The Value of Hydroclimatic Teleconnections for Snow-based Seasonal Streamflow Forecasting
Preprint
- 26 July 2024
- preprint
- Published by Copernicus GmbH in EGUsphere
- Vol. 2024, 1-19
Abstract
Due to the long memory of snow processes, statistical seasonal streamflow predictions in snow-dominated catchments typically rely on snowpack estimates. Using mountainous catchments in Central Asia as a case study, we demonstrate how seasonal hydrological forecasts benefit from incorporating large-scale climate oscillations (COs). First, we examine the teleconnections between the major COs and peak precipitation season in eight catchments across the Pamir and Tian-Shan mountains from February to June. We then employ a machine learning framework that incorporates snow water equivalent (SWE) and dominant COs indices as predictors for mean discharge from April to September. Our workflow leverages an ensemble technique that uses multiple SWE estimates from near-time global data sources and diverse types of explainable machine-learning models. We find that the winter states of the El Niño-Southern Oscillation and the North Atlantic Oscillation enhance SWE-based forecasts of seasonal discharge in the study catchments. We identify three instances in which the inclusion of COs as additional predictors could be instrumental for snowpack-based seasonal streamflow forecasting: 1) when forecasts are issued at extended lead times and accumulated SWE is not yet representative of seasonal terrestrial water storage; 2) when climate variability during the forecasted season plays a larger role in shaping seasonal discharge; and 3) SWE estimates for a catchment are subject to larger uncertainty. Our approach provides a novel way to reduce uncertainties in seasonal discharge predictions in data-scarce snowmelt-dominated catchments.Keywords
All Related Versions
- Published version: Hydrology and Earth System Sciences, 29 (14), 3055.
Funding Information
- Volkswagen Foundation (96 264)
This publication has 31 references indexed in Scilit:
- Comparing the performance of high‐resolution global precipitation products across topographic and climatic gradients of Central AsiaInternational Journal of Climatology, 2022
- Ensemble machine learning paradigms in hydrology: A reviewJournal of Hydrology, 2021
- A seasonal forecasting procedure for reservoir inflows in Central AsiaRiver Research and Applications, 2019
- Machine learning of large‐scale spatial distributions of wild turkeys with high‐dimensional environmental dataEcology and Evolution, 2019
- Spatiotemporal characteristics of seasonal precipitation and their relationships with ENSO in Central Asia during 1901–2013Journal of Geographical Sciences, 2018
- Climate change and adaptation of mountain societies in Central Asia: uncertainties, knowledge gaps, and data constraintsRegional Environmental Change, 2018
- MODSNOW-Tool: an operational tool for daily snow cover monitoring using MODIS dataEnvironmental Earth Sciences, 2016
- Regional climate model simulation of winter climate over Central–Southwest Asia, with emphasis on NAO and ENSO effectsInternational Journal of Climatology, 2009
- How ENSO impacts precipitation in southwest central AsiaGeophysical Research Letters, 2007
- Mountains of the world, water towers for humanity: Typology, mapping, and global significanceWater Resources Research, 2007