Among the several tools developed in D-HYDROFLEX project, one of the key tools is the weather and water flow forecasting tool (HYDRO-WFF). This tool is able to first, forecast the weather, specially the rainfall, at a desired region. And second, to use the weather along the river course to predict the runoff river that arrives to a specific hydro power plant (HPP). This tool is crucial for a correct water administration and for an efficient HPP operation planning.

#Description

The flow forecasting is crucial for a correct planning and production forecasting in a HPP as knowing the water availability is key for the management of this resource. One of the solutions of the D-HYDROFLEX project that address this issue is the Weather and water flow forecasting tool, the HYDRO-WFF tool.

The HYDRO-WFF is the combination of three different methodologies: weather downscaling, spatial-temporal downscaling and the rainfall-runoff model, that allows it to provide HPP operators with very useful information to operate the plant and could allow him to optimize the usage of the water resources.

With the objective of obtaining a water flow forecasting, obtaining a weather prediction is mandatory, as the climate conditions are the main aspect to affect the water flow of a river. The HYDRO-WFF is able to predict the climate condition such as precipitation, wind speed, temperature, etc. and use all this information to obtain reliable run off river forecasts.

The weather predictions are obtained following a downscaling strategy. First, General Circulation Models (GCM) prediction and historical data were gathered from open sources. This data is obtained following very complex climate models that, unfortunately can’t give very high-resolution information, both temporal and geographical, but are the best starting point for weather predictions. Then, climate stations are located around the geographical area of interest and climate variables’ historical data are obtained from them.

The weather downscaling consists of mapping the historical GCM data to the station level data, with the help of some other parameters or inputs that help to correlate the both of them. This way, the mapping could extend the GCM prediction data to predict the climate variables’ behaviour at climate station level.

Once the weather downscaling is finished, a downscaling methodology is followed. Similar to the previous one, the spatial-temporal downscaling consists of the obtention of climate historical and prediction data, this time from the climate stations around the area of interests, and then extend them to this specific area.

The data used this time consists on mainly geographical characteristics (coordinates, altitude, etc.) of several climate stations around the area of interest and objective temporal frequency, but some relevant climate data at the climate station could be used. Lastly, a new mapping will be used to predict the climate variables at the geographical area of interest with the time resolution required.

Lastly, a rainfall-runoff model, is used to balance the water flow in the river ecosystem.  This is, the precipitation, the river water flow, the evaporation, the water used for irrigation and the soil moisture. For this, a total of four types of water storages were modeled: the snow storage is the solid water accumulated through the year. The soil moisture storage is the water contained in the first layers of soil in contact with the air and the river. The aquifer storage is the water is the underground water. And lastly, the reservoir storage refers to the water contained within a section of a river, dammed water, natural lakes, etc.

All interactions between these four storages have being carefully considered and modeled. Likewise, the effect of climate conditions is considered in each case. Another important factor is the human action. For instance, crops irrigation near the river side, affects the soil moisture and water concentration.

The tool was validated using the data of one of the demos in the D-HYDROFLEX project: Salto de Touro hybrid HPP in Spain, which provides data enough to validate every step of the HYDRO-WFF tool. Moreover, the combination of this tool with others developed in the project as the HYDRO-SDS has given the big potential of this tool for plant operation, in particular for hybrid plants.