NASA and Google have expanded an existing partnership to help local governments improve air quality monitoring and forecasting for better decision-making. The expanded collaboration focuses on developing advanced machine learning-based algorithms that combine NASA data with Google Earth Engine data streams to generate high-resolution air quality maps in near real-time. “We’re excited about our partnership with NASA to make everyday air quality more effective at the local level,” said Rebecca Moore, director of Google Earth, Earth Engine and Outreach at Google.
“Environmental information such as high-resolution air quality maps can be useful tools for cities and community organizations to take climate and health action in their neighborhoods,” Moore said. “This scientific research partnership with NASA will help us improve the resolution, validation and utility of air quality maps across space and time giving everyone more data to make decisions about cleaner air.” Harmful air pollution affects people and the environment, and according to the World Health Organization, air pollution is responsible for approximately 7 million deaths worldwide each year.
“The World Bank estimates that the global cost of health damage from air pollution is $8.1 trillion, with people in low- and middle-income countries most affected,” said Christoph Keller. He is a senior researcher at Morgan State University working with NASA’s Goddard Space Flight Center and is the lead developer of the GEOS Composition Forecasting System (GEOS-CF). “Mitigating these impacts requires free access to high-quality, high-resolution, near-real-time air quality information, which this partnership will provide.”
The two-year annex agreement between NASA and Google, signed on July 1, 2022, builds on existing research and collaboration under the ongoing Space Law Agreement. The agreement aims to improve NASA science data discovery, access and usability by storing and sharing data on Google Cloud Platform and Google Earth Engine. This expanded collaboration leverages NASA and Google’s technical expertise and data to make daily air quality monitoring and forecasting information more actionable for cities and local communities to inform their climate and air quality management efforts. The results will produce estimates and forecasts of harmful pollutants such as nitrogen dioxide and fine particles in near real time.
To achieve this, as a first step, Google has incorporated two new NASA data sets into the Earth Engine catalog, which are automatically updated daily. These include data from the NASA Goddard Earth Observing System Composition Forecasts (GEOS-CF) and Modern Retrospective Analysis for Research and Applications, Version 2 (MERRA-2). These provide satellite observations of pollutants to help map and predict areas with poor air quality.
NASA scientists are now developing machine learning algorithms in Earth Engine to identify relationships between these newly added datasets and data collected from Google Street View mapping vehicles, surface observation stations and Earth monitoring satellites such as the TROPOspheric Monitoring Instrument (TROPOMI). aboard the European Space Agency’s (ESA) Copernicus Sentinel-5 satellite and on the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA’s Terra and Aqua satellites.
Understanding these relationships will allow the project team to combine atmospheric data from different instruments in new ways to create time-continuous, high-resolution, near-real-time maps of surface air quality. Independent observations will be used to continuously evaluate these maps. Increasing spatial resolution can help infer differences in air pollution exposure across neighborhoods or cities.
To further refine the concentration maps to cover areas even smaller than a city, NASA scientists will work with the Google Accelerated Science team to integrate data sources already available in Google Earth Engine, such as the location of major pollution sources. The team will first apply this new methodology to the San Francisco region and then expand to cities in the Lower Mekong region, Thailand, Cambodia, Laos and Vietnam through SERVIR. SERVIR is a joint initiative of NASA and the United States Agency for International Development (USAID) that works with regional organizations. The Google-NASA partnership ultimately aims to provide full global coverage.
“This partnership is a major step forward in integrating air pollution data from a range of critical sources, from ground-based observations to satellite data, into advanced machine learning algorithms,” said Pawan Gupta, principal scientist at the Universities Space Research Association (USRA). at NASA’s Marshall Space Flight Center. “This can allow us to estimate air pollution at a local level and make it accessible to the communities themselves – including those most affected by poor air quality.” In accordance with NASA’s Free and Open Data Policy, NASA and Google will make all products, algorithms, workflows, case studies, and tutorials developed under this partnership free and open to the public. The Google partnership is part of NASA’s Earth Science Division Global Partnership Program. More information about the program can be found on the Division of Earth Sciences Applied Sciences Program website.