At its core, the FSB is based on a coarse-scale classification and quantification of fire regimes and tunable land management strategies. Fire regime attributes are created from reconstructed fire history, historical fire records, and potential natural vegetation. The FSB also utilizes weather and climate information from global climate model output. The climate projections used in this effort are those from the NCAR Climate Systems Model which contains the CCM3 atmospheric model at T42 resolution, but the FSB methodology is not limited to a particular model or resolution. The FSB utilizes in parallel known statistical relationships between climate, vegetation, and fire to create fire-ignition probabilities and a probability distribution of estimated fire sizes. The FSB is still under development, but we present initial results from forecasts for 1996-2000 and 2046-2050 based on initial vegetation, lightning, and relative humidity relationships. Future work will utilize dynamical downscaling of the GCM output using the MM5 mesoscale model to increase spatial resolution and adapt the FSB to application at multiple spatial scales.