Climate forecasts for the wind energy industry in the Mediterranean Region
Sandro Calmanti, ENEA, Italy, Melanie Davies, IC3, Spain, Alessandro Dell’Aquila, ENEA, Italy, Peter Schmidt, PIK, Germany
According to the World Wind Energy Association the total wind generation capacity worldwide has come close to cover 3% of the world's electricity demand in 2011 (1). Thanks to the enormous resource potential and the relatively low costs of construction and maintenance of wind power plants, the wind energy sector will remain one of the most attractive renewable energy investment options.
Wind energy stakeholders rely on short term forecasting of wind speed on time scales ranging from a few hours to a few days for the daily management of wind farms (2). However, CLIM-RUN stakeholder surveys in Spain, Morocco, Croatia and Cyprus show that wind energy stakeholders do have limited knowledge on how changing climate patterns may influence energy output of wind power plants in the medium- to long-term. Studies reveal that climate variability and change pose a new challenge to the entire renewable energy sector (3).
Stakeholders in the wind energy sector mainly use, if available, site-specific historical climate information to assess wind resources at a given project site. So far, this is the only source of information that investors (e.g., banks) are keen to accept for decisions concerning the financing of wind energy projects. However, one possible wind energy risk at the seasonal scale is the volatility of earnings from year to year investment. Some projects are financed with up to 80% debt, and interest on the debt remains payable every year—whether there is a good climate resource availability or not. The most significant risk is therefore that not enough units of energy (or megawatt hours) can be generated from the project to capture energy sales to pay down debt in any given quarter or year.
On the longer time scale the risk is that a project’s energy yields fall short of their estimated levels, resulting in revenues that consistently come in below their projection, over the life of the project.
The nature of the risk exposure determines considerable interest in wind scenarios, as a potential component of both the planning and operational phase of a renewable energy project. Fundamentally, by using climate projections, the assumption of stationary wind regimes can be compared to other scenarios where large scale changes in atmospheric circulation patterns may affect local wind regimes.
Climate experts of CLIM-RUN are exploring the potential of seasonal to decadal climate forecast techniques (time-frame 2012-2040) and regional climate scenarios (time horizon 2040+) over the Mediterranean Region as a tool for assessing the impact of changes in climate patterns on the energy output of wind power plants. Subsequently, we will give a brief overview of these techniques as well as first results related to wind projections for different sites across the Mediterranean Region.
Seasonal to decadal wind forecasts
The Climate Forecasting Unit (CFU) at IC3 has started to generate seasonal and decadal climate forecasts from which environmental parameters of interest for the wind energy sector can be extracted (e.g. 10 m wind speed).
The spatial resolution of these forecasts range from 120km x 120km for the decadal forecasts and 80km x 80km for the seasonal forecasts over the whole of the Mediterranean, to a few specific sites following statistical downscaling that uses data from selected weather stations, in regions where the climate models show the highest skill.
Wind speed ensemble-mean and probability forecasts have been produced at seasonal timescales using ECMWF S4 for all seasons in 2011/2012 with a 1 month lead time, and at decadal timescales using EC-EARTH for 2011-2016 and 2016-2021. Initial results for wind speed forecasts generally show a moderate, yet significant, model skill in both seasonal and decadal forecasts (anomaly correlations up to 0.4 for 1-5 year wind forecasts and up to 0.5 for 6-10 year wind forecasts; anomaly correlations up to 0.6 for DJF, 0.3 for MAM, 0.4 for JJA and 0.2 for SON).
Seasonal to decadal wind speed predictions at specific sites have relatively low skill over most of Europe, with the highest anomaly correlations of 0.59 and 0.55 obtained at sites in Egypt and Turkey respectively. Statistical downscaling will be carried out for all the main sites across the Mediterranean for wind predictions.
Recent studies have demonstrated that regional climate models have a large potential for enhancing the quality of climate projections in the presence of complex orography (4) and in the proximity of coastal areas (5). Climate scenarios produced within the 6th-Framework Program European project ENSEMBLES (6) provide first examples of how climate models can generate data that are useful for planning purposes in the wind energy sector.
The map below (Figure 1.) shows the projected changes in 10m wind speed over the Mediterranean. The map is obtained by averaging the mean annual wind speed of 16 regional climate models that have produced climate scenarios for the ENSEMBLES project. According to those projections, most of the Mediterranean is expected to see a decrease in mean wind speed which would imply an overall loss of the expected productivity. In particular, a consistent negative trend of the monthly mean wind speed along the coast of North Africa (Figure 2), amounting to 5-10% of the corresponding long term average. Such climate scenarios may be an input for the more insightful and spatially accurate analysis that would be required to assess the potential changes in energy productivity at any specific site across the region.
Climate modelers of the CLIM-RUN consortium are contributing to international coordinated efforts for the production of long-term climate scenarios, such as the Med-CORDEX initiative (www.med-cordex.eu) that will continue to provide more data for the assessment of the impact of climate variability and change on wind energy.
The complex nature of physical factors affecting year to year wind variability poses a challenge to climate modelers in producing climate information that is usable in the wind energy sector. Investors and operators in the wind industry are exposed to risks related to climate variability and change. Therefore, they do have considerable interest in acquiring more knowledge on the medium-to-long-term mechanisms of wind variability. As shown before, initial findings from the climate experts of CLIM-RUN’s energy case study (WP 7) demonstrate interesting potential for meeting this specific user need. This also applies to the solar energy sector – first results from this research will be presented in the next issue of this newsletter. In the meantime, CLIM-RUN will work towards closer collaboration with renewable energy stakeholders to further exchange knowledge and find out about the added value that climate forecasts may bring to the renewable energy sector.
Figure 1:Mean change in wind speed [m s-1] at 10 m height projected by 16 regional climate models. Colours represent the average long term change in wind speed projected for the decade 2040-2050 with respect to 1990-2000. Model data are produced at an horizontal resolution of about 50Km. Hatched areas respresent areas where more than 66% of the model agree in the sign of the long term change.
Figure 2 : Monthly wind speed simulated by 16 regional climate models in the area included between (17°E, 37°N) and (21°E, 33°N) over the coast of North Africa. Thin lines represent the monthly wind speed simulated by each model. The thick line indicates the ensemble average. The shaded area represents one standard deviation of the model spread around the mean.
(1) http://www.wwindea.org: preliminary data gathered by WWEA and published on the occasion of the 3rd International Wind Conference & Exhibition in Coimbatore/India
(2) Marquis et al. (2012). Forecasting the Wind to Reach Significant Penetration Levels of Wind Energy. Bull. of Am. Met. Soc.
(3) For instance, The Economist (2011). Managing the risk in renewable energy. Economist Intelligence Unit on behalf of SwissRe. October 2011.
(4) Artale et al. (2010). An atmosphere-ocean regional climate model for the Mediterranean area: assessment of a present climate simulation. Clim. Dyn.
(5) Winterfeldt J. and R. Weisse (2009). Assessment of value added for surface marine wind speed obtained from two regional climate models. Mon. Wea. Rev.