Disentangling timescales of terrestrial water storage variability

Understanding regional terrestrial water storage (TWS) is essential for efficiently managing water resources in a warming climate. The nearly twenty years of observations from GRACE/GRACE-FO missions have enabled unprecedented investigation into the spatiotemporal variability of TWS. However, it remains unclear to what extent regional TWS trends are persistent fingerprints of anthropogenic warming or natural variability in the Earth system. In particular, ambiguity arises in the potential aliasing of quasi-periodic processes operating on interannual to decadal timescales into observed decadal trends.

To alleviate this ambiguity, we leverage additional datasets that extend the TWS record in time: two statistical reconstructions of GRACE/GRACE-FO that extend the observed time-series into the past, and the suite of modeling results contained in the Community Earth System Model 2 Large Ensemble (LENS) Project. After establishing that these additional data reflect the observed variability, we assess the impact of record length on regional trends. Specifically, we address two questions:

1) Is the magnitude of observed trends anomalous relative to longer record lengths?

2) How does the dominant timescale of variability change as a function of record length and dataset?

Preliminary findings suggest that in most locations around the world, the magnitude of regional trends observed by GRACE/GRACE-FO are high compared to the reconstructions. In much of the Amazon, Central Africa, Southeast Asia, and India, the LENS dataset also shows lower variability than the observations. With spectral analysis, we find that much of these regions are dominated by interannual periodicities. Regression against climate indices gives further evidence that interannual periodicities are more important than decadal in explaining TWS variability. Such results suggest that the strong ENSO events in 2010-2011 and 2015-2016 were an important component of observed regional trends and that further observations are needed to disentangle climate variability from secular change.

© 2022. California Institute of Technology. Government sponsorship acknowledged.