Session 9: Improved Geolocation of Satellite Measurements Using Bayesian Hierarchical Models
Presentation Type
Oral
Track
Other
Abstract
The Global Ecosystem Dynamics Investigation (GEDI) is a light detection and ranging (lidar) instrument onboard the International Space Station (ISS). GEDI was designed to collect near global measurements of forest structure, supporting our understanding of forest ecosystems and the carbon cycle. As the ISS orbits, laser pulses emitted from GEDI impact the Earth’s surface within 25-meter diameter circular “footprints”. The laser pulses reflect back to GEDI as a waveform, providing the vertical distribution of above ground matter within the footprint. However, determining the exact geographic coordinates of a laser pulse's impact is difficult: Subtle vibrations in the ISS and deviations in the orbit path can cause substantial geolocation errors in GEDI measurements. Currently, the geolocation error of GEDI measurements is believed to have a standard deviation of 10 meters, though even determining a general statistic such as this is challenging.
We implement a Bayesian hierarchical model using GEDI measurements and ground-truth data to make inference on the true locations of GEDI measurements and to reveal patterns in the geolocation errors. The refined geolocation improves the quality of to-date GEDI measurements. Further, the revealed patterns in the geolocation errors suggest physical root causes, informing the improved engineering of future instruments.
Start Date
2-6-2024 2:30 PM
End Date
2-6-2024 3:30 PM
Session 9: Improved Geolocation of Satellite Measurements Using Bayesian Hierarchical Models
Pasque 255
The Global Ecosystem Dynamics Investigation (GEDI) is a light detection and ranging (lidar) instrument onboard the International Space Station (ISS). GEDI was designed to collect near global measurements of forest structure, supporting our understanding of forest ecosystems and the carbon cycle. As the ISS orbits, laser pulses emitted from GEDI impact the Earth’s surface within 25-meter diameter circular “footprints”. The laser pulses reflect back to GEDI as a waveform, providing the vertical distribution of above ground matter within the footprint. However, determining the exact geographic coordinates of a laser pulse's impact is difficult: Subtle vibrations in the ISS and deviations in the orbit path can cause substantial geolocation errors in GEDI measurements. Currently, the geolocation error of GEDI measurements is believed to have a standard deviation of 10 meters, though even determining a general statistic such as this is challenging.
We implement a Bayesian hierarchical model using GEDI measurements and ground-truth data to make inference on the true locations of GEDI measurements and to reveal patterns in the geolocation errors. The refined geolocation improves the quality of to-date GEDI measurements. Further, the revealed patterns in the geolocation errors suggest physical root causes, informing the improved engineering of future instruments.