Upcoming Changes to the EPA Photochemical Assessment Monitoring Stations Network; Closing the Gap between Surface-to-Satellite Air Quality James Szykman U.S. EPA Office of Research and Development National Exposure Research Laboratory NOAA Satellite Aerosol Product Workshop for Science and Operational Users National Center for Weather and Climate Prediction (NCWCP) College Park, MD 20740 September 13-14, 2016

EPA Use of Satellite Aerosol Products for Air Quality Exceptional Event Analysis: Direct impact on CAA attainment demonstrates. Regional-Global Model Evaluation: Provides on platform-consistent data set to evaluate models at regional to global scales. Data Fusion for Surface PM2.5 Predictions: Focused on AOD/PM2.5 relationship - Daily AOD is a measure of the true state of the atmosphere for aerosols. Chemical Transport Models require extensive emission inventories for model predictions, and often do not capture high PM2.5 concentrations associated with wildfires.

Hierarchical Autoregressive Model for PM2.5 predictions Use of VIIRS AOT in hierarchical autoregressive model to model daily average PM2.5 concentration across CONUS: VIIRS AOT data - day-specific spatially-varying intercept and coefficient Account for missingness in AOT data via model-based imputation at missing grid cells Autoregressive term based on previous day surface PM2.5 concentrations Meteorological covariates (daily avg. T and RH) Pt(s)= α0,t + β0,t(s) + (α1,t + β1,t(s))Ai,t + Xt(s)γ + ρPt−1(s)+ Et(s) (Schliep et al, 2015, Adv. Stat. Clim. Meteorol. Oceanogr)

Challenges Faced with AOD and Surface PM2.5 in Fusion Models Data sources are temporally and spatially misaligned Extensive missing data in both the monitoring data and satellite data Correlation between the two data sources varies both in time and in space, largely a function of - aerosol type, vertical distribution, and meteorological effects Inclusion of AOD provides limited improvement over current EPA fusion model based on CMAQ output (sigma layer 1) and surface PM2.5 developed for use with EPA-CDC public health tracking system effort.

Challenges with AOD and Surface PM2.5 in Fusion Models Data sources are temporally and spatially misaligned Extensive missing data in both the monitoring data and satellite data Correlation between the two data sources varies both in time and in space, largely a function of - aerosol type, vertical distribution, and meteorological effects GOES-R ABI AOD will help address the first two issues…what about the third issue?

PAMS Network Re-Engineering Forthcoming changes in air quality monitoring requirements under the Photochemical Assessment Monitoring Station (PAMS) program by EPA present an opportunity to better connect air quality relevant satellite measurements with surface measurements for both aerosol and trace gases. -The overall objective of PAMS network is to provide measurements that will assist States in understanding ozone nonattainment problems and to implement strategies to solve the problem. -The PAMS network sites are operated by the state and local air pollution agency – not EPA. -As a result of the changing nature of the ozone problem and the form of the standard since the start of the PAMS program, the EPA has proposed a re-engineered PAMS program. -At numerous locations across the country, PAMS sites will be co-located, or in near proximity to, NCore sites established primarily for particulate matter (PM) monitoring.

Combined NCore and PAMS Measurements: Can Measurements at combined NCore/PAMS Sites help address the AOD/PM2.5 Combined NCore and PAMS Measurements: NO, NO2, NOy, O3 (year round). SO2, CO, PM2.5 mass and speciation (At least 1-in-3 day), PM2.5 continuous, PM10-2.5 mass, basic met. parameters. 33 high priority VOC/Carbonyl Profiling Measurements: Radar wind profiler (RWP)/Radio acoustic sounding system (RASS) (or) Ceilometer for hourly mixing heights Some site contain optional met measurements – Vertical wind speed, solar radiation, precipitation, baro. pressure, delta-T for 2-10m. Current PAMS Network PAMS at NCore Number of Sites 75 43 -Existing 14 -New 19

Vaisala CL-51Ceilometer Vaisala CL-51 Ceilometer Characteristics: Cloud reporting range: 0…43,000 ft (0…13km) Backscatter profiling range: 0…49,200 ft (0…15km) Can operate in all weather Fast measurement - 6 second measurement cycle Reliable automatic operation Good data availability Eye safe diode laser (LIDAR) CL-51 positioned next to Space Science and Engineering Center, University of Wisconsin Mobile Lab Within in the U.S. peer-reviewed literature contains limited evaluation of the CL-51 derived mixing layer height compared to radio-sonde boundary layer height.

Field Evaluation locations for CL-51 NOAA BOA Tower Site (Erie, CO) – start of High Plains Golden NREL Site -On a mesa -intermountain site Both sites low aerosol loading NASA Langley (Hampton, VA) Near sea level – coastal site Low-moderate aerosol loading – with marine influence

CL-51 Comparison was conducted using the Vaisala BL-View Software for the CL-51 MLHs Uses a proprietary gradient method algorithm Identifies up to 3 aerosol layers for consideration of MLH Layers assigned quality index (QI) 1 to 3; 3 highest confidence Use of variable time and altitude averaging Characteristic backscatter curtain plot generated in BL-View for 17-July 2014 Golden, CO

CL-51Mixing Layer Height (aerosol gradient) vs. Planetary Boundary Layer Height (thermal gradient) Total of 53 radio sondes used for evaluation CL-51 data averaged over 5-minutes to account for spatial differences with sondes R = unfiltered R = filtered; 5-minute σ > 0.20 km or RSD > 20% Knepp, T.N., J. S. Szykman, R. Long, R. Duvall, J. Krug, M. Beaver, K. Cavender, K. Kronmiller, M. Wheeler, R. Delgado, R. Hoff, E. J. Welton, E. Olson, R. Clark, D. Wolfe, D. Van Gilst, and D. Neil, Assessment of Mixed-Layer Height Estimation from Single-wavelength Ceilometer Profiles, submitted Atmos. Chem. Phys

June 10, 2015 – Canadian Forest Fires Smoke Plume Can a Ceilometer Provide a Regional Relative Profile Ratio of Backscatter to Scale AOD? June 10, 2015 – Canadian Forest Fires Smoke Plume Characteristic backscatter curtain plot generated in BL-View for 10-June, 2015 at CAPABLE Research Site - Hampton, VA Smoke from the Canadian forest fire was observed by increased backscatter in the 2500 – 4000 m range.

Use of Airborne Science Assets to Characterize Spatial Variability: NASA LaRC Airborne HSRL Mixed Layer Heights Mixed Layer (ML) heights were derived from daytime-only cloud-screened aerosol backscatter profiles measured by the airborne HSRL using a Haar wavelet covariance transform (adapted from Brooks, JAOT, 2003) with multiple wavelet dilations to identify sharp gradients in the backscatter ML heights derived from the ceilometer (CL31) using the Vaisala BL View software During the 2010 CalNex field campaign, comparisons between ML heights from HSRL and a Vaisala ceilometer operated during CalNex were used to evaluate the representativeness of a fixed measurement over a larger region. (Scarino et al., 2014, ACP)

Can a Ceilometer Provide a Regional Relative Profile Ratio of Backscatter to Scale AOD? Due to the ceilometer emission wavelength’s proximity to the near-infrared water vapor bands, there is a known water vapor interference, However, the interference on aerosol profile and MLH estimation is said to be negligible (Wiegner et al. (2014)). Use co-located micro-pulse lidar (MPL) to evaluate the ceilometer backscatter profile.

Potential for PAMS/NCore to serve as Backbone for Satellite Air Quality Ground Validation Ceilometer/lidar Aerosol layers/mixing heights Ground-based FRM/FEMs Ground-based CIMEL radiometer for AOD and PANDORA spectrometer Column density O3, NO2, HCHO, and SO2

Are we finally heading is the right direction…. The Combined NCore/PAMS sites: Begin to address the lingering AOD/PM2.5 issue. Are focused on the areas with the worst air quality problems across the United States. Help foster more significant interaction between satellite data providers and the air quality agencies, and provide a stable infrastructure. Build on existing AQ long term datasets in urban areas, and make satellite data products more relevant to air quality managers. With the addition of CIMEL and PANDORA, can serve as national set of ground validation sites for satellite air quality on an operational basis.

Acknowledgements/Disclaimer ORD Contributors - Rachelle Duvall, Luke Valin, Melinda Beaver (OAQPS) and John Krug NASA Langley – Jim Crawford, Amy Scarino, and Travis Knepp (SSAI) U.S. EPA, OAR/OAQPS/AAMG – Kevin Cavender CDPHE Alion Science and Technology NASA, NOAA, UMBC, Millersville University, SSEC/CIMSS (Univ. of Wisc.), Ball Aerospace, Harvard-Smithsonian - CfA Disclaimer Although this work was reviewed by EPA and approved for presentation, it may not necessarily reflect official Agency policy. Office of Research and Development National Exposure Research Laboratory