Evaluating the performance and mapping of three fuel classification systems using Forest Inventory and Analysis surface fuel measurements

• Three fuel classifications were evaluated for performance and mapping accuracy.
• All three fuel classifications performed poorly in describing fuel loadings.
• All three maps of these classifications had low accuracies.
• Classifications and maps were inaccurate because of high biomass variability across scales.

Fuel Loading Models (FLMs) and Fuel Characteristic Classification System (FCCSs) fuelbeds are used throughout wildland fire science and management to simplify fuel inputs into fire behavior and effects models, but they have yet to be thoroughly evaluated with field data. In this study, we used a large dataset of Forest Inventory and Analysis (FIA) surface fuel estimates (n = 13,138) to create a new fuel classification called Fuel Type Groups (FTGs) from FIA forest type groups, and then keyed an FLM, FCCS, and FTG class to each FIA plot based on fuel loadings and stand conditions. We then compared FIA sampled loadings to the keyed class loading values for four surface fuel components (duff, litter, fine woody debris, coarse woody debris) and to mapped FLM, FCCS, and FTG class loading values from spatial fuel products. We found poor performances (R2 < 0.30) for most fuel component loadings in all three classifications that, in turn, contributed to poor mapping accuracies. The main reason for the poor performances is the high variability of the four fuel component loadings within classification categories and the inherent scale of this variability does not seem to match the FIA measurement scale or LANDFIRE mapping scale.