In projecting climatic adjustments to anthropogenically elevated atmos
pheric carbon dioxide, most global climate models fur biome distributi
on to current geographic conditions. Previous biome maps either remain
unchanging or shift without taking into account climatic feedbacks su
ch as radiation and temperature. We develop a model that examines the
albedo-related effects of biome distribution on global temperature. Th
e model was tested on historical biome changes since 1860 and the resu
lts fit both the observed temperature trend and order of magnitude cha
nge. The model is then used to generate an optimized future biome dist
ribution that minimizes projected greenhouse effects on global tempera
ture. Because of the complexity of this combinatorial search, an artif
icial intelligence method, the genetic algorithm, was employed. The me
thod is to adjust biome areas subject to a constant global temperature
and total surface area constraint. For regulating global temperature,
oceans are found to dominate continental biomes. Algal beds are signi
ficant radiative levers as are other carbon intensive biomes including
estuaries and tropical deciduous forests. To hold global temperature
constant over the next 70 years this simulation requires that deserts
decrease and forested areas increase. The effect of biome change on gl
obal temperature is revealed as a significant forecasting factor.