Interpretation of fire danger rating indices in the U.S. is based on the historic weather of a given area. Index values can be interpreted differently for different regions and for different fuel models in the same region. Therefore, indices are often expressed in relative terms (e.g., 97th and 90th percentile). Fire danger levels based on percentiles depend only on historical weather data for selected years and season. Fire danger levels have been used in fire management operations and planning, for producing fire danger maps for the Wildland Fire Assessment System (WFAS), and for public awareness campaigns. In addition, percentile levels are used to define average and extreme conditions, and associated moisture values have been used for fire behavior modeling. Studies have indicated that changes in wildland fire activity are, at least in part, a product of climate change. Fire danger indices and their interpretation should reflect this change. We consider several issues related to the use of percentiles to define fire danger levels, including: missing weather data, seasonal vs. annual weather records, and the number of years of data included in the record. We also examine possible trends in fire danger rating indices as a result of recent and predicted changes in climate.