Final Report


Biosphere-Atmosphere Interactions

NASA Earth Observing System (EOS) Interdisciplinary Science (IDS) 1991-2000 

Our NASA Earth Observing System- Inter-Disciplinary System (EOS-IDS) team has played a central role in advancing earth system science understanding the past decade. “Biosphere-Atmosphere Interactions” is an EOS-IDS team selected in the original competition in 1991. At the suggestion of NASA management, the team resulted from the merger of Piers Sellers’ group in global scale modeling and observations and Harold Mooney’s group in local-scale biophysical and biogeochemical processes. Principal investigator responsibilities, initially in the hands of Piers Sellers, transferred to Dave Randall in 1996, and rotated to Inez Fung in 1998.

The focus of our work the past 10 years has been atmosphere and biosphere exchanges of energy, water, carbon, and other trace constituents at all space and time scales. The exchanges are dependent on and, in turn, alter the states of the biosphere and the atmosphere. The ten years of research of our IDS project has resulted in tremendous progress in the study of biosphere-atmosphere interactions. The progress has come from global and multi-temporal satellite and in situ observations of ecosystem variations, and the modeling of the biophysics and biogeochemistry on scales compatible with global climate models. Satellite observations have been fundamental to our research.

The research of the IDS team has integrated the diverse scales and approaches of the Sellers and Mooney groups into single-framework investigations. The global-scale multi- temporal NDVI observations is a major team product, and has served as the starting point for the biophysical and biogeochemical modeling of different aspects of biosphere- atmosphere interactions. Satellite data are required to understanding the spatial and temporal variability of biospheric processes, and the modeling studies have stimulated an exploration of their consequences on climate and atmospheric composition.

We have used the 1981-1999 advanced very high resolution radiometer to represent global variations in photosynthetic capacity and related variables through time. The same AVHRR data, augmented by Landsat data, have been used to produce improved descriptions of land cover. The unique contributions of satellite data enabled us to simulate biosphere-atmosphere interactions with unprecedented accuracy and realism. Our work is continuing, as we incorporate improved satellite data streams from the Terra Platform into our studies of biosphere-atmosphere interactions through a partial continuation of our previous work. We are presently working to make the transition to MODIS, MISR, and ASTER data as we continue our studies into the new millennium. Satellite data will continue to be a fundamental component of our biosphere-atmosphere interaction research. It is impossible to capture the spatial and temporal complexity of the biosphere which our advanced coupled models require without using satellite data.

Major Accomplishments of the IDS project “Biosphere-Atmosphere Interactions”

  • Global distributions of land surface properties have been derived from satellite observations for use in GCM studies of energy and water exchange [Defries and Townshend, 1994; Sellers, 1995; Sellers et al., 1995].
  • We have produced a global 20-year time series of NDVI by merging and intercalibrating observations across different instruments on different polar orbiters [Los, 1993 and 1998; Malmstrom et al., 1997]. We have succeeded in significant reductions in errors in the NDVI so that the time series can be used to assess interannual variations in vegetation at the global scale [Tucker and Nicholson, 1999].
  • We have led the development of a third generation SVAT model SiB2 for incorporation into atmospheric GCMs [Randall et al., 1995 and 1996; Sellers et al., 1996b; Sellers et al. 1996c]. SiB2 incorporates realistic biophysics and links the transpiration of water with the assimilation of carbon. A unique feature of our approach is the a priori incorporation of satellite information into the model formulation and data stream.
  • We have developed a new global biogeochemical model CASA that is forced by, inter alia, satellite observations of photosynthetically active radiation and employs distribution of FPAR from NDVI [Potter et al., 1993; Field et al., 1995].
  • We have developed a new approach for more realistic characterizing, from satellite observations, land surface variations as a continuum rather than by discrete biomes [Defries et al., 1995; DeFries et al., 1999].
  • We first hypothesized that climate variability is a non-negligible contributor to variations in annual imbalances in CO2 net flux [Dai and Fung, 1993]. Using the NDVI time series and an inverse model, we showed that an early growing season at high latitudes is directly observed by the NDVI [Myneni et al., 1997] and is corroborated by analysis, via tracer transport modeling, of the changing seasonal cycle of atmospheric CO2 in the Northern Hemisphere [Randerson, et al. 1999].
  • Using SiB2-GCM, we showed that vegetation variability (based on 1981-1990 NDVI) may contribute to the variability in the physical climate [Bounoua et al., accepted in Journal of Climate].
  • Using the SiB2-GCM, we showed, for the first time, that direct effects of increased CO2 on vegetation physiology will lead to a relative reduction in evapotranspiration over the continents, with associated regional warming and drying over that predicted for conventional greenhouse warming effects, particularly in the tropics (Figure 1) [Sellers et al., 1996; Bounoua et al., 1999].
  • Using the SiB2-GCM, we showed that covariation of seasonally varying CO2 fluxes and the height of the planetary boundary layer contributes to a positive CO2 concentration in the PBL in the annual mean, even when fluxes cancel in the annual mean (the rectifier effect )[Denning et al., 1995 and1999]. This finding has significant implications for the magnitudes of CO2 sources and sinks inferred from atmospheric CO2 measurements in the PBL.
  • Using CASA, we have produced the first global model of C13 exchange with the biosphere and first calculation of the isotopic disequilibrium due to the long residence time of carbon in the biosphere [Fung et al., 1997]. The long residence time suggests that C4 vegetation takes up a non-trivial fraction anthropogenic CO2 [Fung et al., 1997] and that CO2 fertilization is not the only mechanism responsible for the uptake [Randerson et al., 1999].
  • In collaboration with Dickinson’s IDS team, we have participated in the inclusion into GCM climate simulations the effects of nitrogen controls on photosynthesis and hence the water and energy cycles (Dickinson et al., 2000).
  • We have initiated modeling of aspects of biosphere-atmosphere interactions other than energy, water and carbon exchange. These include the cycles of oxygen18 in CO2 [Ciais et al., 1997; Peylin et al., 1999], mineral aerosols [Tegen and Fung, 1994; Tegen and Fung, 1995; Tegen et al., 1996], and iron [Fung et al., 2000].