IONTB · FORECAST MODELS
Florida, Gulf & Atlantic outlook
Compare model forecasts and explore how the pattern could evolve.
Loading model data…
Tap or point at the map to inspect a value.
Sea-level pressure is measured in hectopascals (hPa). Lower pressure can indicate unsettled weather; pressure alone does not predict local impacts.
How to read these maps
L / H labels: selected local pressure minima and maxima, with the sampled model pressure in hPa. Labels are recalculated at each forecast time; they do not identify or track a named storm. Broad or weak centers and centers near the edge may not be labeled. Ensemble labels describe the average field, not an individual storm’s central pressure.
GFS also offers simulated composite radar reflectivity in dBZ: a model forecast at each displayed time, not observed or super-resolution radar. Values below 10 dBZ are left unshaded. These maps show sea-level pressure and six-hour precipitation totals in inches for the next five days in six-hour steps. Each single-forecast model includes forecast surface wind speed at 10 meters (about 33 feet), in mph, with direction arrows. Ensemble wind maps are not included. This is not gust speed. These maps do not show storm tracks. Rainfall totals include frozen precipitation as liquid equivalent; no precipitation-type classification is provided.
Single forecast: one model solution. Ensemble: the average of many forecasts. The average can smooth out individual storms. Spread: standard deviation of pressure across members; larger values show greater disagreement, not the probability of a storm.
GFS, ECMWF IFS/AIFS and Canadian GDPS fields are sampled to a 0.5° display grid using the nearest source grid point. ICON uses the nearest cell on DWD’s native global grid, sampled onto the same 0.5° display grid. GEFS uses NOAA’s 0.5° ensemble mean and spread. ECMWF EPS and AIFS ENS each use all 50 perturbed members, excluding the control; IONTB calculates their pressure mean and population standard deviation and mean precipitation totals. ECMWF and ICON six-hour precipitation is calculated by subtracting successive run-total accumulations; NOAA and Canadian six-hour products are used directly. Display colors are interpolated; these are regional model fields, not street-level forecasts.
Runs may become available at different times. The run timestamp always identifies the data on the map. Switching models keeps the closest available forecast time. We check shortly after each provider is expected to release a complete five-day forecast. Most models offer four runs daily; Canadian GDPS offers two. Delayed releases are checked again, and the previous complete forecast remains visible until new data passes validation.