NHC Track and Intensity Models
Updated 1 June 2026
The term "forecast model" refers to any objective tool used to generate predictions of future events, such as the state of the atmosphere. The National Hurricane Center (NHC) uses many models as guidance in the preparation of official track and intensity forecasts. The most commonly used models at NHC are summarized in the tables below.
Forecast models vary tremendously in structure and complexity. They can be simple enough to run in a few seconds on an ordinary computer, or complex enough to require a number of hours on a supercomputer.
- Dynamical models, also known as numerical models, are the most complex and use high-speed computers to solve the physical equations of motion governing the atmosphere.
- Statistical models, in contrast, do not explicitly account for atmospheric physics; instead, they are based on historical relationships between storm behavior and storm-specific details such as location and date.
- Statistical-dynamical models combine dynamical and statistical techniques by forecasting storm behavior using established historical relationships between storm behavior and atmospheric variables provided by dynamical models.
- Trajectory models move a tropical cyclone (TC) along based on the prevailing flow obtained from a separate dynamical model (either the GFS or the ECMWF).
- Artificial Intelligence/Machine Learning (AI/ML) models are trained on reanalysis outputs from dynamical models, as well as other datasets, to reproduce predicted atmospheric conditions and minimize errors in the training dataset.
- Ensemble or consensus models are created by combining forecasts from multiple models/perturbations.
Table 1. Summary of global and regional dynamical models for track, intensity, and wind radii.
| ATCF ID | Global/Regional Model Name |
Horizontal Resolution |
Vertical Levels and Coordinates |
Data Assimilation |
Convective Scheme | Cycle/Run Frequency | NHC Forecast Parameter(s) |
|---|---|---|---|---|---|---|---|
| AVNO/AVNI GFSO/GFSI |
Global Forecast System (FV3-GFS) | Finite Volume Cube Sphere (~13km) | 64 Hybrid Sigma-pressure | GSI/4D-VAR EnKF hybrid | Simplified Arakawa Schubert | 6 hr (240 hr) 00/06/12/18 UTC |
Track, intensity, wind radii |
| *EMX/EMXI/EMX2 | European Centre for Medium-Range Weather Forecasts | Spectral (~9km) | 137 Hybrid Sigma-pressure | 4D-VAR | Tiedke mass flux | 12 hr (240 hr) 00/12 UTC |
Track, intensity, wind radii |
| *UKM/UKMI/UKM2 | U.K. Met Office Global Model | Grid point (~10 km) | 70 Hybrid Sigma-pressure | 4D-VAR Ensemble Hybrid | UKMET | 12 hr (144 hr) 00/12 UTC |
Track, intensity, wind radii |
| CMC/CMCI/CMC2 | Canadian Deterministic Prediction System | Grid point (~15 km) | 80 Hybrid Sigma-pressure | 4D-VAR Ensemble Hybrid | Kain-Fritsch | 12 hr (240 hr) 00/12 UTC |
Track, intensity, wind radii |
| NVGM/NVGI | Navy Global Environmental Model | Spectral (~31km) | 60 Hybrid Sigma-pressure | NAVDAS-AR 4D-VAR | Simplified Arakawa Schubert | 6 hr (144 hr) 00/06/12/18 UTC |
Track, intensity, wind radii |
| HFSA/HFAI | Hurricane Analysis and Forecast System Version A | Nested Grid point (5.4-1.8 km) | 81 Hybrid Sigma-pressure | 4D-Ensemble VAR | Simplified Arakawa Schubert + GFS shallow convection (5.4km) 1.8km nest - none | 6 hr (126 hr) 00/06/12/18 UTC Runs on request from NHC/JTWC |
Track, intensity, wind radii |
| HFSB/HFBI | Hurricane Analysis and Forecast System Version B | Nested Grid point (6-2 km) | 81 Hybrid Sigma-pressure | 4D-Ensemble VAR | Simplified Arakawa Schubert + GFS shallow convection (6km) 2km nest - none | 6 hr (126 hr) 00/06/12/18 UTC Runs on request from NHC/JTWC |
Track, intensity, wind radii |
| HWRF/HWFI | Hurricane Weather Research and Forecast System | Nested Grid point (13.5-4.5-1.5km) | 75 Hybrid Sigma-pressure | 4D-VAR Hybrid GDAS GFS IC/BC | Simplified Arakawa Schubert + GFS shallow convection (6 and 18km) 1.5km nest - none | 6 hr (126 hr) 00/06/12/18 UTC Runs on request from NHC/JTWC |
Track, intensity, wind radii |
| CTCX/CTCI | NRL COAMPS-TC w/ GFS initial and boundary conditions | Nested Grid point (36-12--4 km) | 40 Hybrid Sigma-pressure | 3D-VAR (NAVDAS) Alternate DA - EnKF DART | Deep convective - Kain-Fritsch Shallow convective scheme - Tiedke |
6 hr (126 hr) 00/06/12/18 UTC Runs commence on 1st NHC/JTWC advisory |
Track, intensity, wind radii |
| HMON/HMNI | Hurricane Multi-scale Ocean-coupled Non-hydrostatic model | Nested Grid point (18-6-2km) | 51 Hybrid Sigma-pressure | GFS IC/BC | Simplified Arakawa Schubert + GFS shallow convection (6 and 18km) 2km nest - none | 6 hr (126 hr) 00/06/12/18 UTC Runs on request from NHC/JTWC |
Track, intensity, wind radii |
* Public Access to these models is restricted due to agreements with the data provider.
Table 2. Summary of ensembles and consensus aids for track and intensity.
| ATCF ID | Model Name or Type | Horizontal Resolution | Vertical Levels and Coordinates |
Data Assimilation |
Perturbation or Consensus Methods | Cycle/Run Frequency | Ensemble Members | NHC Forecast Parameter(s) |
|---|---|---|---|---|---|---|---|---|
| AEMN/AEMI | Global Ensemble Forecast System | 25 km | 64 Hybrid Sigma-pressure | GSI/3D-VAR EnKF hybrid | 20 of 80 6 hr DA system hybrid EnKF members per cycle | 6 hr (384 hr) 00/06/12/18 UTC |
31 | Track and Intensity |
| *UEMN/UEMI | U.K. Met Office MOGREPS | ~20 km | 70 Hybrid Sigma-pressure | 4D-VAR EnKF hybrid | 44-member EnKF | 12 hr (168 hr) 00/12 UTC |
17 | Track and Intensity |
| *EEMN/EMNI | ECMWF Ensemble Prediction System (EPS) | 9 km | 91 Hybrid Sigma-pressure | 4D-VAR | Stochastically perturbed parameterizations | 12 hr (360 hr) 00/12 UTC |
50 | Track and Intensity |
| CEMN/CEMI | Canadian Ensemble Mean | 25 km | 84 | 4D-VAR Ensemble Hybrid | Stochastic perturbations | 12 hr (240 hr) 00/12 UTC | 32 | Track and Intensity |
| *FSSE | Florida State Super Ensemble | Corrected consensus | 6 hr (120 hr) 00/06/12/18 UTC |
Track and Intensity | ||||
| HCCA | HFIP Corrected Consensus Approach | Corrected consensus | 6 hr (120 hr) 00/06/12/18 UTC |
AEMI, GFSI, CTCI, DSHP, EGRI, EMN2, EMXI, HFAI, HFBI, HWFI, LGEM | Track and Intensity | |||
| *GFEX | 2 model consensus | Simple consensus | 6 hr (120 hr) 00/06/12/18 UTC |
AVNI, EMXI | Track | |||
| TVCN (Atlantic / TVCA) (Pacific / TVCE) |
Variable consensus | Simple consensus, minimum 2 members | 6 hr (168 hr) 00/06/12/18 UTC |
GDMI, AEMI, EMXI, UKMI, HFAI, CTCI | Track | |||
| RVCN | Wind Radii Consensus | Multi-model wind radii, bias-corrected initial wind | 6 hr (168 hr) 00/06/12/18 UTC |
GDMI, AVNI, HWFI, EMXI, CTCI, HFAI, HFBI | Wind radii | |||
| ICON | Intensity consensus | Simple consensus, all 4 must be present | 6 hr (168 hr) 00/06/12/18 UTC |
DSHP, LGEM, HFAI, HFBI, HWFI, HMNI | Intensity | |||
| IVCN | Intensity variable consensus | Simple consensus, minimum 2 members | 6 hr (168 hr) 00/06/12/18 UTC |
GDMI, DSHP, LGEM, HFAI, HFBI, CTCI, HWFI, HMNI | Intensity | |||
| IVDR | Weighted Consensus | Simple consensus, minimum 2 members, double-weighted HFAI, HFBI, HWFI, HMNI, and CTCI |
6 hr (168 hr) 00/06/12/18 UTC |
HFAI, HFBI, HWFI, HMNI, CTCI, GFSI, DSHP, LGEM | Intensity | |||
| NNIC | Neural Network Intensity Consensus | Neural Network with input models and 5 TC predictors | 6 hr (168 hr) 00/06/12/18 UTC |
DSHP, LGEM, HWRF, GFSI | Intensity |
* Public Access to these models is restricted due to agreements with the data provider.
Table 3. Summary of Artificial Intelligence/Machine Learning Models for track and intensity
| ATCF ID | Model Name or Type | Horizontal Resolution | Vertical Levels and Coordinates |
Training Dataset | Perturbation or Consensus Methods | Cycle/Run Frequency | Ensemble Members | NHC Forecast Parameter(s) |
|---|---|---|---|---|---|---|---|---|
| GAIO / GAII | AI-GFS deterministic |
28 km | 13 Isobaric levels | GFS and ERA5 analyses of historical TCs | 6 hr (168 hr) 00/06/12/18 UTC |
Track and Intensity | ||
| EGMN / EGMI | AI-GEFS Ensemble mean |
28 km | 13 Isobaric levels | GFS and ERA5 analyses of historical TCs | Mean of 30 ensemble members | 6 hr (168 hr) 00/06/12/18 UTC |
31 | Track and Intensity |
| *EAIO / EAII | AIFS- ECMWF deterministic |
28 km | 14 Isobaric levels | ERA5 analyses fine-tuned on IFS | 6 hr (168 hr) 00/06/12/18 UTC |
Track and Intensity | ||
| *EAMN / EAMI | AI-ECMWF ensemble | 28 km | 14 Isobaric levels | ERA5 analyses fine-tuned on IFS | Mean of 50 ensemble members | 6 hr (168 hr) 00/06/12/18 UTC |
50 | Track and intensity |
| GDMN / GDMI | Google DeepMind Ensemble Mean | 28 km | ERA5/IBTrACS analyses of historical TCs (1979-2018) fine-tuned on HRES-fc0 |
Mean of 50 members of the DeepMind ensemble | 6 hr (15 days) 00/06/12/18 UTC |
50 | Track, intensity, wind radii | |
| GENC / GENI | Google GenCast ensemble mean | 28 km | ERA5 reanalyses of historical TCs (1979-2018) | 6 hr (168 hr) 00/06/12/18 UTC |
50 | Track and Intensity | ||
| GRPH / GRPI | Google GraphCast deterministic | 28 km | ERA5 reanalyses of historical TCs (1979-2018) | 6 hr (168 hr) 00/06/12/18 UTC |
Track and Intensity |
* Public Access to these models is restricted due to agreements with the data provider.
Table 4. Summary of statistical models for track, intensity, and wind radii.
| ATCF ID | Model Name or Type | Comments | Prediction Methodology | Cycle/Run Frequency | NHC Forecast Paramter(s) |
|---|---|---|---|---|---|
| CLP5 (OCD5) | CLIPER5 Climatology and Persistence | Used to measure skill in a set of track forecasts | Multiple regression technique. Inputs are current and past TC motion (previous 12-24 hr), forward motion, date, latitude/longitude, and initial intensity | 6 hr (168 hr) 00/06/12/18 UTC |
Track |
| SHF5/DSF5 (OCD5) | Decay-SHIFOR5 Statistical Hurricane Intensity Forecast | Used to measure skill in a set of intensity forecasts, includes land decay rate component | Multiple regression technique using climatology and persistence predictors | 6 hr (168 hr) 00/06/12/18 UTC |
Intensity |
| TCLP | Trajectory-CLIPER | Used to measure skill in a set of track or intensity forecasts | Substitute for CLIPER and SHIFOR; similar predictors, but uses trajectories based on reanalysis fields instead of linear regression | 6 hr (168 hr) 00/06/12/18 UTC |
Track and intensity |
| DRCL | Wind Radii CLIPER | Statistical parametric vortex model | Employs climatology with the parameters determined from 13 coefficients and persistence to produce 34-kt, 50-kt, 64-kt wind radii estimates | 6 hr (168 hr) 00/06/12/18 UTC |
Wind radii |
| SHIP | Statistical Hurricane Intensity Prediction Scheme | Statistical-dynamical model based on standard multiple regression techniques | Climatology, persistence, environmental atmosphere parameters, and an ocean component | 6 hr (168 hr) 00/06/12/18 UTC |
Intensity |
| DSHP | Decay-Statistical Hurricane Intensity Prediction Scheme | Statistical-dynamical model based on standard multiple regression techniques | Climatology, persistence, environmental atmosphere parameters, oceanic input, and an inland decay component | 6 hr (168 hr) 00/06/12/18 UTC |
Intensity |
| LGEM | Logistic Growth Equation Model | Statistical intensity model based on a simplified dynamical prediction framework | A subset of SHIPS predictors, ocean heat content, and variability of the environment used to determine growth rate and wind coefficient | 6 hr (168 hr) 00/06/12/18 UTC |
Intensity |
| *SHPE | Statistical Hurricane Intensity Prediction Scheme (ECMWF version) | Statistical-dynamical model based on standard multiple regression techniques | Climatology, persistence, environmental atmosphere parameters, and an ocean component | 6 hr (168 hr) 00/06/12/18 UTC |
Intensity |
| *DSPE | Decay-Statistical Hurricane Intensity Prediction Scheme (ECMWF version) | Statistical-dynamical model based on standard multiple regression techniques | Climatology, persistence, environmental atmosphere parameters, oceanic input, and an inland decay component | 6 hr (168 hr) 00/06/12/18 UTC |
Intensity |
| *LGME | Logistic Growth Equation Model (ECMWF version) | Statistical intensity model based on a simplified dynamical prediction framework | A subset of SHIPS predictors, ocean heat content, and variability of the environment used to determine growth rate and wind coefficient | 6 hr (168 hr) 00/06/12/18 UTC |
Intensity |
* Public Access to these models is restricted due to agreements with the data provider.
Table 5. Summary of trajectory models for track.
| ATCF ID | Model Name or Type | Vertical Levels and Coordinates |
Dynamical Model Field | Cycle/Run Frequency | NHC Forecast Parameter(s) |
|---|---|---|---|---|---|
| TABS | Shallow Layer Trajectory and Beta Advection | 850-700 mb | GFS | 6 hr (168 hr) 00/06/12/18 UTC |
Track |
| TABM | Medium Layer Trajectory and Beta Advection | 850-400 mb | GFS | 6 hr (168 hr) 00/06/12/18 UTC |
Track |
| TABD | Deep Layer Trajectory and Beta Advection | 850-200 mb | GFS | 6 hr (168 hr) 00/06/12/18 UTC |
Track |
| *TBSE | Shallow Layer Trajectory and Beta Advection | 850-700 mb | ECMWF | 6 hr (168 hr) 00/06/12/18 UTC |
Track |
| *TBME | Medium Layer Trajectory and Beta Advection | 850-400 mb | ECMWF | 6 hr (168 hr) 00/06/12/18 UTC |
Track |
| *TBDE | Deep Layer Trajectory and Beta Advection | 850-200 mb | ECMWF | 6 hr (168 hr) 00/06/12/18 UTC |
Track |
* Public Access to these models is restricted due to agreements with the data provider.
Early versus Late Models
Numerous objective forecast aids (guidance models) are available to help the NHC Hurricane Specialists prepare their official track and intensity forecasts. Guidance models are classified as early or late, depending on whether they are available to the Hurricane Specialist during the forecast cycle. For example, consider the 1200 UTC (12Z) forecast cycle, which begins at 12Z synoptic time and concludes with the release of the official forecast at 15Z. The 12Z run of the NWS/Global Forecast System (GFS) model is not complete and available to the forecaster until about 16Z, or about an hour after the forecast is released – thus the 12Z GFS would be considered a late model since it could not be used to prepare the 12Z official forecast.
Multi-layer dynamical models are generally, if not always, late models. Fortunately, a technique can be used to take the latest available run of a late model and adjust its forecast to apply to the current synoptic time and initial conditions. In the example above, forecast data for hours 6-126 from the previous (06Z) run of the GFS would be adjusted, or shifted, so that the 6-h forecast (valid at 12Z) would exactly match the observed 12Z position and intensity of the tropical cyclone. The adjustment process produces an "early" version of the GFS model for the 12Z forecast cycle, based on the most current available guidance. The adjusted versions of the late models are known, for historical reasons, as interpolated models.
Interpreting Forecast Models
NHC provides detailed information on the verification of its past forecasts through an annual verification report (https://www.nhc.noaa.gov/verification/verify3.shtml). On average, NHC official forecasts have smaller errors than any individual model. An NHC forecast reflects consideration of all available model guidance as well as forecaster experience. Therefore, users should consult the official forecast products issued by the NHC and local National Weather Service Forecast Offices rather than relying solely on forecast model output. Users should also be aware that uncertainty is inherent in every forecast, and that proper interpretation of the NHC forecast must account for it. NHC forecasters typically discuss forecast uncertainty in the Tropical Cyclone Discussion (TCD) product. NHC also prepares probabilistic forecasts that incorporate forecast uncertainty information (https://www.nhc.noaa.gov/aboutnhcprobs.shtml).
NOAA/NWS Models
The National Weather Service produces some of the models used by the National Hurricane Center. These models are run by NOAA/NWS National Centers for Environmental Prediction (NCEP) Central Operations (NCO). Output images from the NOAA/NWS models can be found through NCEP's Model Analyses and Guidance (MAG) interface. Raw data from the models can be found through the NOAA Operational Model Archive and Distribution System (NOMADS).
Other model background information
Read about the Inland Wind Model and the Maximum Envelope Of Winds
NHC on Facebook
NHC on X
NHC on YouTube
NHC Blog: