The outcome of the pilot will be a dedicated web service providing:

• Monthly 90-day seasonal ensemble forecasts of streamflow and hydropower reservoir inflow at end-user (Kemijoki Oy) selected forecast points. A particular focus will be placed on snowmelt driven runoff forecasting.
• Daily 10-day deterministic forecasts (incl. nowcasts and hindcasts) of streamflow and hydropower reservoir inflow at end-user (Kemijoki Oy) selected forecast points. In addition to streamflow forecasts, the service will provide maps of numerous other key hydrological variables used in Hydropower production optimization and general water management operations.
• A unified web service combining data from various other external sources in a single user interface allowing the user to assess forecasts uncertainties by comparing information derived from an FMI hydrological model (HOPS) with external observations and estimates – a particular focus will be placed on providing EO-based observations and in situ measurements of snow conditions and soil freeze/thaw states

A first realease version is at the moment running on hops.fmi.fi

• Most of the preliminarily planned products have been added to the WebApp

An expanded and improved version of the WebApp is currently being developed

• The new and improved service will be made public during the spring of 2022.

Availability: public

• 20-day hindcast maps
• 10-day forecast maps
• Hindcast and deterministic forecast time series plots
• 90-day ensemble forecast time series and probabilistic output

A Web Application service has been developed. The front-end web service is run on an FMI internal server (hops.fmi.fi). The back end of the service is run on another FMI internal server.

A Web Application: hops.fmi.fi

Kemijoki Oy is the most important producer of hydropower and regulating power in Finland. Kemijoki Oy operates 16 hydropower plants and 4 regulated reservoirs in the Kemijoki watercourse area. Currently, hydropower production accounts for approx. 19% of total energy production in Finland, which whilst on-par with European averages is quite low when compared to the neighbouring Scandinavian countries. For example, in Sweden hydropower accounts for nearly half of the country’s electrical energy production. As with other hydropower operations elsewhere, the river’s natural conditions during different seasons of the year are the most important factor affecting the planning of hydropower production. In addition to fluctuations in natural conditions, hydropower plant operations planning involves other holistic considerations such as energy consumption demand and the many users of the rivers, for example, bobbers, fishermen, and boaters. Streamflow forecasts are as such a crucial part in designing reservoir operations and flow regulation plans in the Kemijoki basin and therefore improved forecasts on snowmelt induced peak flow volumes are of considerable interest to both local authorities as well as the governing hydropower company, Kemijoki Oy.

In most basins in Scandinavia and Finland, more than 50 % of the annual precipitation falls as snow. Errors in snowmelt timing and melt rate simulations are one of the largest sources of stream flow prediction errors in snowmelt-dominated watersheds. Snow storage and melting periods have a significant impact on hydropower production reservoirs. Forecasts of reservoir inflow and energy prices are used to schedule the quantity and timing of releases for daily, weekly, and seasonal operations. Assimilation of Earth Observation (EO) based snow observations can significantly reduce hydrological model snowpack simulation uncertainties. Currently Copernicus EO-based Snow Water Equivalent GlobLand SWE (https://land.copernicus.eu/global/products/swe) data is operationally assimilated into the HOPS modelling procedure. The use of EO based snow observations offer two main types of advantages: 1) EO based snow observations are independent from point scale observations (usually used to drive hydrological models), and 2) they provide basin wide information and data on the state of the snowpack as opposed to point scale observations. As part of the E-SHAPE project, assimilation of EO-based snow cover data into the HOPS’ hydrological modelling component is an integral part of the project. FMI’s indirect EO-based SWE assimilation method relies on constraining the simulated SWE to match observed EO-based SWE by manipulating snow season precipitation forcing data.

All data provided by the web service will be made available for download through the web service. The pilot’s source code can be made available on GitHub after the service is finalized (this still needs to be discussed internally at FMI).

If the pilot is deemed of interest for the GEO/GEOSS catalogue, registration will be established, especially in connection with Finnish portals. This will be assessed after full launch of the service, during the Summer of 2022.