Evapotranspiration estimation for fire danger models

A comparison was made between a dryness indexmeasured over an oak-grassland savannah and the KBDI index using a range of annual precipitation inputs. The original KBDI, using input of the actual mean annual precipitation, greatly underestimated fire risk of a oak-grassland savannah in Sierra Nevada foothills in California. This under-estimate was attributed to an under-estimate of potential ET in the original KBDI model. Increasing the mean annual rainfall improved the index, but justification of the change is questionable. Substituting ETo from either the Hargreaves-Samani or the Penman-Monteith equation for ETp in the KBDI model greatly improved performance. None of the KBDI curves was able to provide day-by-day fluctuations in the fire risk index that were comparable with the measured fuel dryness index using the surface renewal method. Further research is needed to determine if use of the Hargreaves-Samani ETo model will improve the KBDI in other climates and over other surfaces.