%0 Report %1 bett2018assessment %A Bett, Philip %A Thornton, Hazel %A de Felice, Matteo %A Suckling, Emma %A Dubus, Laurent %A Saint-Drenan, Yves-Marie %A Troccoli, Alberto %A Goodess, Clare %D 2018 %I Zenodo %K ecem energy renewables seasonal skill verification myown %N D3.4.1 %R 10.5281/zenodo.1295518 %T Assessment Of Seasonal Forecasting Skill For Energy Variables %X This report describes assessments of the skill in seasonal forecasts of energy variables (electricity supply and demand) in European countries, using data from seasonal climate prediction systems available through the Copernicus Climate Change Service (C3S). This work follows on from our previous report on the skill in seasonal forecasts of climate variables (ECEM Deliverable D2.2.1, Bett et al. 2018a), and uses the newly-available historical energy data produced in WP3 of the ECEM project (ECEM Deliverables D3.1.1 and D3.2.1, Dubus et al. 2017a,b). We show that, when examined on the seasonal-average, country-average scale, solar PV power and wind power are very strongly correlated to solar irradiance and wind speed respectively. This means that less post-processing of the climate data is required to obtain the corresponding energy variable, which can greatly simplify the production of seasonal energy forecasts. However, the cases of hydropower and electricity demand are intrinsically more complex. While in many cases they are strongly linked to precipitation and air temperature, it is clear that for some countries, forecasts could benefit from more bespoke, country-specific modelling. It might be possible to improve the skill in some cases by using larger-scale modes of atmospheric variability, such as the North Atlantic Oscillation (NAO), as predictors of the energy variables. While this is an area of ongoing research, we demonstrate that the observed relationships between the NAO and the relevant climate variables show that this is a promising approach. We also demonstrate one way of using the Principal Component Analysis technique to try to identify the relevant modes of variability more generally, and assess their predictive skill for energy variables. The diverse levels of skill we have found for seasonal forecasts of climate variables mean that the skill in forecasting energy variables is also diverse: in some cases it is very promising, whereas in others it is clearly not useful. However, many cases could benefit from more detailed approaches, using more sophisticated modelling between the physical climate system and particular energy variables. Close collaboration between experts in climate science and energy systems in those cases could lead to the provision of more skillful forecasts in the future.

@techreport{bett2018assessment, abstract = {This report describes assessments of the skill in seasonal forecasts of energy variables (electricity supply and demand) in European countries, using data from seasonal climate prediction systems available through the Copernicus Climate Change Service (C3S). This work follows on from our previous report on the skill in seasonal forecasts of climate variables (ECEM Deliverable D2.2.1, Bett et al. 2018a), and uses the newly-available historical energy data produced in WP3 of the ECEM project (ECEM Deliverables D3.1.1 and D3.2.1, Dubus et al. 2017a,b). We show that, when examined on the seasonal-average, country-average scale, solar PV power and wind power are very strongly correlated to solar irradiance and wind speed respectively. This means that less post-processing of the climate data is required to obtain the corresponding energy variable, which can greatly simplify the production of seasonal energy forecasts. However, the cases of hydropower and electricity demand are intrinsically more complex. While in many cases they are strongly linked to precipitation and air temperature, it is clear that for some countries, forecasts could benefit from more bespoke, country-specific modelling. It might be possible to improve the skill in some cases by using larger-scale modes of atmospheric variability, such as the North Atlantic Oscillation (NAO), as predictors of the energy variables. While this is an area of ongoing research, we demonstrate that the observed relationships between the NAO and the relevant climate variables show that this is a promising approach. We also demonstrate one way of using the Principal Component Analysis technique to try to identify the relevant modes of variability more generally, and assess their predictive skill for energy variables. The diverse levels of skill we have found for seasonal forecasts of climate variables mean that the skill in forecasting energy variables is also diverse: in some cases it is very promising, whereas in others it is clearly not useful. However, many cases could benefit from more detailed approaches, using more sophisticated modelling between the physical climate system and particular energy variables. Close collaboration between experts in climate science and energy systems in those cases could lead to the provision of more skillful forecasts in the future.}, added-at = {2018-06-29T14:36:35.000+0200}, author = {Bett, Philip and Thornton, Hazel and de Felice, Matteo and Suckling, Emma and Dubus, Laurent and Saint-Drenan, Yves-Marie and Troccoli, Alberto and Goodess, Clare}, biburl = {https://www.bibsonomy.org/bibtex/2d79941e9d840beb36be9e34f8c0f5e2d/pbett}, description = {Assessment of seasonal forecasting skill for energy variables | Zenodo}, doi = {10.5281/zenodo.1295518}, institution = {Met Office}, interhash = {085f1705c4ea90f446fb3c447430c172}, intrahash = {d79941e9d840beb36be9e34f8c0f5e2d}, keywords = {ecem energy renewables seasonal skill verification myown}, language = {en}, month = jan, number = {D3.4.1}, publisher = {Zenodo}, timestamp = {2018-10-01T09:50:26.000+0200}, title = {Assessment Of Seasonal Forecasting Skill For Energy Variables}, type = {ECEM Deliverable}, year = 2018 }