Ryan Kramer, a 2012 Hollings student at WFO Charleston, SC, studied wind gust factors at several land and marine sites to help us develop a more effective Smart Tool for day-to-day forecasting. Prior to his study our primary means for creating Wind Gust grids was the Wind_Gust_Changer Smart Tool, applying either a percentage or addition factor to Wind Speed. The tool defaulted to a 115% factor, meaning wind gusts equal 1.15 x wind speed at each grid box. Local forecasters were employing a variety of techniques, including adjusting multiplication factor differently between land/marine/lakes, analyzing mixing profiles, time of year, time of day, looking at BUFKIT soundings, etc. However most of these forecaster-specific methods were based on individual preference, experience, and gut-instincts, rather than science or climatology. Also, beyond the first several forecast periods, most forecasters would run the tool with the default 115% factor to save time. Therefore the goal was developing a local wind gust climatology that could be used to improve overall wind gust forecasts in the GFE gridded database.
Wind Gust Definitions
There are differences in the way Wind Gust is defined/calculated between NDFD and observation platforms that should be noted:
- NDFD Gridded Database: the maximum 3-second wind speed (in knots) forecast to occur within a 2-minute interval at a height of 10 meters.
- NDBC Buoys: 5-second peak gust during an 8-minute period at a height of 5 meters.
- C-MAN Station: 5-second peak gust during a 2-minute period at a height of 10 meters.
- ASOS: 3-second peak gust during a 2-minute period at a height of 10 meters (corrected 5/9/13).
- AWOS: 5-second peak gust during a 2-minute period at a height of 10 meters.
Although the NDFD definition for Wind Gust does not match that of the standard observation platforms, it was determined that gusts should be forecast as would be measured by NWS observation equipment.
Ryan analyzed archived surface observations from 2007 to 2011 for 14 land and 2 marine sites within the Charleston forecast area. For each location, observations with gust factors outside of 1 standard deviation from the mean were removed from the study to help ensure that only “fair-weather” scenarios, instead of convectively induced conditions, were being analyzed. Erroneous data was also removed. The surface data was then categorized as coastal or inland and marine or land; and observations containing wind gust reports were identified.
Over land, 13% of observations contain wind gusts, which occur mostly during peak daytime heating hours. At marine locations, 68% of observations contain wind gusts, which occur consistently at all hours. The frequency of wind gusts also varies seasonally.
Sustained wind speed and average wind gust trends correlate at over 95%, suggesting a strong relationship between these variables. Over land, gust factors are inversely related to sustained wind speeds. Gust factor magnitudes at land sites can also be affected by wind direction, and exhibit diurnal and seasonal variability. These trends are less apparent or nonexistent at marine sites. Furthermore, gust factor values at marine sites are generally lower than those found at land sites. Finally, it was found that observation sites directly along a shoreline (Folly Beach, SC and Fort Pulaski, GA) exhibited gust factors similar to offshore marine sites.
The observed trends, and differences between land and marine sites, support previous research findings that propose wind gusts are affected by surface roughness and are dependent on atmospheric boundary layer conditions. The strong correlation between sustained wind speeds and wind gusts suggested that accurate gust forecasts could be created by applying a varying percentage factor based on wind speed.
A more detailed summary of Ryan’s research can be found here.
Contact me (Jonathan Lamb) if you’d like a copy of the Wind Gust Smart Tool I developed that utilizes Ryan’s research.