Supporting NASA Missions with the GEOS Composition Forecast System, “GEOS-CF”
Explore the source record for details and available documents.
GEOS↗
Engineering topics
Joseph Ardizzone
Publications and source records attributed to Joseph Ardizzone.
Google and NASA Air Quality Partnership – A Collaboration Using GEOS-CF Data and Google Earth Engine
NASA and Google have expanded their partnership to create data and tools that help with pollution mitigation and decision making on a local government scale. The goal is to use the technologies available at NASA and Google to create city-scale data estimates and forecasts of air pollutants such as NO2 derived from the GEOS Composition Forecast (GEOS-CF) model. Efforts are also being led by Pawan Gupta to create a downscaled MERRA-2 PM2.5 product.
Callum Wayman↗
Improving Access to the GEOS Composition Forecast Model with API Development and Ingestion into Google Earth Engine
The GEOS Composition Forecast (GEOS-CF) model produces forecast and historical estimates of atmospheric composition and meteorology fields, which provide useful insight into air quality issues and events. In a year for which Canadian wildfires created adverse air quality conditions in the eastern United States, access to model fields such as PM2.5 are in high demand. The GEOS-CF team at the NASA Global Modeling and Assimilation Office (GMAO) first developed in-house solutions to improve data access via the CF API, and recently partnered with Google to ingest a collated set of model diagnostics into the Google Earth Engine (GEE) data repository. GEOS-CF model output is also being ingested into AWS storage. Creating these various open access points to GEOS-CF model diagnostics provides the public with an opportunity to easily interact with air quality information. Users are able to use a temporally consistent global grid of air quality fields in machine learning applications, mapping tools, and data informatics. Hosting GEOS-CF forecasts and the historical timeseries of these chemistry and meteorology fields in GEE allows users to create dynamic JavaScript-based air quality applications in the GEE code editor. GEOS-CF users can also access the model output via the GEE Python application programming interface (API), making it easy to perform various analyses with Python. This presentation will show two examples of accessing the GEOS-CF model through GEE. The first is an example application made in the GEE code editor which allows users to view time series plots and downscaled maps of surface level NO2. The second example exhibits using the GEE Python API to create a machine learning model to temporally gap-fill between air quality observations. These examples are an introduction to the many possible benefits of having open access to the GEOS-CF model through multiple platforms.
Callum Wayman↗
Improving Access to the GEOS Composition Forecast Model with API Development and Ingestion into Google Earth Engine
The GEOS Composition Forecast (GEOS-CF) model produces forecast and historical estimates of atmospheric composition and meteorology fields, which provide useful insight into air quality issues and events. In a year for which Canadian wildfires created adverse air quality conditions in the eastern United States, access to model fields such as PM2.5 are in high demand. The GEOS-CF team at the NASA Global Modeling and Assimilation Office (GMAO) first developed in-house solutions to improve data access via the CF API, and recently partnered with Google to ingest a collated set of model diagnostics into the Google Earth Engine (GEE) data repository. GEOS-CF model output is also being ingested into AWS storage. Creating these various open access points to GEOS-CF model diagnostics provides the public with an opportunity to easily interact with air quality information. Users are able to use a temporally consistent global grid of air quality fields in machine learning applications, mapping tools, and data informatics. Hosting GEOS-CF forecasts and the historical timeseries of these chemistry and meteorology fields in GEE allows users to create dynamic JavaScript-based air quality applications in the GEE code editor. GEOS-CF users can also access the model output via the GEE Python application programming interface (API), making it easy to perform various analyses with Python. This presentation will show two examples of accessing the GEOS-CF model through GEE. The first is an example application made in the GEE code editor which allows users to view time series plots and downscaled maps of surface level NO 2 . The second example exhibits using the GEE Python API to create a machine learning model to temporally gap-fill between air quality observations. These examples are an introduction to the many possible benefits of having open access to the GEOS-CF model through multiple platforms.
Callum Wayman↗
Version 8 of the SMAP Level-4 Soil Moisture Data Assimilation Product
The NASA Soil Moisture Active Passive (SMAP) mission Level-4 Soil Moisture (L4_SM) product provides global, 9-km resolution, 3-hourly surface (0-5 cm) and root-zone (0-100 cm) soil moisture from April 2015 to present with a mean latency of 2.5 days from the time of observation. The product is based on the assimilation of SMAP L-band (1.4 GHz) brightness temperature (Tb) observations into the NASA Catchment land surface model as the model is driven with observations-based precipitation forcing. This presentation discusses the improvements in the forthcoming Version 8 of L4_SM, including updates in the precipitation forcing, the Catchment model parameters, and the L-band microwave radiative transfer model (mwRTM). The precipitation observations used in L4_SM Version 8 outside of North America and the high latitudes are from the latest (Version 7) NASA Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement mission (IMERG) products. Moreover, occasionally excessive precipitation rates in earlier versions of L4_SM along certain longitudes in North America were eliminated by a bug fix in the precipitation corrections algorithm. The Catchment model in L4_SM Version 8 uses climatological snow albedo values based on observations from the Moderate Resolution Imaging Spectroradiometer, replacing the look-up table parameterization of earlier versions. Additionally, corrected soil parameters were implemented for a small region in Argentina that had erroneously been classified as peat because of an error in the ancillary soil data. Finally, the mwRTM in L4_SM Version 8 uses the Mironov soil mixing approach and updated values of the L-band scattering albedo, soil roughness, and vegetation opacity climatology obtained from the latest (Release 19) SMAP Level-2 dual-channel soil moisture retrieval product. During the development of L4_SM Version 8, the change in the mwRTM parameterization resulted in a reduced unbiased RMSE of surface soil moisture when verified against in situ measurements. It also reduced the standard deviation of the Tb observation-minus-forecast residuals by ~0.15 K, highlighting the importance of the mwRTM for successful data assimilation. The bug fixes in the precipitation corrections algorithm and the Catchment model soil parameters in the Argentina region result in locally large improvements of the simulated land surface states. In summary, the ongoing refinements of the L4_SM product continue to improve its science quality and performance for global soil moisture monitoring
Rolf Reichle↗