Advanced Wind Estimation (AWE)

This function should be used with extreme care as it can lower the estimated wind speeds significantly.  While we believe it produces more accurate results, it is highly dependent on accurate Land Cover Database information or correct user input. Like any good estimation calculation, it may in some cases, underestimate the actual winds experienced. If you always want to overestimate wind speeds, do not use this feature. See considerations section at the end of this topic.

The National Hurricane Center's estimated maximum winds and wind field radii issued in their advisories are valid, generally, for open sea exposures only.  They cannot factor local wind characteristics for inland areas.  This is where Advanced Location Wind Estimation "comes in". 

Advanced location wind estimation (AWE) is a function of HURRTRAK that applies a wind speed "correction factor" to the forecast/estimated wind speeds based on the local “site” characteristics.  These include distance from open (ocean) exposures, types and density of vegetation and man made structures, and elevation being analyzed. Established wind dynamics formulae are applied for the location... adjusting the estimated wind speeds (sustained and gusts) to account for frictional, exposure and height considerations of the location.

There are 2 methods of employing this capability.

Site/Location characteristics - With this option the user specifies the characteristics of a specific location. This is useful when creating impact reports.

NLCD (National Land Cover Data) - This option allows the system to determine the characteristics of locations. zip codes and county area. This method is only available for US mainland locations.

All graphics that use AWE, such as wind pattern and wind arrows, utilize the NLCD information while reports can use either. If there is no user entered information for a report location, the system will use the NLCD data.

If you require the most accurate wind speed estimate reports possible, we recommend that you use the site/location method and enter the roughness index information for each of your locations. This can be a time consuming exercise as it requires detailed knowledge of a locations surroundings and the ability to translate this into roughness index values.

We will describe each method below.

NLCD Method :

The NLCD Enhanced AWE has 3 important changes over the original Site/Location characteristics method.

The NLCD database (see image below) has land usage information on every state.  An example of the land usage graphic for part of S. Florida is shown below.  It shows the urban land usage of the Florida east coast as well as the Everglades and Crop areas.  With his information we can estimate the "roughness index" (friction) for any location by direction of wind.  i.e. an East wind into Palm Beach has a different "roughness index" than a west wind.

 

The system then considers these factors to adjust the “raw” derived wind speed (as determined by the NHC forecast advisory) for this site.  Typically the adjusted wind speeds are considerable lower than the “raw” values.  This function affects the forecast and actual location wind impact reports as well as the summary report.

This type of information allows us to create graphics such as the wind pattern shown below.

Site/Location Characteristics Method:

For the most accurate wind estimation reports possible and for areas outside of the US, the user can enter site/location characteristics within the Location Maintenance portion of the System Setup function.

An example of the advanced location data entry screen is shown below.


In this example, we have defined a location, Bay Harbor Island, FL. Based on the site characteristics, we have entered the average roughness indices, for 16 compass directions. We used a distance of about 2 KM “upstream” for each direction to help determine the average RI. This type of analysis requires a good knowledge of your sites and may require some experimentation as real data is analyzed. i.e., if the system is always underestimating easterly winds, your RI values may be too high in that direction.

Many users use aerial images and local knowledge to help determine the best RI values. The image below is for an airport location as shown on Google Earth. In order to determine a point location’s surrounding RI values, these kinds of tools are invaluable.

If your requirements are not quite as strict, the NLCD method does provide good results and you do not need to do this type of site analysis.

 

Advanced Wind Estimations affect on reports:

When creating a wind impact location report, the system will utilize the user entered information, if available.
If no information was entered, it will use the NCLD database data to estimate the friction (RI) values instead.

The following is an impact summary report for Hurricane Wilma in 2005. All of the locations used the NLCD database information.

The Max Wind AWE column represents the maximum sustained winds using Advanced Wind Estimation; the next column shows the “uncorrected” wind speeds and the next column shows the max gusts using AWE. As you can see, the impact of AWE on the report values is significant. Sustained wind speeds along the SE coast of Florida, where the land usage is most dense, have been decreased about 40-50% while the wind speed adjustment in more exposed areas, is much less.

One interesting location on this report is Pahokee. It is located on the SE shore of Lake Okeechobee and thus is more exposed to a NW wind off of the lake. The detail hourly report (below) verifies that indeed their highest winds were likely on the back side of the storm when the winds switched to the NW.

 
Important Advanced Wind Estimation Considerations:

We are pleased with the performance of Advanced Wind Estimation (AWE) but we feel that is important to mention a few things about the accuracy of wind estimation in general and advanced wind estimation specifically.

In the documentation and helps we try to communicate that AWE does a good job of estimating sustained winds in Cat 3 and below storms and also outside of the eye wall. The eye wall is an extremely turbulent region where violent up and down motions of air will bring very strong winds aloft down to the surface.. thus reducing the "normal" effect of friction on wind speed.

In Cat 4 and 5 storms like Charley and Katrina, we suggest that you do not rely solely on AWE for estimating winds in the eye wall of this hurricane. We cannot afford to underestimate the winds in these violent storms near the center.

To understand this better, generally, a perfect site specific wind forecast (location impact report) depends on several things to occur.

1. A perfect forecast of storm track, storm strength and 34, 50 and 64 wind radii.

2. A homogeneous wind field around the storm. i.e. no eyewall vortices, local effects and squally areas of higher winds, etc.

3. For AWE, a perfectly analyzed location with exact roughness index figures in all quadrants.

Let’s address each one.

1. It is obvious we do not have item #1.. While NHC forecast are very good, especially 12 and 24 hours in advance, the average error is still in the 80-90nm range at 24 hours. That aside, forecasting the exact storm strength and wind radii has proven to be nearly impossible.

2. As everyone who experienced hurricanes knows, a homogeneous storm does not exist in nature. Hurricane Frances, for example, caused much higher winds in NE Florida than was expected due to a detached persistent area of squalls that prevailed in this region even though the storm center was near the Florida West Coast. The region inside the eyewall can be a violent area with vortices and other local effects that no software can predict.

3. While we have a good feel for the type of land usage around a location, without a detailed site by site analysis, it is not exact. AWE uses this information to adjust the raw extrapolated winds downwards to a more realistic level.

Due to these 3 factors, Hurrtrak (and any other software) is not able to provide EXACT wind speed forecast.. thus the use of the word estimate. While AWE is a huge improvement, it is not a panacea.

Specific to AWE, because it lowers the location’s wind speed to a more realistic level, It is important to understand that it WILL sometimes UNDERESTIMATE wind speed forecast for a location (as in a location impact report) something that the standard method rarely did. This may be suitable for some customers while other customers might want to make decisions based on the higher end forecast generated from the standard (non-AWE) method. Remember the wind impact reports estimate SUSTAINED (2 minute average) wind speeds at an observation point 10 meters above ground level. Wind gusts can be as much higher and average about 30-50% greater than the average wind speed.

Finally, If you choose to use AWE, I would highly recommend that you do not use AWE for the county or zip code report. This is because AWE is site specific and a zip code/county area average is not representative of a single point. Imagine a county that is 5% coastline and 95% pine forest. The pine forecast friction factors will dominate bringing the calculation of the wind level down to an unrealistic level for the coastal location. The only place you can use AWE on the zip and county report with some confidence is with areas that are entirely inland and uniform in nature.