• Bacour, C., Baret F. , Béal D. , Weiss M. , and Pavageau K. , 2006: Neural network estimation of LAI, fAPAR, fCover and LAIxCabb, from top of canopy MERIS reflectance data: Principles and validation. Remote Sens. Environ., 105, 313325.

    • Search Google Scholar
    • Export Citation
  • Baret, F., and Coauthors, 2007: LAI, fAPAR and fCover CYCLOPES global products derived from VEGETATION: Part 1: Principles of the algorithm. Remote Sens. Environ., 110, 275286.

    • Search Google Scholar
    • Export Citation
  • Bonan, G., 2008: Forests and climate change: Forcings, feedbacks, and the climate benefits of forests. Science, 320, 14441449.

  • Camacho-de-Coca, F., Jiménez-Muñoz J. C. , Martinez B. , Bicheron P. , Lacaze R. , and Leroy M. , 2006: Prototyping of fCover product over Africa based on existing CYCLOPES and JRC products for VGT4Africa. Proc. Second Int. Symp. on Recent Advances in Quantitative Remote Sensing, Valencia, Spain, Universitat de Valencia, 722–727.

    • Search Google Scholar
    • Export Citation
  • De Colstoun, E. C. B., DeFries R. S. , and Townshend J. R. G. , 2006: Evaluation of ISLSCP Initiative II satellite-based land cover data sets and assessment of progress in land cover data for global modelling. J. Geophys. Res., 111, D22S07, doi:10.1029/2006JD007453.

    • Search Google Scholar
    • Export Citation
  • Di Gregorio, A., and Jansen L. , 2000: Land Cover Classifications System (LCCS), Classification Concepts and User Manual. Food and Agriculture Organisation, 179 pp.

    • Search Google Scholar
    • Export Citation
  • Ek, M. B., Mitchell K. E. , Lin Y. , Rogers E. , Grunmann P. , Koren V. , Gayno G. , and Tarpley J. D. , 2003: Implementation of the upgraded Noah land surface model in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108, 8851, doi:10.1029/2002JD003296.

    • Search Google Scholar
    • Export Citation
  • Fang, H., Liang S. , Townshend J. , and Dickinson R. , 2008: Spatially and temporally continuous LAI data sets based on a new filtering method: Examples from North America. Remote Sens. Environ., 112, 7593.

    • Search Google Scholar
    • Export Citation
  • FAO/IIASA/ISRIC/ISSCAS/JRC, cited 2009: Harmonized World Soil Database (version 1.1). [Available online at http://www.iiasa.ac.at/Research/LUC/External-World-soil-database/HTML/.]

    • Search Google Scholar
    • Export Citation
  • Faroux, S., Roujean J.-L. , Kaptué A. , and Masson V. , 2009: The ECOCLIMAP-II land surface database over Europe. CNRM/Météo France Tech. Note 86, 120 pp. [Available from CNRM/Météo France, 42 Av. G. Coriolis, 31057 Toulouse CEDEX 01, France.]

    • Search Google Scholar
    • Export Citation
  • Feddema, J. J., Oleson K. W. , Bonan G. , Mearns L. O. , Buja L. E. , Meehl G. A. , and Washington W. M. , 2005: The importance of land cover change in simulating future climate. Science, 310, 16741678.

    • Search Google Scholar
    • Export Citation
  • Foley, J. A., and Coauthors, 2005: Global consequences of land use. Science, 309, 570574.

  • Fritz, S., and See L. , 2008: Quantifying uncertainty and spatial disagreement in the comparison of global land cover for different applications. Global Change Biol., 14, 10571075.

    • Search Google Scholar
    • Export Citation
  • Garrigues, S., Allard D. , Baret F. , and Morisette J. , 2008a: Multivariate quantification of landscape spatial heterogeneity using variogram models. Remote Sens. Environ., 112, 216230.

    • Search Google Scholar
    • Export Citation
  • Garrigues, S., and Coauthors, 2008b: Validation and intercomparison of global leaf area index products derived from remote sensing data. J. Geophys. Res., 113, G02028, doi:10.1029/2007JG000635.

    • Search Google Scholar
    • Export Citation
  • Gómez, J. M., Valladares F. , and Puerta-Piñero C. , 2004: Differences between structural and functional environmental heterogeneity caused by seed dispersal. Funct. Ecol., 18, 787792.

    • Search Google Scholar
    • Export Citation
  • Gutman, G., 1999: On the use of long-term global data of land reflectances and vegetation indices derived from the advanced very high resolution radiometer. J. Geophys. Res., 104, 62416255.

    • Search Google Scholar
    • Export Citation
  • Gutman, G., Tarpley D. , Ignatov A. , and Olson S. , 1995: The enhanced NOAA global land dataset from the Advanced Very High Resolution Radiometer. Bull. Amer. Meteor. Soc., 76, 11411156.

    • Search Google Scholar
    • Export Citation
  • Hall, F. G., and Coauthors, 2006: The ISLSCP Initiative II global data sets: Surface boundary conditions and atmospheric forcings for land-atmosphere studies. J. Geophys. Res., 111, D22S01, doi:10.1029/2006JD007366.

    • Search Google Scholar
    • Export Citation
  • Houldcroft, C. J., Grey W. M. F. , Barnsley M. , Taylor C. M. , Los S. O. , and North P. R. J. , 2009: New vegetation albedo parameters and global fields of soil background albedo derived from MODIS for use in a climate model. J. Hydrometeor., 10, 183198.

    • Search Google Scholar
    • Export Citation
  • Imhoff, M. L., Bounoua L. , Ricketts T. , Loucks C. , Harriss R. , and Lawrence W. T. , 2004: Global patterns in human consumption of net primary production. Nature, 429, 870873.

    • Search Google Scholar
    • Export Citation
  • Jin, M., and Liang S. , 2006: An improved land surface emissivity parameter for land surface models using global remote sensing observations. J. Climate, 19, 28672881.

    • Search Google Scholar
    • Export Citation
  • Justice, C. O., Townshend J. R. G. , Vermote E. F. , Masuoka E. , Wolfe R. E. , Saleous N. , Roy D. P. , and Morisette J. T. , 2002: An overview of MODIS Land data processing and product status. Remote Sens. Environ., 83, 315.

    • Search Google Scholar
    • Export Citation
  • Kaptué, T. A. T., 2010: Mapping ecosystem and biophysical satellite parameters for the study of hydric fluxes over the African continent. Ph.D dissertation, Paul Sabatier University of Toulouse, 192 pp. [Available online at http://thesesups.ups-tlse.fr/994/.]

    • Search Google Scholar
    • Export Citation
  • Kaptué, T. A. T., De Jong S. M. , Roujean J.-L. , Favier C. , and Mering C. , 2010a: Ecosystem mapping at the African continent scale using a hybrid clustering approach based on 1-km resolution multi-annual data from SPOT/VEGETATION. Remote Sens. Environ., 115, 452464, doi:10.1016/j.rse.2010.09.015.

    • Search Google Scholar
    • Export Citation
  • Kaptué, T. A. T., Roujean J.-L. , and De Jong S. M. , 2010b: Comparison and relative quality assessment of the GLC2000, GLOBCOVER, MODIS and ECOCLIMAP land cover data sets at the African continental scale. Int. J. Appl. Earth Obs. Geoinf., 13, 207219, doi:10.1016/j.jag.2010.11.005.

    • Search Google Scholar
    • Export Citation
  • Kaptué, T. A. T., Roujean J.-L. , and Faroux S. , 2010c: ECOCLIMAP-II: An ecosystem classification and land surface parameter database of western Africa at 1 km resolution for the Africa Monsoon Multidisciplinary Analysis (AMMA) project. Remote Sens. Environ., 114, 961976.

    • Search Google Scholar
    • Export Citation
  • Kolasa, J., and Rollo C. D. , 1991: Introduction: The heterogeneity of heterogeneity: A glossary. Ecological Heterogeneity, J. Kolasa and S. T. A. Pickett, Eds., Springer, 1–23.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., Suarez M. J. , Ducharne A. , Stieglitz M. , and Kumar P. , 2000: A catchment-based approach to modeling land surface processes in a general circulation model 1. Model structure. J. Geophys. Res., 105, 24 80924 822.

    • Search Google Scholar
    • Export Citation
  • Liu, J., and Coauthors, 2009: Validation of Moderate Resolution Imaging Spectroradiometer (MODIS) albedo retrieval algorithm: Dependence of albedo on solar zenith angle. J. Geophys. Res., 114, D01106, doi:10.1029/2008JD009969.

    • Search Google Scholar
    • Export Citation
  • Los, S. O., and Coauthors, 2000: A global 9-yr biophysical land surface dataset from NOAA AVHRR data. J. Hydrometeor., 1, 183199.

  • Lucht, W., Schaaf C. B. , and Strahler A. H. , 2000: An algorithm for the retrieval of albedo from space using semi-empirical BRDF models. IEEE Trans. Geosci. Remote Sens., 38, 977998.

    • Search Google Scholar
    • Export Citation
  • Masson, V., Champeaux J.-L. , Chauvin F. , Meriguer C. , and Lacaze R. , 2003: A global database of land surface parameters at 1-km resolution in meteorological and climate models. J. Climate, 16, 12611282.

    • Search Google Scholar
    • Export Citation
  • Miller, J., Barlage M. , Zeng X. , Wei H. , Mitchell K. and Tarpley D. , 2006: Sensitivity of the NCEP/Noah land surface model to the MODIS green vegetation fraction data set. Geophys. Res. Lett., 33, L13404, doi:10.1029/2006GL026636.

    • Search Google Scholar
    • Export Citation
  • Moody, E. G., King M. D. , Platnick S. , Schaaf C. B. , and Gao F. , 2005: Spatially complete global spectral surface albedos: Value-added datasets derived from Terra MODIS land products. IEEE Trans. Geosci. Remote Sens., 43, 144158.

    • Search Google Scholar
    • Export Citation
  • Morisette, J. T., and Coauthors, 2006: Validation of global moderate-resolution LAI products: A framework proposed within the CEOS land product validation subgroup. IEEE Trans. Geosci. Remote Sens., 44, 18041814.

    • Search Google Scholar
    • Export Citation
  • Myneni, R. B., and Coauthors, 2002: Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data. Remote Sens. Environ., 83, 214231.

    • Search Google Scholar
    • Export Citation
  • Noilhan, J., and Mahfouf J.-F. , 1996: The ISBA land surface parametrisation scheme. Global Planet. Change, 13, 145159.

  • Patz, J. A., and Coauthors, 2004: Unhealthy landscapes: Policy recommendations on land use change and infectious disease emergence. Environ. Health Perspect., 112, 10921098.

    • Search Google Scholar
    • Export Citation
  • Pitman, A. J., 2003: The evolution of, and revolution in, land surface schemes designed for climate models. Int. J. Climatol., 23, 479510.

    • Search Google Scholar
    • Export Citation
  • Rechid, D., Raddatz T. J. , and Jacob D. , 2009: Parameterization of snow-free land surface albedo as a function of vegetation phenology based on MODIS data and applied in climate modelling. Theor. Appl. Climatol., 95, 245255.

    • Search Google Scholar
    • Export Citation
  • Román, M. O., and Coauthors, 2009: The MODIS (Collection V005) BRDF/albedo product: Assessment of spatial representativeness over forested landscapes. Remote Sens. Environ., 113, 24762498.

    • Search Google Scholar
    • Export Citation
  • Román, M. O., and Coauthors, 2010: Assessing the coupling between surface albedo derived from MODIS and the fraction of diffuse skylight over spatially-characterized landscapes. Remote Sens. Environ., 114, 738760.

    • Search Google Scholar
    • Export Citation
  • Roujean, J.-L., and Lacaze R. , 2002: Global mapping of vegetation parameters from POLDER multiangular measurements for studies of surface-atmosphere interactions: A pragmatic method and validation. J. Geophys. Res., 107, 4150, doi:10.1029/2001JD000751.

    • Search Google Scholar
    • Export Citation
  • Running, S. W., Loveland T. R. , Pierce L. L. , Nemani R. , and Hunt E. R. , 1995: A remote sensing based vegetation classification logic for global land-cover analysis. Remote Sens. Environ., 51, 3948.

    • Search Google Scholar
    • Export Citation
  • Salmun, H., and Molod A. , 2006: Progress in modelling the impact of land cover change on the global climate. Prog. Phys. Geogr., 30, 737749.

    • Search Google Scholar
    • Export Citation
  • Schaaf, C. B., and Coauthors, 2002: First operational BRDF, albedo nadir reflectance products from MODIS. Remote Sens. Environ., 83, 135148.

    • Search Google Scholar
    • Export Citation
  • Schaaf, C. B., Cihlar J. , Belward A. , Dutton E. , and Verstraete M. , 2009: Albedo and reflectance anisotropy, ECV-T8. GTOS Assessment of the Status of the Development of Standards for the Terrestrial Essential Climate Variables, R. Sessa, Ed., FAO, 16 pp.

    • Search Google Scholar
    • Export Citation
  • Schaaf, C. B., Liu J. , Gao F. , and Strahler A. H. , 2011: Aqua and Terra MODIS albedo and reflectance anisotropy products. Land Remote Sensing and Global Environmental Change: NASA’s Earth Observing System and the Science of ASTER and MODIS, B. Ramachandran, C. Justice, and M. Abrams, Eds., Remote Sensing and Digital Image Processing Series, Vol. 11, Springer-Verlag, 549–561.

    • Search Google Scholar
    • Export Citation
  • Schaepman-Strub, G., Schaepman M. E. , Painter T. H. , Dangel S. , and Martonchik J. V. , 2006: Reflectance quantities in optical remote sensing—Definitions and case studies. Remote Sens. Environ., 103, 2742.

    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., Los S. O. , Tucker C. J. , Justice C. O. , Dazlich D. A. , Collatz G. J. , and Randall D. A. , 1996a: A revised land surface parameterization (SiB2) for atmospheric GCMs. Part II: The generation of global fields of terrestrial biophysical parameters from satellite data. J. Climate, 9, 706737.

    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., and Coauthors, 1996b: The ISLSCP Initiative I global datasets: Surface boundary conditions and atmospheric forcings for land–atmosphere studies. Bull. Amer. Meteor. Soc., 77, 19872005.

    • Search Google Scholar
    • Export Citation
  • Strahler, A. H., and Coauthors, 2006: Global land cover validation: Recommendations for evaluation and accuracy assessment of global land cover maps. GOFC-GOLD Rep. 25, 51 pp.

    • Search Google Scholar
    • Export Citation
  • Strugnell, N., and Lucht W. , 2001: An algorithm to infer continental-scale albedo from AVHRR data, land cover class, and field observations of typical BRDFs. J. Climate, 14, 13601376.

    • Search Google Scholar
    • Export Citation
  • Sutherland, W. J., and Coauthors, 2009: One hundred questions of importance to the conservation of importance to the conservation of global biological diversity. Conserv. Biol., 23, 557567.

    • Search Google Scholar
    • Export Citation
  • Trigo, I. F., and Coauthors, 2010: The satellite application facility on land surface analysis. Int. J. Remote Sens., 32, 27252744, doi:10.1080/01431161003743199.

    • Search Google Scholar
    • Export Citation
  • Tucker, C. J., 1979: Red and photographic infrared linear combinations for monitoring vegetation. Remote Sens. Environ., 8, 127150.

  • Tucker, C. J., Townshend J. R. G. , and Goff T. E. , 1985: African land-cover classification using satellite data. Science, 227, 369374.

    • Search Google Scholar
    • Export Citation
  • Turner, W. R., Katrina B. , Brooks T. M. , Costanza R. , Da Fonseca G. A. B. , and Portela R. , 2007: Global conservation of biodiversity and ecosystem services. Bioscience, 57, 868873.

    • Search Google Scholar
    • Export Citation
  • Williams, C. W., Hanan N. P. , Neff J. C. , Scoles R. J. , Berry J. A. , Denning A. S. , and Baker D. F. , 2007: Africa and the global carbon cycle. Carbon Balance Manage., 2, 113.

    • Search Google Scholar
    • Export Citation
  • Wilson, M. F., and Henderson-Sellers A. , 1985: A global archive of land cover and soils data for use in general circulation models. J. Climatol., 5, 119143.

    • Search Google Scholar
    • Export Citation
  • Yang, W., Shabanov N. V. , Huang D. , Wang W. , Dickinson R. E. , Nemani R. R. , Knyazikhin Y. , and Myneni R. B. , 2006: Analysis of leaf area index products from combination of MODIS Terra and Aqua data. Remote Sens. Environ., 104, 297312.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    The simplified seven-class ECOCLIMAP-II land cover map (adapted from Kaptué et al. 2010a). The number in brackets refers to the percentage of the coverage area.

  • View in gallery

    Dominant major soil groups (FAO/IIASA/ISRIC/ISSCAS/JRC 2009) mapped for Africa. Note that anthrosols, greyzems, and urban have less than 1500 pixels while the other soil groups have more than 19 500 pixels.

  • View in gallery

    Topsoil (left) sand and (right) clay fraction maps for Africa (FAO/IIASA/ISRIC/ISSCAS/JRC 2009).

  • View in gallery

    Calculated bare soil albedo maps in (left) the visible and (right) the near-infrared parts of the spectrum.

  • View in gallery

    Histograms showing the bare soil albedos for eight major soil groups in the visible (dark gray) and near-infrared (light gray) parts of the spectrum. The x axis represents the albedo values and the y axis the number of pixels.

  • View in gallery

    Calculated vegetation albedo maps in (left) the visible and (right) the near-infrared parts of the spectrum.

  • View in gallery

    Differences between the annual mean of fractional vegetation coverage of (left) CYCLOPES and ECOCLIMAP-I and (right) ECOCLIMAP-II. The numbers r, RMSE, and AB respectively indicate the correlation coefficient, the RMSE, and the absolute bias between both versions of ECOCLIMAP and CYCLOPES, calculated over vegetated land (i.e., forest/mixed forest, woodland/shrubland, cropland, and grassland).

  • View in gallery

    Spatial pattern of differences of bare soil albedo of ECOCLIMAP-II with (top) ECOCLIMAP-I, (middle) Swansea University, and (bottom) ISLSCP-II in the (left) visible and (right) near-infrared parts of the spectrum.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 88 78 5
PDF Downloads 43 42 4

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

View More View Less
  • 1 CNRM/GAME URA 1357, Météo France, Toulouse, France
  • | 2 UMR TETIS, CIRAD, Montpellier, France
  • | 3 Department of Geography, Swansea University, Swansea, United Kingdom
  • | 4 CNRM/GAME URA 1357, Météo France, Toulouse, France
  • | 5 Ecole Supérieure Polytechnique, UCAD, Dakar, Sénégal
© Get Permissions
Full access

Abstract

Information related to land surface is immensely important to global change science. For example, land surface changes can alter regional climate through its effects on fluxes of water, energy, and carbon. In the past decades, data sources and methodologies for characterizing land surface heterogeneity (e.g., land cover, leaf area index, fractional vegetation cover, bare soil, and vegetation albedos) from remote sensing have evolved rapidly. The double ECOCLIMAP database—constituted of a land cover map and land surface variables and derived from Advanced Very High Resolution Radiometer (AVHRR) observations acquired between April 1992 and March 1993—was developed to support investigations that require information related to spatiotemporal dynamics of land surface. Here is the description of ECOCLIMAP-II: a new characterization of the land surface heterogeneity based on the latest generation of sensors, which represents an update of the ECOCLIMAP-I database over Africa. Owing to the many features of the Moderate Resolution Imaging Spectroradiometer (MODIS) sensors (more accurate in spatial resolution and spectral information compared to the AVHRR sensor), a variety of methods have been developed for an extended period of 8 yr (2000–07) to strengthen consistency between land surface variables as required by the meteorological and ecological communities. The relative accuracy (or performance) quality of ECOCLIMAP-II was assessed (i.e., by comparison with other global datasets). Results illustrate a substantial refinement; for instance, the fractional vegetation cover resulting in a root-mean-square error of 34% instead of 64% in comparison with the original version of ECOCLIMAP.

Corresponding author address: Armel Kaptué, CNRM, 42 Av. G. Coriolis, 31057 Toulouse CEDEX 01, France. E-mail: armel.kaptue@cnrm.meteo.fr

Abstract

Information related to land surface is immensely important to global change science. For example, land surface changes can alter regional climate through its effects on fluxes of water, energy, and carbon. In the past decades, data sources and methodologies for characterizing land surface heterogeneity (e.g., land cover, leaf area index, fractional vegetation cover, bare soil, and vegetation albedos) from remote sensing have evolved rapidly. The double ECOCLIMAP database—constituted of a land cover map and land surface variables and derived from Advanced Very High Resolution Radiometer (AVHRR) observations acquired between April 1992 and March 1993—was developed to support investigations that require information related to spatiotemporal dynamics of land surface. Here is the description of ECOCLIMAP-II: a new characterization of the land surface heterogeneity based on the latest generation of sensors, which represents an update of the ECOCLIMAP-I database over Africa. Owing to the many features of the Moderate Resolution Imaging Spectroradiometer (MODIS) sensors (more accurate in spatial resolution and spectral information compared to the AVHRR sensor), a variety of methods have been developed for an extended period of 8 yr (2000–07) to strengthen consistency between land surface variables as required by the meteorological and ecological communities. The relative accuracy (or performance) quality of ECOCLIMAP-II was assessed (i.e., by comparison with other global datasets). Results illustrate a substantial refinement; for instance, the fractional vegetation cover resulting in a root-mean-square error of 34% instead of 64% in comparison with the original version of ECOCLIMAP.

Corresponding author address: Armel Kaptué, CNRM, 42 Av. G. Coriolis, 31057 Toulouse CEDEX 01, France. E-mail: armel.kaptue@cnrm.meteo.fr

1. Introduction

Land surface properties play an important role in various aspects of global change studies including biogeochemical processes (carbon cycle), biogeophysical processes (evapotranspiration), and biogeographical processes (phenology) (Feddema et al. 2005; Bonan 2008). Land surface changes can alter regional climates through its effects on the net radiation, the hydrologic cycle, the division of energy balance at the surface–atmosphere interface into sensible and latent heat, and the partitioning of precipitation into soil water, evapotranspiration, and runoff (Tucker et al. 1985; Foley et al. 2005). However, the accurate representation of these phenomena attuned to uses other than climate is sometimes necessary. For instance, changes in vegetation and soil characteristics can affect several aspects of human health (Patz et al. 2004) and the land’s capacity to support human life (Sutherland et al. 2009). Changes in land surface can also cause declines in biodiversity (Imhoff et al. 2004; Turner et al. 2007). Therefore, in order to correctly describe the earth system, it is first necessary to have information on landscape heterogeneity (Salmun and Molod 2006; Garrigues et al. 2008a). This can be obtained by characterizing spatial and temporal variability of global surface biophysical variables; for example, leaf area index (LAI) and fraction of absorbed photosynthetically active radiation, surface albedo and reflectance anisotropy (Schaaf et al. 2009), and land cover (Strahler et al. 2006).

Although the “heterogeneity” concept is a notion that is complex to be defined unequivocally (Kolasa and Rollo 1991), state of the art describes it in a structural or functional manner (e.g., Gómez et al. 2004). According to the structural approach, a phenomenon is heterogeneous if one or many variables representing it vary in space; the heterogeneity here is considered as a static property (e.g., a land cover class). On the other hand, according to the functional approach, a phenomenon is heterogeneous if the intensity of subjacent variables varies in space in response to structural variations of the system (e.g., the density of the vegetation within a land cover class). In the following, the characterization of the land surface heterogeneity will be referred to both identification of structural and functional descriptors of surfaces as well as the description of their seasonal evolution and interannual variation for scales ranging from local to global.

For studying the impact of land surface heterogeneity on the Earth system, the most commonly used tools are land surface models (LSMs). This category of models simulates global-scale land surface water and energy fluxes. They can be driven by an atmospheric forcing dataset, or can be coupled to an atmospheric general circulation model. A comprehensive review of the development of LSMs over the last decades can be found in Pitman (2003). Since the scale of the land surface heterogeneities may be smaller than the grid scale in LSMs, a most commonly used technique in current LSMs to attempt to account for the subgrid-scale heterogeneity is the “tile approach” (Salmun and Molod 2006). Separate surface heat and moisture balance equations are solved for each functional type contained in a LSM grid cell and the resulting heat and moistures fluxes are aggregated linearly. Nonetheless, LSM modeling needs a large number of inputs and variables that depend on functional types. These relatively complex models need to be calibrated through the prescription of a number of measured variables describing notably vegetation characteristics (albedo, leaf area index, and fractional vegetation cover). These key biophysical variables are used in LSMs in controlling fluxes of water, energy, and carbon dioxide (Baret et al. 2007; Hall et al. 2006). LAI controls the plant transpiration and the absorbed solar radiation, while veg (the fractional vegetation cover) indicates the horizontal heterogeneity of vegetation and affects division between transpiration and evaporation and other fluxes. Albedo determines the radiation balance by controlling the available solar energy to the earth-surface atmosphere (Sellers et al. 1996b).

With a quasi-continuous and exhaustive mapping of Earth’s surface over the last 30 years, satellite remote sensing has been, and is being, explored as a unique source of information to provide representative measurements over large areas. 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). The strong chlorophyll absorption in the red and the high reflectance in the near-infrared portion for most soil types provide a unique contrast, which permits us to distinguish vegetation from other Earth-surface components. 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). The consistency of biophysical variables derived by different methods (even when they are produced from the same instrument) is an essential requirement arising from the meteorological (and ecological) communities. It still represents today a challenge to the remote sensing community (De Colstoun et al. 2006).

In this regard, the assessment of new and improved land surface datasets is central to a number of programs and experiments—for example, the data collections of the International Satellite Land Surface Climatology Program (ISLSCP; Sellers et al. 1996b; Hall et al. 2006) and the ECOCLIMAP database (Masson et al. 2003). The aim of the ECOCLIMAP database is to answer the needs of the meteorological community in investigating natural and managed functionally homogeneous ecosystems in connection with modeling climate change extent and impacts. ECOCLIMAP provides, per land cover, and in a tabular form, datasets of surface variables: albedo, LAI, fraction of vegetation cover, roughness length, minimum stomatal resistance, and root zone in order to initialize LSMs. With this database, modelers can download ECOCLIMAP products to ensure a broad use in the context of their applications. From the basic resolution of 1/120° (i.e., approximately 1 km), the biophysical variables provided by ECOCLIMAP can be aggregated to their respective model resolutions.

The aim of this paper is to present the new database ECOCLIMAP-II—a more realistic characterization of the land surface heterogeneity—by describing the changes and the improvements that were made since the first publication of the ECOCLIMAP database (ECOCLIMAP-I hereafter). We focused our experiences on the African mainland, which has been identified in a recent review by Williams et al. (2007) as “one of the weakest links in our understanding of the global carbon cycle.” With an area covering 30 221 532 km2, Africa represents 20% of the continental surfaces. The region encompasses a large variety of ecosystems from humid tropical forest to semiarid and arid grass and shrub communities. Weaknesses in the first database were addressed. First is the use of the NOAA AVHRR normalized difference vegetation index (NDVI), which was 1.1 km at the equator and was reprocessed to obtain the final NDVI product at 1 km resolution. Secondly, the land covers were defined into preexisting climatic zones, which do not necessarily correspond to limits between ecosystems.

At the heart of the approach to derive the new database ECOCLIMAP-II is a new stratification of the continent into ecoregions that facilitates studies and mapping of vegetation at local to regional scales. The technical changes to the dataset and the methods developed to refine the consistency between the biophysical products are described in section 2. It includes the 8-yr period from 2000 to 2007, the actual production of LAI, veg, bare soil albedo, and vegetation albedo. Section 3 presents the new characterization of the land surface heterogeneity, while the assessment of its quality by comparison with existing land surface databases is shown in section 4. Finally, results are summarized and short-term perspectives are stressed in section 5.

2. Preparation of the new database

The biophysical variables were estimated from the following two datasets: 1) the recent ecosystem classification map prepared by Kaptué et al. (2010a) and 2) the earth observation MODIS data, which was used as the primary data source. The approach builds on the method first developed in Masson et al. (2003), and can also been seen as an application to the whole African continent of the methods implemented only over the western African region in Kaptué et al. (2010c).

a. ECOCLIMAP-II classification

The 1-km ECOCLIMAP-II classification (Kaptué et al. 2010a) derived from the VEGETATION NDVI data spanning an 8-yr period from 1 January 2000 to 31 December 2007 was used for the estimation of biophysical variables. This new land cover map was built using an unsupervised hybrid classification with successive refinements on the basis of ecoregion interpretations. This map also updates and complements existing continental products. Since it is of prime importance to identify and quantify uncertainty and spatial disagreement in land cover maps (Fritz and See 2008), comparison with the other classifications GLC2000, GLOBCOVER, and MODIS, also derived from remotely sensed data acquired in the 2000s, show that the per-pixel agreement at 1 km resolution was ranging between 56.3% and 68.8% (Kaptué et al. 2010b).

The 73 classes of the ECOCLIMAP-II classification were gathered into 7 aggregated classes (see Fig. 1) following the requirements of the Land Cover Classification Scheme (LCCS) of Di Gregorio and Jansen (2000) at level I: (i) the forest/mixed forest class contains evergreen or deciduous forest that is either closed or open, flooded forest, forest/savanna mosaic, and forest/cropland mosaic; (ii) the woodland/shrubland class includes lands with woody savanna or shrub canopy cover; (iii) the grassland class consists of herbaceous-type cover; (iv) the cropland class is land covered with rainfed, irrigated, or postflooded crops; (v) the urban and built-up class is covered by buildings and other man-made structures; (vi) the bare land class consists of barren or sparsely vegetated lands covered by sand, rocks, and dunes; and (vii) the inland water class includes natural and artificial inland water bodies. The latter aggregated classes are well suited for a plant functional type classification like the one of Running et al. (1995) used for the production of MODIS LAI data.

Fig. 1.
Fig. 1.

The simplified seven-class ECOCLIMAP-II land cover map (adapted from Kaptué et al. 2010a). The number in brackets refers to the percentage of the coverage area.

Citation: Journal of Hydrometeorology 12, 6; 10.1175/JHM-D-11-020.1

b. MODIS data

This study relies on an 8-yr data record (from 1 January 2000 to 31 December 2007) acquired by MODIS (Justice et al. 2002), which is an instrument on board Terra and Aqua satellites. The Collection 5 of MODIS LAI (Myneni et al. 2002; Yang et al. 2006) and white sky albedo (WSA; Schaaf et al. 2002; 2011) products are freely available through the Warehouse Inventory Search Tool (WIST; https://wist.echo.nasa.gov/api/) at a spatial resolution of 1/120°. The MODIS definition of LAI is the ratio of one-sided green foliage area per unit horizontal ground area (Yang et al. 2006). MODIS LAI is computed on a lookup table approach using the eight-biome classification of Running et al. (1995) with the radiative transfer approach of Myneni et al. (2002). The MODIS WSA is the dimensionless ratio of incident to reflected solar radiation by the earth’s surface under diffuse conditions, which is calculated by integration of the Bidirectional Reflectance Distribution Function (BRDF) over all viewing and illumination directions (Lucht et al. 2000; Schaepman-Strub et al. 2006).

Combined data from the observations provided by both Terra and Aqua (MCD15A2 and MCD43B3), when available, were preferred rather than the one produced solely using either Terra observations (MOD15A2 and MOD43B3) or Aqua observations (MYD15A2 and MYD43B3). MODIS Collection 5 LAI values are based on an 8-day period, versus 16 days for the surface albedo products. It is worth mentioning that two successive retrieval periods of albedo have an overlapping period of eight days. This resulted in the same number of 368 available samples of high-quality MODIS albedo retrievals during the ECOCLIMAP-II study (when excluding gap periods).

c. Disentangling broadband albedo into bare soil and vegetation albedos

In many LSMs (e.g., Noilhan and Mahfouf 1996; Koster et al. 2000; Ek et al. 2003), surface energy fluxes and some biophysical variables are calculated separately for each surface (vegetation and bare soil) and weighted by their fractional areas (Kaptué et al. 2010c). Hence the variable veg is the basis for the computation of many other biophysical variables in LSMs. For instance, surface albedo (A) is considered as a partition between visible (0.3–0.7 μm) and near-infrared (0.7–5.0 μm) spectral ranges, which are computed separately for vegetation and bare soil and then combined together.

For the visible and near-infrared spectral bands, four steps were carried out to estimate bare soil and vegetation albedos using time series of the fractional vegetation cover (veg) and total broadband surface albedo (A) based on a method developed by Kaptué et al. (2010c), which ensured consistency between LAI and surface albedo using the following equations:
e1
e2
The reconstructed land surface albedo is then computed by determining the adequate coefficients of the following equation:
e3
where Avis and Anir are respectively the surface (vegetation and bare soil) broadband albedo in the visible and the near-infrared parts of the spectrum.

The MODIS quality assessment data layers were used to retain in the time series only LAI values generated with high confidence by the main algorithm (Yang et al. 2006). Secondly, spatial interpolations were performed to fill gaps in the data and temporal interpolation to match LAI on the same temporal resolution of albedo (see next section). Thirdly, bare soil albedos were computed using an exponential relationship between veg and LAI and by applying a linear regression between A and veg on a pixel-by-pixel basis [see Eqs. (1)(2)]. Fourthly, the latter bare soil albedos were used to derive 8-day values of vegetation albedo. Note that all these steps are explained in great detail in Kaptué et al. (2010c). The multiyear values of vegetation albedo are averaged to obtain a temporally constant fixed value for each pixel (i.e., independently of season). In this context, veg includes all the seasonality of canopies. After a discussion about the possible assumptions of representation of vegetation (whether fractional vegetation coverage and/or other biophysical variables such as vegetation albedo or LAI vary seasonally or are kept constant) in LSMs, Miller et al. (2006) conclude that only veg or LAI can been used to plot the seasonality of the other variables.

d. Spatial and temporal interpolations

Even after some corrections performed on MODIS products during the retrieval and inversion process, there is residual noising in earth observation time series. For instance, cloud contamination, persistent clouds, and other suboptimal atmospheric or illumination conditions can reduce data quality. This can lead to data dropouts or data gaps, which is not desirable given that LSMs need continuous data input. The following spatial and temporal interpolations were applied to the MODIS products.

A spatial interpolation of MODIS products was performed by replacing missing or poor-quality observations for albedo and LAI datasets. A pixel with less than 27% clear data (i.e., less than about 100 of the 368 eight-day composites) is removed. A total number of 1 033 744 pixels representing about 2.65% of total land pixels (not water inland or oceanic) is discarded in this way. Then LAI values are linearly interpolated over time to comply with the temporal resolution of albedo. The reader can refer to Kaptué et al. (2010c) for further details about the gap-filling method. Note that some gap-filling methods have already been developed by the members of the MODIS Land Discipline Group and they are only applied over the North American mainland (Moody et al. 2005; Fang et al. 2008).

e. Quality assessment of land surface properties

The quality of remotely sensed products over large areas can be assessed absolutely (i.e., by comparison with in situ measurements) or relatively (i.e., by comparison with other datasets) (Morisette et al. 2006; Garrigues et al. 2008b; Kaptué et al. 2010b). In this study, only the second approach was used.

The new database was compared with other existing land surface databases presented in Table 1 such as ECOCLIMAP-I, CYCLOPES, and MODIS. The obtained bare soil albedo maps were compared with the bare soil albedo datasets of ECOCLIMAP-I (Masson et al. 2003), ISLSCP-II (Hall et al. 2006), and the Swansea University albedo (Houldcroft et al. 2009). To this end, the 1° ISLSCP-II bare soil albedos and 0.5° Swansea University bare soil albedos were reprojected over to the 1-km grid of ECOCLIMAP. We related the four bare soil albedo datasets to specific soil and geological features because information on soil type and composition—like the global database prepared by Wilson and Henderson-Sellers (1985)—is widely used in many LSMs to compute bare soil albedo. Major soil groups (Fig. 2) and their topsoil variables like sand and clay fraction (Fig. 3) come from the Harmonized World Soil Database (HWSD) (FAO/IIASA/ISRIC/ISSCAS/JRC 2009).

Table 1.

Presentation of the main characteristics of the 1-km databases used in this study.

Table 1.
Fig. 2.
Fig. 2.

Dominant major soil groups (FAO/IIASA/ISRIC/ISSCAS/JRC 2009) mapped for Africa. Note that anthrosols, greyzems, and urban have less than 1500 pixels while the other soil groups have more than 19 500 pixels.

Citation: Journal of Hydrometeorology 12, 6; 10.1175/JHM-D-11-020.1

Fig. 3.
Fig. 3.

Topsoil (left) sand and (right) clay fraction maps for Africa (FAO/IIASA/ISRIC/ISSCAS/JRC 2009).

Citation: Journal of Hydrometeorology 12, 6; 10.1175/JHM-D-11-020.1

3. Presentation of the new database

In this section we present only the results concerning the bare soil and vegetation albedos in the visible and infrared parts of the spectrum, obtained from solving Eqs. (1) and (2) in the two wave bands, respectively. The LAI and veg variables are not presented or discussed in this section, as the LAI [and consequently veg; see Eq. (2)] of ECOCLIMAP-II is equal to the MODIS LAI per land cover class.

a. Bare soil albedo

The obtained bare soil albedo patterns on which considerable spatial variability can be clearly seen are shown in Fig. 4. The highest values of bare soil albedo generally coincide with arid regions—in particular in the Sahara Desert but also in the Kalahari Desert and in the Horn of Africa. Lowest values mainly surround the Democratic Republic of Congo, especially over the transition zone between regions having a low percentage of available data and regions having a high percentage of available data.

Fig. 4.
Fig. 4.

Calculated bare soil albedo maps in (left) the visible and (right) the near-infrared parts of the spectrum.

Citation: Journal of Hydrometeorology 12, 6; 10.1175/JHM-D-11-020.1

The overlaying of texture, soil type, and bare soil albedo maps (shown in Figs. 2, 3, and 4) revealed that the most reflective soil types are shifting sands and dunes, followed by calcisols, rock debris, and gypsisols, while the lowest reflective soils types are ferralsols, greysems, histosols, and acrisols. Some of these soil types (e.g., gypsisols) have bimodal albedo histograms in the near infrared, and the difference between the modes of near-infrared and visible wave bands in ECOCLIMAP-II vary between 0.15 and 0.25 according to the soil type (Fig. 5). Nevertheless, the spatial soil type distribution of the Harmonized World Soil Database should be taken with caution since (i) Table 2 revealed that some soil types like acrisols have a sand percentage greater than the one of the dunes, which should be theoretically the most sandy; and (ii) Fig. 2 illustrated that there are many gaps in the sand and clay fraction. Also, in many cases the leaf litter layer covers the soil and is therefore more important for the soil albedo than the underlying soil type (e.g., Sellers et al. 1996a).

Fig. 5.
Fig. 5.

Histograms showing the bare soil albedos for eight major soil groups in the visible (dark gray) and near-infrared (light gray) parts of the spectrum. The x axis represents the albedo values and the y axis the number of pixels.

Citation: Journal of Hydrometeorology 12, 6; 10.1175/JHM-D-11-020.1

Table 2.

Examples of mean values and standard deviation values of topsoil sand and clay fraction for eight soil groups (FAO/IIASA/ISRIC/ISSCAS/JRC 2009).

Table 2.

b. Vegetation albedo

Figures 6a,b show the spatial pattern of the temporally fixed derived vegetation albedo values over the African continent. High values are mainly located over croplands and grasslands while low values of vegetation albedo occur over forests. This can be understood by the fact that (i) over croplands and grasslands, low values of fractional vegetation coverage are observed and hence the obtained vegetation albedo values are still contaminated by the remaining effect of bare soil; and (ii) in the lower altitudes, low solar energy input and low surface temperatures slow down the decomposition of plant material, and so much organic material is stored in the soils. Therefore, the new vegetation albedo maps show realistic results, except for some barren lands in Algeria and Libya where we denote high uncertainties of the MODIS LAI values. Indeed, over such locations, the analysis of the quality assessment of the MODIS LAI datasets reveals that the main algorithm was used under reflectance saturation conditions. Nonetheless, the fractional vegetation cover is null over these regions since they are barren, so that the reconstructed surface albedo (vegetation and soil) is judged to be realistic.

Fig. 6.
Fig. 6.

Calculated vegetation albedo maps in (left) the visible and (right) the near-infrared parts of the spectrum.

Citation: Journal of Hydrometeorology 12, 6; 10.1175/JHM-D-11-020.1

4. Evaluation of the new database

In this section we present the results concerning the fractional vegetation cover, the bare soil and vegetation albedos in the visible and infrared parts of the spectrum, and the reconstructed land surface albedo. The LAI variable is not presented here, as the LAI of ECOCLIMAP-II is equal to the MODIS LAI per land cover class.

a. Fractional vegetation coverage

Annual averages of fractional cover for both versions of ECOCLIMAP were compared with CYCLOPES fractional cover. Agreement was found in most areas; for discussion, we focus only on the differences (Fig. 7). ECOCLIMAP-II has higher veg values than CYCLOPES, excepted in the Congo basin where ECOCLIMAP-II is greater by up to 0.4 and the absolute bias (AB) over the aggregated class forest/mixed (shown in Fig. 1) is 0.27. The above biases may be attributed by the lack of the clumping representation over dense canopies in the CYCLOPES algorithm, which mainly occurred over forested areas. Indeed, Camacho-de-Coca et al. (2006) have shown that the magnitude of high values of veg is underestimated in CYCLOPES data. Thus, higher values in the ECOCLIMAP-II veg derived via the exponential relationship defined by Eq. (2) appear more realistic.

Fig. 7.
Fig. 7.

Differences between the annual mean of fractional vegetation coverage of (left) CYCLOPES and ECOCLIMAP-I and (right) ECOCLIMAP-II. The numbers r, RMSE, and AB respectively indicate the correlation coefficient, the RMSE, and the absolute bias between both versions of ECOCLIMAP and CYCLOPES, calculated over vegetated land (i.e., forest/mixed forest, woodland/shrubland, cropland, and grassland).

Citation: Journal of Hydrometeorology 12, 6; 10.1175/JHM-D-11-020.1

In contrast, veg of ECOCLIMAP-I shows higher inconsistencies with CYCLOPES in spatial distribution, although the mean values are roughly in agreement (r = 0.23 and RMSE = 0.63). The spatial inconsistency can in part be explained by the inadequate land cover representation in ECOCLIMAP-I. For a pixel dominated by high vegetation (forest or woodland), ECOCLIMAP-I gives almost a constant value around 0.9. This results in a high bias of 0.4 for areas with high vegetation cover (Fig. 7a). ECOCLIMAP-II and CYCLOPES veg are believed to be more reliable because of higher quality input satellite data than those derived from AVHRR of ECOCLIMAP-I on which a residual noise is observed (Gutman 1999).

b. Bare soil and vegetation albedo

Per-pixel comparisons of bare soil albedo of ECOCLIMAP-I and -II, ISLSCP-II, and Swansea University are shown in Fig. 8. On average, ECOCLIMAP-II gives visible and near-infrared bare soil albedos that are very approximately by about 0.02 with the ones of Swansea University. This situation may be due to the fact that among the four bare soil datasets, only ECOCLIMAP-II and Swansea University have been derived solely with remotely sensed data. Nonetheless, pronounced differences between ECOCLIMAP-II and Swansea University in the near infrared are observed over adjoining pixels of gaps encountered in Swansea University data, more precisely over the aggregated class forest/mixed forest where the contribution of vegetation albedo dominates the remote sensing signal.

Fig. 8.
Fig. 8.

Spatial pattern of differences of bare soil albedo of ECOCLIMAP-II with (top) ECOCLIMAP-I, (middle) Swansea University, and (bottom) ISLSCP-II in the (left) visible and (right) near-infrared parts of the spectrum.

Citation: Journal of Hydrometeorology 12, 6; 10.1175/JHM-D-11-020.1

Altogether, most vegetation canopies feature albedo values between 0.08 and 0.10 in the visible and between 0.22 and 0.24 in the near infrared. By using parameters defined in Kaptué et al. (2010c) of α = 0.02, β = 0.461, and γ = 0.455 for Eq. (3) to compute the total shortwave vegetation albedo, we determined that most vegetation albedo values reside between 0.14 and 0.16; this is in agreement with the study of Rechid et al. (2009). Therefore, the retrieval of bare soil albedo and vegetation albedo using Eqs. (1) and (2) is rigorous and at this moment, the new bare soil albedo and vegetation albedo maps of ECOCLIMAP-II are the ones having the highest spatial resolution in comparison with the other existing datasets.

c. Reconstructed land surface albedo

Table 3 shows that differences between MODIS monthly averaged and ECOCLIMAP monthly reconstituted land surface albedo [using Eq. (3)] are considerably reduced when we pass from ECOCLIMAP-I to ECOCLIMAP-II. Over bare lands, for example, in January and July in the visible wave band, the use of revised albedo values instead of the existing yield to the decreasing of the RMSE from 0.072 to 0.010 and from 0.069 to 0.009, respectively. This poor representation of the land surface albedo in ECOCLIMAP-I is explained by the lack of precision of bare soils albedo maps (see Fig. 8).

Table 3.

Differences (expressed in RMSE) between MODIS and ECOCLIMAP vegetation and bare soil land surface albedos in the visible and near-infrared parts of the spectrum, calculated monthly and yearly over the 2000–07 period. Note that urban and built-up areas were excluded for the comparison.

Table 3.

Table 4 shows the percentage of pixels of both versions of ECOCLIMAP, which are not statistically different from MODIS. Recalling that since 0.02 is the absolute bias of MODIS Collection 5 snow-free albedo (Liu et al. 2009; Román et al. 2009), the difference is not statistically significant when it is lower than ±0.02 by taking MODIS as reference. Since one of the objectives of this work is to update land surface variables via remote sensing science, the difference MODIS − ECOCLIMAP-I illustrates the biases while the difference MODIS − ECOCLIMAP-II illustrates the improvement of the new database. In January for instance, 60% of the land satisfies this condition for ECOCLIMAP-II versus only 23% in ECOCLIMAP-I.

Table 4.

Comparison of the accuracy (in percentage) of both versions of ECOCLIMAP in terms of monthly and yearly surface broadband albedos (over the 2000–07 period), calculated in the visible and near-infrared parts of the spectrum, and using MODIS Collection 5 snow-free albedo as reference.

Table 4.

5. Conclusions and future plans

With the aim of improving earth system model performance, this paper presents the updates of the ECOCLIMAP database that were made to better characterize the land surface heterogeneity at the African continental scale. The 1-km database includes the ecosystem map (Kaptué et al. 2010a); the bare soil and vegetation albedo maps, which are both pixel dependent but season independent; and the leaf area index (LAI) data and the fractional vegetation coverage (veg), both of which are season and pixel dependent. In summary, all these data were derived using MODIS and VEGETATION data, and represent a significant improvement over those currently used in ECOCLIMAP-I. Evaluating the value-added benefits of using the new ECOCLIMAP database at the African continental scale in numerical weather prediction (NWP) models will be investigated in a future research work. Nonetheless, preliminary investigations of this issue tend to show that impact on the modeled land water and energy budget over the western African region is largely in favor of the new landscape variables (Kaptué 2010). These datasets should help to reduce the model biophysical variable biases in land surface models. An update of the ECOCLIMAP database is planned at the global scale and this is already done for the European continent (Faroux et al. 2009).

Optimally, the use of MODIS observations or other remotely sensed observation (such as those from the SEVIRI) with our method is currently developed with a fully dynamic phenology to parameterize surface albedo in dynamic vegetation models. Here, we make the assumption that bare soil albedos and vegetation albedos are stationary variables and that seasonal variations of albedo are then controlled only by the one of the fractional vegetation coverage. How to integrate the fact that changes in bare soil albedo and vegetation albedo are mainly caused by soil moisture and chlorophyll content, respectively, are questions that needed to be addressed. Though the spectral dependency was considered by distinguishing visible and near-infrared surface albedos, this study only considers white sky albedo. It will be interesting to evaluate the solar zenith angle dependence of the model albedo (and hence the one of the bare soil albedo) since the broadband albedo is theoretically a linear combination of white sky and black sky values (Román et al. 2010). To advance the land surface parameterization, this study outlines that efforts may be undertaken to produce high-quality soil texture information of the thin surface layer (first millimeters) because this information is used in many LSMs to produce bare soil albedo. The latter information will induce the improvement of other biophysical variables such as the emissivity of natural land surface, which is also parameterized like albedo over bare soil and canopy (Jin and Liang 2006). Narrow band imaging spectroscopy in optical and thermal regions of the electromagnetic spectrum will provide comprehensive insight into various aspects of soil and their properties and to answer the quantitative aspects of soil science—namely, soil mineralogy, soil fertility, soil organic matter, soil moisture, and thermal properties of soils.

Acknowledgments

The authors thank the science team members of the MODIS Land Discipline Group, the Harmonized World Soil Database (HWSD), and the CYCLOPES projects for sharing data. Additional thanks go to Stéphanie Faroux for help in implementing the algorithms for the rasterization of the HWSD.

APPENDIX

Soil Taxonomy According to the Food and Agricultural Organization

Acrisols Soils with subsurface accumulation of low activity clays and low base saturation

Alisols Soils with subsurface accumulation of high activity clays, rich in exchangeable aluminum

Andosols Young soils formed from volcanic deposits

Anthrosols Soils in which human activities have resulted in profound modification of their properties

Arenosols Sandy soils featuring very weak or no soil development

Calcisols Soils with accumulation of secondary calcium carbonates

Cambisols Weakly to moderately developed soils

Chernozems Soils with a thick, dark topsoil, rich in organic matter with a calcareous subsoil

Ferralsols Deep, strongly weathered soils with a chemically poor but physically stable subsoil

Fluvisols Young soils in alluvial deposits

Greyzems Acid soils with a thick, dark topsoil rich in organic matter

Gypsisols Soils with accumulation of secondary gypsum

Histosols Soils that are composed of organic materials

Kastanozems Soils with a thick, dark brown topsoil, rich in organic matter and a calcareous or gypsum-rich subsoil

Leptosols Very shallow soils over hard rock or in unconsolidated very gravelly material

Lixisols Soil with subsurface accumulation of low activity clays and high base saturation

Luvisols Soils with subsurface accumulation of high activity clays and high base saturation

Nitisols Deep, dark red, brown, or yellow claylike soils having a pronounced shiny, nut-shaped structure

Phaeozems Soils with a thick, dark topsoil rich in organic matter and with evidence of removal of carbonates

Planosols Soils with a bleached, temporarily water-saturated topsoil on a slowly permeable subsoil

Plinthosols Wet soils with an irreversibly hardening mixture of iron, clay, and quartz in the subsoil

Podzols Acid soils with a subsurface accumulation of iron–aluminum–organic compounds

Regosols Soils with very limited soil development

Solonchaks Strongly saline soils

Solonetz Soils with subsurface clay accumulation; rich in sodium

Vertisols Dark-colored cracking and swelling clays

REFERENCES

  • Bacour, C., Baret F. , Béal D. , Weiss M. , and Pavageau K. , 2006: Neural network estimation of LAI, fAPAR, fCover and LAIxCabb, from top of canopy MERIS reflectance data: Principles and validation. Remote Sens. Environ., 105, 313325.

    • Search Google Scholar
    • Export Citation
  • Baret, F., and Coauthors, 2007: LAI, fAPAR and fCover CYCLOPES global products derived from VEGETATION: Part 1: Principles of the algorithm. Remote Sens. Environ., 110, 275286.

    • Search Google Scholar
    • Export Citation
  • Bonan, G., 2008: Forests and climate change: Forcings, feedbacks, and the climate benefits of forests. Science, 320, 14441449.

  • Camacho-de-Coca, F., Jiménez-Muñoz J. C. , Martinez B. , Bicheron P. , Lacaze R. , and Leroy M. , 2006: Prototyping of fCover product over Africa based on existing CYCLOPES and JRC products for VGT4Africa. Proc. Second Int. Symp. on Recent Advances in Quantitative Remote Sensing, Valencia, Spain, Universitat de Valencia, 722–727.

    • Search Google Scholar
    • Export Citation
  • De Colstoun, E. C. B., DeFries R. S. , and Townshend J. R. G. , 2006: Evaluation of ISLSCP Initiative II satellite-based land cover data sets and assessment of progress in land cover data for global modelling. J. Geophys. Res., 111, D22S07, doi:10.1029/2006JD007453.

    • Search Google Scholar
    • Export Citation
  • Di Gregorio, A., and Jansen L. , 2000: Land Cover Classifications System (LCCS), Classification Concepts and User Manual. Food and Agriculture Organisation, 179 pp.

    • Search Google Scholar
    • Export Citation
  • Ek, M. B., Mitchell K. E. , Lin Y. , Rogers E. , Grunmann P. , Koren V. , Gayno G. , and Tarpley J. D. , 2003: Implementation of the upgraded Noah land surface model in the National Centers for Environmental Prediction operational mesoscale Eta model. J. Geophys. Res., 108, 8851, doi:10.1029/2002JD003296.

    • Search Google Scholar
    • Export Citation
  • Fang, H., Liang S. , Townshend J. , and Dickinson R. , 2008: Spatially and temporally continuous LAI data sets based on a new filtering method: Examples from North America. Remote Sens. Environ., 112, 7593.

    • Search Google Scholar
    • Export Citation
  • FAO/IIASA/ISRIC/ISSCAS/JRC, cited 2009: Harmonized World Soil Database (version 1.1). [Available online at http://www.iiasa.ac.at/Research/LUC/External-World-soil-database/HTML/.]

    • Search Google Scholar
    • Export Citation
  • Faroux, S., Roujean J.-L. , Kaptué A. , and Masson V. , 2009: The ECOCLIMAP-II land surface database over Europe. CNRM/Météo France Tech. Note 86, 120 pp. [Available from CNRM/Météo France, 42 Av. G. Coriolis, 31057 Toulouse CEDEX 01, France.]

    • Search Google Scholar
    • Export Citation
  • Feddema, J. J., Oleson K. W. , Bonan G. , Mearns L. O. , Buja L. E. , Meehl G. A. , and Washington W. M. , 2005: The importance of land cover change in simulating future climate. Science, 310, 16741678.

    • Search Google Scholar
    • Export Citation
  • Foley, J. A., and Coauthors, 2005: Global consequences of land use. Science, 309, 570574.

  • Fritz, S., and See L. , 2008: Quantifying uncertainty and spatial disagreement in the comparison of global land cover for different applications. Global Change Biol., 14, 10571075.

    • Search Google Scholar
    • Export Citation
  • Garrigues, S., Allard D. , Baret F. , and Morisette J. , 2008a: Multivariate quantification of landscape spatial heterogeneity using variogram models. Remote Sens. Environ., 112, 216230.

    • Search Google Scholar
    • Export Citation
  • Garrigues, S., and Coauthors, 2008b: Validation and intercomparison of global leaf area index products derived from remote sensing data. J. Geophys. Res., 113, G02028, doi:10.1029/2007JG000635.

    • Search Google Scholar
    • Export Citation
  • Gómez, J. M., Valladares F. , and Puerta-Piñero C. , 2004: Differences between structural and functional environmental heterogeneity caused by seed dispersal. Funct. Ecol., 18, 787792.

    • Search Google Scholar
    • Export Citation
  • Gutman, G., 1999: On the use of long-term global data of land reflectances and vegetation indices derived from the advanced very high resolution radiometer. J. Geophys. Res., 104, 62416255.

    • Search Google Scholar
    • Export Citation
  • Gutman, G., Tarpley D. , Ignatov A. , and Olson S. , 1995: The enhanced NOAA global land dataset from the Advanced Very High Resolution Radiometer. Bull. Amer. Meteor. Soc., 76, 11411156.

    • Search Google Scholar
    • Export Citation
  • Hall, F. G., and Coauthors, 2006: The ISLSCP Initiative II global data sets: Surface boundary conditions and atmospheric forcings for land-atmosphere studies. J. Geophys. Res., 111, D22S01, doi:10.1029/2006JD007366.

    • Search Google Scholar
    • Export Citation
  • Houldcroft, C. J., Grey W. M. F. , Barnsley M. , Taylor C. M. , Los S. O. , and North P. R. J. , 2009: New vegetation albedo parameters and global fields of soil background albedo derived from MODIS for use in a climate model. J. Hydrometeor., 10, 183198.

    • Search Google Scholar
    • Export Citation
  • Imhoff, M. L., Bounoua L. , Ricketts T. , Loucks C. , Harriss R. , and Lawrence W. T. , 2004: Global patterns in human consumption of net primary production. Nature, 429, 870873.

    • Search Google Scholar
    • Export Citation
  • Jin, M., and Liang S. , 2006: An improved land surface emissivity parameter for land surface models using global remote sensing observations. J. Climate, 19, 28672881.

    • Search Google Scholar
    • Export Citation
  • Justice, C. O., Townshend J. R. G. , Vermote E. F. , Masuoka E. , Wolfe R. E. , Saleous N. , Roy D. P. , and Morisette J. T. , 2002: An overview of MODIS Land data processing and product status. Remote Sens. Environ., 83, 315.

    • Search Google Scholar
    • Export Citation
  • Kaptué, T. A. T., 2010: Mapping ecosystem and biophysical satellite parameters for the study of hydric fluxes over the African continent. Ph.D dissertation, Paul Sabatier University of Toulouse, 192 pp. [Available online at http://thesesups.ups-tlse.fr/994/.]

    • Search Google Scholar
    • Export Citation
  • Kaptué, T. A. T., De Jong S. M. , Roujean J.-L. , Favier C. , and Mering C. , 2010a: Ecosystem mapping at the African continent scale using a hybrid clustering approach based on 1-km resolution multi-annual data from SPOT/VEGETATION. Remote Sens. Environ., 115, 452464, doi:10.1016/j.rse.2010.09.015.

    • Search Google Scholar
    • Export Citation
  • Kaptué, T. A. T., Roujean J.-L. , and De Jong S. M. , 2010b: Comparison and relative quality assessment of the GLC2000, GLOBCOVER, MODIS and ECOCLIMAP land cover data sets at the African continental scale. Int. J. Appl. Earth Obs. Geoinf., 13, 207219, doi:10.1016/j.jag.2010.11.005.

    • Search Google Scholar
    • Export Citation
  • Kaptué, T. A. T., Roujean J.-L. , and Faroux S. , 2010c: ECOCLIMAP-II: An ecosystem classification and land surface parameter database of western Africa at 1 km resolution for the Africa Monsoon Multidisciplinary Analysis (AMMA) project. Remote Sens. Environ., 114, 961976.

    • Search Google Scholar
    • Export Citation
  • Kolasa, J., and Rollo C. D. , 1991: Introduction: The heterogeneity of heterogeneity: A glossary. Ecological Heterogeneity, J. Kolasa and S. T. A. Pickett, Eds., Springer, 1–23.

    • Search Google Scholar
    • Export Citation
  • Koster, R. D., Suarez M. J. , Ducharne A. , Stieglitz M. , and Kumar P. , 2000: A catchment-based approach to modeling land surface processes in a general circulation model 1. Model structure. J. Geophys. Res., 105, 24 80924 822.

    • Search Google Scholar
    • Export Citation
  • Liu, J., and Coauthors, 2009: Validation of Moderate Resolution Imaging Spectroradiometer (MODIS) albedo retrieval algorithm: Dependence of albedo on solar zenith angle. J. Geophys. Res., 114, D01106, doi:10.1029/2008JD009969.

    • Search Google Scholar
    • Export Citation
  • Los, S. O., and Coauthors, 2000: A global 9-yr biophysical land surface dataset from NOAA AVHRR data. J. Hydrometeor., 1, 183199.

  • Lucht, W., Schaaf C. B. , and Strahler A. H. , 2000: An algorithm for the retrieval of albedo from space using semi-empirical BRDF models. IEEE Trans. Geosci. Remote Sens., 38, 977998.

    • Search Google Scholar
    • Export Citation
  • Masson, V., Champeaux J.-L. , Chauvin F. , Meriguer C. , and Lacaze R. , 2003: A global database of land surface parameters at 1-km resolution in meteorological and climate models. J. Climate, 16, 12611282.

    • Search Google Scholar
    • Export Citation
  • Miller, J., Barlage M. , Zeng X. , Wei H. , Mitchell K. and Tarpley D. , 2006: Sensitivity of the NCEP/Noah land surface model to the MODIS green vegetation fraction data set. Geophys. Res. Lett., 33, L13404, doi:10.1029/2006GL026636.

    • Search Google Scholar
    • Export Citation
  • Moody, E. G., King M. D. , Platnick S. , Schaaf C. B. , and Gao F. , 2005: Spatially complete global spectral surface albedos: Value-added datasets derived from Terra MODIS land products. IEEE Trans. Geosci. Remote Sens., 43, 144158.

    • Search Google Scholar
    • Export Citation
  • Morisette, J. T., and Coauthors, 2006: Validation of global moderate-resolution LAI products: A framework proposed within the CEOS land product validation subgroup. IEEE Trans. Geosci. Remote Sens., 44, 18041814.

    • Search Google Scholar
    • Export Citation
  • Myneni, R. B., and Coauthors, 2002: Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data. Remote Sens. Environ., 83, 214231.

    • Search Google Scholar
    • Export Citation
  • Noilhan, J., and Mahfouf J.-F. , 1996: The ISBA land surface parametrisation scheme. Global Planet. Change, 13, 145159.

  • Patz, J. A., and Coauthors, 2004: Unhealthy landscapes: Policy recommendations on land use change and infectious disease emergence. Environ. Health Perspect., 112, 10921098.

    • Search Google Scholar
    • Export Citation
  • Pitman, A. J., 2003: The evolution of, and revolution in, land surface schemes designed for climate models. Int. J. Climatol., 23, 479510.

    • Search Google Scholar
    • Export Citation
  • Rechid, D., Raddatz T. J. , and Jacob D. , 2009: Parameterization of snow-free land surface albedo as a function of vegetation phenology based on MODIS data and applied in climate modelling. Theor. Appl. Climatol., 95, 245255.

    • Search Google Scholar
    • Export Citation
  • Román, M. O., and Coauthors, 2009: The MODIS (Collection V005) BRDF/albedo product: Assessment of spatial representativeness over forested landscapes. Remote Sens. Environ., 113, 24762498.

    • Search Google Scholar
    • Export Citation
  • Román, M. O., and Coauthors, 2010: Assessing the coupling between surface albedo derived from MODIS and the fraction of diffuse skylight over spatially-characterized landscapes. Remote Sens. Environ., 114, 738760.

    • Search Google Scholar
    • Export Citation
  • Roujean, J.-L., and Lacaze R. , 2002: Global mapping of vegetation parameters from POLDER multiangular measurements for studies of surface-atmosphere interactions: A pragmatic method and validation. J. Geophys. Res., 107, 4150, doi:10.1029/2001JD000751.

    • Search Google Scholar
    • Export Citation
  • Running, S. W., Loveland T. R. , Pierce L. L. , Nemani R. , and Hunt E. R. , 1995: A remote sensing based vegetation classification logic for global land-cover analysis. Remote Sens. Environ., 51, 3948.

    • Search Google Scholar
    • Export Citation
  • Salmun, H., and Molod A. , 2006: Progress in modelling the impact of land cover change on the global climate. Prog. Phys. Geogr., 30, 737749.

    • Search Google Scholar
    • Export Citation
  • Schaaf, C. B., and Coauthors, 2002: First operational BRDF, albedo nadir reflectance products from MODIS. Remote Sens. Environ., 83, 135148.

    • Search Google Scholar
    • Export Citation
  • Schaaf, C. B., Cihlar J. , Belward A. , Dutton E. , and Verstraete M. , 2009: Albedo and reflectance anisotropy, ECV-T8. GTOS Assessment of the Status of the Development of Standards for the Terrestrial Essential Climate Variables, R. Sessa, Ed., FAO, 16 pp.

    • Search Google Scholar
    • Export Citation
  • Schaaf, C. B., Liu J. , Gao F. , and Strahler A. H. , 2011: Aqua and Terra MODIS albedo and reflectance anisotropy products. Land Remote Sensing and Global Environmental Change: NASA’s Earth Observing System and the Science of ASTER and MODIS, B. Ramachandran, C. Justice, and M. Abrams, Eds., Remote Sensing and Digital Image Processing Series, Vol. 11, Springer-Verlag, 549–561.

    • Search Google Scholar
    • Export Citation
  • Schaepman-Strub, G., Schaepman M. E. , Painter T. H. , Dangel S. , and Martonchik J. V. , 2006: Reflectance quantities in optical remote sensing—Definitions and case studies. Remote Sens. Environ., 103, 2742.

    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., Los S. O. , Tucker C. J. , Justice C. O. , Dazlich D. A. , Collatz G. J. , and Randall D. A. , 1996a: A revised land surface parameterization (SiB2) for atmospheric GCMs. Part II: The generation of global fields of terrestrial biophysical parameters from satellite data. J. Climate, 9, 706737.

    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., and Coauthors, 1996b: The ISLSCP Initiative I global datasets: Surface boundary conditions and atmospheric forcings for land–atmosphere studies. Bull. Amer. Meteor. Soc., 77, 19872005.

    • Search Google Scholar
    • Export Citation
  • Strahler, A. H., and Coauthors, 2006: Global land cover validation: Recommendations for evaluation and accuracy assessment of global land cover maps. GOFC-GOLD Rep. 25, 51 pp.

    • Search Google Scholar
    • Export Citation
  • Strugnell, N., and Lucht W. , 2001: An algorithm to infer continental-scale albedo from AVHRR data, land cover class, and field observations of typical BRDFs. J. Climate, 14, 13601376.

    • Search Google Scholar
    • Export Citation
  • Sutherland, W. J., and Coauthors, 2009: One hundred questions of importance to the conservation of importance to the conservation of global biological diversity. Conserv. Biol., 23, 557567.

    • Search Google Scholar
    • Export Citation
  • Trigo, I. F., and Coauthors, 2010: The satellite application facility on land surface analysis. Int. J. Remote Sens., 32, 27252744, doi:10.1080/01431161003743199.

    • Search Google Scholar
    • Export Citation
  • Tucker, C. J., 1979: Red and photographic infrared linear combinations for monitoring vegetation. Remote Sens. Environ., 8, 127150.

  • Tucker, C. J., Townshend J. R. G. , and Goff T. E. , 1985: African land-cover classification using satellite data. Science, 227, 369374.

    • Search Google Scholar
    • Export Citation
  • Turner, W. R., Katrina B. , Brooks T. M. , Costanza R. , Da Fonseca G. A. B. , and Portela R. , 2007: Global conservation of biodiversity and ecosystem services. Bioscience, 57, 868873.

    • Search Google Scholar
    • Export Citation
  • Williams, C. W., Hanan N. P. , Neff J. C. , Scoles R. J. , Berry J. A. , Denning A. S. , and Baker D. F. , 2007: Africa and the global carbon cycle. Carbon Balance Manage., 2, 113.

    • Search Google Scholar
    • Export Citation
  • Wilson, M. F., and Henderson-Sellers A. , 1985: A global archive of land cover and soils data for use in general circulation models. J. Climatol., 5, 119143.

    • Search Google Scholar
    • Export Citation
  • Yang, W., Shabanov N. V. , Huang D. , Wang W. , Dickinson R. E. , Nemani R. R. , Knyazikhin Y. , and Myneni R. B. , 2006: Analysis of leaf area index products from combination of MODIS Terra and Aqua data. Remote Sens. Environ., 104, 297312.

    • Search Google Scholar
    • Export Citation
Save