Centre for Resource and Environmental Studies Machine VP
Co-Investigators Jay W Larson, Robert J Oglesby and S Marshall
Centre for Resource and Environmental Studies, Purdue University, USA and University of North Carolina, USA
Climatic Effects of Biomass Burning
Biomass burning occurs as a result of natural causes or through the actions of humans. Every year, the world-wide emission of carbon from biomass burning is estimated to fall in the range 3.0-6.2 PgC (peta-grams carbon). The net flux of carbon to the atmosphere from changes in land use is of the order 1.0-2.0 PgC with nearly all of this flux occurring as a result of tropical deforestation. IPCC projections forecast continued high levels of deforestation well into the 21st century. The effect of increasing CO2 in the atmosphere as a result of deforestation and fossil fuel combustion has been extensively studied. The indirect cooling effect of the smoke released from deforestation has, however, received little attention.
Taylor and Zimmerman developed a model to predict the distribution and intensity of biomass burning at a resolution of 2.5o latitude by 2.5o longitude. This model has been used in a number of studies of the sources and sinks of key greenhouse gases, e.g., CO2, CH4 and N2O. The model predicts that more than half of the emissions will occur in the southern hemisphere, which implies that biomass burning may be the most significant anthropogenic source of particulates affecting climate in that hemisphere. It should also be noted that biomass burning occurs predominantly during the dry season when natural cloud formation will generally be at its lowest level. The Taylor and Zimmerman model will be used to generate, on a month-by-month basis, an estimate of the flux of particulates to the atmosphere as input into climate modelling studies of the impact of smoke on climate.
What are the basic questions addressed?
How does biomass burning at the global scale effect climate?
What are the results to date and the future of the work?
Considerable progress was made towards implementing the necessary modifications of CCM1, which enabled us to make several short runs to explore the impacts that the biomass smoke made in the climate model. We found noticeable cooling in the regions where the biomass smoke was released (especially portions of South America and Africa) and much smaller effects elsewhere. These results were presented at the Western Pacific Regional Geophysics Meeting in Hong Kong, July 1994. Difficulties in getting a stable short wave heating profile in the model have prevented us, however, from making the long simulations necessary to properly evaluate the effects. Because of this, we have decided to switch from CCM1 to the newly released, more sophisticated CCM2 (which is now running on the VP2200). The manner in which we modify the presence of clouds is essentially the same as the CCM1, but because the code is written in a much more systematic fashion, it has proven considerably easier to make the necessary modifications. This change also provides the additional advantages that (i) we can better specify the optical properties of the biomass smoke clouds and (ii) we can use the model to explicitly move the clouds around.
What computational techniques are used and why is a supercomputer required?
Even highly vectorized general circulation models are computationally expensive. For example, CCM2 requires ~18 CPU hours per model year. Model runs need to be over tens of years in order to accumulate appropriate climate statistics.