PWS-Severe_Weather_Tool_Upgrade-Ver_2b.pdf

PERFORMANCE WORK SATEMENT – 2b

FOR

UPGRADE SUMMER SEVERE WEATHER TOOL USED BY 45TH WEATHER SQUADRON

1. Title: Upgrade Summer Severe Weather Tool Used by 45th Weather Squadron (45 WS)

2. Definitions:

• East Central Florida - the following six counties: Brevard,Volusia, Seminole, Orange, Osceola, and Indian River.

• Summer – the months of May through September.

• Severe Thunderstorms - straight-line winds of 50 kt or greater, tornadoes of any intensity, or hail of 1 inch or greater. Tropical cyclones do not qualify as severe thunderstorms for this project.

3. Description: Upgrade the summer severe weather tool used by 45 WS. The current tool uses the 1500 UTC sounding from Cape Canaveral Air Force Station (CCAFS) and two user inputs (existence and relative location of 200-mb speed maximum, and lightning flow regime) to predict the probability of severe weather that day in east-central Florida. In this tool, summer is defined to be May-September and east-central Florida is Brevard and adjacent counties (Volusia, Seminole, Orange, Osceola, and Indian River). It was developed using data collected in the period 1989-2012. The upgrade will consist of the following actions:

a. The data from summer 1989-2012 from the original tool should be available for reuse in this project.

- Note: Including the data from the summers of 2013-2016 was originally considered, but the 1500 UTC soundings were not released often in 2013 and 2014 due to budget cuts and so would not be useful to this project. While the 1500 UTC soundings were usually released in summer 2015-2016, adding just the two additional years would not be cost-effective. Therefore the project is restricted to just the preexisting 1989-2012 data.

b. Redevelop the predictive equations using an optimized regression as opposed to the method in the current tool.

The new regression will likely be a non-linear regression. Logistic regression is often used for predicting probability and should be considered. Also, a modified logistic regression may be better, where the maximum probability approaches a value less than 100% while maintaining the sigmoid shape. In any event, the choice is left to the discretion of the researcher, but should be done in consultation with the 45 WS. The predictors should be selected to maximize independence between the variables and maximize the variance explained. A method similar to that used in the AMU lightning probability tool should be considered, where predictors were chosen in order of the most new additional variance explained, until some small threshold was met, e.g. ≤ 0.5% new variance, or other statistical tests indicated that no new predictors should be added. But as with the regression method, the choice is left to the discretion of the researcher, but should be done in consultation with 45 WS.

c. Compare the performance of the new tool to the current tool using a verification data set. This is needed to determine if the new tool performs better than the current tool and should be implemented into 45 WS operations.

The verification data to be excluded from the training data will be selected randomly based on a desired percentage of days (~20%) from the available data as opposed to randomly selected seasons. The performance metrics should include Probability Of Detection, Probability Of False Alarm, True Skill Statistic, and Briar Skill Score comparing skill relative to climatology.

d. Implement the new tool in the Meteorological Interactive Data Display System (MIDDS) used by the 45 WS in the Morrell Operations Center on Cape Canaveral Air Force Station. The implementation will include a GUI similar to that in the original tool.

e. Period Of Performance: 20 Sep 2016

4. Deliverables:

a. Written report documenting the development process and results.

b. Briefing of the development process and results to the 45 WS.

c. Implementation of tool into MIDDS, if justified.

5. References:

Bauman, W., 2013: Severe Weather Tool Using 1500 UTC Cape Canaveral Air Force Station Soundings. NASA Contractor Report

CR-2013-217920, Kennedy Space Center, FL, 36pp. [Available from ENSCO, Inc., 1980 N. Atlantic Ave., Suite 830, Cocoa Beach, FL, 32931 and online at science.ksc.nasa.gov/amu/final-reports/severe-tool-15z.pdf.]

Lambert, W., 2007: Objective Lightning Probability Forecasting for Kennedy Space Center and Cape Canaveral Air Force Station, Phase II. NASA Contractor Report CR-2007-214732, Kennedy Space Center, FL, 59 pp. [Available from ENSCO, Inc., 1980 N. Atlantic Ave., Suite 830, Cocoa Beach, FL 32931, and online at science.ksc.nasa.gov/amu/final-reports/objective-ltg-fcst-phase3.pdf].

Watson, L., 2011: Upgrade Summer Severe Weather Tool. NASA Contractor Report CR-2011-216299, Kennedy Space Center, FL, 29 pp. [Available from ENSCO, Inc., 1980 N. Atlantic Ave., Suite 830, Cocoa Beach, FL, 32931 and online at science.ksc.nasa.gov/amu/final-reports/severe-tool-upgrade-II.pdf.]