Improving Cost Effective Air Improving Cost Effective Air Quality Forecasting Quality Forecasting UPDATE 9/25/2008: UPDATE 9/25/2008: • Revised NAQFS Analysis Revised NAQFS Analysis • Addition of 2007 NAQFS data Addition of 2007 NAQFS data • Prelim. Background research Prelim. Background research – 2007 Model Evaluation 2007 Model Evaluation – MD/VA Free Ride Program numbers MD/VA Free Ride Program numbers
23
Embed
Improving Cost Effective Air Quality Forecasting UPDATE 9/25/2008:
Improving Cost Effective Air Quality Forecasting UPDATE 9/25/2008:. Revised NAQFS Analysis Addition of 2007 NAQFS data Prelim. Background research 2007 Model Evaluation MD/VA Free Ride Program numbers. Revised NAQFS Analysis. Update Yorks, 2007 with 2007 NAQFS 1hr and 8hr forecasts - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Improving Cost Effective Air Improving Cost Effective Air Quality ForecastingQuality Forecasting
UPDATE 9/25/2008:UPDATE 9/25/2008:
•Revised NAQFS AnalysisRevised NAQFS Analysis•Addition of 2007 NAQFS dataAddition of 2007 NAQFS data•Prelim. Background researchPrelim. Background research
–2007 Model Evaluation2007 Model Evaluation–MD/VA Free Ride Program numbersMD/VA Free Ride Program numbers
Revised NAQFS AnalysisRevised NAQFS Analysis
• Update Yorks, 2007 with 2007 NAQFSUpdate Yorks, 2007 with 2007 NAQFS
– 1hr and 8hr forecasts1hr and 8hr forecasts
• 2007 forecast data inconsistent with 04-062007 forecast data inconsistent with 04-06
• Reviewed methodsReviewed methods
• 8hr average forecast values used as 1hr!8hr average forecast values used as 1hr!
– 8hr averages 8hr averages 1hr average 1hr average
– 8hr averaged 8hr averages 8hr averaged 8hr averages 8hr average 8hr average
• 04-06 maximum forecasts appeared low04-06 maximum forecasts appeared low
Original NAQFS Analysis (Yorks, 2007)Original NAQFS Analysis (Yorks, 2007)
Updated NAQFS AnalysisUpdated NAQFS Analysis
Original NAQFS Analysis (Yorks, 2007)Original NAQFS Analysis (Yorks, 2007)
Updated NAQFS AnalysisUpdated NAQFS Analysis
Original NAQFS Analysis (Yorks, 2007)Original NAQFS Analysis (Yorks, 2007)
Updated NAQFS AnalysisUpdated NAQFS Analysis
Original NAQFS Analysis (Yorks, 2007)Original NAQFS Analysis (Yorks, 2007)
Updated NAQFS AnalysisUpdated NAQFS Analysis
NAQFS Model Evaluation (Eder, 2008?)NAQFS Model Evaluation (Eder, 2008?)
• Categorical EvaluationCategorical Evaluation– Ozone threshold exceeded?Ozone threshold exceeded?– Was it forecasted to exceed?Was it forecasted to exceed?
ForecastForecast ThresholdThreshold
• a a MISSMISS HIGHHIGH• b b HITHIT HIGHHIGH• c c HITHIT LOWLOW• d d MISSMISS LOWLOW
•CSICSI Critical Success Index (%) Critical Success Index (%)
•FARFAR False Alarm Rate (%) False Alarm Rate (%)
•WSIWSI Weighted Severity Index (%) Weighted Severity Index (%)
(Kang, 2007)(Kang, 2007)
NAQFS Model Evaluation (Eder, 2008?)NAQFS Model Evaluation (Eder, 2008?)
•High Accuracy misleading! Large number of correctly forecasted non-exceedence High Accuracy misleading! Large number of correctly forecasted non-exceedence daysdays
•High FAR, Low CSI, and ~40% increase on WSI over CSI High FAR, Low CSI, and ~40% increase on WSI over CSI Lots of mis-forecasted Lots of mis-forecasted days, all very close to thresholddays, all very close to threshold
•Bias >1 Bias >1 Forecast generally higher than observed Forecast generally higher than observed
NAQFS Model Evaluation (Eder, 2008?)NAQFS Model Evaluation (Eder, 2008?)
•Low correlation days associated with cloud cover daysLow correlation days associated with cloud cover days
NAQFS Model Evaluation (Eder, 2008?)NAQFS Model Evaluation (Eder, 2008?)
MD/VA Free Ride Code Red ProgramMD/VA Free Ride Code Red Program
• Free bus rides on forecasted “AQI Red” daysFree bus rides on forecasted “AQI Red” days
• Attempt to reduce emission of ozone Attempt to reduce emission of ozone
precursorsprecursors
• Promotes awareness of air quality issuesPromotes awareness of air quality issues
• Educates public on steps to improve their airEducates public on steps to improve their air
• Less health problemsLess health problems
MD/VA Free Ride Code Red ProgramMD/VA Free Ride Code Red Program
• Northern VANorthern VA
– Funded 100% by Congestion Mitigation and Air Funded 100% by Congestion Mitigation and Air
• How can we incorporate said findings into How can we incorporate said findings into the model?the model?
• Use AirNow/CASTNet/other AQ data Use AirNow/CASTNet/other AQ data sources to help verify forecastssources to help verify forecasts
• Determine data neededDetermine data needed– Temporal rangeTemporal range
– Locations to look atLocations to look at
– Types of cost dataTypes of cost data
• Determine the best method of evaluating Determine the best method of evaluating the value of an AQ forecastthe value of an AQ forecast
ReferencesReferences
• Desimone, J., 2008; Memorandum: Desimone, J., 2008; Memorandum: Free Ride on Code Red Program.Free Ride on Code Red Program. Provided by Joan Rholfs, MWCOG.Provided by Joan Rholfs, MWCOG.
• Eder, B., D. Kang, R. Mathur, S. Yu, K. Schere, 2006. An Operational Eder, B., D. Kang, R. Mathur, S. Yu, K. Schere, 2006. An Operational Evaluation of the Eta-CMAQ Air Quality Forecast Model. Evaluation of the Eta-CMAQ Air Quality Forecast Model. Atmospheric Atmospheric EnvironmentEnvironment 4040, 4894 – 4905, 4894 – 4905
• Eder, B., D. Kang, R. Mathur, J. Pleim, S. Yu, 2008?. An Evaluation of the Eder, B., D. Kang, R. Mathur, J. Pleim, S. Yu, 2008?. An Evaluation of the National Air Quality Forecast Capability for the Summer of 2007. National Air Quality Forecast Capability for the Summer of 2007.
• Kang, D., R. Mathur, K. Schere, S. Yu, B. Eder, 2007. New Categorical Kang, D., R. Mathur, K. Schere, S. Yu, B. Eder, 2007. New Categorical Metrics for Air Quality Model Evaluation. Metrics for Air Quality Model Evaluation. Journal of Applied MeteorologyJournal of Applied Meteorology. . 4646, 549-55, 549-55