European hydropower potential has been explored at length in recent years but a thorough analysis of pumped storage potential has never really been addressed. Although research into potential pumped storage capacity, or the transformation to pumped storage, has been carried out by private companies or at regional or national levels, this analysis hasn’t been carried out for the whole of Europe.
As environmental considerations, a lack of adequate sites and certain social acceptance issues limit the development of further conventional hydropower in Europe, attention is now turning to pumped storage.
New pumped storage schemes are subject to similar limitations as conventional hydro projects. However, this is unlikely to be the case for pumped storage resulting from the transformation of existing hydropower and non-hydropower reservoirs. The reasons for this include that an existing reservoir candidate for pumped storage has already been established so there are no new environmental effects, and is currently forming part of a more stable system where – it is hoped – ecosystems have adapted and any problems have been alleviated.
Pumped storage is only as renewable as the electricity used to pump the water up for the first time, minus cycle efficiency losses of 15-30%. So while it may not necessarily add more electricity of a renewable origin, its greatest asset is its ability to allow the integration of the variable forms of renewable generation.
Energy storage is one of the three main options to enable a higher uptake of variable renewable power such as wind and solar into the energy system (the other two are improved interconnections and more flexible conventional power plants). Pumped storage is currently seen as the only energy storage technology able to provide the large storage needed for accommodating renewable electricity under 2020 EU energy targets. It is the largest and most mature form of energy storage currently available but the capital costs required can be very high and the availability of suitable sites is decreasing. Therefore the identification of any remaining sites is becoming vital so that the most beneficial location is chosen in terms of capacity and economics.
In the absence of a Europe-wide assessment on pumped storage potential, a new methodology has recently been developed. This has been produced under the umbrella of the European Commission’s Strategic Energy Technologies Information System (SETIS), by the Joint Research Commission (JRC) in collaboration with University College Cork in Ireland. The goal has been to develop a geographical information system (GIS)-based methodology and a model to identify the potential for transforming single reservoirs into pumped storage systems. It will then assess the additional energy storage which these new schemes could contribute to electricity systems. The methodology was applied to case studies in Croatia and Turkey. Currently the Joint Research Centre is applying the model to the whole of Europe and expects to obtain preliminary results by the end of the year. The authors believe that future improvements to the model could be effectively applied to the whole of the EU with minimum effort, thus obtaining a more accurate assessment.
Published in 2012 and entitled Pumped hydro energy storage: potential for transformation from single dams, the report is co-authored by Roberto Lacal Arantegui from the Institute of Energy and Transport at the EC’s JRC, along with Niall Fitzgerald and Paul Leahy from the University College Cork’s Sustainable Energy Research Group.

Pumped storage transformation

The report defines a methodology for the transformation to pumped storage capabilities. It was decided that two topologies could be developed in order to analyse if this could be achieved:
– Topology A (TA) is when a reservoir already exists and a second reservoir needs to be added, normally at a higher elevation, along with the penstock and equipment. An example of a TA is Turlough Hill scheme in Ireland, where a natural lake formed the lower reservoir and an upper reservoir was constructed.
– Topology B (TB) is when two reservoirs already exist and are within suitable distance and difference in elevation. A TB consists of adding penstock, generation and pumping equipment between them. Existing natural lakes might be considered one of the two reservoirs and these dams might be in the same river or in parallel valleys. The 480MW Limberg II plant in Austria is an example of a TB transformation.
Croatia was considered a suitable case study due to its size and the high penetration of hydropower in the country. Croatian hydropower accounts for 2076MW of installed power, supplying 31% of annual power consumption. The country already has three pumped storage plants in operation:
– RHE Velebit – 276MW with a pumping capacity of 240MW.
– Fuzine – 4.6MW with a pumping capacity of 4.8MW.
– Lepenica – 1.14MW with a pumping capacity of 1.25MW.
Meanwhile it also has a strong commitment to a 20% share of renewable energy in total consumption by 2020. Another advantage is the country’s numerous electricity interconnections. It is directly interconnected to Slovenia, Bosnia and Herzegovina, Serbia and Hungary, which creates the potential to store surplus wind generation from these neighbouring countries.
Turkey was selected as the second suitable case study country due to its large number of dam sites (260 large dams and 413 small dams in operation, with 1418 planned hydro plants). This could equate to a large number of potential transformation sites. Although hydropower currently supplies 20% of Turkish power demands, it has been suggested that only 35% of estimated economic hydropower potential is utilised in the country. The Turkish government hopes that hydro capacity will increase to 35,000MW by 2020. In addition, government plans to further increase wind power to 20,000MW by 2023 make Turkey an ideal pumped storage transformation country.
The methodology used for the transformation of existing reservoirs into pumped storage, using both TA and TB, is shown in figure 1. First of all topography and physical characteristics for transformation must be defined and assumptions made on environmental factors, distance to protected natural spaces and electricity grids etc. Each site can then be assessed in a uniform and consistent manner.
Dams and hydropower schemes with a water storage capacity of less than 1Mm3 and a nominal electricity capacity of less than 1MW were excluded from the study. All dam types were considered suitable for transformation and a head of 150m or more was required otherwise transformation to pumped storage would not be viable. A distance of 5km or more between the two reservoirs is required and an indicative minimum reservoir surface is 50,000m2 (70,000m2 for TA). Other factors which may deem a potential development unviable include:
– Presence of an inhabited site within 200m of a new construction (penstock or reservoir). For TB where both dams already exist there should be no generation, pumping and penstock placed within 200m of an inhabited site.
– Transport infrastructure present within 100m of a transformation site.
– Suitable grid infrastructure located more than 20km away from a non-hydro dam.
Geographic Information Systems (GIS) shapefiles were used to build up a full country map for each of the proposed countries. Digital terrain maps provide topographic information but additional information is also required on river and water bodies, elevation data, inhabited sites, environmental sensitivity, and electricity grids both at transmission and distribution levels.
Although it is not geo-referenced, ICOLD’s World Register of Dams was considered to be the most comprehensive database available and proved to be the primary data source for this study. Missing data gaps were addressed using the Global Reservoir and Dam (GRanD) database. However as only 20% and 15% of dams in Croatia and Turkey are geo-referenced through GRanD, the remaining dams had to be done manually using Google Earth. This was a time consuming exercise and although care was taken there may be errors in the data.
Remotely sensed elevation data was obtained from the Shuttle Radar Topography dataset. GIS shapefiles of electricity infrastructure in Croatia and Turkey could not be obtained. Although maps were obtained from the Global Energy Network Institute these are of a limited accuracy for the purpose of this study.

Country potential
Twenty-three dams have a reservoir capacity of more than 1Mm3 in Croatia, and all of these were analysed in the GIS model. More dams are also at an elevation of 101-200m then at any other range. Results for Croatia found that there are no solutions for TB sites under the constraints applied. Any dams within 5km of each other do not have an elevation difference of 150m or greater to provide sufficient head. The methodology’s constraints have to be relaxed to allow a buffer distance of 11km before a TB solution can be found.
Within Croatia there are 13 prospective TA sites with up to 60GWh of energy storage. Under the limitations of this study it was concluded that the potential for transformation to pumped storage in the country was at least three times the capacity of existing pumped storage plants.
A total of 612 reservoirs larger than 1Mm3 were analysed in Turkey, and a large proportion of dams are at elevations of 0-400m and between 801-1200m. A total of 448 potential TA sites with an energy storage of 4372GWh existed before nature protection areas were excluded. With environmental constraints this total dropped to 444 potential sites and a significant loss of 555GWh of energy storage.
Under topology B analysis (with nature protection areas) there were only two potential sites from a total of 612 dams under analysis with an average head of 294m and energy storage of 3.04GWh.


The model developed in the study identified transformation sites based on head difference, distance between existing and potential sites, flatness of the topography surrounding the proposed second reservoir and reservoir volume. It also implemented constraints related to the construction of new reservoirs in relation to inhabited sites etc. However the model is unable to analyse potential sites based on their geology and hydrology – both of which could be potential barriers to development of pumped storage. Indeed the following have all been identified as possible barriers to development:
– Geology: A detailed geologic analysis of each potential transformation site would need to be performed to assess its feasibility for transformation to pumped storage. Construction of underground penstocks may also be hindered by geology.
– Hydrology: A lack of surface water at or near to the potential transformation site could be a potential barrier. An analysis of the hydrology of the existing reservoir is needed to identify if there are seasonal variations in the supply and level of the water. This is a critical criterion in areas and countries with dry climates, such as those in Southern Europe. For example dams in Cyprus are rarely full or even near full – they are oversized to maximise water collection in rainy years and use it as storage for dry years. In these cases the addition of a second reservoir would increase the volume of water stored at the peak rainy season. However during this period the reservoir could not be used for pumped storage but as permanent water storage.
Incoming water sediment loads to any existing or new reservoir would also need to be assessed, as silting may pose a problem by reducing reservoir volumes over time. A pumped storage plant cannot also use all the existing reservoir volume otherwise silt and debris would be drawn up the pump turbine.
– Infrastructure: The analysis of transport and grid infrastructure requires a thorough look at roads in the region to see if a potential transformation site could support high volumes of traffic and heavy construction machinery.
Electricity grid infrastructure may also require upgrading to provide two-way power flows to facilitate pumping as well as generation. For non-hydro dams the grid infrastructure has to be looked at in more detail. For example issues to consider include proximity to the substation and the availability of spare capacity here, as well as the feasibility of upgrading any inadequate substations and so forth.
– Economics: The uncertainty of capital costs can be a barrier to transformation. In pumped storage projects there is great variability in capital costs. Construction costs are very much site and country specific due to the high labour and material intensity of this type of construction.
New technological developments may allow some barriers to be overcome. Variable speed, reversible pump-turbines will increase the operational flexibility of planned pumped storage facilities, and will better equip them to support the integration of variable renewable generation.
New concepts such as coastal seawater pumped storage, where the sea acts as the lower reservoir, would open up a greater number of potential sites. The only such plant in operation in the world at the current time is the 30MW Okinawa demonstration facility in Japan. Pumped storage using an underground cavern as the lower reservoir has also been proposed and if successful would eliminate any environmental problems associated with constructing reservoirs on the surface.
At the end of the report the authors noted that the modelling exercise is, to their knowledge, the first approach to identifying and quantifying the potential for transformation to pumped storage in European countries based on one or two existing dams. However, it was also acknowledged that as the exercise belongs to the fields of research, its results may be some stages away from the accuracy and definition required for an actual project feasibility study. The authors believe that this is important as the ultimate goal of such a study should be dual:
– To the feed decision-making process with sound science.
– Reduce the costs of transformation for all actors involved – such as governmental spatial planning agencies, engineering companies and pumped storage developers.
In April 2012 as a follow-up to the report, JRC organised a workshop. More than 30 experts took part with the objectives to:

– Validate the JRC methodology and model. An expected outcome was a set of recommendations for the improvement of the effectiveness and efficiency of the methodology, focusing on the approach followed, software and data sources used, multi-criteria elements, costing methodology, sensitivity analysis and envisaged output.
– Address the issue of data availability and collection in the Member States, since the JRC intention is to apply the methodology to the EU and Norway.
– Share and disseminate the methodology among relevant stakeholders.

A report, entitled SETIS expert workshop on the assessment of the potential of pumped hydropower storage, documented the discussions which took place. It also included recommendations and sources of data that could be used by any individual or organisation interested in assessing the potential for pumped storage using a GIS-based approach.

Various suggestions for refining the methodology are contained within the report. Included amongst these is an analysis of the screening method for potential transformation sites. For example it was considered that the minimum reservoir size of 1Mm3 was inappropriate as many reservoirs with less capacity than this can yield high hydropower generation. It was proposed that initial thresholds should be variable and generally from 0.5-1Mm3. A 5km horizontal distance between the two reservoirs was also considered too restrictive and 20km is considered a more sensible figure depending on the difference in elevation. Proposals also suggested using a minimum head of between 20-200m and could extend to 1500m in order to avoid excluding any possible sites.

Discussions on the GIS files suggested:

– Software plug-ins and tools.
– Alternative elevation data sources such as LIDAR surveys, military geodetic maps and data, plus the use of Japan Space Systems.
– Additional data sources for lakes and reservoirs such as JRC’s Catchment Characterisation and Modelling, European Environmental Agency’s European Catchments and Rivers Network Systems (ECRINS), EuroGlobalMap, and the European lakes, Dams and Reservoirs Database.

Recommendations for additional layers to the GIS files looked at the inclusion of geological data such as checking tunnels in the area, the possibility of mines and seismic hazards etc. The European Environmental Agency has developed GIS layers and databases that could be available for pumped storage assessment. It was also suggested that the model could export specific results that would enable further cost estimation in a separate tool.

Quality and detailed data

The use of GIS models is effective, efficient and convenient for assessing pumped storage potential for a project proposal or for country or regional assessments, the workshop concluded. What differs is the intensity of the use of the tools, the detail of the data needed and the assumptions behind the model and methodology.
The country and European assessments are heavily dependent on the assumptions taken. For example the sensitivity analysis showed that enlarging the maximum distance between two reservoirs from 5-20km multiplies the theoretical potential for Croatia by a factor of 10. Obtaining adequate quality and detailed data forms a significant part of the effort required for this kind of assessment.

This article has been compiled from following reports.
TRC Scientific and Technical reports: Pumped-hydro energy storage: potential for transformation from single dams. Analysis of the potential for transformation of non-hydropower dams and reservoir hydropower schemes into pumping hydropower schemes in Europe by Roberto Lacal Arántegui, Institute for Energy and Transport, Joint Research Centre of the European Commission, the Netherlands; and Niall Fitzgerald and Paul Leahy, Sustainable Energy Research Group, University College Cork, Ireland. 2012. The report can be downloaded from

JRC Technical Report: SETIS expert workshop on the assessment of the potential of pumped hydropower storage. Held in Petten, the Netherlands, 2-3 April 2012. By Roberto Lacal Arántegui and Evangelos Tzimas. Other JRC-IET contributors: Andrei Bocin-Dumitriu and Alyona Zubaryeva. The report can be downloaded from: