Discharge Estimation Using Satellite Gravity During Flood Seasons at the Óbidos Gauge Station, Amazon River

Reliable measurement of river discharge is important for management of water resource and understanding of hydrological cycles particularly associated with global and regional climate changes. Practically, to obtain continuous time series of river discharge, regression analysis of an empirical relationship between accumulated water level and discharge data is used. During wet season, however, the relationship includes more uncertainty due to the difficulty of accurate discharge measurement. This is particularly true for the Amazon River because significant amount of water flows outside river channel during flooding.

For an alternative way to estimate river discharge, we use GRACE time-varying gravity measurement from January 2003 to December 2012. We first apply Empirical Orthogonal Function (EOF) for GRACE time-varying gravity fields in Amazon and successfully isolate gravity signal in the main stream. The EOF time series represents relative river discharge variations without larger uncertainty during flooding season compared to conventional in-situ discharge estimate. Estimates of Amazon River discharge based on GRACE data are very close to those from observed at gauge stations during dry seasons. However, our estimates are larger than in-situ data in high water seasons, and the difference is the maximum at the 2009 flooding. This is probably because in-situ observation underestimates river discharge during wet season due possibly to detoured water in river pathway developed during flooding while GRACE observes integrated water mass variations in river channels.