Testing existing models for predicting hourly variation in fine fuel moisture in eucalypt forests

One of the simple empirical fuel moisture prediction models worked well on the litter fuels for which it was intended, but another of the simple models in common use, predicted poorly, particularly on the damp site. One of the simple models was modified, and gave acceptable predictions for profile litter. Another simple model was calibrated to give good predictions for near surface, bark and litter fuels in the middle of the day. A more physically-based 'bookkeeping' model for estimating response time and EMC worked very well on all strata but the litter profile. Response times of all fuel strata were of the order of 1 h. A process based model worked slightly better than the EMC estimation model for near surface, litter and bark fuel and considerably better for profile litter, especially on the damp site. On the whole these models performed better than the simple empirical models, but they were tuned to various degrees by site information.