Home   About Us   Contact   Links

StormWinds™ is a wind and pressure analysis method for historical hurricanes on a storm-following grid that can drive surge and wave models.   Wind vectors and pressures are produced at user-specified grid resolution and time intervals.  The data is produced in NWS 12 format for easy input into the ADCIRC model. Also available in the PBL model NWS 4 format, if desired. Typical spatial resolution is 6 km and temporal resolution is 15 minutes. 

Example StormWinds™ Analysis for Gustav.
Click for larger version (20 MB).

Gustav Winds

At the core of StormWinds™ are multiple sources of objective analysis.  These analyses integrate actual surface observations at sea, on land, from satellites, and Hurricane Hunter aircraft adjusting each source for observing height, exposure, and averaging time.  They also assign confidence levels based on past performance of different observation platforms and conditions at the time of the observation.  The wind values from these analyses are usually produced as one-minute sustained vector components at a 10 meter elevation and are usually available every three or six hours in near-realtime.

Several realistic features of objective analyses make them desirable to drive ocean models for tropical cyclones relative to applying modeled winds, alone.  As real world storms evolve, each of the following are observed in nearly every hurricane, typhoon, or cyclone: changing radii of maximum winds, wind asymmetries around the eyewall, and secondary eyewall maxima.

Despite these advantages of the analyses, direct use of them, alone, for hydrology modeling efforts has several deficiencies.  First, the spatial domain of the analysis grids is limited in area and typically only encompasses 500 km from the center of the storm, at most, which may not allow for effective modeling of a days-long event.  Second, most available surface analyses also only provide a temporal resolution of 3 or 6 hours, which may cause significant loss of wind magnitude in Eulerian temporal interpolation schemes. A third limitation of directly leveraging surface wind analyses is the lack of an accompanying sea level pressure. Finally, upon landfall sustained winds usually degrade, partly, due to drag as a function of land roughness, which depends on the specific type of terrain.  Objective wind analysis data may or may not include roughness-induced drag and, if so, tends to apply it using a less sophisticated algorithm and only at analysis spatial resolutions. All of these factors can profoundly influence modeled surge and waves.

StormWinds™ overcomes these deficiencies by

  • using mesoscale wind analyses from the National Weather Service, and other sources, to extend the tropical cyclone analysis grid and by providing pressures using a parametric model.  These multiple wind sources are blended into a single domain covering the storm’s track through its life cycle with weighting of the sources by location relative to the tropical cyclone’s center and by known biases. Final domain size is 2000 km by 2000 km.
  • performing temporal interpolation between the analysis and observation times, using a storm following grid to faithfully retain features present in the storm.  The features remain due to the winds being interpolated, in time, at consistent locations relative to center, thus having consistent direction; hence temporal interpolation does not result in the aforementioned loss of wind magnitude.
  • providing marine exposure winds to allow ADCIRC users to apply the nodal attribute directional effective roughness length (Westerink, et al., 2008) at spatial resolutions of their meshes, rather than limiting detail to wind input spatial resolutions or less sophisticated land roughness algorithms.
  • performing an analysis of wind parameters at every time step and builds an atmospheric pressure field matching, in time and space, that of the wind field. This is conducted by running a parametric model algorithm.

Example Pressure Analysis for Galveston's Pleasure Pier During Hurricane Ike in 2008.

StormWinds™ results are representative of a 10-minute average appropriate for surge and wave modeling.

Though typical wind analyses of late have been produced in 15 minute intervals, this is easily variable. Hourly snapshots of wind forcing may be appropriate for slow-moving cyclones; surge simulations of cyclones with a relatively higher translational velocity may benefit from shorter intervals. 5, 10, 15, 20, 30, and 60 minute interval input files for ADCIRC are available.

For more information, please contact Benjamin Jelley.



Westerink, J.J., R.A. Luettich , J.C. Feyen, J.H. Atkinson, , C. Dawson, H.J. Roberts, M.D. Powell, J.P. Dunion, E.J. Kubatko, H. Pourtaheri, “A Basin to Channel Scale Unstructured Grid Hurricane Storm Surge Model Applied to Southern Louisiana,” Monthly Weather Review, 136, 3, 833-864, 2008.