Although the projected increase in RSDS largely compensates such TAS-induced effects, these results call for future efforts to reduce the dependence of the PV cells performance on the ambient temperature, as already stated in ref. Under such a fixed scenario of high-PV penetration, a general and progressive decline of the generated PV power is found in all regions along the entire period Fig.
In Southern Europe, changes are smaller see also Fig. Note that the fact that changes are expressed here in relative terms partly contributes to obtain the smallest changes over regions and seasons see also Supplementary Fig. Thirty-year running mean time series of the estimated PV power production anomalies under both the RCP4. Shadows show the ensemble spread. If the ensemble mean change exceeds such a quantity, a thin black line is superimposed on the ensemble mean series. Changes projected to the end of this century — versus — in a the mean, b the annual variability, c the monthly variability and d the daily variability of the PV power production series of each region under both the RCP4.
The individual signals corresponding to each ensemble member are depicted by circles: filled if they are significant, empty if not. Boxes represent ensemble mean signals: coloured, white or grey if robust, negligible or uncertain, respectively. Except for the regions 4 and 9, the mean projected changes under RCP8. However, in all cases the magnitude of the ensemble mean projected changes, even for those noted as robust in Fig. Only for region 1, and especially under RCP8. It drops to even below 0.
To better understand the aforementioned uncertainty, a two-step one-way analysis of variance 19 is applied to the set of time series of PV power production anomalies to discern the roles played by the RCP, the GCM and the RCM choice. Results shown in the small subplots of Fig. Overall, the influence of the GCM prevails in the northernmost regions, while in the southern and central regions the RCM plays a dominant role. These latter regions, namely 4, 6, 7, 8 and 9, are located in South-central Europe, where regional feedbacks involving temperature, soil moisture and clouds are particularly important 20 , 21 , 22 , inducing a higher variety of model responses than in northern areas with weakly constrained land—atmosphere interactions Besides, these southernmost areas have a highly complex orography, which has a strong impact on the simulated regional climates.
Worth to note that, while RCMs thus appear more tightly constrained by their driving GCM northwards than southwards, RCMs are able to turn the sign of the projected change in the resource from positive to negative in some northern areas Supplementary Fig. However, as it occurs similarly for all RCMs driven by the same GCM, it is not reflected in the results of the analysis of variance. Finally, we focus on a key aspect of all weather-dependent renewable energies concerning their grid integration and management: the temporal stability of their supply.
Our results indicate that changes in the monthly and annual variability of the PV production series see Methods are overall non-significant Fig.
- chapter and author info!
- Sex, Power, Conflict: Evolutionary and Feminist Perspectives.
- Solar Panels Electricity | Energy Saving Trust.
- A grammar of the Yoruba language?
This is most evident in region 4, where it occurs in all seasons but winter Supplementary Fig. In the central region 6, however, a robust positive signal for changes in daily variability is found assuming RCP8. These changes would occur progressively along the entire period, being consistently smaller by mid-century Supplementary Fig. In southern areas, in contrast, slight increases in the mean PV supply and in its daily stability were found.
Uncertainties still remain in our assessment though, due for instance to indirect effects of natural and anthropogenic aerosols 24 and to land-use changes, both features being currently poorly represented or totally ignored in RCMs, and to the tilt of PV panels, a fact not considered here. Also, whereas the effect of ambient conditions on the PV cell temperature and, thereby, performance has been explicitly accounted for here, other factors that may affect the outdoor performance of PV modules, such as the solar spectrum distribution and the airmass effect on it 25 , 26 , are absent in our analysis.
These unrepresented factors may have an impact on the reported signals for northern areas, which will suffer more from the low irradiance performance of PV modules than sites in the south of Europe where the solar irradiance retains higher values and the PV cell temperature effect prevails Besides, we found somewhat conflicting signals when comparing GCM and RCM-downscaled results regarding PVpot projections, which needs further investigation. Nevertheless, and although it should be acknowledged that climate models tend to underestimate the multidecadal variations in RSDS 28 , 29 , indicating the potential for larger changes than currently projected, none of the simulations analysed so far, here or elsewhere 2 , 3 , 7 , 8 , 9 , 15 , 24 , provides strong changes.
Not even, or even less, if we incorporate the assumption of a well-developed PV power supply system across Europe. Actually, the fact that PV systems are foreseen to largely expand over the 21st century, together with other technological and politico-economical aspects for example, increased lifetime of PV modules, module price decrease and appropriate policies supporting PV system deployment —all non-physical aspects determining the evolution of the PV energy market 30 , 31 , as discussed in ref.
Therefore, climate change hardly compromises the European development of PVs. Two ensembles of regional climate simulations spanning the period — were used in this study. The ensemble mean climatologial patterns of all the magnitudes utilized in this study, that is, mean values and variability of several variables at various timescales, are provided in Supplementary Figs 9 and 10 as a referential baseline. A concise evaluation exercise of the simulated RSDS, further supported by the results presented in ref.
All ensemble members are well able to capture both the spatial distribution and variability of the observed mean RSDS pattern, which is mainly dominated by a latitudinal gradient both in models and observations. Excessive RSDS is a long-standing problem in climate modelling 34 , However, while the spatial distribution of the simulated monthly and daily variability patterns is acceptable, it is not the case for the annual variability simulated pattern.
PEG - The revolution in utility-scale PV power
Nonetheless, taking into account 1 the small number of observational points, 2 the very nature of models that perform a spatial discretization of the simulation domain, along with the actual spatial representativeness of ground-based solar radiation measurements 36 , 3 the still moderate accuracy of models to properly reproduce cloudiness 37 and 4 the lack of a dynamically modelling of aerosols concentration levels, the results of this validation exercise are within an acceptable range, thus allowing a moderately high confidence on the analysed simulations.
The PV power output at a site depends on two factors: its PV power generation potential PVpot and the installed capacity. As defined and used in this study, PVpot is a dimensionless magnitude accounting for the performance of the PV cells with respect to their nominal power capacity according to the actual ambient conditions. Therefore, PVpot multiplied by the nominal installed watts of PV power capacity gives instantaneous PV power production.
PVpot mainly involves the amount of the resource RSDS but also the influence that other atmospheric variables may have on the efficiency of the PV cells, which diminishes as their temperature increases According to the literature 39 , it can be expressed as:. CLIMIX performs an optimization exercise based on the resource availability for allocating PV plants that is, by default, the preferred locations are those where RSDS is most abundant , but without overloading the best locations with huge deployments of PV installations, discarding sites previously identified as forestry or inaccessible areas and taking into account the population distribution.
Using the PVpot time series estimated as described above and the spatial scenario of PV power-installed capacity, 3-h time series of PV power generation were obtained for each grid cell of the domain holding PV power units. These series were finally aggregated per region, considering those regions for which the aforementioned targets were specified Fig. As a result, nine 3-h time series of PV power generation spanning the period — were obtained, one per region, from each one of the available simulations. Ensemble mean values of their means and standard deviations as computed here to assess changes in time variability are presented in Supplementary Fig.
Time variability is assessed here at various timescales daily, monthly and annual as follows.
The impact of climate change on photovoltaic power generation in Europe | Nature Communications
First, the 3-h series of PV power generation are daily, monthly and yearly aggregated and then detrended to avoid capturing long-term changes rather than those occurring at the scales of interest. In addition, multi-year monthly and daily means are removed from the monthly and daily series, respectively, to avoid the masking effect of the annual cycle of PVpot. The normalization consists on dividing by the climatology of the historical period that is, the — mean values and was not performed in the case of the TAS series due to the very nature of this variable.
The described procedure was applied either for the whole series or for the seasonally split series. Climate change signals from each individual ensemble member are computed as the difference between the value of the assessed magnitude in a year-long future period namely, — and its value in the reference period — The ensemble mean E mean signals are computed as the arithmetic mean of the set of individual signals within each ensemble.
The following criteria are used to consider them as either uncertain or negligible: at least two significant individual signals differ in the sign of the projected change uncertain ; the uncertainty condition is not fulfilled and less than a half of the individual signals are significant negligible. Otherwise, the ensemble mean signals are referred to as robust.
In addition, ensemble maximum and minimum E max and E min , respectively signals are computed as the maximum and minimum values, respectively, of the whole set of individual signals.
- Art and Love in Renaissance Italy (Metropolitan Museum of Art);
- Benefits of solar electricity;
- BELECTRIC – Ground-mounted Solar Power Plants and PV Roof Systems?
- Workbook for Introductory Statistics for the Behavioral Sciences.
- Stefan C.W. Krauter (Author of Solar Electric Power Generation - Photovoltaic Energy Systems);
The ensemble spread denotes the range between these two limit values. Thereby, changes in PVpot are given by:. This is the procedure adopted here, as it was done in previous works 2 , 3 , 8. However, note that the cross-products in the last two terms of equation 5 make it impossible to fully isolate the contribution of each single field. Two additional measures were employed to elucidate the forcefulness of the ensemble mean projected changes in PV power generation.
The signal-to-noise ratio S2N refers to the ratio between the magnitude of the ensemble mean change and the ensemble spread. Detectability describes the ratio between the magnitude of the ensemble mean change and the ensemble mean annual variability of the PV power generation series. Therefore, a value of this ratio larger than unity implies that the magnitude of the ensemble mean change falls beyond the expected natural variability of the series. How to cite this article: Jerez, S. The impact of climate change on photovoltaic power generation in Europe. Press Crook, J.
Climate change impacts on future photovoltaic and concentrated solar power energy output. Wild, M. Projections of long-term changes in solar radiation based on CMIP5 climate models and their influence on energy yields of photovoltaic systems. Energy , 12—24 Coupled Model Intercomparison Project Phase 5. Last accessed on 5 November Moss, R. The next generation of scenarios for climate change research and assessment. Nature , — Burnett, D. The UK solar energy resource and the impact of climate change.
Energ 71 , — Panagea, I. Climate change impact on photovoltaic energy output: the case of Greece. Wachsmuth, J. How will renewable power generation be affected by climate change? The case of a Metropolitan Region in Northwest Germany. Energy 58 , — Tobin, I. Change , 99— Jacob, D.
- Reader Interactions;
- 1. Investigate Your Home's Energy Efficiency!
- Primary Sidebar.
- Advantages of Photovoltaic Cells.
- External Liberalization, Economic Performance and Social Policy?
- A Kiss from Maddalena;
- Knowledge Science, Engineering and Management: Third International Conference, KSEM 2009, Vienna, Austria, November 25-27, 2009. Proceedings;
Change 14 , — Jerez, S. The impact of the North Atlantic Oscillation on renewable energy resources in southwestern Europe. Cambridge Univ. Assessment of climate change impacts on energy generation from renewable sources in Croatia.
Patt, A. Vulnerability of solar energy infrastructure and output to climate change. Change , 93— European Climate Foundation. Roadmap A practical guide to a prosperous, low-carbon Europe.
Von Storch, H. Seneviratne, S. Land-atmosphere coupling and climate change in Europe. Zampieri, M. Hot European summers and the role of soil moisture in the propagation of Mediterranean drought. The role of the land-surface model for climate change projections over the Iberian Peninsula. Stegehuis, A.