Using Pacific Ocean Sea Surface Temperatures to Improve Lead Time for Streamflow Estimates
Abstract
Due to the increased climate variability, there is stress on the available water resources. Improving streamflow forecast lead time aid water managers in providing useful information for the planning and management of water resources. In the current study, we propose a time lagged analyses between the Pacific Ocean sea surface temperatures (SSTs) and water year streamflow volume for improving the forecast lead time. Singular Value Decomposition (SVD) statistical technique is used to identify coupled regions of SST and four streamflow unimpaired gages in Truckee River Basin, located in the western United States, for a 50-year period (1960-2009). The significant SST regions along with the predefined climate patterns i.e., Pacific Decadal Oscillation and El Niño-Southern Oscillation are used as predictors in a data-driven model, Support Vector Machine (SVM), for providing water year streamflow volumes with 8-11 months lead time. The SVD 1st temporal expansion series explained 86% of the variability in streamflow. Additionally, the results indicated improved streamflow forecasts using SST information compared to using only predefined climate patterns. The SVD based SVM technique is simple and parsimonious and could result in improvement in lead time for streamflow estimates for other river basins.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFM.H21C1190K
- Keywords:
-
- 1807 HYDROLOGY / Climate impacts;
- 1833 HYDROLOGY / Hydroclimatology;
- 1860 HYDROLOGY / Streamflow;
- 3305 ATMOSPHERIC PROCESSES / Climate change and variability