Transcript Slide 1
Assessing U.S Air Quality with Remote Sensing Data via Goddard Interactive Online Visualization ANd aNalysis Infrastructure NASA Goddard Space Flight Center http://giovanni.gsfc.nasa.gov/ Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 1 1 Content • What is Giovanni? • Giovanni for Air Quality • Trace gases in Giovanni • Interoperability of Giovanni • Case studies • Giovanni and Google Earth Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 2 2 About Giovanni • Giovanni is a Web-based application developed by the NASA Goddard Earth Sciences Data and Information Services Center (GES DISC). • Giovanni provides a simple and easy way to explore, visualize, analyze, and access vast amounts of Earth science remote sensing and model data. http://giovanni.gsfc.nasa.gov/ Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 3 3 10-6 ppmv Data Inputs Aerosol from GOCART model AIRS MODIS MISR Parasol Particulate Matter (PM 2.5) from AIRNow Aerosol from MODIS and GOCART model Carbon Monoxide from AIRS MODIS vs SeaWiFS Chlorophyll Ozone Hole from OMI Giovanni Instances CloudSat CALIOP TOMS OMI MLS HIRDLS HALOE TRMM AMSR-E SeaWiFS Models and more… Water Vapor from AIRS Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 4 4 Basic (one-parameter): • Area plot – averaged or accumulated over any data period for any rectangular area (various map projections) • Time plot – time series averaged over any rectangular area • Hovmöller plots –longitude-time or latitude-time cross sections • ASCII output – for all plot types (can be used with GIS apps) • Image animation – for area plot • Vertical profiles • Vertical cross-sections, zonal means Beyond basics: • Area plot - geographical intercomparison between two parameters • Time plot - an X-Y time series plot of several parameters • Scatter plot of parameters in selected area and time period • Scatter plot of area averaged parameters - regional (i.e., spatially averaged) relationship between two parameters • Temporal correlation map - relationship between two parameters at each grid point in the selected spatial area • Temporal correlation of area averaged parameters - a single value of the correlation coefficient of a pair of selected parameters • Difference plots • Anomaly plots • Acquiring parameter and spatial subsets in a batch mode through Giovanni Gregory Leptoukh & Frank Lindsay, AQ Data Summit http://giovanni.gsfc.nasa.gov/ Capabilities RTP, NC 5 5 Only a Web browser is needed. No need to learn data formats and programming. No need to download large amounts of data. Customized data and analyses can be obtained with only a few mouse clicks. Caution: Giovanni is an exploration tool! Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC http://giovanni.gsfc.nasa.gov/ The Power of Simplicity 6 6 AOT for June 2006 Terra MODIS Parasol POLDER Aqua MODIS Envisat MERIS Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 7 7 AOT Differences for June 2006 Terra MODIS – Aqua MODIS Aqua MODIS - POLDER Terra MODIS – MERIS MERIS – POLDER Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 8 8 Scatter plots Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 9 9 Time series Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 10 10 Maps of NO2 Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 11 11 Maps of CO Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 12 12 Profile Data Comparisons HIRDLS MLS HIRDLS and MLS ozone (top) and temperature (bottom) profiles acquired March 12, 2007, over France during the passage of a weather front. Note the tropopause fold (arrow) in the ozone profiles. MLS vertical resolution is ~3 km, HIRDLS vertical resolution is ~1 km. Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 13 13 California fires by MODIS Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 14 14 PM2.5 from AirNow in Giovanni Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 15 15 Visualizing California’s Wildfires from Space 23-27 October 2007 Data from NASA’s Aura OMI (Tropospheric NO2 and UV Aerosol Index), Aqua AIRS (Total Column CO) and Terra MODIS (Aerosol Small Fraction, Cloud Optical Thickness and Aerosol Mass Concentration Over Land) Tropospheric NO2 OMI OMI Aerosol Small Mode Fraction MODIS UV Aerosol Index Cloud Optical Thickness MODIS Gregory Leptoukh & Frank Lindsay, AQ Data Summit Total Column CO AIRS Aerosol Mass over Land MODIS RTP, NC 16 16 Data Fusion (prototype) in Terra Terra + Aqua Aqua Dust event, MayAQ 23, 2007 Gregory Leptoukh & Frank Lindsay, Data Summit RTP, NC 17 17 Interoperability Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 18 18 Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 19 19 MODIS Aerosol Data TOMS already in Giovanni 3.06 WMS OMI Protocols Binary SATs in Giovanni Sea WIFS HDF 4/5 NASA GES DISC KML Format IDL MAPSS ESDC ASCII NetCDF Open DAP WCS Data Fed CALIPSO MATLAB etc. ICARE Interoperability BAMGOMAS AMAPS SERVIR etc. Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 20 20 OGC and other m2m protocols • WMS: – Via Map Server – Via Giovanni • WCS: – Via WebGIS – Via Giovanni • OPeNDAP Hyrax 4 serving data on 12 machines. Sample URLs: http://acdisc.sci.gsfc.nasa.gov/opendap/ http://atrain.sci.gsfc.nasa.gov/opendap/ • WMS sample URL to get WMS data for 'LAYER=AIRX3STD_TOTO3_A‘: http://g0hep12u.ecs.nasa.gov/mapservbin/wms_ogc?TARGET_SRS=EPSG:4326&Service=WMS&VERSION=1.1. 1&REQUEST=GetMap&SRS=EPSG:4326&WIDTH=768&HEIGHT=512&BB OX=-180,-90,180,90&LAYERS=AIRX3STM_TOTO3_A,coastline • GDS (Grads-DODS): http://agdisc.gsfc.nasa.gov/dods/ Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 21 21 WMS and WCS in Serving MODIS data via WMS: Maps (decorated and undecorated), time-series, Hovmoller, time-averaged maps, difference maps Example of Maps: http://giovanniplus-ts1.sci.gsfc.nasa.gov/daac-bin/G3/giovanniwms.cgi?SERVICE=WMS&WMTVER=1.0.0&REQUEST=GetMap&SRS=EPSG:4326&EXCEPT IONS=INIMAGE&FORMAT=GIF&BBOX=-130,24,-60,52&TIME=2006-0201T00:00:00Z&WIDTH=800&HEIGHT=400&LAYERS=MOD08_D3.005::Optical_Depth_Land_A nd_Ocean_Mean Serving data via WCS: Here is a get capabilities url: http://giovanniplus-ts1.sci.gsfc.nasa.gov/daac-bin/G3/giovanniwcs.cgi?SERVICE=WCS&WMTVER=1.0.0&REQUEST=GetCapabilities Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 22 22 WCS support outside of OMI NO2 (Level 3): getCapabilities and describeCoverage requests http://acdisc.sci.gsfc.nasa.gov/daac-bin/wcsNO2?service=wcs&version=1.0.0&request=describeCoverage http://acdisc.sci.gsfc.nasa.gov/daac-bin/wcsNO2?service=wcs&version=1.0.0&request=getCapabilities • getCoverage example request: http://acdisc.sci.gsfc.nasa.gov/daacbin/wcsNO2?service=WCS&version=1.0.0&request=getCoverage&CRS=WGS84&resx=0.5&resy=0.5&cov erage=NO2Total&bbox=-179.75,-89.75,179.75,89.75&TIME=2006-08-01/2006-08-04&format=netCDF AIRS X2RET (Level 2 collection 5): getCapabilities and describeCoverage requests http://g0dup05u.ecs.nasa.gov/cgi-bin/ceopAIRX2RET?service=wcs&version=1.0.0&request=getCapabilities http://g0dup05u.ecs.nasa.gov/cgibin/ceopAIRX2RET?service=wcs&version=1.0.0&request=describeCoverage • getCoverage request: http://g0dup05u.ecs.nasa.gov/cgibin/ceopAIRX2RET?service=WCS&version=1.0.0&request=getCoverage&coverage=H2OMMRStd&crs=W GS84&bbox=-179.75,-89.75,179.75,89.75&resX=0.5&resY=0.5&time=2004-07-28&format=netCDF • The server now supports 28 variables, including both 2D and 3D fields: TSurfAir, TAirStd, GP_Height, GP_Surface, PSurfStd, TSurfStd, totH2OStd, H2OMMRStd, H2OMMRSat, H2OMMRSat_liquid, O3VMRStd, totO3Std, PCldTopStd, TCldTopStd, olr, clrolr, CO_total_column, CO_VMR_eff, CO_eff_press, CH4_total_column, CH4_VMR_eff, CH4_eff_press, GP_Height_MWOnly, sfcTbMWStd, EmisMWStd, totH2OMWOnlyStd, totCldH2OStd, and numCloud. • OMI UVB and O3 (Level 3) coming shortly Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 23 23 Air Quality Tools and Datasets on aerosols Available Science Data Sets (examples) • PM2.5 station data - EPA AirNow (via WCS) DataFed (aggregated and gridded) (via WCS) Giovanni • MODIS TERRA and AQUA total and Fine mode Aerosol Optical Depth • CALIOP Aerosol Feature Mask curtain plots • OMI NO2 Tropospheric column and Aerosol Index Useful Tools for Air Quality Applications • AOD/ PM2.5 scatter plots, correlation maps, time series and difference plots • AOD and PM2.5 loops for examining long range transport of aerosols Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 24 24 Prototyping PM25 data in PM2.5 (EPA DataFed Giovanni) The standard MODIS AOT Deep Blue MODIS Aerosol Optical Depth GOCART AOT Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 25 25 Giovanni Air Quality Data (July 7th, 2006) EPA AirNow PM2.5 (ug/m3) OMI Aerosol Index • MODIS and OMI imagery show smoke aerosols over the northeast, southeast and Great Lakes. Level-3 MODIS AQUA AOD CALIOP Aerosol Flag (yellow) • CALIOP Aerosol Flag (yellow) confirms that aerosols are above the boundary layer • EPA AirNow PM2.5 doesn’t show anything around Great Lakes, i.e. aerosols are primarily above the boundary layer Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 26 26 Giovanni Air Quality Services: AOD/PM2.5 Correlation Maps and Time Series May 2007 AOD/PM2.5 correlation map over the U.S Moderate to good correlation in the eastern U.S No significant differences were found when using the Fine Mode MODIS AOD. May 2007- AOD and PM2.5 Time series over the southeast Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 27 27 CALIPSO: Elevated Smoke Layers over the US Midwest Giovanni MODIS Terra AOD map Giovanni PM2.5 Map Smoke Smoke in Great Lakes region moving east Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 28 28 Giovanni Air Quality tools: Understanding AOD/PM2.5 correlations Level-3 MODIS AOD May 22nd, 2007: Smoke over North Carolina. High AOD and low PM2.5 (r=0.54). There is also haze in the southeast EPA PM2.5 (ug/m3) • Improved correlation over this region when excluding smoke areas (r=0.80) • Giovanni data sets and tools help provide a more complete understanding of the origin, evolution, and vertical distribution of aerosol pollution over the continental U.S. Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 29 29 Giovanni Air Quality tools: Understanding AOD/PM2.5 correlations EPA PM2.5 (ug/m3) • CALIOP Aerosol Flag (yellow) for examining the vertical aerosol distribution. Level-3 MODIS AOD • Aerosols in Georgia and Alabama from the surface to 4 km, AOD/PM2.5 correlation is moderately good • Aerosols in the northeast are above boundary layer, AOD/PM2.5 correlation is poor Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 30 30 July 31st, 2007 In Canada and the north-central US, MODIS and OMI show thick aerosols plumes. CALIOP overpass has a plume above the boundary layer Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 31 31 July 31, 2007: Haze over the south eastern US • In the southeast (Tennessee, Mississippi, Alabama and Arkansas) MODIS and PM2.5 show good spatial agreement and have moderately good correlation (see scatter plot) • Low OMI Aerosol Index and CALIPSO Aerosol flag (see previous slide) also indicate aerosols are primarily confined to the boundary layer in these states Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 32 32 July 31, 2007: Haze over the south eastern US OMI contours over MODIS AOD. White lines indicated CALIPSO overpass Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 33 33 Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 34 34 Importing Giovanni Data into Google Earth Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 35 35 A-Train in Google Earth via Giovanni: Calipso Lidar Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 36 36 Analysis of U.S Air Quality Via • Examined AOD and PM2.5 maps, correlation maps and time series plots. Fine Mode AOD also available • High AOD/PM2.5 correlation indicates the MODIS algorithm is capturing aerosols at the surface in addition to elevated aerosols (if any) • High AOD and low PM2.5 may indicate the presence of aerosol plumes above the boundary layer • CALIPSO overpass (if available) together with AOD/PM2.5 correlations and scatter plots to qualitatively assess the vertical distribution of aerosols • OMI measurements are less sensitive to aerosols in the boundary layer, so if OMI doesn’t show high aerosol while MODIS does, it may indicate aerosol being in the boundary layer • MODIS algorithm issues (e.g. retrieval problems over bright surfaces) may affect correlations Gregory Leptoukh & Frank Lindsay, AQ Data Summit RTP, NC 37 37 Test Case - NO2 Air Pollution Data from Aura OMI Iamges Courtesy of Mark O. Wenig, Eric J. Bucsela, Edward A. Celarier, James F. Gleason, NASA J. Pepijn Veefkind, K. Folkert Boersma, Ellen Brinksma, KNMI