Automated short-term hydrological forecast system

While working in State Hydrological Institute I developed automated system for water levels and discharges forecast at small rivers of North-West Russia.

System is based on opensource models HBV, GR4J and SIMHYD, which were calibrated with machine learning techniques.

WRF forecast was used for precipitation and temperature predictions - they are needed as inputs for hydrological models.

Public web-frontend provides operative information (forecast for 3 days) and all retrospective data. Also email notifications for dangerous water levels were implemented.

Some details can be found here: Automated short-term forecast system based on open-source hydrological models for the Tikhvinka river (Leningrad region of Russia)