Quality Assessment Activities Jennifer Mahoney NOAAForecast Systems Laboratory

Quality Assessment Activities Jennifer Mahoney NOAA/Forecast Systems Laboratory October 2003

Topics • • PIREPs CCFP Accuracy for 2003 Convective Activity and CCFP Confidence Level Information on ADDS – Graphical Turbulence Guidance (GTG) algorithm

PIREPS Needed • Verification data: PIREPS • Subjective • Spatially and temporally biased • Non-systematic Ø Just another interesting verification challenge

CCFP Accuracy 2 -h Coverage (2 March – 31 August) Actual Coverage High Medium Low Medium High Forecast Category

CCFP Accuracy 2 -h Coverage/Probability (2 March – 31 August) Actual Coverage High Medium Low Medium Forecast Category High

CCFP Accuracy 2 March – 31 August 2002 and 2003 PODy Bias 2002 2003

CCFP Accuracy 2 March – 31 August 2002 and 2003 Heidke Skill Score Gilbert Skill Score 2002 2003

CCFP Accuracy Summary Ø Overall, forecast skill improved in 2003. Ø Overforecasting of the convection increased slightly in 2003. Ø The 2 -h lead typically contained the correct amount of coverage for all categories of coverage. Largest improvement in the forecast occurred for the high coverage category. Ø At the 4 - and 6 -h leads, the observed coverage was generally a category lower than the forecast coverage. Ø The likelihood of obtaining the correct coverage increased with increasing probability at the 2 -h lead. This was not apparent at the 4 - or 6 -h leads.

Convective Activity Key: Defining convective activity that impacts the flow of traffic 10 nm surrounding 4 -km NCWD

Convective Activity Key: Defining convective activity that impacts the flow of traffic Compute coverage over 3, 000 sq. area

Convective Activity and CCFP

Confidence Level Statistics Confidence Level Information for GTG Annual, Seasonal, Regional

Questions?