Operational Discharge Forecasting in Central Asia using Deep Learning

Accurate short-term streamflow forecasts are critical for water resources management, hydropower operations, and early warning of hydrological hazards. This need is particularly pronounced in Central Asia, where water is predominantly stored as seasonal snow and glacier ice in the high mountain region and released during the warm season, sustaining extensive irrigated agriculture and hydropower production in the region.  

The Swiss Agency for Development and Cooperation supports the strengthening of the operational hydrological forecasting capabilities of National Hydrometeorological Services across Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, and Uzbekistan (SAPPHIRE Central Asia project). SAPPHIRE co-designs, develops, and deploys operationally open-source forecasting tools that integrate existing and machine-learning-based forecasting methods in these organizations. 

As part of this work, we evaluate the performance of three state-of-the-art time-series forecasting architectures—the Temporal Fusion Transformer (TFT), the Time-Series Dense Encoder (TiDE), and the Time-Series Mixer (TSMixer)—for operational 10-day-ahead streamflow prediction across more than 100 gauges in Kazakhstan, Kyrgyzstan, and Tajikistan. 

The models are trained jointly across all basins within each country to enhance spatiotemporal generalization. Probabilistic forecasts are produced using a quantile loss function, thereby representing aleatoric uncertainty. Model skill is assessed against observed discharge and benchmarked against periodic linear regression models for both 5-day and 10-day averaged forecasts. 

Results indicate that all three deep learning models consistently outperform the existing benchmark approaches, with particularly pronounced improvements at the  
10-day forecast horizon. For example, in Kyrgyzstan and Tajikistan, mean absolute errors get reduced by 30% - 37%. The auto-regressive information from past discharge emerges as the most influential predictor, underscoring its central role in snow- and glacier-melt-dominated runoff regimes of high-mountain Central Asia. 

These advances directly strengthen the forecasting capacity of the Central Asian Hydrometeorological Services and improve the quality of information available to their diverse user base—including national water management authorities responsible for irrigation allocation, hydropower operators optimizing reservoir releases, agencies managing climate-sensitive infrastructure such as roads and airports, and transboundary water management institutions like the Interstate Commission for Water Coordination (ICWC). By demonstrating the operational viability of modern deep learning approaches within existing institutional frameworks, this work contributes to more reliable and actionable hydrological information across one of the world's most water-stressed transboundary regions.