IDEA for GOES-R ABI. Presented by S. Kondragunta, NESDIS/STAR. Team Members: R. Hoff and H. Zhang, UMBC

Transcription

1 IDEA for GOES-R ABI Presented by S. Kondragunta, NESDIS/STAR Team Members: R. Hoff and H. Zhang, UMBC 1

2 Project Summary Use operational MODIS, GOES Aerosol Optical Depth (AOD) products, and OMI/GOME-2 Aerosol Index (AI) to provide near-real-time air quality monitoring and forecasting guidance. Research and development work done under this project will investigate the usefulness of satellite measurements in improving air quality forecast guidance and pave the way for using enhanced aerosol products from GOES-R ABI Operational GOES AOD, MODIS AOD, OMI/GOME-2 AI data GOES-R ABI like retrievals obtained from MODIS radiances Tasks Develop and evaluate new GOES AOD retrieval algorithm (MAIAC) Adapt IDEA to GOES-R ABI retrievals, CONUS views and full disk views Expected Outcome Improved IDEA product Implementation of the new GOES AOD algorithm into IDEA Demonstration of improved air quality predictions 2

3 Infusing satellite Data into Environmental Applications (IDEA) Developed by NASA Langley and U. Wisconsin in cooperation with EPA and NOAA (Al-Saadi et al., BAMS, 2006) To improve air quality assessment, management, and prediction by infusing satellite measurements into analyses for public benefit IDEA tool Near real-time, integrates satellite and ground data, and generates forward trajectories for dissemination to users. MODIS Terra AOD was the main data source Ran continuously at UW from 2004 to

4 IDEA transition and enhancement Transition IDEA from Wisconsin to NESDIS from research to pre-operational system UMBC was funded by GOES-R risk reduction program and NASA 3D-AQS project to enable this transition and add GOES components. Additional data sources incorporated into IDEA : GOES-E AOD, GOES-W AOD, MODIS Aqua AOD, OMI aerosol index Integrated IDEA with the original MODIS Terra product and made it switchable between different satellite products More components are added IDEA or a system similar to it will use GOES-R ABI aerosol data once GOES-R is launched 4

5 IDEA webpage at NESDIS National and regional views of MODIS/GOES AODs and true color image (MODIS) or reflectance of visible band (GOES) Aerosol trajectory initiated from MODIS/GOES AOD and OMI AI 3-day composite data animation Time series, scatter plots, histograms of MODIS/GOES AOD and PM 2.5 for each hourly PM 2.5 site National correlation summary between MODIS/GASP AOD and surface measurements of PM2.5 PM2.5 estimates from AOD 5

6 New data source included: GOES AOD national & regional map animation EPA Region 1-3 Can switch to view MODIS AOD, GASP AOD map, and GASP AOD animation Can select from the 10 EPA geographical regions to view regional map 6

7 Incorporation of OMI AI into 48 hr forward trajectories from MODIS Aqua AOD High OMI aerosol index usually indicates high level of aerosol layer It can also detect aerosol above cloud OMI onboard Aura is eight minutes behind Aqua In trajectory initialization height depends on OMI AI value: OMI AI >2.0, trajectory initialized at 3 km. Otherwise, trajectory initialized close to surface. 7

8 Statistics of MODIS/GOES AOD and surface PM 2.5 Statistics plots for each hourly PM 2.5 site 8

9 Daily PM 2.5 estimates from AOD Calculate daily average AOD from MODIS Terra, Aqua, and GASP East. Ignore GASP AOD if MODIS AOD is available. Compute PM 2.5 from daily average AOD using linear regression relationship between AOD and PM 2.5 The linear regression relationship varies seasonally and regionally ( four season and ten EPA geographic regions, Zhang et al., JAWMA 2009) 9

10 Visualization of MODIS AOD Data in Google Earth We set up MODIS AOD near-real-time image in KML format, which is available as a link on IDEA. The AOD images can be viewed along with the RGB images (from MODIS Today, in Google Earth. *Thanks Liam Gumley of CIMSS for the help on the set up 10

11 Aerosol wind -- an experimental product Wind fields derived from GOES AOD imagery A portion of the GOES smoke concentration algorithm code that was developed by STAR provides observed wind fields (speed and direction) from GOES AOD imagery. This code was adapted for the IDEA website to display 3-hr movie loops of observed wind fields A survey of the users revealed that while they like these observed wind fields, they are of not great value for forecasting applications Product could become useful to forecasters in GOES-R era because of 5-minute refresh rate. 11

12 IDEA user survey Obtained feedback from a focus group comprising of 20 state and local air quality forecasters on IDEA User feedback was very favorable to IDEA product. An excerpt from one user is shown in the adjacent panel These trajectories are one of the most important forecast tools that are available for PM2.5. Bear in mind, forecasters have very few tools for PM forecasting. Statistical models are not good and numerical models are experimental. We have to depend on persistence and transport for forecast guidance to a much larger degree than ozone forecasting. Having observations tied to forecast transport and presented in an elegant manner as on IDEA are a critical tool for us. Bill Ryan Penn State Forecaster for Philadelphia 12

13 Air quality monitoring during Beijing Olympics We provided near real-time MODIS AOD images over northeastern China during Beijing Olympics for smog blog 13

14 Port MAIAC for GOES data Apply Lyapustin s MAIAC (Multi-Angle Implementation of Atmospheric Correction) algorithm for MODIS to GOES AOD retrieval MAIAC algorithm derive surface BRDF and AOD from time sequences of MODIS images Assumptions: AOD is homogeneous over 25x25 km 2 area Surface BRDF in blue channel and red channel is proportional to 2.1 μm channel 14

15 Comparison with GASP, MODIS Terra, Aqua MAIAC GASP MAIAC retrieval has more details than MODIS and larger area than GASP and MODIS Terra Aqua 15

16 Interactions With AWG/PG This project work on adapting MAIAC algorithm is done with close interaction with the AWG team. Products/tools developed under this project provide high visibility to AWG algorithms/products when this or a similar tool is adapted for Air Quality Proving Ground 16

17 Summary Transition of IDEA from UW to NOAA/NESDIS IDEA enhancements Incorporate AOD data from GOES East, GOES West, MODIS Aqua into components of IDEA Incorporate aerosol index (AI) data from OMI in trajectory forecast Scatter plots, histograms of AOD and PM2.5 at each hourly PM2.5 stations over North America PM2.5 estimates from AOD MODIS true color imagery MODIS AOD in KML format, ready to be viewed in google Earth Collaborations 3D-AQS data product: MODIS, GASP AOD matchup dataset Air quality monitoring for Beijing Olympic SERVIR, air quality monitoring for Central America Development of new AOD retrieval algorithm for GOES GOES image co-registration Validation 17

18 Future Work Complete testing of MAIAC algorithm implementation for GOES Develop GOES-R ABI tutorials for AMS/AGU conference presentations Work hand-in-hand with air quality proving ground to bring IDEA experience for the development of a visualization tool for ABI products 18