Presentation Type
Poster
Faculty Mentor’s Full Name
Hilary Martens
Faculty Mentor’s Department
Geosciences
Abstract / Artist's Statement
Surface mass loads, such as oceans, atmosphere, glaciers, snowpack, and freshwater exert forces on Earth’s surface, causing the crust to change shape, also known as crustal deformation. The response of Earth’s crust to these mass loads is often easily predicted and removed from crustal measurements, except atmospheric response, which is much more variable from weather conditions. The Global Positioning System (GPS), which transmits radio signals between satellites and ground receivers, can monitor millimeter changes to the crust’s shape. This data is recorded by over 1200 GPS stations from the Plate Boundary Observatory network all across the western United States. However, methods for processing raw GPS data differ among analysis centers, producing different positions. One difference in GPS processing methods is the treatment of satellite signals delayed in the troposphere, the region of the atmosphere with the most mass. This research is significant to helping develop a set of standard methods for GPS processing and therefore making GPS more accurate. We investigated inconsistencies in three-dimensional (east, north, up) positions from five datasets created by the following three processing centers: NASA’s Jet Propulsion Laboratory, the Nevada Geodetic Laboratory, and UNAVCO consortium. We used the software program LoadDef to model the response of Earth’s surface to changes in atmospheric pressure at each station, which were then compared with observed GPS positions. We find a trend that GPS datasets produced with highly simplified assumptions about tropospheric delays have smaller reductions in root-mean-square scatter after correcting for atmospheric loading. Up to 50 % of the scatter in the residual GPS time series normally considered noise can be explained by crustal deformation responses from atmospheric mass loading. We conclude that differences in GPS time-series solutions can be partially explained by the temporal treatment of tropospheric signal delays during initial processing of raw GPS data.
Category
Physical Sciences
Exploring Effects of GPS Processing on Atmospheric Responses of Earth Deformation
Surface mass loads, such as oceans, atmosphere, glaciers, snowpack, and freshwater exert forces on Earth’s surface, causing the crust to change shape, also known as crustal deformation. The response of Earth’s crust to these mass loads is often easily predicted and removed from crustal measurements, except atmospheric response, which is much more variable from weather conditions. The Global Positioning System (GPS), which transmits radio signals between satellites and ground receivers, can monitor millimeter changes to the crust’s shape. This data is recorded by over 1200 GPS stations from the Plate Boundary Observatory network all across the western United States. However, methods for processing raw GPS data differ among analysis centers, producing different positions. One difference in GPS processing methods is the treatment of satellite signals delayed in the troposphere, the region of the atmosphere with the most mass. This research is significant to helping develop a set of standard methods for GPS processing and therefore making GPS more accurate. We investigated inconsistencies in three-dimensional (east, north, up) positions from five datasets created by the following three processing centers: NASA’s Jet Propulsion Laboratory, the Nevada Geodetic Laboratory, and UNAVCO consortium. We used the software program LoadDef to model the response of Earth’s surface to changes in atmospheric pressure at each station, which were then compared with observed GPS positions. We find a trend that GPS datasets produced with highly simplified assumptions about tropospheric delays have smaller reductions in root-mean-square scatter after correcting for atmospheric loading. Up to 50 % of the scatter in the residual GPS time series normally considered noise can be explained by crustal deformation responses from atmospheric mass loading. We conclude that differences in GPS time-series solutions can be partially explained by the temporal treatment of tropospheric signal delays during initial processing of raw GPS data.