This research area involves David Schimel and collaborators Bobby Braswell, Berrien Moore III, and Ernst Linder (all of University of New Hampshire). In earlier modeling work, we suggested, based on the responses of the Century ecosystem model, that a substantial portion of the response of ecosystems to temperature anomalies should be lagged relative to the forcing. These lags occur in the model because of the long turnover times of soil organic matter and deep soil moisture compartments. In an analysis of the effects of the Mt. Pinatubo climate anomaly, Charles Keeling (Scripps Institution of Oceanography) showed that growth rates of atmospheric carbon dioxide (CO₂) dropped after the eruption, suggesting anomalous terrestrial carbon storage. Keeling also showed that the drop in growth rate began before the Mt. Pinatubo eruption, and he suggested lagged effects of the anomalous warm years of the late 1980s. We conducted a cross-correlation analysis of CO₂ growth rate anomalies versus satellite-derived (Microwave Sounding Unit, MSU) temperature anomalies. We found a weak, instantaneous, positive correlation between temperature and growth rate, and a significant anti-correlation lagged 1.5 to 3 years. The sign and timing of this correlation are consistent with model-predicted responses. We conducted a further analysis of the satellite vegetation index, derived from the NOAA Advanced Very High Resolution Radiometer (AVHRR) instrument. Using geostatistical techniques, we were able to show significant lagged correlations between temperature and "greenness" but with the patterns of correlation varying significantly between different vegetation types. Northern ecosystems become "greener" in warm years, while arid systems become less "green" in warm years. The lagged effects are once again of opposite sign, suggesting feedbacks through nutrient cycling. This work provided important observational evidence for the mechanisms linking temperature to ecosystem dynamics but also showed that different biomes (northern versus tropical forests and grasslands) behave differently with respect to climate change. This work was published in Science in October, 1997.

The launch of the EOS AM-1 satellite is planned for June, 1998. AM-1 will carry three new instruments that will provide dramatic improvements in remote sensing of the land. The algorithms being developed by the instruments focus on evolutionary advances based on techniques developed for the AVHRR and Land Remote Sensing Satellite (LANDSAT) thematic mapper instruments. The Ecosystem Dynamics and the Atmosphere Section (EDAS) hosts an Earth Observing System (EOS) Interdisciplinary Team, with collaborators Carol Wessman (University of Colorado) and Braswell (University of New Hampshire), that is developing next generation retrieval techniques based on the direct inversion of a radiative transfer model and the use of spectral mixture modeling (a technique that estimates the proportions of various components, vegetation, bare soil, etc., within a pixel). The algorithm combines data from the Moderate Resolution Imaging Spectrometer (MODIS) with data from Multi-angle Imaging Spectroradiometer (MISR) to take advantage of the anisotropy of the earth's reflectance (the "Bi-directional Reflectance Distribution" function) to aid in model inversion. The algorithm retrieves key land surface and land-atmosphere interaction variables: albedo, leaf area, interception of photosynthetically-active radiation, and timing of leaf growth and death. The algorithm was first applied as part of the Atmospheric Chemistry Division's (ACD) Experiment for Regional Sources and Sinks of Oxidants (EXPRESSO) mission and is now being applied in Texas, Colorado-Kansas, and Brazil as part of several investigations. This effort represents the culmination of a long-term development effort (12 years) by Schimel and Wessman.

During the development of the Climate System Model (CSM), it has become apparent that a from-the-ground-up effort was needed to develop a new land model. Linking the existing NCAR codes (Land Surface Model (LSM) of Bonan and Century of Schimel and collaborators), while complementary in the processes and timescales they address, presented major conceptual problems. In addition, neither model addresses changes to vegetation type. As a consequence, a collaboration has developed with Jon Foley (University of Wisconsin) to develop a new version of Foley's IBIS model that would combine the science in LSM and Century. This model will be adapted for studying climate change, land use change, and paleoclimate interactions. It will be formulated for long-term climatological applications of the CSM, while Century and LSM will remain as alternate approaches with compatible mechanisms to the integrated model.