Integrating satellite and pollen data with biogeochemical modelling to reconstruct long-term trends in the productivity and carbon sequestration of terrestrial ecosystems

The annual net primary productivity (NPP) of terrestrial ecosystems directly regulates the amount of CO2 in the atmosphere (Huntet al., 1996; IPCC, 1995). The advent of satellite-based remote sensing within the past two decades has enabled us to assess global patterns of phenology, NPP, and carbon sequestration. However, the satellite record is far too short to determine centennial or longer trends. Therefore we must look to new avenues if we are to place the 20-yearrecord of satellite-measured terrestrial NPP within the context of the steady adjustments of plant populations in response to long-term changes in climate and atmospheric CO2. Fossil pollen data provide continuous records spanning millennia that demonstrate the large shifts in the ranges and abundances in plant populations that have occurred since the last glacial maximum (LGM), 21,000 years ago. Maps of modern pollen abundances clearly show that the pollen percentage of a plant taxon reflects its abundance on the landscape. Although there have been several attempts to estimate the past amounts of carbon stored in the terrestrial biosphere, most of these attempts have not made use of process-based biogeochemistry models and their estimates have ranged widely.

I suggest a pioneering approach to determine the past productivity and carbon sequestration of the terrestrial biosphere. I plan to synthesize the information contained in the pollen and satellite vegetation records to estimate the leaf area index (LAI) of past ecosystems, then use the inferred LAI, maps of past land cover already published (Williams et al., 1998; Williams et al., in press), and paleoclimatic simulations to run BIOME-BGC, a sophisticated biogeochemistry model that predicts NPP and carbon sequestration by directly simulating the processes of photosynthesis, respiration, and carbon allocation (Running and Hunt, 1993). The goal is to produce maps of LAI, NPP, and carbon storage for key time periods in the past. These maps will be highly relevant to those seeking to understand the present-day interactions between vegetation productivity, climate, and atmospheric CO2, and the potential for the system to change.