GOES-R Program Research Opportunities in Space and Earth Sciences (ROSES)
Fiscal Year 2020 New Starts

All fiscal year 2020 awards have a three-year period of performance.

Development and implementation of a set of new enhanced GOES-R ABI snow cover products
Peter Romanov, City College of New York

This project involves using the GOES-16/17 Advanced Baseline Imager (ABI) to retrieve information about snow cover. Snow information includes a snow mask, factional snow cover, and snow depth at an hourly cadence. This information is then used for a variety of applications, including better initializing numerical weather prediction models, ultimately leading to more accurate weather forecasts


Realizing LEO Sounder Products at GEO Imager Spatial and Temporal Resolution
Elisabeth Weisz, SSEC/Univ. of Wisconsin

This project takes data from Low Earth Orbiting (LEO) satellites and combines it with data from Geostationary (GEO) satellites in order to gain the best of both: high spatial and temporal resolution from GEO and high spectral resolution from LEO. The resulting products have the potential to significantly benefit weather forecasting and nowcasting of severe storms as well as environmental monitoring operations.


Development of a Next-Generation Science-Quality Geostationary Satellite Active Fire Product
Louis Giglio, Univ. of Maryland

The project aims to develop an optimized, automated fire detection algorithm for the newest generation of geostationary imagers, including the GOES-16/17 Advanced Baseline Imager. The team will build on the experience gained during development and validation of fire detection products from Low Earth Orbiting satellites. Special attention will be devoted to minimizing false alarms, improving the detection envelope, optimizing processing to ensure near real-time availability, and delivering sub-pixel fire characterization.


GeoRing-ProxyVisible Satellite Imagery: Turning Night into Day with Machine Learning
Galina Chirokova, CIRA/CSU

Visible satellite imagery from geostationary satellites is used for a variety of applications, including identification of low cloud motions in and around the center of hurricanes. Since visible imagery relies in reflected sunlight, it’s only available during the day. This project seeks to use a machine learning method to bring in information from both geostationary and polar orbiting satellites to approximate visible imagery during the nighttime, then apply it to operational geostationary imagers around the globe. This will provide forecasters with a 24x7 source for both high and low cloud identification.


New Fused GEO+LEO Multi-Satellite Product: Stereo-Winds from Collocated ABI and VIIRS Datasets
Jaime Daniels, NOAA/NESDIS/STAR

The ability to view the same cloud from two different satellites at different look angles allows an accurate estimation of the cloud height using stereo techniques. The same method can be done with a combination of geostationary and low Earth orbiting satellites. This project aims to implement that methodology in regions where GOES-16, GOES-17, and the VIIRS instrument on NOAA-20 and SNPP overlap. The clouds can then be tracked in time to estimate winds with accurate height assignments, which in turn can be assimilated into numerical weather prediction models, leading to better weather forecasts.


Enhancing forecast applications of the GOES-R GLM in tropical cyclones using multi-platform data fusion and AI to assess environment and storm structure
Stephanie Stevenson, NOAA/NWS/NHC

This project involves use of the GOES-16/17 Geostationary Lightning Mapper, with the help of machine learning, to better understand the relationship between lightning and tropical cyclone behavior. If successful, the GLM information may improve the prediction of hurricane strength and intensity, which has the potential to save lives of people in hurricane-prone areas


Determination of Exospheric Neutral Hydrogen Density from GOES-R
Janet Machol, Univ. of Colorado

This project will use data from the Extreme Ultraviolet and X-Ray Irradiance Sensors (EXIS) from GOES-16/17 to determine daily neutral hydrogen (H) density distributions of the terrestrial upper atmosphere. Knowledge of the neutral H density distribution at the bottom edge of the exosphere (the exopause) is needed to set the upper boundary for the empirical MSISE model which provides the neutral temperature and densities in Earth's atmosphere from ground to thermospheric heights. MSISE is the international standard model used to estimate satellite drag for orbit calculations. Improved estimates would greatly help satellite operators and atmospheric modeling communities.


Utilizing geostationary satellite observations to develop a next generation ice cloud optical property model in support of JCSDA Community Radiative Transfer Model (CRTM) and JPSS CAL/VAL
Ping Yang, Texas A&M Univ

This project will use observations from the GOES-16/17 Advanced Baseline Imager to develop a state-of-the-art ice optical property model. This model will then be incorporated into the Community Radiative Transfer Model (CRTM), which is used by NOAA operational models as part of its data assimilation scheme. An improved CRTM will result in better weather forecasts via more accurate numerical weather prediction model forecasts.


Advanced Concepts Enabling Situational and Hazards Awareness via Imagery (ACES HAI)
Steve Miller, CIRA/CSU

This project aims to build on innovative work in imagery visualization techniques using the GOES-16/17 Advanced Baseline Imager. Some of the planned new applications include detection of lofted dust and snow, snow cover at night, improved nighttime low cloud detection, and spatial and temporal sharpening to approximate a mesoscale sector in regions where 1-minute data is not being collected. These improvements in ultimately result in better weather forecasts.


Use of modern geostationary data to improve a global diurnal warming model for multi-satellite data fusion
Andy Harris, Univ. of Maryland

This project will use Sea Surface Temperature (SST) data from current state-of-the-art imagers aboard geostationary satellites to develop a model for the daily rise and fall of SST. Frequent observations from geostationary sensors afford this capability. The diurnal heating model will in turn be used for a variety of applications, including applying it to the long-term satellite SST record in order to correct for time-of-day observations. The resulting dataset may be used for climate studies


Probabilistic nowcasting of aviation turbulence using deep learning applied to advanced geostationary imagery
Tony Wimmers, CIMSS/Univ. of Wisconsin

Aircraft turbulence is an important but difficult phenomenon to anticipate, particularly when it occurs in clear skies and fair weather conditions. This project will apply machine learning to data from the newest images aboard geostationary satellites in order to provide a probabilistic outlook for turbulence. Pilots can then use this information to alter their flight paths and avoid flying through regions where turbulence is expected.


Improved Monitoring of the Rapidly-Evolving Upper-Tropospheric Wind Fields over the Core of Hurricanes from High Spatiotemporal Resolution Geostationary Satellite Observations
Chris Velden, CIMSS/Univ. of Wisconsin

Geostationary satellite data from the newest generation of advanced imagers can be used to track the motions of clouds and calculate an estimate for the wind speed and direction in the atmosphere. This project aims to improve these Atmospheric Motion Vectors in the vicinity of tropical cyclones. The derived wind information will then be assimilated into numerical weather prediction models, resulting in a more accurate forecast for the track and intensity of the storm.


Enhancing Evapotranspiration and Evaporative Stress Index Data Products from GOES-16/17 Advance Baseline Imagers for NOAA NWP, NWM and Drought Monitoring Operations
Jerry Zhan, NOAA/NESDIS/STAR

Evapotranspiration (ET) is a significant component of the global and regional water cycle, and the GOES-16/17 Advanced Baseline Imager is capable of retrieving estimates of ET. This project involves building on previous work to develop a GOES-R ET and Drought product. The new product can then be used by numerical weather prediction models, resulting in an improved weather forecasts and improved drought monitoring.


Downscaling of GLM Lightning Observations Using ISS-LIS Data
Daile Zhang, Univ. of Maryland

The Lightning Imaging Sensor (LIS) currently rides aboard the International Space Station (ISS) collecting optical lightning data from storms below. At the same time, the GOES-16/17 Geostationary Lightning Mapper (GLM) provides constant hemispheric coverage, though at a more coarse spatial resolution thanks to its farther distance from Earth. This project will use machine learning techniques to combine lightning information from the two sensors when the ISS under-flies GOES-16/17, resulting in more robust and consistent lightning data for use by forecasters.


GOES High cadence Operational Total Irradiance (GHOTI)
Martin Snow, Univ. of Colorado

The Extreme Ultraviolet and X-ray Irradiance Sensors (EXIS) aboard the GOES-R series includes a quadrant diode that can be used to estimate Total Solar Irradiance (TSI). Having this capability in geostationary orbit allows for a continuous TSI dataset, avoiding gaps that are necessary with TSI measured by satellites in low Earth orbit. This project will create a long-term dataset of TSI which is important for monitoring climate.


Probabilistic Quantitative Precipitation Estimation with Geostationary Satellites
Pierre Kirstetter, Univ. of Oklahoma

This project builds on previous work to combine precipitation estimates from ground-based radars and the GOES-16/17 Advanced Baseline Imager. The ground-based radar data is part of the Multi-Radar/Multi-Sensor (MRMS) dataset. A machine learning methodology is used to produce a probabilistic precipitation estimation product that can be used by forecasters to help in their forecasts for heavy rain.


Assimilating GOES-R Latent Heating in FV3 using Machine Learning
Kyle Hilburn, CIRA/CSU

This project will use data from the GOES-16/17 Advanced Baseline Imager and Geostationary Lightning Mapper to estimate precipitation as indicated by ground-based radar reflectivity. The machine learning technique is trained using radar data and has been shown to produce realistic reflectivity estimates. Reflectivity is currently used to derive latent heating profiles and is assimilated into the RAP/HRRR operational numerical weather prediction model. Satellite estimates of latent heating will allow expanded coverage of the data that is assimilated. The team plans to work toward assimilation into the FV3 model, the future operational NOAA modeling system.


Assimilation of radiance tendency of water vapor bands from geostationary satellites using FV3GFS
Jim Jung, Univ. of Wisconsin

Traditionally, satellite information is assimilated into numerical weather prediction models using the observed radiances. This method requires a bias correction, which can reduce or compromise useful information in the observations. This project involves investigation of assimilating the radiance tendency, or time rate of change of the water vapor channel radiances on the newest generation geostationary imagers. It has the distinct advantage of avoiding the bias correction step, and may lead to more accurate weather forecasts from the models.