Skip to yearly menu bar Skip to main content


Poster
in
Workshop: Tackling Climate Change with Machine Learning: Global Perspectives and Local Challenges

Attention-based Domain Adaption Forecasting of Streamflow in Data Sparse Regions

Roland Oruche · Fearghal O'Donncha

Keywords: [ Time-series analysis ] [ Disaster management and relief ] [ Meta- and transfer learning ]


Abstract:

Streamflow forecasts are critical to guide water resource management, mitigate drought and flood effects, and develop climate-smart infrastructure and industries. Many global regions, however, have limited streamflow observations to guide evidence-based management strategies. In this paper, we propose an attention-based domain adaptation streamflow forecaster for data-sparse regions. Our approach leverages the hydrological characteristics of a data-rich source domain to induce effective 24h lead-time streamflow prediction in a limited target domain. Specifically, we employ a deep-learning framework leveraging domain adaptation techniques to simultaneously train streamflow predictions and discern between both domains using an adversarial method. Experiments against baseline cross-domain forecasting models show improved performance for 24h lead-time streamflow forecasting.

Chat is not available.