About this Tool
Logistic regression models estimate the probability that future groundwater levels will be at or less than the lower 10th percentile groundwater level at selected wells in the northeastern U.S. The ability of forecast models to estimate the probability of groundwater levels at or less than the 10th percentile varies among the wells analyzed, and as a result, predictions may be available from one to three months into the future for any month of the year. The explanatory variables used in the forecast models represent current conditions at the time of forecasting. For example, for a one-month forecast model, the current month-end values for groundwater levels and Palmer Drought Severity Index provide the input data to forecast the probability that a groundwater level will be at or less than the 10th percentile threshold at the end of the next month. Forecast methods are described in the article "Forecasting the Probability of Future Groundwater Levels Declining Below Specified Low Thresholds in the Conterminous U.S."
Custom symbology developed in ArcMap to display three summer-month groundwater-drought probabilities for each groundwater well. Clicking the left-most square (actual well location) displays pop-up information. Drought-probability values are shown using a color-coded scale of 7 probability classes. An 8th no-equation class for months without a forecast, no satisfactory forecast model was derived.
Numbers (1,2,3) in custom symbology represents drought probabilities for next three months (first, second and third month).