Continued improvements of air quality forecasting through emission adjustments using surface and satellite data Georgia Institute of Technology Yongtao Hu 1, M. Talat Odman 1, Michael E. Chang 2 and Armistead G. Russell 1 1 School of Civil & Environmental Engineering, 2 Brook Byers Institute of Sustainable Systems Georgia Institute of Technology With thanks to Pius Lee and the NOAA ARL Forecasting Team 13 th Annual CMAS Conference, October 27 th, 2014

Objective Improve air quality forecasting accuracy using near real time measurements through dynamic adjustments of emissions inventories –Air quality forecasting is an integral part of air quality management. –With source impacts forecasting in addition to air quality forecasting, operational forecasting can provide more information critical for dynamic air quality management. –Current forecasting accuracy calls for improvement. Forecasting with 3-D models relies on accuracy of emissions. Emission inventories are typically at least 4 years behind and “growth factors” are outdated. Wildland fires are becoming an increasingly important contributor to PM and ozone. –Dynamic adjustments of emissions can help improve both air quality and source impact forecasting Ultimate goal is to provide better information for the purpose of dynamic air quality management Georgia Institute of Technology

Current Hi-Res Forecasting System Based on SMOKE, WRF and CMAQ models Forecasting ozone and PM 2.5 since hour forecast at 4-km resolution for Georgia and 12-km for most of Eastern US Used by GA EPD assisting their AQI forecasts for Atlanta, Columbus and Macon Potentially useful for other states Georgia Institute of Technology Hi-Res Modeling Domains

Georgia Institute of Technology Hi-Res performance during ozone seasons for Metro Atlanta Ozone PM 2.5 MNB20% MNE25% MNB-10% MNE32%

Newly Built Hi-Res2 Forecasting System Georgia Institute of Technology Forecasting Air Quality for CONUS ( to be open on November 28 th ) Updated base emissions to 2011NEI WRF3.6.1 and CMAQv5.02 used 72-hour forecasts at 4-km resolution for Georgia and surrounding states, 12-km for most of Eastern state and 36-km for the rest of CONUS

Georgia Institute of Technology Forecasting Source Impacts at 4-km for Georgia ( to be open on November 28 th )

Georgia Institute of Technology E VOC Ozone -EA-EA  ∆C∆C EAEAEAEA CACACACA Brute Force vs. DDM-3D and HDDM-3D CBCBCBCB EBEBEBEB B A + + Forecasting Source Impacts using DDM-3D DDM-3D is just right for dynamic management purpose due to small adjustments from the current emission level

Uncertainties in Emissions hurt both air quality and source impacts forecasting! Georgia Institute of Technology

Currently working with PM 2.5 measurements at ~20 sites in Georgia An emissions and air quality auto-correction system utilizing near real-time satellite and surface observations Minimizes the differences between forecasted and observed concentrations (or fused concentrations with AOD and PM measurements) With minimum adjustment to source emissions Using impacts of emission sources calculated by CMAQ- DDM-3D –Source impacts can be used for dynamic air quality management.(e.g., traffic and fires) Georgia Institute of Technology Hi-Res2: Online Auto-Emissions-Adjustment Inverse Modeling Approach for Adjusting Emissions

Solve for R j that minimizes  2 Georgia Institute of Technology uncertainties number of obs number of sources DDM-3D calculated sensitivity of concentration i to source j emissions emission adjustment ratio weigh for the amount of change in source strengths Inverse Model Formulation L-BFGS algorithm is used for the optimization (R package nloptr)

Georgia Institute of Technology Example of selecting appropriate Г <- L-curve plot

Off-line tests using “real-time” PM 2.5 observations Surface PM 2.5 data from six sites in Atlanta –Direct use of satellite data (AOD) was problematic because of much larger uncertainties compared to surface data. –AOD will be “fused” to PM 2.5 concentration fields to provide “real-time” spatial patterns. Georgia Institute of Technology

DDM-3D sensitivities calculated for week1: Dec. 1-7, 2013 Georgia Institute of Technology SourceAreaOn-roadNon-roadPoint Dec. 1-7, Obtained emissions adjustments ratios (Rj) Shown for select day Dec. 2, 2013

Georgia Institute of Technology PM 2.5 Forecasting Performance for week 2: Dec , 2013 Obs (ug/m3)Sim (ug/m3)NFENFB Dec. 11, % Emis adjusted8.4524%2% Dec. 8-14, %85% Emis adjusted5.6254%39% without emissions adjustments Dec. 11, 2013 PM 2.5 Concentration with emissions adjustments Dec.11, 2013 PM 2.5 Concentration

DDM-3D sensitivities calculated for week1: Jul. 6-12, 2011 Georgia Institute of Technology Emission adjustments ratios (Rj) Shown for Jul. 11, 2011 SourceAreaOn-roadNon-roadPoint Jul. 6-12,

Georgia Institute of Technology PM 2.5 Forecasting Performance of week2: Jul , 2011 Obs (ug/m3)Sim (ug/m3)NFENFB Jul. 15, %-94% Emis adjusted7.2350%-40% Jul , %-44% Emis adjusted %7% without emissions adjustments Jul. 15, 2011 PM 2.5 Concentration with emissions adjustments Jul.15, 2011 PM 2.5 Concentration

Concluding Remarks Hi-Res2 is in operation with source impact forecasting –New in air quality forecasting practice –Promote dynamic air quality management through providing source specific information –Currently for traffic and power plant emissions only, can add more –Fire emission impact forecasts underway for dynamic prescribed-burn management Hi-Res2 is also in operation with dynamic adjustments of emissions –Performance evaluation is ongoing –Dynamic emissions adjustment dramatically improving PM forecast accuracy in off-line testing –Expansion to include other species measurements underway –Improved approach to assimilating AOD and PM measurements underway Georgia Institute of Technology

Acknowledgements NASA Georgia EPD Georgia Forestry Commission US Forest Service – Scott Goodrick, Yongqiang Liu, Gary Achtemeier Strategic Environmental Research and Development Program Joint Fire Science Program (JFSP) Environmental Protection Agency (EPA)

Georgia Emission Totals (tons/yr) Georgia Institute of Technology