Dynamic-statistical models giving spatially-continuous predictions lightning probability in three-hour intervals to 48 hours in the warm months have run at the Canadian Meteorological Center (CMC) since 2003 (Burrows et al., WAF, 2005). Since the original development it became possible to include important predictors not previously available. New models were developed and have run daily at CMC since April 2006. Model resolution is 15 km. Training data consists of just 1 day per month March to September 2005 for a domain covering the northern United States and much of southern Canada. There are two predictands derived from observations by the North American Lightning Detection Network: (1) “time-area coverage” of lightning (similar to probability), and (2) number of flashes per three-hours. Several predictors from convective parameterization in CMC's GEM regional NWP model are included, plus important environment predictors. Calculations are on a moveable 9*9 grid centered on each grid point at four times in each three-hour diurnal period (t, t+1, t+2, and t+3 hours). Derived predictors are calculated as statistics for the 324 data points, e.g. the minimum Showalter index; the fraction of points with upward convection velocity greater than 20 m/s; mean convective rain rate where convection is activated. Data reduction keeps the number of predictors to less than 40. Tree-structured regression, a modern data-mining technique, is used to build models. Cross-validation shows the trees fit 80-90% of expected predictand variance. Trees have 300-700 nodes, allowing for quasi-continuous predictions across the whole domain. Prediction will be year-round and forecast coverage is extended to regions not included in the training data such as northern Canada, Alaska, and the southern United States. Use of the forecasts by Canadian forecasters is widespread for thunderstorm prediction in public forecasts, assessing areas where convection is likely to be initiated and the extent to which it will develop, and specifying convection boundaries in aviation area forecasts. There has been considerable interest from forestry groups for using the forecasts in 1-2 day fire likelihood predictions.