Ocean mass variations from GRACE and tsunami gauges

Abstract

We compared the mass variations from Gravity Recovery and Climate Experiment (GRACE) with those measured by bottom pressure recorders (BPRs) deployed as tsunami gauges. Prior to the comparison, we estimated the degree‐1 Stokes coefficients by the degree‐1 loading method using both the GPS and GRACE data. We successfully recovered the degree‐1 Stokes coefficients that are consistent with those from the satellite laser ranging analysis and from the loading model. We found that the use of the GRACE data in the degree‐1 loading method significantly mitigates the aliasing effect from the unmodeled higher‐degree loading deformations. Then, we compared the mass variations from GRACE with those from the five BPRs that have been deployed as part of the Deep‐ocean Assessment and Reporting of Tsunamis (DART) system. We observed a fairly good (over 0.4) correlation except for one station that is close to the U.S. mainland. However, we found that the GRACE‐derived mass variations underestimate those from BPRs by nearly a factor of 2. Through numerical simulation, we found that the artificial leakage from the landmass is responsible for the low correlation observed at the station, and the spatial averaging in the GRACE processing may be responsible for the underestimation of the mass variations from BPRs by GRACE. These results illustrate the importance of performing a simulation for these effects with numerical models before comparing the GRACE‐derived mass variations with those from pointwise measurements.