An NCEP Global Ensemble Forecast System (GEFS) was recently implemented to improve probabilistic forecasts. More importantly, the North American Ensemble Forecast System (NAEFS) has operationally generated products since 1200 UTC on May 30, 2006. Anomaly forecast by percentile is one of the NAEFS products made after a bias correction from the current analysis. Based on the NCEP/NCAR 40-year reanalysis, a daily climatological distribution (PDF) has been built up for 19 atmospheric variables such as height, temperature, winds, etc.. This anomaly uncertainty information allows users to create climatological percentiles that correspond to any percentile forecast value, such as 5 or 10 (low extreme), 50 (medium), and 90 or 95 (high extreme) bias corrected values. For example, a low extreme (5%) of the ensemble 2 meter temperature forecast could be much above climatological normals (or a high percentile), which indicates high probability of a high temperature extreme. This could be a heat wave in the Northern Hemisphere summer season if combined with high humidity. If downscaled from the current one by one degree forecast to 5 kilometer resolution by applying local climatology, this product could be a future part of the NDGD (National Digital Guidance Database) uncertainty information for users.