ON THE INFLUENCE OF BIOMASS BURNING ON THE SEASONAL CO2 SIGNAL AS OBSERVED AT MONITORING STATIONS

We investigated the role of biomass burning in simulating the seasonal signal in both prognostic and diagnostic analyses. The prognostic ana lysis involved the High-Resolution Biosphere Model, a prognostic terre strial biosphere model, and the coupled vegetation fire module, which together produce a prognostic data set of biomass burning. The diagnos tic analysis involved the Simple Diagnostic Biosphere Model (SDBM) and the Hao and Liu [1994] diagnostic data set of biomass burning, which have been scaled to global 2 and 4 Pg C yr(-1), respectively. The mont hly carbon exchange fields between the atmosphere and the biosphere wi th a spatial resolution of 0.5 degrees x 0.5 degrees, the seasonal atm osphere-ocean exchange fields, and the emissions from fossil fuels hav e been coupled to the three-dimensional atmospheric transport model TM 2. We have chosen eight monitoring stations of the National Oceanic an d Atmospheric Administration network to compare the predicted seasonal atmospheric CO2 signals with those deduced from atmosphere-biosphere carbon exchange fluxes without any contribution from biomass burning. The prognostic analysis and the diagnostic analysis with global burnin g emissions of 4 Pg C yr(-1) agree with respect to the change in the a mplitude of the seasonal CO2 concentration introduced through biomass burning. We find that the seasonal CO2 signal at stations in higher no rthern latitudes (north of 30 degrees N) is marginally influenced by b iomass burning. For stations in tropical regions an increase in the CO 2 amplitude of more than 1 ppmv (up to 50% with respect to the observe d trough to peak amplitude) has been calculated. Biomass burning at st ations farther south accounts for an increase in the CO2 amplitude of up to 59% (0.6 ppmv). A change in the phase of the seasonal CO2 signal at tropical and southern stations has been shown to be strongly influ enced by the onset of biomass burning in southern tropical Africa and America. Comparing simulated and observed seasonal CO2 signals, we fin d higher discrepancies at southern tropical stations if biomass burnin g emissions are included. This is caused by the additional increase in the amplitude in the prognostic analysis and a phase shift in a diagn ostic analysis. In contrast, at the northern tropical stations biomass burning tends to improve the estimates of the seasonal CO2 signal in the prognostic analysis because of strengthening of the amplitude. Sin ce the SDBM predicts the seasonal CO2 signal reasonably well for the n orthern hemisphere tropical stations, no general improvement of the fi t occurs if biomass burning emissions are considered.