As recently articulated in the NIDIS implementation plan, decision-makers in drought sensitive sectors would gain substantial utility from enhanced, objectively-derived drought outlooks. This presentation reports on new statistical-dynamical approaches (emphasis on the former) that are being developed in the area of drought prediction on monthly to annual time scales. Of particular importance to the seasonal drought prediction problem is the initial drought state, which can be identified via multiple drought indicators to provide the most utility to a variety of users. Given the initial conditions, transitional drought probabilities can then be computed as a function of location, season, lead time, SST state, and climate model inputs. These methods allow for probabilistic drought transition probability assessments even in locations and seasons where there is little or no skill in seasonal climate predictions.

Output from these tools can be displayed in multiple ways which include 1) current drought condition maps for multiple indicators (SPI, PDSI, SC-PDSI, etc. and updated in real time), 2) maps of marginal probability distributions for a variety of indicators, 3) maps of probability of exceeding user-specified drought indicator thresholds, and 4) time series of historical drought index values and "plume charts" of predicted conditions from one month to a year for an individual location. Examples from a prototype tool being developed will be presented.