• Baldocchi, D., 2008: “Breathing” of the terrestrial biosphere: Lessons learned from a global network of carbon dioxide flux measurement systems. Aust. J. Bot., 56, 1–26, doi:10.1071/BT07151.

    • Search Google Scholar
    • Export Citation
  • Baldocchi, D., and et al. , 2001: FLUXNET: A new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities. Bull. Amer. Meteor. Soc., 82, 24152434, doi:10.1175/1520-0477(2001)082<2415:FANTTS>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Beer, C., and et al. , 2010: Terrestrial gross carbon dioxide uptake: Global distribution and covariation with climate. Science, 329, 834838, doi:10.1126/science.1184984.

    • Search Google Scholar
    • Export Citation
  • Bonan, G. B., 1998: The land surface climatology of the NCAR land surface model coupled to the NCAR Community Climate Model. J. Climate, 11, 13071326, doi:10.1175/1520-0442(1998)011<1307:TLSCOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bonan, G. B., , S. Levis, , L. Kergoat, , and K. W. Oleson, 2002: Landscapes as patches of plant functional types: An integrating concept for climate and ecosystem models. Global Biogeochem. Cycles, 16, 1021, doi:10.1029/2000GB001360.

    • Search Google Scholar
    • Export Citation
  • Bonan, G. B., , P. J. Lawrence, , K. W. Oleson, , S. Levis, , M. Jung, , M. Reichstein, , D. M. Lawrence, , and S. C. Swenson, 2011: Improving canopy processes in the Community Land Model version 4 (CLM4) using global flux fields empirically inferred from FLUXNET data. J. Geophys. Res., 116, G02014, doi:10.1029/2010JG001593.

    • Search Google Scholar
    • Export Citation
  • Buermann, W., , J. Dong, , X. Zeng, , R. B. Myneni, , and R. E. Dickinson, 2001: Evaluation of the utility of satellite-based vegetation leaf area index data for climate simulations. J. Climate, 14, 35363550, doi:10.1175/1520-0442(2001)014<3536:EOTUOS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Carlson, T. N., , and D. A. Ripley, 1997: On the relation between NDVI, fractional vegetation cover, and leaf area index. Remote Sens. Environ., 62, 241252, doi:10.1016/S0034-4257(97)00104-1.

    • Search Google Scholar
    • Export Citation
  • Carroll, M. L., , J. R. Townshend, , C. M. DiMiceli, , P. Noojipady, , and R. A. Sohlberg, 2009: A new global raster water mask at 250 m resolution. Int. J. Digital Earth, 2, 291308, doi:10.1080/17538940902951401.

    • Search Google Scholar
    • Export Citation
  • Chen, M., , W. Shi, , P. Xie, , V. B. S. Silva, , V. E. Kousky, , R. W. Higgins, , and J. E. Janowiak, 2008: Assessing objective techniques for gauge-based analyses of global daily precipitation. J. Geophys. Res., 113, D04110, doi:10.1029/2007JD009132.

    • Search Google Scholar
    • Export Citation
  • Choi, H. I., , and X.-Z. Liang, 2010: Improved terrestrial hydrologic representation in mesoscale land surface models. J. Hydrometeor., 11, 797809, doi:10.1175/2010JHM1221.1.

    • Search Google Scholar
    • Export Citation
  • Choi, H. I., , P. Kumar, , and X.-Z. Liang, 2007: Three-dimensional volume-averaged soil moisture transport model with a scalable parameterization of subgrid topographic variability. Water Resour. Res.,43, W04414, doi:10.1029/2006WR005134.

  • Choi, H. I., , X.-Z. Liang, , and P. Kumar, 2013: A conjunctive surface–subsurface flow representation for mesoscale land surface models. J. Hydrometeor., 14, 1421–1442, doi:10.1175/JHM-D-12-0168.1.

    • Search Google Scholar
    • Export Citation
  • Chou, M.-D., , and M. J. Suarez, 1999: A solar radiation parameterization for atmospheric studies. NASA Tech. Memo. NASA/TM-1999-104606, Vol. 15, 40 pp.

  • Chou, M.-D., , M. J. Suarez, , X.-Z. Liang, , and M. M.-H. Yan, 2001: A thermal infrared radiation parameterization for atmospheric studies. NASA Tech. Memo. NASA/TM-2001-104606, Vol. 19, 56 pp

  • Dai, Y., , R. E. Dickinson, , and Y.-P. Wang, 2004: A two-big-leaf model for canopy temperature, photosynthesis, and stomatal conductance. J. Climate, 17, 22812299, doi:10.1175/1520-0442(2004)017<2281:ATMFCT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Daly, C., , M. Halbleib, , J. I. Smith, , W. P. Gibson, , M. K. Doggett, , G. H. Taylor, , J. Curtis, , and P. P. Pasteris, 2008: Physiographically sensitive mapping of climatological temperature and precipitation across the conterminous United States. Int. J. Climatol., 28, 20312064, doi:10.1002/joc.1688.

    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, doi:10.1002/qj.828.

    • Search Google Scholar
    • Export Citation
  • De Kauwe, M. G., , M. I. Disney, , T. Quaife, , P. Lewis, , and M. Williams, 2011: An assessment of the MODIS collection 5 leaf area index product for a region of mixed coniferous forest. Remote Sens. Environ., 115, 767780, doi:10.1016/j.rse.2010.11.004.

    • Search Google Scholar
    • Export Citation
  • Dickinson, R. E., , K. W. Oleson, , G. Bonan, , F. Hoffman, , P. Thornton, , M. Vertenstein, , Z.-L. Yang, , and X. Zeng, 2006: The Community Land Model and its climate statistics as a component of the Community Climate System Model. J. Climate, 19, 23022324, doi:10.1175/JCLI3742.1.

    • Search Google Scholar
    • Export Citation
  • Fang, H., , S. Wei, , and S. Liang, 2012: Validation of MODIS and CYCLOPES LAI products using global field measurement data. Remote Sens. Environ., 119, 4354, doi:10.1016/j.rse.2011.12.006.

    • Search Google Scholar
    • Export Citation
  • Fang, H., and et al. , 2013: Characterization and intercomparison of global moderate resolution leaf area index (LAI) products: Analysis of climatologies and theoretical uncertainties. J. Geophys. Res. Biogeosci., 118, 529548, doi:10.1002/jgrg.20051.

    • Search Google Scholar
    • Export Citation
  • Goward, S. N., , B. Markham, , D. G. Dye, , W. Dulaney, , and J. Yang, 1991: Normalized difference vegetation index measurements from the Advanced Very High Resolution Radiometer. Remote Sens. Environ., 35, 257277, doi:10.1016/0034-4257(91)90017-Z.

    • 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, doi:10.1029/1998JD200106.

    • Search Google Scholar
    • Export Citation
  • Gutman, G., , and A. Ignatov, 1998: The derivation of the green vegetation fraction from NOAA/AVHRR data for use in numerical weather prediction models. Int. J. Remote Sens., 19, 15331543, doi:10.1080/014311698215333.

    • Search Google Scholar
    • Export Citation
  • Hansen, M. C., , R. S. DeFries, , J. R. G. Townshend, , M. Carroll, , C. Dimiceli, , and R. A. Sohlberg, 2003: Global percent tree cover at a spatial resolution of 500 meters: First results of the MODIS vegetation continuous fields algorithm. Earth Interact., 7, doi:10.1175/1087-3562(2003)007<0001:GPTCAA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Henderson-Sellers, A., , and M. F. Wilson, 1983: Surface albedo data for climatic modeling. Rev. Geophys., 21, 17431778, doi:10.1029/RG021i008p01743.

    • Search Google Scholar
    • Export Citation
  • Hill, M. J., , U. Senarath, , A. Lee, , M. Zeppel, , J. M. Nightingale, , R. J. Williams, , and T. R. McVicar, 2006: Assessment of the MODIS LAI product for Australian ecosystems. Remote Sens. Environ., 101, 495518, doi:10.1016/j.rse.2006.01.010.

    • Search Google Scholar
    • Export Citation
  • Holder, C., , R. Boyles, , A. Syed, , D. Niyogi, , and S. Raman, 2006: Comparison of collocated automated (NCECONet) and manual (COOP) climate observations in North Carolina. J. Atmos. Oceanic Technol., 23, 671682, doi:10.1175/JTECH1873.1.

    • Search Google Scholar
    • Export Citation
  • Holtslag, A. A. M., , and B. A. Boville, 1993: Local versus nonlocal boundary-layer diffusion in a global climate model. J. Climate, 6, 18251842, doi:10.1175/1520-0442(1993)006<1825:LVNBLD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Jung, M., and et al. , 2010: Recent decline in the global land evapotranspiration trend due to limited moisture supply. Nature, 467, 951954, doi:10.1038/nature09396.

    • Search Google Scholar
    • Export Citation
  • Jung, M., and et al. , 2011: Global patterns of land-atmosphere fluxes of carbon dioxide, latent heat, and sensible heat derived from eddy covariance, satellite, and meteorological observations. J. Geophys. Res., 116, G00J07, doi:10.1029/2010JG001566.

    • Search Google Scholar
    • Export Citation
  • Ke, Y., , L. R. Leung, , M. Huang, , A. M. Coleman, , H. Li, , and M. S. Wigmosta, 2012: Development of high resolution land surface parameters for the Community Land Model. Geosci. Model Dev., 5, 13411362, doi:10.5194/gmd-5-1341-2012.

    • Search Google Scholar
    • Export Citation
  • Lawrence, D. M., , and A. G. Slater, 2008: Incorporating organic soil into a global climate model. Climate Dyn., 30, 145160, doi:10.1007/s00382-007-0278-1.

    • Search Google Scholar
    • Export Citation
  • Lawrence, D. M., and et al. , 2011: Parameterization improvements and functional and structural advances in version 4 of the Community Land Model. J. Adv. Model. Earth Syst., 3, M03001, doi:10.1029/2011MS000045.

    • Search Google Scholar
    • Export Citation
  • Lawrence, P. J., , and T. N. Chase, 2007: Representing a new MODIS consistent land surface in the Community Land Model (CLM 3.0). J. Geophys. Res., 112, G01023, doi:10.1029/2006JG000168.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , and F. Zhang, 2013: The cloud–aerosol–radiation (CAR) ensemble modeling system. Atmos. Chem. Phys., 13, 83358364, doi:10.5194/acp-13-8335-2013.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , K. E. Kunkel, , and A. N. Samel, 2001: Development of a regional climate model for U.S. Midwest applications. Part I: Sensitivity to buffer zone treatment. J. Climate, 14, 43634378, doi:10.1175/1520-0442(2001)014<4363:DOARCM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , L. Li, , K. E. Kunkel, , M. F. Ting, , and J. X. L. Wang, 2004: Regional climate model simulation of U.S. precipitation during 1982–2002. Part I: Annual cycle. J. Climate, 17, 35103529, doi:10.1175/1520-0442(2004)017<3510:RCMSOU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , H. I. Choi, , K. E. Kunkel, , Y. Dai, , E. Joseph, , J. X. Wang, , and P. Kumar, 2005a: Surface boundary conditions for mesoscale regional climate models. Earth Interact., 9, doi:10.1175/EI151.1.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., and et al. , 2005b: Development of land surface albedo parameterization based on Moderate Resolution Imaging Spectroradiometer (MODIS) data. J. Geophys. Res., 110, D11107, doi:10.1029/2004JD005579.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , M. Xu, , J. Zhu, , K. Kunkel, , and J. X. L. Wang, 2005c: Development of the Regional Climate–Weather Research and Forecasting Model (CWRF): Treatment of topography. 2005 WRF/MM5 Users’ Workshop, Boulder, CO, NCAR, 9.3. [Available online at http://www2.mmm.ucar.edu/wrf/users/workshops/WS2005/abstracts/Session9/3-Liang.pdf.]

  • Liang, X.-Z., , M. Xu, , W. Gao, , K. R. Reddy, , K. Kunkel, , D. L. Schmoldt, , and A. N. Samel, 2012a: A distributed cotton growth model developed from GOSSYM and its parameter determination. Agron. J., 104, 661674, doi:10.2134/agronj2011.0250.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , M. Xu, , W. Gao, , K. R. Reddy, , K. Kunkel, , D. L. Schmoldt, , and A. N. Samel, 2012b: Physical modeling of U.S. cotton yields and climate stresses during 1979 to 2005. Agron. J., 104, 675683, doi:10.2134/agronj2011.0251.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., and et al. , 2012c: Regional Climate–Weather Research and Forecasting Model. Bull. Amer. Meteor. Soc., 93, 13631387, doi:10.1175/BAMS-D-11-00180.1.

    • Search Google Scholar
    • Export Citation
  • Ling, T.-J., , X.-Z. Liang, , M. Xu, , Z. Wang, , and B. Wang, 2011: A multilevel ocean mixed-layer model for 2-dimension applications. Acta Oceanol. Sin., 33, 110.

    • Search Google Scholar
    • Export Citation
  • Liu, Y., , R. Liu, , and J. M. Chen, 2012: Retrospective retrieval of long-term consistent global leaf area index (1981–2011) from combined AVHRR and MODIS data. J. Geophys. Res., 117, G04003, doi:10.1029/2012JG002084.

    • Search Google Scholar
    • Export Citation
  • Myneni, R. B., , R. Ramakrishna, , R. Nemani, , and S. W. Running, 1997: Estimation of global leaf area index and absorbed PAR using radiative transfer models. IEEE Trans. Geosci. Remote Sens., 35, 13801393, doi:10.1109/36.649788.

    • Search Google Scholar
    • Export Citation
  • Myneni, R. B., and et al. , 2002: Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data. Remote Sens. Environ., 83, 214231, doi:10.1016/S0034-4257(02)00074-3.

    • Search Google Scholar
    • Export Citation
  • Oleson, K. W., and et al. , 2008: Improvements to the Community Land Model and their impact on the hydrological cycle. J. Geophys. Res., 113, G01021, doi:10.1029/2007JG000563.

    • Search Google Scholar
    • Export Citation
  • Park, S., , and C. S. Bretherton, 2009: The University of Washington shallow convection and moist turbulence schemes and their impact on climate simulations with the Community Atmosphere Model. J. Climate, 22, 34493469, doi:10.1175/2008JCLI2557.1.

    • Search Google Scholar
    • Export Citation
  • Pielke, R. A., Sr., , J. Adegoke, , A. Beltrán-Przekurat, , C. A. Hiemstra, , J. Lin, , U. S. Nair, , D. Niyogi, , and T. E. Nobis, 2007: An overview of regional land-use and land-cover impacts on rainfall. Tellus, 59B, 587601, doi:10.1111/j.1600-0889.2007.00251.x.

    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., , T. Corti, , E. L. Davin, , M. Hirschi, , E. B. Jaeger, , I. Lehner, , B. Orlowsky, , and A. J. Teuling, 2010: Investigating soil moisture–climate interactions in a changing climate: A review. Earth Sci. Rev., 99, 125161, doi:10.1016/j.earscirev.2010.02.004.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., and et al. , 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., doi:10.5065/D68S4MVH.

  • Strahler, A. H., and et al. , 1999: MODIS BRDF/albedo product: Algorithm theoretical basis document version 5.0, 53 pp. [Available online at modis.gsfc.nasa.gov/data/atbd/atbd_mod09.pdf.]

  • Tao, W.-K., and et al. , 2003: Microphysics, radiation and surface processes in the Goddard Cumulus Ensemble (GCE) model. Meteor. Atmos. Phys., 82, 97137.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., and et al. , 2004a: Comparison of seasonal and spatial variations of leaf area index and fraction of absorbed photosynthetically active radiation from Moderate Resolution Imaging Spectroradiometer (MODIS) and Common Land Model. J. Geophys. Res., 109, D01103, doi:10.1029/2003JD003777.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., , R. Dickinson, , L. Zhou, , R. Myneni, , M. Friedl, , C. Schaaf, , M. Carroll, , and F. Gao, 2004b: Land boundary conditions from MODIS data and consequences for the albedo of a climate model. Geophys. Res. Lett., 31, L05504, doi:10.1029/2003GL019104.

    • Search Google Scholar
    • Export Citation
  • Tucker, C., , J. Pinzon, , M. Brown, , D. Slayback, , E. Pak, , R. Mahoney, , E. Vermote, , and N. El Saleous, 2005: An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. Int. J. Remote Sens., 26, 44854498, doi:10.1080/01431160500168686.

    • Search Google Scholar
    • Export Citation
  • Vermote, E. F., , and N. Z. Saleous, 2006: Calibration of NOAA16 AVHRR over a desert site using MODIS data. Remote Sens. Environ., 105, 214220, doi:10.1016/j.rse.2006.06.015.

    • Search Google Scholar
    • Export Citation
  • Wang, Q., , J. Tenhunen, , N. Q. Dinh, , M. Reichstein, , T. Vesala, , and P. Keronen, 2004: Similarities in ground- and satellite-based NDVI time series and their relationship to physiological activity of a Scots pine forest in Finland. Remote Sens. Environ., 93, 225237, doi:10.1016/j.rse.2004.07.006.

    • Search Google Scholar
    • Export Citation
  • Yang, W., and et al. , 2006: MODIS leaf area index products: from validation to algorithm improvement. IEEE Trans. Geosci. Remote Sens., 44, 18851898, doi:10.1109/TGRS.2006.871215.

    • Search Google Scholar
    • Export Citation
  • Yuan, H., , Y. Dai, , Z. Xiao, , D. Ji, , and W. Shangguan, 2011: Reprocessing the MODIS leaf area index products for land surface and climate modelling. Remote Sens. Environ., 115, 11711187, doi:10.1016/j.rse.2011.01.001.

    • Search Google Scholar
    • Export Citation
  • Yuan, X., , and X.-Z. , 2011: Evaluation of a Conjunctive Surface–Subsurface Process Model (CSSP) over the contiguous United States at regional–local scales. J. Hydrometeor., 12, 579599, doi:10.1175/2010JHM1302.1.

    • Search Google Scholar
    • Export Citation
  • Yucel, I., 2006: Effects of implementing MODIS land cover and albedo in MM5 at two contrasting U.S. regions. J. Hydrometeor., 7, 10431060, doi:10.1175/JHM536.1.

    • Search Google Scholar
    • Export Citation
  • Zeng, X., , R. E. Dickinson, , A. Walker, , M. Shaikh, , R. S. DeFries, , and J. Qi, 2000: Derivation and evaluation of global 1-km fractional vegetation cover data for land modeling. J. Appl. Meteor., 39, 826839, doi:10.1175/1520-0450(2000)039<0826:DAEOGK>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zeng, X., , M. Shaikh, , Y. Dai, , R. E. Dickinson, , and R. Myneni, 2002: Coupling of the Common Land Model to the NCAR Community Climate Model. J. Climate, 15, 18321854, doi:10.1175/1520-0442(2002)015<1832:COTCLM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zhang, P., , B. Anderson, , M. Barlow, , B. Tan, , and R. B. Myneni, 2004: Climate-related vegetation characteristics derived from Moderate Resolution Imaging Spectroradiometer (MODIS) leaf area index and normalized difference vegetation index. J. Geophys. Res., 109, D20105, doi:10.1029/2004JD004720.

    • Search Google Scholar
    • Export Citation
  • Zhao, X., and et al. , 2013: The Global Land Surface Satellite (GLASS) remote sensing data processing system and products. Remote Sens., 5, 24362450, doi:10.3390/rs5052436.

    • Search Google Scholar
    • Export Citation
  • Zhu, Z., and et al. , 2013: Global data sets of vegetation leaf area index (LAI)3g and fraction of photosynthetically active radiation (FPAR)3g derived from Global Inventory Modeling and Mapping Studies (GIMMS) normalized difference vegetation index (NDVI3g) for the period 1981 to 2011. Remote Sens., 5, 927948, doi:10.3390/rs5020927.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    The geographic distributions of (a) old and (b) new USGS-based land-cover category data, and (c) new FVC and (d) FVC differences (new minus old) in the CWRF computational domain. The eight boxes in (a) are the eight regions that are studied to assess model ability to simulate precipitation and surface air temperature annual cycles.

  • View in gallery

    (top) Seasonal geographic distributions of the new LAI and (bottom) differences between the new and old LAI (new minus old). (from left to right) The seasons are winter (DJF), spring (MAM), summer (JJA), and autumn (SON).

  • View in gallery

    Geographic distributions of seasonal (winter, spring, summer, and autumn) mean (top half) latent heat flux (W m−2) and (bottom half) sensible heat flux (W m−2) biases (departures from MTE flux data) averaged during 1999–2008 as simulated by CTL and EXP as well as their differences (CTL minus EXP). Biases and differences with statistical significance greater than 95% are stippled.

  • View in gallery

    Geographic distributions of seasonal (winter, spring, summer, and autumn) mean differences (EXP-CTL) averaged during 1999–2008 for (top) albedo (%), (middle) top 0.1 m soil moisture (mm), and (bottom) planetary boundary layer height (m).

  • View in gallery

    As in Fig. 3, but for (top half) mean precipitation (mm day−1) and (bottom half) 2-m air temperature (°C). Biases and differences with statistical significance greater than 95% are stippled.

  • View in gallery

    The (a) 1999–2008 monthly mean precipitation (mm day−1) from observations as well as the CTL and EXP simulations, and (b) 2-m air temperature (°C) biases (departures from observations) simulated by CTL and EXP averaged over the eight regions showed in Fig. 1.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 30 30 9
PDF Downloads 24 24 11

MODIS Consistent Vegetation Parameter Specifications and Their Impacts on Regional Climate Simulations

View More View Less
  • 1 Earth System Science Interdisciplinary Center, University of Maryland, College Park, College Park, Maryland
  • | 2 Earth System Science Interdisciplinary Center, and Department of Atmospheric and Oceanic Science, University of Maryland, College Park, College Park, Maryland
  • | 3 Department of Geography, Bowling Green State University, Bowling Green, Ohio
  • | 4 UV-B Monitoring and Research Program, Natural Resource Ecology Laboratory, and Department of Ecosystem Science and Sustainability, Colorado State University, Fort Collins, Colorado
© Get Permissions
Full access

Abstract

A consistent set of Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation parameters, including leaf and stem area index (LAI and SAI, respectively), land-cover category (LCC), fractional vegetation cover (FVC), and albedo parameterization are developed, and their impacts on North American regional climate are evaluated based on 10-yr Climate–Weather and Research Forecasting Model (CWRF) simulations. As compared with the previous Advanced Very High Resolution Radiometer (AVHRR) set, MODIS LCC increases grassland and cropland fractions in the central Great Plains and Midwest, respectively. Evergreen needleleaf forest converts to mixed forest in the Southeast, and mixed forest converts to evergreen needleleaf in Canada. FVC decreases by 0.05–0.3 over the central Great Plains but increases by 0.1–0.35 over the northern Rocky Mountains, Canada, and the U.S. Southeast. MODIS LAI is less than AVHRR by 2–6, except in the central Great Plains, eastern Rocky Mountains, and central Mexico. LCC and FVC changes over the central Great Plains reduce CWRF warm biases by 0.71°C and wet biases by 0.36 mm day−1. Large LAI reductions cause latent and sensible heat fluxes to decrease by 0.78–5.81 and 0.91–6.54 W m−2, respectively. They also lessen cold biases over the Gulf States and Southeast and wet biases over the North American monsoon region and Canada during summer. In densely vegetated regions including eastern Canada, the Ohio Valley, and the mid-Atlantic region, spring and summer precipitation decreases and temperature increases result from LAI reductions that cause positive evapotranspiration–precipitation–soil moisture feedbacks. Conversely, precipitation and temperature decreases in sparely vegetated regions, such as the Great Plains, result from FVC reductions that cause negative albedo–evapotranspiration–precipitation–soil moisture feedbacks.

Corresponding author address: Dr. Xin-Zhong Liang, Earth System Science Interdisciplinary Center, University of Maryland, 5825 University Research Court, Suite 4001, College Park, MD 20740. E-mail: xliang@umd.edu

Abstract

A consistent set of Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation parameters, including leaf and stem area index (LAI and SAI, respectively), land-cover category (LCC), fractional vegetation cover (FVC), and albedo parameterization are developed, and their impacts on North American regional climate are evaluated based on 10-yr Climate–Weather and Research Forecasting Model (CWRF) simulations. As compared with the previous Advanced Very High Resolution Radiometer (AVHRR) set, MODIS LCC increases grassland and cropland fractions in the central Great Plains and Midwest, respectively. Evergreen needleleaf forest converts to mixed forest in the Southeast, and mixed forest converts to evergreen needleleaf in Canada. FVC decreases by 0.05–0.3 over the central Great Plains but increases by 0.1–0.35 over the northern Rocky Mountains, Canada, and the U.S. Southeast. MODIS LAI is less than AVHRR by 2–6, except in the central Great Plains, eastern Rocky Mountains, and central Mexico. LCC and FVC changes over the central Great Plains reduce CWRF warm biases by 0.71°C and wet biases by 0.36 mm day−1. Large LAI reductions cause latent and sensible heat fluxes to decrease by 0.78–5.81 and 0.91–6.54 W m−2, respectively. They also lessen cold biases over the Gulf States and Southeast and wet biases over the North American monsoon region and Canada during summer. In densely vegetated regions including eastern Canada, the Ohio Valley, and the mid-Atlantic region, spring and summer precipitation decreases and temperature increases result from LAI reductions that cause positive evapotranspiration–precipitation–soil moisture feedbacks. Conversely, precipitation and temperature decreases in sparely vegetated regions, such as the Great Plains, result from FVC reductions that cause negative albedo–evapotranspiration–precipitation–soil moisture feedbacks.

Corresponding author address: Dr. Xin-Zhong Liang, Earth System Science Interdisciplinary Center, University of Maryland, 5825 University Research Court, Suite 4001, College Park, MD 20740. E-mail: xliang@umd.edu

1. Introduction

The performance of regional climate models (RCMs) strongly depends upon the accurate specification of land surface boundary conditions (SBCs), given that the surface interacts with the atmosphere through hydrometeorological, biogeophysical, and biogeochemical processes (Liang et al. 2012c). Inaccurate SBC specifications generally result in large biases and uncertainties in the prediction of surface energy, carbon, and related fluxes as well as hydrology (Yuan and Liang 2011). Land SBCs describe the characteristics and properties of soil and vegetation (Liang et al. 2005a,b; Lawrence and Chase 2007; Lawrence and Slater 2008; Lawrence et al. 2011). The vegetation parameters or properties in land SBCs typically include leaf area index (LAI), stem area index (SAI), fraction of vegetation cover (FVC), land-cover category (LCC), and albedo parameterization, where each plays an important role in regulating surface energy and mass balance exchanges (Zeng et al. 2002; Liang et al. 2005a). Additionally, data source consistency and the realistic specification of seasonal and interannual variations of these parameters, both in terms of phenology and physiology, are necessary to produce reliable global and regional climate model simulations (Lawrence and Chase 2007; Liang et al. 2005a; Zeng et al. 2002).

Global remote sensing products are typically used to specify these parameters and improve the representation of land–atmosphere interactions (Henderson-Sellers and Wilson 1983; Zeng et al. 2002). Advanced Very High Resolution Radiometer (AVHRR) Normalized Difference Vegetation Index (NDVI) data have been widely used to derive the vegetation parameters used in various land surface models, including the National Center for Atmospheric Research (NCAR) Land Surface Model (Bonan 1998; Buermann et al. 2001), Common Land Model (CoLM; Zeng et al. 2002; Dai et al. 2004), Community Land Model (CLM) versions 2, 3, and 3.5 (Bonan et al. 2002; Dickinson et al. 2006; Oleson et al. 2008) and Conjunctive Surface and Subsurface Process (CSSP) model (Liang et al. 2005a,b; Yuan and Liang 2011; Liang et al. 2012c). In addition, nearly half of the global circulation models (GCMs) used in the Coupled Model Intercomparison Project phase 5 (CMIP5) have either previously utilized or continue to utilize vegetation data directly or partially derived from AVHRR data for their land surface models (LSMs).

AVHRR data possess major deficiencies when compared with AVHRR’s successor, the Moderate Resolution Imaging Spectroradiometer (MODIS) including 1) quality degradation due to atmospheric effects, satellite drift, and changeover (Gutman 1999), 2) larger uncertainties from the calibration, geometric registration, and cloud screen (e.g., Goward et al. 1991; Zhang et al. 2004), and 3) lower spatial, spectral, and temporal resolutions. Additionally, all AVHRR vegetation parameters (i.e., LAI, LCC, and FVC) are derived from NDVI data that have limited available information and redundant variations (Zeng et al. 2002). These data will be saturated in dense vegetation and cause substantial errors (Zhang et al. 2004).

Numerous studies have indicated that substantial differences exist between the global land surface mapping products from MODIS and AVHRR. This is especially true for FVC, LAI, the fraction of photosynthetically active radiation (FPAR), and land surface albedo (Tian et al. 2004a,b; Wang et al. 2004; Liang et al. 2005a; Lawrence and Chase 2007; Oleson et al. 2008; Ke et al. 2012). Therefore, we need to develop a consistent set of MODIS vegetation data and perform a comprehensive evaluation of their impacts on climate before replacing AVHRR data in current climate models.

The use of MODIS vegetation data is expected to improve climate simulations (Tian et al. 2004b). However, simulation quality is highly dependent upon the physical representations and coupling strength of various land–atmosphere interactions in LSMs (Liang et al. 2005b; Yuan and Liang 2011). For example, albedo is an important parameter that normally depends on the angular and spectral distributions of solar radiation and is determined by surface characteristics that include LAI, SAI, and fractional vegetation cover (Liang et al. 2005b).

Liang et al. (2005b) developed a dynamic and statistical surface albedo parameterization by optimizing MODIS broadband black-sky (direct) and white-sky (diffuse) albedo in visible and near infrared broadbands using both AVHRR and MODIS vegetation data and the North American and Global Land Data Assimilation System outputs of soil moisture during 2000–03. There were, however, three inconsistencies in optimizing the albedo parameterization with the MODIS albedo product in Liang et al. (2005b). First, because of large differences between the AVHRR and MODIS LAI and SAI, the MODIS LAI and SAI were scaled to AVHRR for each LCC. Second, because the LCC and FVC were derived from only one year of AVHRR NDVI data (1992–93), they produced large uncertainties when treated as climatology. Third, because the data used in the study were collection version 4 and only available for a short period (2000–03), snow-free pixel values with an embedded mandatory quality flag of “processed” (QA = 0 and 1) were chosen rather than those with the flag of “best quality” (QA = 0) in order to maximize useable data for the optimization. This may have degraded the data quality and resulted in large albedo parameterization biases.

At present, the new collection version 5 of MODIS contains continuous data from 2000 (to present) that is of higher quality due to a combination of observations from the Terra and Aqua satellites with many refinements, retrieval algorithm improvements, atmosphere correction, and cloud screen and mask1 in 8-day composites. Thus, these data are well suited for albedo optimization and construction of the vegetation climatology to be used in the LSMs. The purpose of this study is to replace the AVHRR-derived vegetation parameters (LAI, SAI, FVC, and LCC) with MODIS-derived data in RCM SBCs and to evaluate the climatic impacts of these modifications. The surface albedo parameterization is also reoptimized using the new MODIS-consistent SBCs. This study provides an overview of new vegetation data and the methods used to derive these data from MODIS products as well as a comprehensive evaluation of the climate impacts that result from the use of these new data. The RCM used in this study is the Climate–Weather and Research Forecasting Model (CWRF; Liang et al. 2012c). Although our results may be somewhat model dependent, they are applicable to RCMs in general through consistent physical analyses, and thus will benefit others when they replace the conventional AVHRR with the recent MODIS data in their climate models.

Section 2 gives a brief description of the CWRF climate model and experimental design. In addition, the processing methods used to derive the new vegetation parameters and observational data for comparisons are described. The results will be presented in section 3 and focus on differences between the MODIS-derived and AVHRR-derived vegetation parameters. The regional surface climates simulated by CWRF using the different parameters will also be compared and discussed in this section. This will be followed by the conclusions in section 4.

2. Model and method

a. Brief description of CWRF

The CWRF model (available at cwrf.umd.edu) is a climate extension of the Weather Research and Forecasting Model (WRF; Skamarock et al. 2008) that inherits all WRF functionalities while enhancing its capability to predict climate by incorporating a grand ensemble of alternate schemes for major physical processes (Liang et al. 2012c). In particular, CWRF incorporates seven important components. The first is an advanced cloud–aerosol–radiation ensemble modeling system that is developed from radiation packages in global and regional climate models to simulate the interplay among clouds, aerosols, and radiation (Liang and Zhang 2013). The second is the core land surface model CSSP that was developed from CoLM and has terrestrial hydrology improvements that account for comprehensive SBCs (Liang et al. 2005a), subgrid topographic controls, and realistic bedrock depth constraints on soil moisture (Choi et al. 2007; Yuan and Liang 2011). CSSP also has an explicit treatment of surface–subsurface lateral flow interactions (Choi and Liang 2010; Choi et al. 2007, 2013). The third is the incorporation of grid and subgrid orographic effects on surface radiation, turbulence stress, and gravity wave drag (Liang et al. 2005c). The fourth is an upper ocean mixing layer module (UOM) to simulate ocean effects (Ling et al. 2011). The fifth is the implementation of two new planetary boundary layer (PBL) schemes and six new cumulus schemes. The sixth is a suite of crop growth modules for climate impact and feedback studies (Liang et al. 2012a,b). Finally, CWRF is fully coupled across all of its components with plug-and-play interfaces and improvements on the integration of external (top, surface, lateral) forcing conditions (Liang et al. 2012c).

The CWRF model has demonstrated significant skill enhancement over the driving GCM for U.S. regional precipitation seasonal forecasts (Yuan and Liang 2011). Moreover, a comparison of CWRF model skill based on a continuous integration during the period 1982–2004, with the WRF and other RCMs, shows that CWRF better simulates radiation and terrestrial hydrology (Liang et al. 2012c). However, Liang et al. (2012c) found that CRWF produces systematic warm surface air temperature biases across the central Great Plains for the entire year and large wet biases over eastern Canada during summer. They suspected that these biases may partially result from the surface albedo parameterization and vegetation parameters developed from combinations of earlier MODIS and AVHRR products. We will show in section 3 that the biases are substantially reduced when the latest and most consistent MODIS data are used.

The CWRF North American computational domain (Fig. 1) for this study is centered at 37.5°N, 95.5°W, covers the entire continental United States with a 30-km grid, and has produced skillful simulations of U.S. regional climate variations resulting from interactions between the planetary circulation (via lateral forcing) and regional surface processes, including orography, soil, vegetation, and adjacent ocean areas (Liang et al. 2012c). The buffer zones are located along the four edges of the domain and have a width of 14 grids, where lateral boundary conditions are temporally and spatially interpolated from the driving GCM or global reanalysis data (Liang et al. 2001, 2012c).

Fig. 1.
Fig. 1.

The geographic distributions of (a) old and (b) new USGS-based land-cover category data, and (c) new FVC and (d) FVC differences (new minus old) in the CWRF computational domain. The eight boxes in (a) are the eight regions that are studied to assess model ability to simulate precipitation and surface air temperature annual cycles.

Citation: Journal of Climate 27, 22; 10.1175/JCLI-D-14-00082.1

b. New vegetation parameters derived from MODIS products

The MODIS data used in this study to derive the new SBC vegetation parameters are the Terra yearly water mask and vegetation continuous field products for the year 2000, and the combined Terra and Aqua yearly LCC (MCD12Q1), 8-day composites (i.e., 1–8 January, 9–16 January, etc.) for LAI (MCD15A2), and albedo parameter (MCD43B1) products in collection version 5 during 2000–10. The products are first aggregated from their original spatial resolutions to a 1-km sinusoidal projection and then remapped using a nearest-neighbor interpolation to a 1-km map projection within a computation domain defined by users. Pixels are processed only where the embedded quality control (QC) flag indicates best quality. The remaining pixels are assigned a missing value. To remove the missing value pixels and minimize cloud and snow contamination, multiyear data are used to construct a climatology of yearly data and 8-day composites. Finally, these 1-km data are spatially aggregated to the RCM grids. The MODIS vegetation parameter derivation details for CWRF are as follows.

1) Land-cover category

The LCC for CWRF is defined to be the vegetation type in the U.S. Geological Survey (USGS) 24-category land-cover classification system that has the largest fractional area within a grid (Liang et al. 2005a). We use MODIS LCC yearly data from 2000–10 with a 500-m spatial resolution to derive a new 1-km LCC in the CWRF map projection by employing the spatial aggregation and map projection mentioned above. The climatology of 1-km LCC is defined to be the vegetation type that occurs most frequently during the 11-yr period. The missing value pixels are effectively eliminated after the climatology calculation. Because the MODIS land-cover category type product provides only five land-cover classification systems [i.e., International Geosphere–Biosphere Programme (IGBP), University of Maryland (UMD), MODIS-derived LAI/FPAR, MODIS-derived Net Primary Production (NPP), and Plant Functional Type (PFT)], but does not include the USGS 24-category classification used by CWRF, a conversion from the MODIS classification systems to the USGS 24-category classification is required before spatial aggregations to the CWRF grids may be performed.

Strahler et al. (1999) provided an example to translate the IGBP to other classification systems (e.g., USGS) that are compatible with those used by the modeling community. They directly remapped one or more IGBP types to their equivalents by relabeling them. Yucel (2006) adopted a similar relabeling (“cross-walking”) approach to translate the IGBP to the USGS and simple biosphere (SiB) classifications used by the fifth-generation Pennsylvania State University–NCAR Mesoscale Model version 5 (MM5).

Liang et al. (2005a) proposed a method to consistently convert LCC between the USGS and IGBP classifications by using corresponding conversion coefficients calculated from LCC data in both the USGS and IGBP classifications that were derived from AVHRR NDVI composites from April 1992 to March 1993. After intersecting the two LCC maps using GIS tools, the fractional areas of all contributing categories within each CWRF grid, and thus the USGS–IGBP conversion coefficients, were determined (Liang et al. 2005a).

In this study, we combine the cross-walking and Liang et al. (2005a) methods. For tree types, since the IGBP and USGS classifications have an exact one-to-one correspondence, we relabel the IGBP types directly to the corresponding USGS types. For other types, including herbaceous vegetation types, the 1-km resolution LCC climatology in the IGBP classification system from MODIS is translated to the LCC in the USGS 24-category classification using the conversion coefficients (Liang et al. 2005a). To make the conversion coefficients more spatially representative, we calculate the conversion coefficient for a 1-km pixel within a 90 km × 90 km spatial window that centers on that pixel rather than within a CWRF grid.

2) Fractional vegetation cover

The FVC is a crucial vegetation parameter that largely affects surface albedo (Tian et al. 2004b). FVC is assumed to be static (i.e., not vary with time in the CWRF SBCs), and all vegetation canopy seasonality is attributed to LAI seasonal variations, following Zeng et al. (2000) and Liang et al. (2005a). This concept was used in the land surface models including CSSP, CoLM, and CLM, and was a basic assumption during model development. The FVC is typically derived from satellite NDVI data using a linear model (Gutman and Ignatov 1998; Zeng et al. 2000; Liang et al. 2005a) or a quadratic model (Carlson and Ripley 1997). However, because of the difficulty in determining the bare soil NDVI and its substantial impact on the derived FVC, the FVC used here is defined to be one minus the sum of the water fraction and bare soil fraction.

The MODIS 250-m water mask dataset (Carroll et al. 2009) for the year 2000 is used to derive the water fraction by computing the approximate proportion of 250-m grids within a CWRF gird. The bare soil fraction is directly calculated from the MODIS 500-m vegetation continuous field dataset (Hansen et al. 2003).

3) Leaf and steam area indices

In climate applications, LAI indicates the vertical vegetative canopy structure and is defined to be the single-side green leaf area per unit ground area in broadleaf canopies and the projected needle leaf area in coniferous canopies (Tian et al. 2004a; Liang et al. 2005a). MODIS Terra and Aqua combined LAI products (Myneni et al. 2002) were used to define the LAI and SAI vegetation parameters. The missing value pixels that resulted from the QC flag filter application are spatially filled using the average of nearby data pixels having the same land-cover type within a certain radius, following Liang et al. (2005a). Moreover, if a pixel is missing at least one 8-day composite in a year, but has at least two measurements per season, then a temporal filling for the missing value(s) within a year is applied by a cubic spline fitting. Finally the 8-day LAI composite climatology is divided by the FVC derived above to calculate the green LAI with respect to the vegetated area used in CWRF (Liang et al. 2005a). The 1-km, 8-day SAI data during 2000–10 are derived from the 1-km, 8-day LAI generated above following the method of Zeng et al. (2002). Both the LAI and SAI are spatially aggregated within a CWRF grid and temporally averaged to generate a monthly climatology.

Yuan et al. (2011) constructed a continuous and consistent global LAI dataset by applying the modified temporal spatial filter method and a Savitzky–Golay filter to fill the MODIS LAI gaps and process lower-quality data (QC > 0). Ke et al. (2012) followed the same methodology to produce a global LAI dataset from MODIS Terra and Aqua combined products in collection version 5. In our study, we use only the best-quality LAI data (QC = 0) to derive the monthly LAI climatology in the North American domain. No LAI inconsistencies or gaps were found using the method described above. We plan to use the global LAI data from Yuan et al. (2011) as a backup in those areas outside of the North American domain where our LAI data may be inconsistent and/or have significant gaps.

4) Land surface albedo

Liang et al. (2005b) developed an improved dynamic and statistical parameterization of nonsnow land surface albedo based on MODIS albedo products. The snow-free albedos (direct and diffuse visible and near infrared) are functions of several vegetation parameters including LAI, SAI, FVC, and LCC, and top 0.1-m volumetric soil moisture. The coefficients in the functions for each LCC are determined from the optimization solutions through advanced inverse modeling that minimizes the calculation error for MODIS albedos across all grids in that LCC (Liang et al. 2005b). Here, we follow their method and use the latest MODIS albedo, LAI, FVC, LCC, and North American Regional Reanalysis (NARR) soil moisture data to derive the nonsnow albedo parameterization at specific CWRF grids.

We hereafter refer to the MODIS-based LAI, SAI, FVC, LCC, and albedo derived from these quantities as “new” vegetation data or parameters, while the corresponding AVHRR-based values will be referred to as “old” vegetation data.

c. Model evaluation data

For model evaluation, daily total precipitation and daily mean surface air temperature (average of maximum and minimum at a screen height of 1.25–2 m above the ground) data from 7325 National Weather Service cooperative stations across the United States are mapped onto the CWRF grid following an objective analysis (Liang et al. 2004) with the topographic adjustment of Daly et al. (2008). Over Canada and Mexico, precipitation and temperature data are based on the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center (CPC) 0.5° daily analysis (Chen et al. 2008) and the Climatic Research Unit (CRU) TS3.0 0.5° monthly-mean analysis (www.cru.uea.ac.uk/cru/data/hrg/), respectively, both from station measurements and with no topographic adjustment.

The upscaled sensible and latent heat fluxes are derived from FLUXNET (Baldocchi 2008; Baldocchi et al. 2001) observations using the data-oriented approach and model tree ensembles (MTE). The derived data can be used to validate the land surface process models at scales that are generally equal to the grid spacing of climate models (Beer et al. 2010; Bonan et al. 2011; Jung et al. 2010, 2011). Cross-validation shows that the MTE produces realistic geographical and seasonal latent and sensible heat flux variability (Jung et al. 2011). However, the MTE flux data are still model outputs from the upscaling of the Lund–Potsdam–Jena Managed Land (LPJmL) model constrained by FLUXNET tower data and cannot be considered as accurate as those from field measurements. These data are mapped onto the CWRF grid using a bilinear spatial interpolation for model performance evaluations below.

d. Experimental design

For simplicity, the CWRF simulation using the old vegetation data and derived albedo is hereafter referred to as the control run (CTL) while that using the new data and derived albedo is the experimental run (EXP). Two 11-yr (1998–2008) simulations in the CWRF domain (Fig. 1), one for CTL and the other for EXP, are conducted with CWRF version 3.1.1, where the physical configurations are identical to those proposed by Liang et al. (2012c). These include the Goddard Space Flight Center (GSFC) radiation scheme (Chou and Suarez 1999; Chou et al. 2001), CSSP land surface model, the Community Atmosphere Model planetary boundary layer scheme (Holtslag and Boville 1993), Ensemble Cumulus Parameterization modified from G3, GSFC Goddard Cumulus Ensemble microphysics scheme (Tao et al. 2003) and University of Washington (UW) shallow convection scheme (Park and Bretherton 2009). Both simulations are driven by the global Interim European Centre for Medium-Range Weather Forecasting (ECMWF) Re-Analysis (ERA-Interim) data (Dee et al. 2011). The initial year is considered as a spinup, while the simulations from the last 10 years (1999–2008) are used in the evaluation below.

3. Results and discussion

a. New land-cover category data

Figures 1a and 1b compare the old and new USGS-based dominant LCC geographic distributions over the North American domain, given a 30-km grid spacing. A comparison of the two maps shows that the new LCC derived from the latest MODIS land-cover product considerably alters dominant vegetation covers over several regions, including the west coast of Mexico (from a mixture of deciduous broadleaf, evergreen needleleaf, and mixed forests to shrubland), the northern Rocky Mountains (mainly from shrubland to grassland), the eastern Dakotas and western Minnesota (from cropland/grassland mosaic to dryland cropland and pasture), eastern Canada (from mixed forest to evergreen needleleaf forest), the U.S. Southeast (from evergreen needleleaf forest to mixed forest), and southeastern Texas (from cropland/grassland mosaic to grassland). In the new LCC map, the dryland cropland type covers much of the U.S. Midwest and eastern Plains and extends into south central Canada, while the shrubland and grassland types prevail over the western Plains and mountainous regions of the western United States and Mexico. The majority forest type in the southeast United States (including the Gulf Coast states) and eastern Canada is altered to mixed forest and evergreen needleleaf forest, respectively.

Overall, western Mexico is the only region where the conversion is from tree to herbaceous types (from deciduous broadleaf, evergreen needleleaf, and mixed forests to shrubland). The MODIS IGBP LCC map reveals that woody savanna and open shrubland are the two major types in this region. When they are remapped onto the USGS LCC classification using the above remapping method, they are mostly reclassified as shrubland. Because of the distinct biophysical and biochemical characteristic differences between trees and shrubland, the climate in this region is expected to be substantially altered. We will discuss the effect of this LCC change on precipitation in section 3d.

Table 1 summarizes the total percentage coverage of each land-cover category across all grids in the domain for both the old and new LCC. The grassland type in the new LCC has the largest increase (4%) in coverage when compared with the old LCC. In addition, the new LCC increases the coverage of urban and built-up areas, irrigated and dryland cropland and pasture, shrubland, mixed shrubland/grassland, savanna, evergreen broadleaf forest, and barren or sparsely vegetated types. Conversely, coverage decreases substantially in the new LCC for the cropland/grassland mosaic and mixed forest types (−2.5% and −2%, respectively) with smaller decreases for water bodies, cropland/woodland mosaic, deciduous broadleaf forest, evergreen needleleaf forest, and wooded wetland. After tabulating differences across all categories, total forest coverage in the new LCC decreases by 3.73% while total cropland (including dryland and irrigated cropland) and other herbaceous coverage increases by 1.24% and 2.44%, respectively.

Table 1.

Summary of new and old vegetation parameters in USGS classifications.

Table 1.

b. New fractional vegetation cover

The new FVC spatial distributions, as well as their differences (new minus old), are illustrated in Figs. 1c and 1d, respectively. The new FVC derived from MODIS is 0.05–0.30 lower over regions located within an arc that extends from south central Canada, through the Great Plains to the desert Southwest. Conversely, FVC values are 0.05–0.35 greater across much of the eastern half of Canada as well as substantial areas of Oregon, Nevada, the U.S. Southeast, and western Mexico.

Table 1 lists the mean FVC for each land-cover category across all land grids for the old and new data. Values increase (old to new) from 0.77 to 0.93 for wooded wetland, 0.42 to 0.59 for shrubland, and 0.73 to 0.88 for wooded tundra. This corresponds to FVC increases in the land-cover types in the eastern half of Canada, the U.S. Southeast, mountainous regions of the western U.S., and areas adjacent to Hudson Bay, respectively, in Fig. 1d. Conversely, the dryland cropland and pasture, grassland, and evergreen broadleaf forest FVCs decrease by 0.04, 0.03, and 0.03, respectively. These decreases occur within an arc shaped region that extends from south central Canada to the desert Southwest in Fig. 1d. For the other land-cover categories, FVC increases are relatively small and range from 0.01 to 0.1.

For a given CWRF grid, the category with the largest fractional coverage is chosen to be the LCC for that grid, and the fraction of the cell that it occupies is assumed to be the FVC (Liang et al. 2005a). Consequently, for a given land cell (i.e., water body fraction is zero), 1.0 − FVC is the bare soil fraction. The new bare soil fraction is 9.6% over the U.S. domain and is 2.1% less than the old bare soil fraction. Ke et al. (2012) found that the bare soil fraction derived from MODIS used in CLM4 is still considerably larger than that from the 2006 National Land Cover Database. Further study on the FVC specification is needed in the future.

c. New LAI and SAI parameters

Both the AVHRR and MODIS LAIs are quite small over the U.S. Intermountain West as well as regions with arid and semiarid climates, while values are much larger across the northwestern and eastern regions of the United States and much of Canada (Fig. 2). In addition, both patterns exhibit substantial seasonal variability over cropland grids in the U.S. Midwest, and mixed forest grids in southern Canada and the U.S. Southeast. Although the LAI spatial pattern derived from MODIS is similar to that from AVHRR, the magnitude is much smaller (Fig. 2). During winter [December–February (DJF)], the MODIS LAI is 0.5–2.5 less than AVHRR over the U.S. Northwest and Southeast as well as southern Canada. The difference patterns during spring [March–May (MAM)], summer [June–August (JJA)], and autumn [September–November (SON)] are similar, when the MODIS LAI is smaller than AVHRR by more than 2.5 over virtually the entire domain, with the exception of the U.S. Great Plains. Differences are particularly large during the summer, when the MODIS LAI is less than AVHRR by more than 4.0 over the U.S. Midwest and southern Canada. The MODIS cropland LAIs over the U.S. Midwest are also much smaller during all seasons except winter. Unlike the LAI parameter, the magnitude of the MODIS SAI parameter (not shown) is close to that of AVHRR in spring and summer but substantially smaller (by 1.8–2.5) over the U.S. Midwest and Northeast as well as southern Canada during autumn and winter.

Fig. 2.
Fig. 2.

(top) Seasonal geographic distributions of the new LAI and (bottom) differences between the new and old LAI (new minus old). (from left to right) The seasons are winter (DJF), spring (MAM), summer (JJA), and autumn (SON).

Citation: Journal of Climate 27, 22; 10.1175/JCLI-D-14-00082.1

Tian et al. (2004b) found that MODIS LAI values are 0.5–1.5 smaller than the AVHRR values used in CLM2 over the U.S. Southeast in July. Tian et al. (2004a) also found that MODIS values were smaller than the AVHRR values used in CoLM by as much as 2 over most of the Northern Hemisphere, especially boreal regions, during March, and that a similar spatial difference pattern occurred in July. Liang et al. (2005a) found that the MODIS LAI is only about half that of the AVHRR LAI during summer over croplands while differences for other land-cover categories, especially trees, are smaller. Lawrence and Chase (2007) identified similar results. Ke et al. (2012) found that MODIS LAI values were much smaller than AVHRR LAI throughout the year, with maximum (minimum) differences over the Northern Hemisphere boreal region during summer (winter). MODIS SAI values are smaller than AVHRR SAI over all land areas throughout the year with the largest differences over the Northern Hemisphere boreal region in autumn (Ke et al. 2012). Our results are consistent with those of the studies described above.

The AVHRR global LAI map is derived generally from the simple relationship of NDVI and LAI as described in Myneni et al. (1997) and is used, without extensive validations, in many land surface models including CLM, CoLM, and CSSP. Because the low quality of AVHRR NDVI data results from the various deficiencies described in section 1 above, the quality of the NDVI derived LAI data also remains low. Recent studies have attempted to correct AVHRR NDVI bias and therefore increase the quality of the LAI derived from it. Tucker et al. (2005) constructed new long-term Global Inventory Monitoring and Modeling Study (GIMMS) NDVI data by correcting the AVHRR NDVI from 1981–2002 in a consistent and quantitatively comparable manner with MODIS and Système Pour l’Observation de la Terre (SPOT)-4 data. The corrections addressed sensor degradation and intercalibration differences, view geometry errors due to satellite drift, biases from lack of atmospheric corrections, and aerosol effects (Tucker et al. 2005). Zhu et al. (2013) developed a neural network algorithm between the third generations of the GIMMS NDVI and best-quality Terra MODIS LAI products for the overlapping period 2000–09 and then generated corresponding LAI data from 1981–2011 using a trained neural network. Their results showed that the improved AVHRR LAI is consistent with the MODIS LAI and is in good agreement (RMSE = 0.68) with observations from 29 sites that represent all major biomes (Zhu et al. 2013). Liu et al. (2012) also generated a consistent long-term LAI product from GIMMS NDVI using the relationship between LAI and the ratio of near-infrared to red band land surface reflectance. Their results indicated that the LAIs from corrected AVHRR NDVI and MODIS are consistent, where LAI differences are less than 0.6 in 99.0% of all vegetated pixels. Additionally, Zhao et al. (2013) used corrected AVHRR surface reflectance data instead of NDVI from NASA’s Land Long Term Data Record project to derive the global LAI product using an artificial neural network algorithm and found that their LAI product is consistent with MODIS LAI data and is both spatially complete and temporally continuous.

In contrast to AVHRR LAI, MODIS LAI is derived from MODIS surface reflectance with various sensor and atmospheric corrections (Myneni et al. 2002) using a lookup table method by an inversion modeling of the three-dimensional radiative transfer in the vegetation canopy. They are typically used to calibrate and validate AVHRR data in the overlapping time (Vermote and Saleous 2006; Liu et al. 2012; Zhu et al. 2013). The MODIS data have been well validated against site measurements (Hill et al. 2006; Yang et al. 2006; Fang et al. 2012) and are continuously improved with a refined algorithm (De Kauwe et al. 2011). Fang et al. (2013) compared five major global LAI products and found that the average uncertainties and relative uncertainties of the MODIS LAI are 0.17% and 11.5%, respectively, which are the smallest among the five products. They further stated that the discrepancies are due mainly to differences between definitions, retrieval algorithms, and input data (Fang et al. 2013).

d. Impacts of new vegetation parameters on regional climate

In this section, the impacts of the new vegetation parameters on regional climate are examined by comparing latent and sensible heat fluxes, precipitation, and surface air temperature, as simulated by CTL and EXP with the old and new vegetation parameters during 1999–2008, respectively. Additionally, their mean bias (MB) and root-mean-square (RMS) errors in the CONUS and eight key regions illustrated in Fig. 1a are compared in Table 2. For precipitation and surface air temperature, we also compare the annual cycles of their biases in the eight regions. A Student’s t test is applied to monthly data to assess the statistical significance of biases from observations and differences between CTL and EXP. Hereafter, the biases and differences with statistical significance greater than 95% are stippled in their geographical distribution maps.

Table 2.

Statistics of the precipitation, surface air temperature, and latent and sensible heat fluxes in eight key regions and continental U.S. (CONUS) land grids. The numbers in the table are mean biases (without parentheses) and root-mean-square errors (with parentheses) between observed and CTL and EXP simulated monthly variations of precipitation and surface air temperature, respectively. For sensible heat flux (SH) and latent heat flux (LH), the statistics are calculated between MTE flux data and CTL and EXP simulations. An asterisk (*) indicates that mean biases or root-mean-square errors are significant at a 95% confidence level.

Table 2.

1) Surface latent and sensible heat fluxes

The top panel of Fig. 3 shows seasonal mean latent heat flux biases (W m−2) from FLUXNET MTE data for the CTL and EXP cases as well as their differences (EXP minus CTL). The CTL simulation produces significant biases almost everywhere during all seasons, with the largest (smallest) absolute biases occurring during summer (winter). The largest positive biases (>25 W m−2) exist over the central Great Plains, southern portions of the Midwest, the Mississippi Valley, and mid-Atlantic regions in spring, and over Mexico, the Rocky Mountains, the Midwest, and central Canada in summer. Significant negative biases (10–28 W m−2) also occur along the U.S. West Coast during both summer and autumn, and the central and southern Great Plains in summer.

Fig. 3.
Fig. 3.

Geographic distributions of seasonal (winter, spring, summer, and autumn) mean (top half) latent heat flux (W m−2) and (bottom half) sensible heat flux (W m−2) biases (departures from MTE flux data) averaged during 1999–2008 as simulated by CTL and EXP as well as their differences (CTL minus EXP). Biases and differences with statistical significance greater than 95% are stippled.

Citation: Journal of Climate 27, 22; 10.1175/JCLI-D-14-00082.1

Although the EXP and CTL patterns are similar, EXP generally produces smaller latent heat fluxes over most regions for all seasons due to its smaller LAI (Fig. 2). Hence, EXP significantly reduces the positive CTL biases over eastern Canada and the northern Rocky Mountains in winter and spring; the Midwest, central Canada, and a broad region that extends from Mexico through the Rocky Mountains to western Canada during summer; and Mexico, the central Great Plains, the Rocky Mountains, and the Midwest in autumn. The improvements are greatest during summer, particularly over the Midwest, the Mississippi Valley, and the region that extends from central Mexico to Arizona and New Mexico, where the area averaged EXP latent heat flux is 16.8 W m−2 smaller than CTL. However, CTL negative biases are enhanced over the central and southern Great Plains, especially eastern Texas, due to latent heat flux decreases in EXP.

Compared to latent heat flux biases, sensible heat flux biases generally have similar spatial patterns with reversed signs (Fig. 3, bottom). The CTL has significant negative biases (3–36 W m−2) over central Canada, central Mexico, the U.S. Northeast, and mountainous regions along the California–Arizona border for all seasons, where the largest biases occur during summer. In addition, significant negative biases are located in the region that extends in an arc from southern portions of the Midwest to the U.S. Northeast during spring. These negative biases are generally reduced in EXP by 2–12 W m−2, especially over Canada, where significant negative biases (20–30 W m−2) in CTL are reduced by nearly half during summer.

Significant positive biases (13–22 W m−2) are produced by CTL for the Great Plains and U.S. West Coast during all seasons except winter. These biases decrease in EXP by approximately 12 and 15 W m−2 in spring and summer, respectively. During autumn, EXP again reduces the magnitude of positive CTL biases over the Great Plains. However, in much of the Midwest, CTL produces negative biases while the EXP generates positive biases.

The latent and sensible heat flux differences between EXP and CTL described above are mainly attributed to discrepancies between MODIS and AVHRR vegetation parameters, especially LAI and FVC. Substantial LAI decreases in densely vegetated regions will generally lead to reduced canopy transpiration and enhanced sensible heat fluxes (Lawrence and Chase 2007; Pielke et al. 2007). In addition, smaller LAI will generally reduce canopy evaporation, due to a decrease in its interception of precipitation, and enhance runoff and soil evaporation (Lawrence and Chase 2007). Compared with CTL, EXP latent heat fluxes are reduced in the Midwest, eastern Canada, eastern Texas, and the U.S. Southeast during summer, while sensible heat fluxes (Fig. 3, bottom panel) increase. In general, enhanced sensible heat fluxes will cause PBL heights to increase as illustrated in Fig. 4. Our results confirm the effects that LAI changes have on sensible and latent heat fluxes, and we will further examine this relationship for precipitation and surface air temperature below.

Fig. 4.
Fig. 4.

Geographic distributions of seasonal (winter, spring, summer, and autumn) mean differences (EXP-CTL) averaged during 1999–2008 for (top) albedo (%), (middle) top 0.1 m soil moisture (mm), and (bottom) planetary boundary layer height (m).

Citation: Journal of Climate 27, 22; 10.1175/JCLI-D-14-00082.1

On the other hand, both the EXP latent and sensible heat fluxes decrease over sparely vegetated areas, including the Great Plains and mountainous areas of the western United States during summer (Fig. 3). This is especially true for the sensible heat fluxes, which diminish during all seasons, when compared with CTL. The LAI differences are not the primary cause for the latent and sensible heat reductions because the EXP and CTL LAIs are very small (less than 1) and have negligible differences (Fig. 2). The fractional vegetation covers in these regions, however, are very small (about 0.35 on average), so that the surface albedo is strongly influenced by the soil albedo. The reduced FVC in EXP in these regions (Fig. 1d) will typically cause surface albedo to increase (Fig. 4) because soil generally is brighter (has higher albedo) than vegetation (Tian et al. 2004b). Increased surface albedo will reduce solar radiation absorption, and consequently decrease both surface sensible and latent heat fluxes. In addition, reduced sensible heat fluxes will cause PBL heights to decrease (Fig. 4). Our results for the Great Plains and U.S. western mountainous areas are consistent with FVC effects on latent and sensible heat fluxes in sparsely vegetated regions.

The positive sensible heat flux MB (RMS) errors in CTL over the central Great Plains, the Southeast, the North American monsoon (NAM) region, and the continental United States (CONUS) are significantly reduced in EXP by 7.54 (3.80), 6.23 (2.13), 3.02 (2.70), and 2.29 (1.82) W m−2, respectively (Table 2). In addition, the positive latent heat flux MB (RMS) errors in CTL over the Midwest, the Northeast, the NAM region, and CONUS are significantly reduced in EXP by 3.78 (2.82), 1.03 (1.47), 5.81 (5.91), and 3.36 (1.72) W m−2, respectively.

2) Precipitation

Figure 5 (top panel) compares seasonal precipitation bias (simulation minus observation) patterns for the CTL and EXP cases averaged during the 10-yr period 1999–2008, as well as their differences (EXP-CTL). The CTL overestimates precipitation almost everywhere throughout the year, except over the west coast of Mexico in summer and coastal areas of the U.S. Northeast in autumn, when dry biases of 0.8–1.6 mm day−1 occur. The overall wet bias and its significance exhibit both seasonal and regional variations. During winter, significant biases occur over the northern United States and western Canada, although the absolute magnitudes are small (less than 0.8 mm day−1). A similar pattern occurs in spring although bias magnitudes increase over central Mexico, the U.S. Midwest, and eastern Canada. Wet biases during summer decrease substantially over the northern Great Plains but increase over the NAM region, Canada, southern Texas, and the Ohio Valley, where values range from 1.2 to 3.4 mm day−1. When compared with the other seasons, CTL bias magnitudes during autumn are smaller and significant only over scattered areas of eastern Canada, the U.S. upper Midwest, and the northern Rockies.

Fig. 5.
Fig. 5.

As in Fig. 3, but for (top half) mean precipitation (mm day−1) and (bottom half) 2-m air temperature (°C). Biases and differences with statistical significance greater than 95% are stippled.

Citation: Journal of Climate 27, 22; 10.1175/JCLI-D-14-00082.1

When compared with CTL during summer, EXP significantly reduces positive rainfall biases over eastern Canada, the NAM region, the Ohio Valley, and Gulf Coast states by 0.90, 0.60, 0.80, and 1.40 mm day−1, respectively. In addition, EXP wet biases during spring over the central Great Plains, upper Midwest, and southern Mexico are less than those in CTL by 0.25, 0.30, and 0.20 mm day−1, respectively. However, they are not statistically different from CTL during spring. During autumn, EXP slightly reduces CTL wet biases over eastern Canada and the southern Great Plains.

As discussed above, in densely vegetated regions, reduced EXP LAI typically decreases evapotranspiration. When a positive evapotranspiration–precipitation feedback exists, evapotranspiration decreases are followed by reduced local cloudiness and precipitation as well as enhanced downwelling solar radiation (Seneviratne et al. 2010). The overall reduction of EXP wet biases in eastern Canada, the Ohio Valley, and the mid-Atlantic during spring and summer confirms the above feedback. Reduced precipitation leads to decreased soil moisture (Fig. 4) and evapotranspiration (Fig. 3), which completes the positive evapotranspiration–precipitation–soil moisture feedback loop.

On the other hand, decreased FVC in sparely vegetated areas, including the broad region that extends from central Mexico to the Intermountain West and Great Plains, will lead to the increased surface albedo and decreased evapotranspiration, sensible and latent heat fluxes described above. The reduced EXP precipitation (Fig. 5) and soil moisture (Fig. 4) illustrated in our results over this area confirm the important contribution of decreased evapotranspiration. In addition, drier soil further increases surface albedo, since soil albedo increases with decreasing top layer soil moisture, especially over dry soil (Liang et al. 2005a). Thus, this set of relationships completes a negative albedo–evapotranspiration–precipitation–soil moisture feedback.

Figure 6a shows the seasonal precipitation cycle for the eight regions. The CTL is generally wetter than EXP over all regions throughout the year except the Northeast in autumn. In addition, EXP occasionally produces dry biases (e.g., the central Great Plains during June through September). The EXP generally reduces CTL wet biases by an average of 0.38 mm day−1 for all regions during winter and spring, while its performance in summer and autumn is region dependent. For example, in summer, EXP reduces CTL wet biases by 0.28, 0.21, 0.21, and 0.41 mm day−1 in the Northeast, Midwest, Southeast, and Gulf States, respectively. However, EXP reverses CTL wet biases to dry biases and underestimates observations by 0.11 and 0.20 mm day−1 over the northern Rockies and central Great Plains, respectively. Both CTL and EXP capture the seasonal cycle for all regions with the exception of the Northeast, where the simulated precipitation maximum lags observations by two months.

Fig. 6.
Fig. 6.

The (a) 1999–2008 monthly mean precipitation (mm day−1) from observations as well as the CTL and EXP simulations, and (b) 2-m air temperature (°C) biases (departures from observations) simulated by CTL and EXP averaged over the eight regions showed in Fig. 1.

Citation: Journal of Climate 27, 22; 10.1175/JCLI-D-14-00082.1

Liang et al. (2012c) argued that CTL peak rainfall overestimations in the NAM region are likely to be realistic since the observational analysis has no topographic adjustment and is expected to underestimate peak amount along mountain ranges. Further study reveals that CTL wet bias reductions in the NAM region in EXP are caused mainly by LCC changes from tree types to shrubland. A sensitivity study of a simulation from 1 January to 31 December 1999 found that, by retaining the tree types in the new LCC, the rainfall peak amount was overestimated by approximately 0.4 mm day−1.

Table 2 compares monthly precipitation MB and RMS errors during 1999–2008 for both the CTL and EXP simulations. The greatest improvements in EXP are found in the NAM, Gulf states, Southeast, Midwest, and central Great Plains regions as well as the entire continental United States, where MB (RMS) error reductions, relative to CTL, are 0.28 (0.23), 0.59 (0.31), 0.42 (0.20), 0.21 (0.25), 0.37 (0.14), and 0.38 (0.12) mm day−1, respectively. For the three remaining regions (the Cascades, northern Rockies, and Northeast), there is no statistically significant improvement when EXP is compared to CTL.

3) Surface temperature

Figure 5 (bottom panel) shows the seasonal cycle of 2-m temperature (T2M) biases (simulations minus observations) for both CTL and EXP. The CTL simulation produces a large area of statistically significant warm biases (1.2°–3.5°C) over the middle of the continent during all seasons except autumn. Maximum extent occurs in winter, when significant biases extend from southern Texas to western Canada, while significant biases are limited to the central Great Plains in spring. Although neither CTL nor EXP produces significant biases over the central Great Plains in autumn, EXP warm biases are 1.0°–1.5°C less than CTL. Significant cold biases of 1°–4°C occur in CTL during winter (summer) over the Great Lakes region, California, and much of Mexico (U.S. Southeast and Gulf states).

A comparison of the two simulations shows that EXP bias magnitudes are substantially less than those produced by CTL during all seasons except in summer. The largest and most significant warm bias reduction (1°–3°C) occurs in the central Great Plains during the cold seasons (winter and spring). In summer, both EXP and CTL have warm biases in the northern Rocky Mountains, central Great Plains, and western Canada. In addition, EXP is about 1.2°C warmer than CTL in the upper Midwest. On the other hand, EXP slightly reduces CTL cold biases in the eastern United States and Gulf states during summer.

As discussed in the previous section, the reduction of CTL warm biases in EXP over the Great Plains during winter and spring results mainly from increased albedo due to decreased FVC. The increased albedo causes reductions in solar radiation absorption, and therefore reduces surface sensible heat fluxes and air temperature. Our results are consistent with the FVC effects in sparsely vegetated areas discussed above. On the other hand, in densely vegetated regions such as eastern Canada, the Midwest, and the Southeast, the enhanced sensible heat flux caused by the reduction of EXP LAI causes the EXP surface temperature to be warmer than CTL. The largest EXP surface temperature increases occur over eastern Canada and the Midwest during summer when the largest LAI differences between EXP and CTL exist.

An analysis of CTL and EXP T2M biases for the eight regions (Fig. 6b) reveals that EXP bias magnitudes are generally 0.2°–1.1°C less than CTL values in winter. The EXP simulation effectively diminishes CTL warm biases over all regions except the Northeast and NAM region, where it slightly increases CTL cold bias magnitudes. Conversely, during summer, EXP biases are generally 0.1°–1.6°C higher than CTL. This produces greater warmth over the Cascades, northern Rockies, and the Midwest, but reduces cold biases over the Gulf Coast states and NAM region. Overall, for the entire year, EXP improves T2M simulations over the central Great Plains and Gulf states by reducing CTL MB (RMS) errors by 0.70° (0.63°) and 0.07°C (0.20°C), respectively (Table 2). For the CONUS, it cuts CTL MB and RMS errors by 0.12° and 0.13°C, respectively.

4. Conclusions

MODIS data have been found to be more realistic than AVHRR and are, thus, preferred for use in climate models. Moreover, because land-use changes continue to occur across large areas of the globe, MODIS (when compared with AVHRR) data better represent current vegetation and are more suitable for current climate simulations. However, before replacing AVHRR data widely used in current climate models, it is necessary to compare the vegetation parameters derived from MODIS and AVHRR and to evaluate their regional climate impacts.

This study first derives the new vegetation parameters from MODIS data during 2000–10 and then replaces the old ones derived from AVHRR in CWRF. When compared with AVHRR vegetation parameters, 1) MODIS LCC greatly alters dominant vegetation covers over the Mexican west coast, the northern Rocky Maintains, eastern Canada, and the U.S. Southeast; 2) MODIS LAI is clearly reduced during all seasons in densely vegetated regions with forest land cover including the U.S. Northeast, the Gulf Coast states, the U.S. Southeast, and Canada; and 3) MODIS FVC is reduced over regions located within an arc that extends from south central Canada, through the Great Plains to the desert Southwest. Because there exist large AVHRR and MODIS vegetation parameter differences, the CWRF simulated regional climates during 1999–2008 differ substantially.

Our results illustrate that the incorporation of MODIS parameters into CWRF substantially increases model performance in precipitation, surface air temperature as well as latent and sensible heat fluxes over several regions, particularly the U.S. central Great Plains, the Southeast, the Gulf states, and the NAM region. Improvements include 1) sensible and latent heat flux bias reductions of 0.78–5.81 and 0.91–6.54 W m−2 across eight regions and 3.46 and 2.29 W m−2 for the CONUS, respectively; 2) wet bias reductions over all regions of 0.12–0.59 mm day−1, depending on region, and 0.28 mm day−1 for the CONUS; and 3) warm temperature bias reductions over the U.S. Great Plains (entire United States) of 0.71°C (0.12°C) averaged over all seasons.

We found that CWRF has strong land–atmosphere coupling that realistically reflects the impacts of land surface parameters on regional climate and is capable of simulating complex land–atmosphere interactions. In densely vegetated areas, including eastern Canada, the Ohio Valley, and the mid-Atlantic region, the reduced LAI causes positive evapotranspiration–precipitation–soil moisture feedbacks that lead to decreased spring and summer precipitation. Reduced LAI also increases the sensible heat flux by decreasing evapotranspiration, which then causes surface temperature to increase during spring, summer, and autumn. Conversely, in sparely vegetated areas, such as the Great Plains, decreased FVC causes negative albedo–evapotranspiration–precipitation–soil moisture feedbacks which lead to reduced precipitation during summer. Decreased FVC also lowers surface temperature during winter and spring due to the sensible heat flux reduction that results from the albedo increase.

Despite the improvements described above for the EXP simulation, there still exist some persistent biases in both the CTL and EXP simulations, including 1) large wet biases over the U.S. Midwest and eastern Canada during spring and summer as well as large dry biases over the U.S. Northeast during autumn and 2) large warm biases over the northern Great Plains during winter and the central Great Plains and eastern Canada during summer as well as large cold biases over the Great Lakes region in winter and Mexico for all seasons. In addition to observational data uncertainties (Holder et al. 2006), these biases may be attributed to errors in physical representations and/or physical or empirical parameters specified in the model. For example, we find that cold biases in the Great Lakes region are caused by extremely cold lake surface temperatures simulated by the CWRF lake model, where the representations of water phase changes, surface physics, and turbulent mixing are unrealistic. We will further investigate these issues to improve regional climate simulations using the new and consistent vegetation parameters derived from MODIS.

Acknowledgments

We thank Dr. Martin Jung for providing us with the MTE data and the Land Processes Distributed Active Archive Center (LP DAAC) for making the MODIS collection 5 data available. This research was supported by the USDA UV-B Monitoring and Research Program, Colorado State University; USDA-CSREES-2009-34263-19774 (G-1449-1); USDA-NIFA-2012-34263-19736 (G-1486-2); USDA-NIFA-2013-34263-20931 (G-7799-1); and NASA TE project NNH12AU03 subaward to the University of Maryland at College Park. We are grateful to Dr. Fengxue Qiao for compiling the observed precipitation and surface temperature data, Dr. Ligang Chen for preparing the CWRF input data for the simulations, and Mr. Shenjian Su for high performance computing technical support. All simulations were conducted on the super computing system (ZEUS) at the NOAA Environmental Security Computing Center (NESCC). We thank two anonymous reviewers for their constructive comments and suggestions. The views expressed are those of the authors and do not necessarily reflect those of sponsoring agencies.

REFERENCES

  • Baldocchi, D., 2008: “Breathing” of the terrestrial biosphere: Lessons learned from a global network of carbon dioxide flux measurement systems. Aust. J. Bot., 56, 1–26, doi:10.1071/BT07151.

    • Search Google Scholar
    • Export Citation
  • Baldocchi, D., and et al. , 2001: FLUXNET: A new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities. Bull. Amer. Meteor. Soc., 82, 24152434, doi:10.1175/1520-0477(2001)082<2415:FANTTS>2.3.CO;2.

    • Search Google Scholar
    • Export Citation
  • Beer, C., and et al. , 2010: Terrestrial gross carbon dioxide uptake: Global distribution and covariation with climate. Science, 329, 834838, doi:10.1126/science.1184984.

    • Search Google Scholar
    • Export Citation
  • Bonan, G. B., 1998: The land surface climatology of the NCAR land surface model coupled to the NCAR Community Climate Model. J. Climate, 11, 13071326, doi:10.1175/1520-0442(1998)011<1307:TLSCOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bonan, G. B., , S. Levis, , L. Kergoat, , and K. W. Oleson, 2002: Landscapes as patches of plant functional types: An integrating concept for climate and ecosystem models. Global Biogeochem. Cycles, 16, 1021, doi:10.1029/2000GB001360.

    • Search Google Scholar
    • Export Citation
  • Bonan, G. B., , P. J. Lawrence, , K. W. Oleson, , S. Levis, , M. Jung, , M. Reichstein, , D. M. Lawrence, , and S. C. Swenson, 2011: Improving canopy processes in the Community Land Model version 4 (CLM4) using global flux fields empirically inferred from FLUXNET data. J. Geophys. Res., 116, G02014, doi:10.1029/2010JG001593.

    • Search Google Scholar
    • Export Citation
  • Buermann, W., , J. Dong, , X. Zeng, , R. B. Myneni, , and R. E. Dickinson, 2001: Evaluation of the utility of satellite-based vegetation leaf area index data for climate simulations. J. Climate, 14, 35363550, doi:10.1175/1520-0442(2001)014<3536:EOTUOS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Carlson, T. N., , and D. A. Ripley, 1997: On the relation between NDVI, fractional vegetation cover, and leaf area index. Remote Sens. Environ., 62, 241252, doi:10.1016/S0034-4257(97)00104-1.

    • Search Google Scholar
    • Export Citation
  • Carroll, M. L., , J. R. Townshend, , C. M. DiMiceli, , P. Noojipady, , and R. A. Sohlberg, 2009: A new global raster water mask at 250 m resolution. Int. J. Digital Earth, 2, 291308, doi:10.1080/17538940902951401.

    • Search Google Scholar
    • Export Citation
  • Chen, M., , W. Shi, , P. Xie, , V. B. S. Silva, , V. E. Kousky, , R. W. Higgins, , and J. E. Janowiak, 2008: Assessing objective techniques for gauge-based analyses of global daily precipitation. J. Geophys. Res., 113, D04110, doi:10.1029/2007JD009132.

    • Search Google Scholar
    • Export Citation
  • Choi, H. I., , and X.-Z. Liang, 2010: Improved terrestrial hydrologic representation in mesoscale land surface models. J. Hydrometeor., 11, 797809, doi:10.1175/2010JHM1221.1.

    • Search Google Scholar
    • Export Citation
  • Choi, H. I., , P. Kumar, , and X.-Z. Liang, 2007: Three-dimensional volume-averaged soil moisture transport model with a scalable parameterization of subgrid topographic variability. Water Resour. Res.,43, W04414, doi:10.1029/2006WR005134.

  • Choi, H. I., , X.-Z. Liang, , and P. Kumar, 2013: A conjunctive surface–subsurface flow representation for mesoscale land surface models. J. Hydrometeor., 14, 1421–1442, doi:10.1175/JHM-D-12-0168.1.

    • Search Google Scholar
    • Export Citation
  • Chou, M.-D., , and M. J. Suarez, 1999: A solar radiation parameterization for atmospheric studies. NASA Tech. Memo. NASA/TM-1999-104606, Vol. 15, 40 pp.

  • Chou, M.-D., , M. J. Suarez, , X.-Z. Liang, , and M. M.-H. Yan, 2001: A thermal infrared radiation parameterization for atmospheric studies. NASA Tech. Memo. NASA/TM-2001-104606, Vol. 19, 56 pp

  • Dai, Y., , R. E. Dickinson, , and Y.-P. Wang, 2004: A two-big-leaf model for canopy temperature, photosynthesis, and stomatal conductance. J. Climate, 17, 22812299, doi:10.1175/1520-0442(2004)017<2281:ATMFCT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Daly, C., , M. Halbleib, , J. I. Smith, , W. P. Gibson, , M. K. Doggett, , G. H. Taylor, , J. Curtis, , and P. P. Pasteris, 2008: Physiographically sensitive mapping of climatological temperature and precipitation across the conterminous United States. Int. J. Climatol., 28, 20312064, doi:10.1002/joc.1688.

    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, doi:10.1002/qj.828.

    • Search Google Scholar
    • Export Citation
  • De Kauwe, M. G., , M. I. Disney, , T. Quaife, , P. Lewis, , and M. Williams, 2011: An assessment of the MODIS collection 5 leaf area index product for a region of mixed coniferous forest. Remote Sens. Environ., 115, 767780, doi:10.1016/j.rse.2010.11.004.

    • Search Google Scholar
    • Export Citation
  • Dickinson, R. E., , K. W. Oleson, , G. Bonan, , F. Hoffman, , P. Thornton, , M. Vertenstein, , Z.-L. Yang, , and X. Zeng, 2006: The Community Land Model and its climate statistics as a component of the Community Climate System Model. J. Climate, 19, 23022324, doi:10.1175/JCLI3742.1.

    • Search Google Scholar
    • Export Citation
  • Fang, H., , S. Wei, , and S. Liang, 2012: Validation of MODIS and CYCLOPES LAI products using global field measurement data. Remote Sens. Environ., 119, 4354, doi:10.1016/j.rse.2011.12.006.

    • Search Google Scholar
    • Export Citation
  • Fang, H., and et al. , 2013: Characterization and intercomparison of global moderate resolution leaf area index (LAI) products: Analysis of climatologies and theoretical uncertainties. J. Geophys. Res. Biogeosci., 118, 529548, doi:10.1002/jgrg.20051.

    • Search Google Scholar
    • Export Citation
  • Goward, S. N., , B. Markham, , D. G. Dye, , W. Dulaney, , and J. Yang, 1991: Normalized difference vegetation index measurements from the Advanced Very High Resolution Radiometer. Remote Sens. Environ., 35, 257277, doi:10.1016/0034-4257(91)90017-Z.

    • 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, doi:10.1029/1998JD200106.

    • Search Google Scholar
    • Export Citation
  • Gutman, G., , and A. Ignatov, 1998: The derivation of the green vegetation fraction from NOAA/AVHRR data for use in numerical weather prediction models. Int. J. Remote Sens., 19, 15331543, doi:10.1080/014311698215333.

    • Search Google Scholar
    • Export Citation
  • Hansen, M. C., , R. S. DeFries, , J. R. G. Townshend, , M. Carroll, , C. Dimiceli, , and R. A. Sohlberg, 2003: Global percent tree cover at a spatial resolution of 500 meters: First results of the MODIS vegetation continuous fields algorithm. Earth Interact., 7, doi:10.1175/1087-3562(2003)007<0001:GPTCAA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Henderson-Sellers, A., , and M. F. Wilson, 1983: Surface albedo data for climatic modeling. Rev. Geophys., 21, 17431778, doi:10.1029/RG021i008p01743.

    • Search Google Scholar
    • Export Citation
  • Hill, M. J., , U. Senarath, , A. Lee, , M. Zeppel, , J. M. Nightingale, , R. J. Williams, , and T. R. McVicar, 2006: Assessment of the MODIS LAI product for Australian ecosystems. Remote Sens. Environ., 101, 495518, doi:10.1016/j.rse.2006.01.010.

    • Search Google Scholar
    • Export Citation
  • Holder, C., , R. Boyles, , A. Syed, , D. Niyogi, , and S. Raman, 2006: Comparison of collocated automated (NCECONet) and manual (COOP) climate observations in North Carolina. J. Atmos. Oceanic Technol., 23, 671682, doi:10.1175/JTECH1873.1.

    • Search Google Scholar
    • Export Citation
  • Holtslag, A. A. M., , and B. A. Boville, 1993: Local versus nonlocal boundary-layer diffusion in a global climate model. J. Climate, 6, 18251842, doi:10.1175/1520-0442(1993)006<1825:LVNBLD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Jung, M., and et al. , 2010: Recent decline in the global land evapotranspiration trend due to limited moisture supply. Nature, 467, 951954, doi:10.1038/nature09396.

    • Search Google Scholar
    • Export Citation
  • Jung, M., and et al. , 2011: Global patterns of land-atmosphere fluxes of carbon dioxide, latent heat, and sensible heat derived from eddy covariance, satellite, and meteorological observations. J. Geophys. Res., 116, G00J07, doi:10.1029/2010JG001566.

    • Search Google Scholar
    • Export Citation
  • Ke, Y., , L. R. Leung, , M. Huang, , A. M. Coleman, , H. Li, , and M. S. Wigmosta, 2012: Development of high resolution land surface parameters for the Community Land Model. Geosci. Model Dev., 5, 13411362, doi:10.5194/gmd-5-1341-2012.

    • Search Google Scholar
    • Export Citation
  • Lawrence, D. M., , and A. G. Slater, 2008: Incorporating organic soil into a global climate model. Climate Dyn., 30, 145160, doi:10.1007/s00382-007-0278-1.

    • Search Google Scholar
    • Export Citation
  • Lawrence, D. M., and et al. , 2011: Parameterization improvements and functional and structural advances in version 4 of the Community Land Model. J. Adv. Model. Earth Syst., 3, M03001, doi:10.1029/2011MS000045.

    • Search Google Scholar
    • Export Citation
  • Lawrence, P. J., , and T. N. Chase, 2007: Representing a new MODIS consistent land surface in the Community Land Model (CLM 3.0). J. Geophys. Res., 112, G01023, doi:10.1029/2006JG000168.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , and F. Zhang, 2013: The cloud–aerosol–radiation (CAR) ensemble modeling system. Atmos. Chem. Phys., 13, 83358364, doi:10.5194/acp-13-8335-2013.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , K. E. Kunkel, , and A. N. Samel, 2001: Development of a regional climate model for U.S. Midwest applications. Part I: Sensitivity to buffer zone treatment. J. Climate, 14, 43634378, doi:10.1175/1520-0442(2001)014<4363:DOARCM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , L. Li, , K. E. Kunkel, , M. F. Ting, , and J. X. L. Wang, 2004: Regional climate model simulation of U.S. precipitation during 1982–2002. Part I: Annual cycle. J. Climate, 17, 35103529, doi:10.1175/1520-0442(2004)017<3510:RCMSOU>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , H. I. Choi, , K. E. Kunkel, , Y. Dai, , E. Joseph, , J. X. Wang, , and P. Kumar, 2005a: Surface boundary conditions for mesoscale regional climate models. Earth Interact., 9, doi:10.1175/EI151.1.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., and et al. , 2005b: Development of land surface albedo parameterization based on Moderate Resolution Imaging Spectroradiometer (MODIS) data. J. Geophys. Res., 110, D11107, doi:10.1029/2004JD005579.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , M. Xu, , J. Zhu, , K. Kunkel, , and J. X. L. Wang, 2005c: Development of the Regional Climate–Weather Research and Forecasting Model (CWRF): Treatment of topography. 2005 WRF/MM5 Users’ Workshop, Boulder, CO, NCAR, 9.3. [Available online at http://www2.mmm.ucar.edu/wrf/users/workshops/WS2005/abstracts/Session9/3-Liang.pdf.]

  • Liang, X.-Z., , M. Xu, , W. Gao, , K. R. Reddy, , K. Kunkel, , D. L. Schmoldt, , and A. N. Samel, 2012a: A distributed cotton growth model developed from GOSSYM and its parameter determination. Agron. J., 104, 661674, doi:10.2134/agronj2011.0250.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., , M. Xu, , W. Gao, , K. R. Reddy, , K. Kunkel, , D. L. Schmoldt, , and A. N. Samel, 2012b: Physical modeling of U.S. cotton yields and climate stresses during 1979 to 2005. Agron. J., 104, 675683, doi:10.2134/agronj2011.0251.

    • Search Google Scholar
    • Export Citation
  • Liang, X.-Z., and et al. , 2012c: Regional Climate–Weather Research and Forecasting Model. Bull. Amer. Meteor. Soc., 93, 13631387, doi:10.1175/BAMS-D-11-00180.1.

    • Search Google Scholar
    • Export Citation
  • Ling, T.-J., , X.-Z. Liang, , M. Xu, , Z. Wang, , and B. Wang, 2011: A multilevel ocean mixed-layer model for 2-dimension applications. Acta Oceanol. Sin., 33, 110.

    • Search Google Scholar
    • Export Citation
  • Liu, Y., , R. Liu, , and J. M. Chen, 2012: Retrospective retrieval of long-term consistent global leaf area index (1981–2011) from combined AVHRR and MODIS data. J. Geophys. Res., 117, G04003, doi:10.1029/2012JG002084.

    • Search Google Scholar
    • Export Citation
  • Myneni, R. B., , R. Ramakrishna, , R. Nemani, , and S. W. Running, 1997: Estimation of global leaf area index and absorbed PAR using radiative transfer models. IEEE Trans. Geosci. Remote Sens., 35, 13801393, doi:10.1109/36.649788.

    • Search Google Scholar
    • Export Citation
  • Myneni, R. B., and et al. , 2002: Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data. Remote Sens. Environ., 83, 214231, doi:10.1016/S0034-4257(02)00074-3.

    • Search Google Scholar
    • Export Citation
  • Oleson, K. W., and et al. , 2008: Improvements to the Community Land Model and their impact on the hydrological cycle. J. Geophys. Res., 113, G01021, doi:10.1029/2007JG000563.

    • Search Google Scholar
    • Export Citation
  • Park, S., , and C. S. Bretherton, 2009: The University of Washington shallow convection and moist turbulence schemes and their impact on climate simulations with the Community Atmosphere Model. J. Climate, 22, 34493469, doi:10.1175/2008JCLI2557.1.

    • Search Google Scholar
    • Export Citation
  • Pielke, R. A., Sr., , J. Adegoke, , A. Beltrán-Przekurat, , C. A. Hiemstra, , J. Lin, , U. S. Nair, , D. Niyogi, , and T. E. Nobis, 2007: An overview of regional land-use and land-cover impacts on rainfall. Tellus, 59B, 587601, doi:10.1111/j.1600-0889.2007.00251.x.

    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., , T. Corti, , E. L. Davin, , M. Hirschi, , E. B. Jaeger, , I. Lehner, , B. Orlowsky, , and A. J. Teuling, 2010: Investigating soil moisture–climate interactions in a changing climate: A review. Earth Sci. Rev., 99, 125161, doi:10.1016/j.earscirev.2010.02.004.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., and et al. , 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp., doi:10.5065/D68S4MVH.

  • Strahler, A. H., and et al. , 1999: MODIS BRDF/albedo product: Algorithm theoretical basis document version 5.0, 53 pp. [Available online at modis.gsfc.nasa.gov/data/atbd/atbd_mod09.pdf.]

  • Tao, W.-K., and et al. , 2003: Microphysics, radiation and surface processes in the Goddard Cumulus Ensemble (GCE) model. Meteor. Atmos. Phys., 82, 97137.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., and et al. , 2004a: Comparison of seasonal and spatial variations of leaf area index and fraction of absorbed photosynthetically active radiation from Moderate Resolution Imaging Spectroradiometer (MODIS) and Common Land Model. J. Geophys. Res., 109, D01103, doi:10.1029/2003JD003777.

    • Search Google Scholar
    • Export Citation
  • Tian, Y., , R. Dickinson, , L. Zhou, , R. Myneni, , M. Friedl, , C. Schaaf, , M. Carroll, , and F. Gao, 2004b: Land boundary conditions from MODIS data and consequences for the albedo of a climate model. Geophys. Res. Lett., 31, L05504, doi:10.1029/2003GL019104.

    • Search Google Scholar
    • Export Citation
  • Tucker, C., , J. Pinzon, , M. Brown, , D. Slayback, , E. Pak, , R. Mahoney, , E. Vermote, , and N. El Saleous, 2005: An extended AVHRR 8-km NDVI dataset compatible with MODIS and SPOT vegetation NDVI data. Int. J. Remote Sens., 26, 44854498, doi:10.1080/01431160500168686.

    • Search Google Scholar
    • Export Citation
  • Vermote, E. F., , and N. Z. Saleous, 2006: Calibration of NOAA16 AVHRR over a desert site using MODIS data. Remote Sens. Environ., 105, 214220, doi:10.1016/j.rse.2006.06.015.

    • Search Google Scholar
    • Export Citation
  • Wang, Q., , J. Tenhunen, , N. Q. Dinh, , M. Reichstein, , T. Vesala, , and P. Keronen, 2004: Similarities in ground- and satellite-based NDVI time series and their relationship to physiological activity of a Scots pine forest in Finland. Remote Sens. Environ., 93, 225237, doi:10.1016/j.rse.2004.07.006.

    • Search Google Scholar
    • Export Citation
  • Yang, W., and et al. , 2006: MODIS leaf area index products: from validation to algorithm improvement. IEEE Trans. Geosci. Remote Sens., 44, 18851898, doi:10.1109/TGRS.2006.871215.

    • Search Google Scholar
    • Export Citation
  • Yuan, H., , Y. Dai, , Z. Xiao, , D. Ji, , and W. Shangguan, 2011: Reprocessing the MODIS leaf area index products for land surface and climate modelling. Remote Sens. Environ., 115, 11711187, doi:10.1016/j.rse.2011.01.001.

    • Search Google Scholar
    • Export Citation
  • Yuan, X., , and X.-Z. , 2011: Evaluation of a Conjunctive Surface–Subsurface Process Model (CSSP) over the contiguous United States at regional–local scales. J. Hydrometeor., 12, 579599, doi:10.1175/2010JHM1302.1.

    • Search Google Scholar
    • Export Citation
  • Yucel, I., 2006: Effects of implementing MODIS land cover and albedo in MM5 at two contrasting U.S. regions. J. Hydrometeor., 7, 10431060, doi:10.1175/JHM536.1.

    • Search Google Scholar
    • Export Citation
  • Zeng, X., , R. E. Dickinson, , A. Walker, , M. Shaikh, , R. S. DeFries, , and J. Qi, 2000: Derivation and evaluation of global 1-km fractional vegetation cover data for land modeling. J. Appl. Meteor., 39, 826839, doi:10.1175/1520-0450(2000)039<0826:DAEOGK>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zeng, X., , M. Shaikh, , Y. Dai, , R. E. Dickinson, , and R. Myneni, 2002: Coupling of the Common Land Model to the NCAR Community Climate Model. J. Climate, 15, 18321854, doi:10.1175/1520-0442(2002)015<1832:COTCLM>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zhang, P., , B. Anderson, , M. Barlow, , B. Tan, , and R. B. Myneni, 2004: Climate-related vegetation characteristics derived from Moderate Resolution Imaging Spectroradiometer (MODIS) leaf area index and normalized difference vegetation index. J. Geophys. Res., 109, D20105, doi:10.1029/2004JD004720.

    • Search Google Scholar
    • Export Citation
  • Zhao, X., and et al. , 2013: The Global Land Surface Satellite (GLASS) remote sensing data processing system and products. Remote Sens., 5, 24362450, doi:10.3390/rs5052436.

    • Search Google Scholar
    • Export Citation
  • Zhu, Z., and et al. , 2013: Global data sets of vegetation leaf area index (LAI)3g and fraction of photosynthetically active radiation (FPAR)3g derived from Global Inventory Modeling and Mapping Studies (GIMMS) normalized difference vegetation index (NDVI3g) for the period 1981 to 2011. Remote Sens., 5, 927948, doi:10.3390/rs5020927.

    • Search Google Scholar
    • Export Citation
Save