Supported Weather Models

Six surface fields for temperature, humidity, pressure and wind. Four daily cycles, with 00 and 12 UTC forecasts through seven days.

Statistically post-processed blend that calibrates many guidance sources into one reliable forecast. This entry is the CONUS core grid. Strong baseline for threshold decisions in energy, emergency management and public works.

NBM core grid for Alaska. Same calibrated blend as CONUS, exposed as a separate dataset code with its own native grid.

NBM core grid for Hawaii. Ideal for island-scale forecasts without reprojecting CONUS output or clipping global grids.

NBM core grid for Puerto Rico. Use for tropical weather monitoring, energy load forecasting, and high-resolution island extraction.

NBM core grid for Guam. Great for tropical cyclone season workflows and high-resolution local extraction in the western Pacific.

NBM oceanic core grid with coarse resolution optimized for marine and wide-area workflows. Exposed as a separate dataset with its own native grid.

Quantile-mapped deterministic branch of the operational CONUS NBM. Useful when a workflow needs calibrated deterministic fields alongside the NBM core and probabilistic products.

Experimental public parallel feed for NBM v5 on the CONUS core grid. Only the last 5 days are retained for side-by-side validation against operational NBM.

Experimental Alaska core grid for the public NBM v5 parallel feed, with only the last 5 days retained for comparison workflows.

Experimental Hawaii core grid for the public NBM v5 parallel feed, with only the last 5 days retained for comparison against operational Hawaii guidance.

Experimental Puerto Rico core grid for the public NBM v5 parallel feed, with only the last 5 days retained for tropical and coastal validation against operations.

Experimental Guam core grid for the public NBM v5 parallel feed, with only the last 5 days retained for western Pacific comparison work.

Experimental Oceanic core grid for the public NBM v5 parallel feed, with only the last 5 days retained for marine validation against operational NBM Oceanic.

Experimental parallel QMD feed for NBM v5, with only the last 5 days retained for CONUS-side comparison against the operational QMD product.

Rapid, native-grid NOAA radar and multi-sensor analyses for composite reflectivity, MESH hail size, precipitation rate, one-hour Pass 2 QPE, and radar quality. Built for severe-weather monitoring, alert enrichment, event reconstruction, and quantitative verification.

The native 0.005° low-level azimuthal-shear field for tracking radar-derived storm rotation across CONUS. Its roughly 500 m grid and rapid updates support mesocyclone monitoring, rotation histories, alert context, and severe-storm research.

Current GOES-West record with 46 source-native fields from 19 full-disk ABI Level 2 families, spanning 16-band imagery and derived cloud, aerosol, fire, rainfall, radiation, land, and ocean observations.

Current GOES-West moisture, stability, cloud-layer, cloud-height, and land-temperature record with 15 source-native fields from five ABI Level 2 families on one fixed grid.

Current GOES-East record with 46 source-native fields from 19 full-disk ABI Level 2 families, spanning 16-band imagery and derived cloud, aerosol, fire, rainfall, radiation, land, and ocean observations.

Current GOES-East moisture, stability, cloud-layer, cloud-height, and land-temperature record with 15 source-native fields from five ABI Level 2 families on one fixed grid.

One hourly near-global mosaic combining operational geostationary satellites, with visible, shortwave infrared, water-vapor, and longwave infrared imagery.

NOAA's 1 km land-surface forecast with soil saturation, snow cover, snow depth, snow water equivalent, snow temperature, and accumulated evapotranspiration for hours 1 through 18.

Six time-lagged, GFS-forced members on NOAA's 1 km land grid, carrying snowpack, near-surface soil saturation, and accumulated evapotranspiration guidance into the 8.5-to-10-day range.

Objective analysis that fuses dense surface observations with short-range guidance to create a best-estimate grid. Designed as analysis of record for verification and ML training. Ideal truth set for bias correction and backfilling missing data.

Near real-time surface analysis built with variational data assimilation. Ingests METAR networks and other sensors to map current conditions at high cadence. Best suited for monitoring, rapid QA and initializing very short-range systems.

15-minute cadence near-real-time surface analysis on the 2.5 km CONUS NDFD grid. Designed for intrahour monitoring, ramp detection, operational QA, and high-frequency verification where hourly actuals are too coarse.

Climatologically calibrated one-hour precipitation analysis on the CONUS NDGD 2.5 km grid. Useful for QPF verification, hydrology, event reconstruction, and precipitation truth fields.

Global numerical weather prediction using the FV3 dynamical core and modern physics suites. Serves as the backbone for synoptic guidance and boundary conditions. Widely used for aviation planning, long-range scenarios and downstream ensembles.

Global data-assimilation analysis that underpins the GFS initial conditions. Useful for near-real-time “what happened” fields, verification, and short lead-time baselines. Complements GFS by focusing on the analysis cycle.

Hourly cycling short-range model tuned for mesoscale detail. Assimilates radar, aircraft and surface data to capture evolving boundaries and jets. A go-to for aviation operations and fast-changing weather in the next few hours.

NOAA aviation turbulence diagnostics for CONUS, including eddy dissipation, clear-air turbulence, mountain-wave turbulence, and convectively induced turbulence guidance on a 3 km grid.

CONUS in-flight icing probability, severity, and supercooled large droplet potential guidance for route-aware aviation weather applications.

Alaska in-flight icing probability, severity, and supercooled large droplet potential guidance on a 3 km polar stereographic grid.

Fifteen-minute CONUS aviation turbulence nowcast analyses from NOAA GTGN, exposing Eddy Dissipation Parameter values by altitude surface while the upstream feed is in NOMADS parallel.

DMI HARMONIE DINI surface subset for Northwestern Europe: solar radiation, 10 m and 100 m wind, cloud, temperature, humidity, pressure, visibility, and accumulated precipitation context on a 2 km regional grid.

NOAA/NCEP Air Quality Model guidance for ozone and PM2.5 over the contiguous United States, including hourly and 8-hour average products where published upstream.

Regional Alaska AQM ozone and PM2.5 guidance on the native polar stereographic domain for exposure, smoke-context, and local air-quality applications.

Hawaii AQM ozone and PM2.5 guidance on the native Mercator domain for island-scale environmental exposure and air-quality decision support.

Global five-day guidance for particulate matter, aerosol optical depth, total-column chemical species, and biologically effective UV dose. Built for air-quality context, dust and smoke transport, solar operations, and environmental risk.

One-hour thunderstorm probability guidance from SPC post-processed HREF output. Built for short-fuse lightning-aware operations, event safety, utilities, and aviation ground workflows.

Four-hour thunderstorm probability layer for planning windows where convective risk matters more than exact minute-by-minute initiation timing.

Lightning-density probability guidance for outdoor safety, venue operations, aviation ramps, and utility workflows where nearby lightning is the trigger.

Calibrated hail probability guidance for severe-weather dashboards, agriculture, insurance exposure, rooftop solar, and fleet protection workflows.

Calibrated tornado probability field for severe-weather situational awareness, emergency-management triage, and risk scoring ahead of warnings.

Calibrated damaging convective-wind probability guidance for utilities, transportation, event safety, emergency management, and logistics workflows.

Taiwan Central Weather Administration regional WRF guidance over East Asia and the western Pacific. A useful regional companion to global models for typhoon-region, maritime, aviation, energy, and logistics applications.

DWD's approximately 7 km deterministic forecast for Europe, with eight daily cycles, full runs through 120 hours, and 69 single-level fields for weather, energy, transport, agriculture, and regional model comparison.

DWD convection-permitting guidance for Germany and neighboring countries at approximately 2.2 km, with eight daily cycles, hourly lead times through 48 hours, and 63 fields for weather, energy, hydrology, transport, and regional model comparison.

Météo-France high-resolution deterministic guidance for metropolitan France and nearby countries, with eight daily cycles, hourly lead times through 51 hours, and surface fields for local weather and convective applications.

High-resolution ItaliaMeteo regional forecast guidance over Italy and nearby Alpine, Adriatic, Tyrrhenian, and Mediterranean areas, with 2.2 km coverage and 72-hour deterministic cycles.

Rapid-update ItaliaMeteo guidance over the same 2.2 km Italy regional domain, with 3-hourly cycles for first-day dashboards, energy operations, transport, and short-range monitoring.

12 km North American Mesoscale (NAM) CONUS guidance covering the gap between global flow and storm-scale refresh models. Useful when GFS is too coarse and HRRR is too local or too short, especially for regional temperature, wind, moisture, and precipitation structure.

12 km North American Mesoscale (NAM) CONUS slice on the same native grid as namawphys, but with a much richer surface, soil, column, radiative, and canopy diagnostic set. Best companion when you want more than the upper-air focused awphys stream.

3 km convection-allowing North American Mesoscale (NAM) nest for CONUS. This is the highest-resolution fixed-domain NAM stream and the closest physics-based counterpart to HRRR when you want a longer mesoscale horizon out to 60 hours.

GOES-like simulated brightness temperatures on the 12 km North American Mesoscale (NAM) CONUS grid. Specialized, but useful for cloud-top, moisture plume, and satellite-analogue workflows without leaving the NAM family.

Storm-resolving, convection-allowing configuration with hourly radar-informed cycling. Captures squall lines, outflow boundaries and sharp wind ramps. Favored for severe timing, road weather and renewable ramp forecasting.

High-frequency HRRR surface output at 15-minute valid times. Built for intraday ramps: wind and solar power, aviation winds/ceilings, and fast-evolving convection signals in the next few hours.

CONUS RRFS pressure-level fields for storm-scale upper-air analysis: temperature, wind, humidity, geopotential height, vertical motion, vorticity, and related isobaric-level forecast variables.

CONUS RRFS 2D fields for surface weather, near-surface wind and temperature, precipitation, clouds, radiation, soil state, fire-weather signals, and severe-storm diagnostics.

Historical experimental RRFS individual-member forecasts. This dataset is deprecated and receives no new forecasts; NOAA's replacement feed publishes separate derived REFS products.

Operational deterministic AI forecast from DWD on the approximately 13 km native ICON triangular grid, with surface weather and 13 atmospheric model levels at 3-hour steps.

Operational NOAA/NCEP GraphCast-based surface forecast fields available from the NOAA EAGLE AWS Open Data bucket. Covers 10 m wind, 2 m temperature, mean sea-level pressure, and precipitation on a 0.25 degree global grid.

Operational NOAA/NCEP GraphCast-based pressure-level fields available from the NOAA EAGLE AWS Open Data bucket. Provides height, temperature, humidity, wind, and vertical velocity at standard pressure levels.

NOAA/NCEP 31-member GraphCast-based AI ensemble surface fields from the NOAA EAGLE AWS Open Data bucket. Provides member-level wind, temperature, humidity, pressure, and precipitation.

NOAA/NCEP 31-member GraphCast-based AI ensemble pressure-level fields from the NOAA EAGLE AWS Open Data bucket. Provides member-level height, temperature, humidity, wind, and vertical velocity.

Historical GraphCastGFS archive for AI-vs-physics comparison and backtesting. GribStream currently offers this experimental NOAA/Google feed through the 2026-05-05 00 UTC run; use AIGFS or AIGEFS for current NOAA AI guidance.

Historical FourCastNetGFS archive for AI-model comparison and backtesting. GribStream currently offers this experimental NOAA/NVIDIA feed through the 2026-03-01 12 UTC run; use AIGFS or AIGEFS for current NOAA AI guidance.

CIRA and NOAA GSL Pangu-Weather forecasts initialized from GFS analyses, with nine variables spanning the surface and 13 pressure levels on a complete 0.25 degree global grid through 10 days.

Pangu-Weather forecasts initialized from ECMWF IFS analyses, with 69 atmospheric fields spanning the surface and 13 pressure levels through 10 days.

FourCastNet v2-small forecasts initialized from GFS analyses, with 73 atmospheric fields including 100 metre wind and total-column water vapour.

The same FourCastNet v2-small architecture initialized from IFS analyses, enabling direct study of sensitivity to a different starting analysis.

GraphCast forecasts initialized from GFS analyses, with 83 atmospheric fields including vertical velocity and 6-hour accumulated precipitation.

GraphCast initialized from IFS analyses, with 83 selectors including vertical velocity and 6-hour accumulated precipitation. Atmospheric fields begin at hour 6 and precipitation at hour 12.

Couples meteorology with aerosol composition to simulate dust, smoke and other species. Uses emissions and chemistry modules to predict transport and deposition. Supports visibility planning, health risk and dispersion studies.

Global ultraviolet exposure guidance for public-health, recreation, tourism, outdoor labor, school sports, and solar-exposure analytics.

Probabilistic wave guidance derived from an ensemble of wave simulations. Provides exceedance chances for heights and periods along routes and coasts. Used for offshore safety windows and marine risk management.

Ensemble mean that filters member noise to reveal persistent signals. Highlights pattern shifts and teleconnections for planning. A practical starting point before drilling into percentiles or clusters.

Set of perturbed atmospheric members that express forecast spread. Uses stochastic physics and initial-state perturbations to sample uncertainty. Essential for percentiles, clusters and scenario risk analysis.

Seasonal shortwave and longwave flux fields from a coupled ocean-atmosphere system. Useful for energy budget studies, agriculture and hydrology at multi-month horizons. Connects surface budgets to ENSO and MJO variability.

Seasonal pressure-level fields from the atmospheric component of CFS. Good for anomaly composites, stratospheric diagnostics and teleconnection studies. Supports aviation climatologies and hemispheric pattern analysis.

Fields on isentropic surfaces suited to transport and potential vorticity analyses. Reveals cross-tropopause exchange and jet dynamics on seasonal scales. Valuable for research, education and specialty diagnostics.

Ocean component of CFS providing temperature, currents and heat content. Captures basin-scale variability and subsurface structure. Applied to ENSO tracking and seasonal marine planning.

Climate-oriented global assimilation fields on a 1° grid. Useful for retrospective baselines, anomaly studies and large-scale machine-learning inputs. Complements CFS by providing analysis-first context.

Deterministic forecast from ECMWF with advanced 4D-Var data assimilation. High skill on synoptic and planetary scales supports many industries. Often used as the anchor run in multi-model decision workflows.

Deterministic wave model tightly coupled to the atmosphere. Resolves swell propagation and locally forced seas for ports and offshore assets. Supports routing, safety windows and coastal hazard outlooks.

ECMWF ensemble prediction system for the medium range. Perturbs initial conditions and model tendencies to sample uncertainty. Trusted for probabilities, percentiles and cluster analysis.

Wave ensemble paired with the atmospheric EPS to quantify marine uncertainty. Delivers probabilities for crossing safe limits along routes and coasts. Used for risk envelopes, port logistics and offshore planning.

ECMWF’s AI forecast trained to reproduce IFS behavior at a fraction of the cost. Enables rapid cycles and wider what-if exploration without heavy compute. Complements physics models for medium-range tasks and decision support.

Ensemble of AIFS runs that provides fast probabilistic guidance. Good for low-cost spread estimates and sensitivity testing. Pairs well with traditional ensembles in hybrid workflows.

Hourly ECMWF reanalysis fields served as valid-time actuals. Useful for verification, machine-learning labels, backtesting, bias correction, and stitching observed conditions against IFS and AIFS forecasts.