As part of the WATExR project, we have developed workflows for seasonal forecasting of river discharge, lake water temperature, ecology and fish phenology. The aim was for stakeholders to have access to probabilistic aquatic forecasts driven by state-of-the-art seasonal climate projections. Seasonal forecasts provide an indication of the expected average environmental conditions during the coming 1 to 9 months, and have the potential to assist in strategic management of catchments, lakes and reservoirs.
The project’s current webpage is here, where you can find out more information about the case study sites, development team, as well as follow our blog. This website hosts the tools and code developed as part of the project, and will remain active after the official project webpage becomes inactive.
Forecasting tools were developed for five pilot case sites, one in Australia and four in Europe. At most sites, seasonal forecasts were only produced for a historic hindcast period (1993-2016/2019) to evaluate their potential usefulness. However, at several sites there is an intention to continue development to operational tools - see details for the individual Case Study sites below.
Watch this short video for an introduction to the project.
Forecasts were produced by driving freshwater “impact” models using downscaled seasonal climate model forecasts.
A more detailed flow chart detailing the workflow for producing forecasts for a hindcast period is shown below (click to enlarge).
Accessing and processing seasonal climate data
We used ECMWF’s SEAS5 seasonal climate model forecasts. These can be accessed directly through the Copernicus Climate Data Store, and you will find an example script for downloading data directly from Copernicus here). However, in WATExR, seasonal climate data were primarily downloaded directly from the Santander Meteorology Group’s User Data Gateway, using R scripts in this folder. These scripts make use of the Climate4R package both for data download and post processing. This data is only historic and was used for the development and evaluation of the forecasting tools, and therefore cannot be used for operational forecasting. Many of the functions in the scripts in this folder can however be reused for operational forecasting.
Statistical and process-based models were used to produce seasonal forecasts for variables that were relevant in the various case study sites. These included, for example, catchment hydrology, lake temperature, lake ice cover, lake water quality and ecological status, and the timing of seaward fish migration.
Case study sites
Burrishoole catchment, Ireland
The challenge: Better understanding and management of diadromous fish stocks, in particular the timing of fish migration.
Forecasting tool: All code and documentation required for data access, pre-processing and statistical analyses are available via download of the EcoCountForecastR R package available at https://github.com/as-french/EcoCountForecastR
Developer and co-developer: Marine Institute
Main contacts: Andrew French, Elvira de Eyto
Sau reservoir, Spain
The challenge: Improved reservoir management to reduce flooding and improve water quality for drinking water and to meet ecological targets.
Forecasting tool: A QGIS plugin is under development (expected first version summer 2021).
Developer and co-developer: ICRA, Catalan Water Agency
Main contacts: Daniel Mercado Bettín, Rafael Marcé
Lake Vansjø, Norway
The challenge: Manage lake water levels and farming practices in the catchment to improve water quality and achieve water quality and ecology targets, including prevention of toxic cyanobacterial blooms.
Forecasting tool: https://watexr.data.niva.no. This was developed using a voila app; the underlying code is in this GitHub repository
Developer and co-developer: Norwegian Institute for Water Research (NIVA), Morsa
Main contacts: Leah Jackson-Blake, François Clayer
Wupper reservoir, Germany
The challenge: Improved reservoir operations to meet requirements for flood protection, recreation and improved water quality both in the reservoir and downstream.
Forecasting tool: R shiny app coming soon (expected summer 2021)
Developer and co-developer: UFZ, WUPPERVERBAND
Main contacts: Muhammed Shikhani, Karsten Rinke
Mount Bold reservoir, Southern Autralia
The challenge: Improve management of the largest reservoir in South Australia to reduce pumping costs and improve water quality.
Forecasting tool: not developed for this site
Developer and co-developer: Dundalk Institute of Technology, SA Water and University of Adelaide
Main contacts: Tadhg Moore, Eleanor Jennings
To find out more…
During the last year of the project, we organized a webinar (February 2021) in which we introduced the aims of the project and showcased some of the main outcomes and results. The webinar included a round table with stakeholders to discuss opportunities and limitations of the use of seasonal forecasting to support water management. The complete recording is available at http://u.pc.cd/3MNctalK.
A number of papers are currently published/submitted, and provide details on the workflows developed at each site, as well as lessons learned throughout the project in terms of key opportunities and barriers for seasonal forecasting to support water management. All will be published open access:
- Mercado et al. (2021), “Forecasting water temperature in lakes and reservoirs using seasonal climate prediction”, Water Research: A description of the workflows developed to forecast discharge and lake temperature.
- Clayer et al. (in prep): Exploration of the sources of skill in seasonal forecasts.
- Jackson-Blake et al. (submitted): Assessment of how useful forecasts are for supporting decision making in the water sector.
Deliverables produced during the course of the project are available in this folder.
This is a contribution of the WATExR project (watexr.eu/), which is part of ERA4CS, an ERA-NET initiated by JPI Climate, and funded by MINECO-AEI (ES), FORMAS (SE), BMBF (DE), EPA (IE), RCN (NO), and IFD (DK), with co-funding by the European Union (Grant 690462). MINECO-AEI funded this research through projects PCIN-2017-062 and PCIN-2017-092.