to serve international assessment programs such as the Millennium Ecosystem Assessment (MA) and the United Nation's Ecosystem-related International Conventions (GLC, 2003). The MODIS land cover was prepared for NASA's Earth Observing System (EOS) MODIS land science team.

These two global land cover data sets were prepared using different data sources, classification schemes, and methodologies, but using the same spatial resolution (1 km) satellite data. Similar classification efforts in the past show wide variations in the estimation of global land cover (Hansen & Reed, 2000, Townshend et al., 1991). This is not surprising given the fact that quantitative analyses of complex land cover types remains an arduous task (Running et al., 1994, Zhu & Walter, 2003). Nevertheless, with the availability of improved spatial, spectral, geometric, and radiometric resolution satellite data (e.g., MODIS and VEGETATION), ground-truth data, and improved classification algorithms, it is possible to produce comprehensive and geospatially consistent global land cover data sets (Justice et al., 2002, Friedl et al., 2002).

The recent release of the GLC-2000 and MODIS land cover data sets call for a comparative analysis to examine their similarities and differences in terms of both methodology and results. It is also critical that both the data producers and users be aware of strengths and limitations of these data products. For example, the global area totals of land cover types could be similar; however, their spatial agreement could be vastly different. Furthermore, both area totals and spatial agreement could vary from region to region. From the producers' perspective, it is important to identify both areas of spatial agreement and disagreement. The areas of spatial agreement could be used as one of the ancillary data sets during training areas selection, and areas of disagreement could help identify issues for further improvement in future land cover characterization and mapping. In-depth understanding of similarities and differences will help users make informed decisions regarding the selection of global land cover data needed for their specific application. Similarly, users could conveniently utilize the data in the areas of agreement, whereas, in the areas of disagreement, users might need to verify the information with the help of secondary information.

The objective of this research is to summarize the similarities and differences in methodology and results of the GLC-2000 and MODIS land cover, and to identify areas of spatial agreement and disagreement.

Section snippets

Methodological similarities and differences

Before comparing the results, it is essential to understand the methodological similarities and differences between the GLC-2000 and MODIS land cover, which are summarized in Table 1.

The GLC-2000 was based primarily on SPOT VEGETATION daily 1-km data. However, some other data sources such as SPOT VEGETATION 10-day mosaic (S-10) data (of Southeast Asia), and the Normalized Difference Vegetation Index (NDVI), radar, and Defense Meteorological Satellite Program (DMSP) data (of Africa) were also

Data sources and methodology

The GLC-2000 v1.1 data were downloaded from the worldwide web at URL http://www.gvm.sai.jrc.it/glc2000/ (last accessed 24 February 2004). We acquired the data in the Geographic Coordinate projection system, and reprojected it to Interrupted Goode Homolosine projection system. The MODIS land cover data were acquired from the Earth Observing System (EOS) Data Gateway, http://edcimswww.cr.usgs.gov/pub/imswelcome/ (last accessed 12 February 2004). The data, which are available in Hierarchical Data

Results and discussion

This section first compares the global area totals obtained from GLC-2000 and MODIS global land cover data for both generalized and IGBP land cover classes. Then a per-pixel comparison of aggregated classes and spatial comparisons by biome are discussed. Finally, strengths and weaknesses of both data sets are discussed.

Conclusions

The foregoing comparative analyses provide insight for both data produces and users. For data producers, the identified areas of agreement may serve as a reference data for training areas selection. Likewise, areas of disagreement may receive special attention in future land cover characterization and mapping. Users also will have an opportunity to examine the similarities and differences in their area of interest, and make informed decisions based on their thematic applications. For example,

Acknowledgements

We would like to thank Craig Walters for editing the manuscript for language and clarity.

This work was made possible in part by Science Applications International Corporation under U.S. Geological Survey contract 1434-CR-97-CN-40274.

References (22)

  • C. Giri et al.

    Land cover characterization and mapping of continental Southeast Asia using multi-resolution satellite sensor data

    International Journal of Remote Sensing

    (2003)
  • Cited by (0)

    View full text