The Impacts of Solar Resource Variability on Regional Aggregate Photovoltaic Power Time Series

Introduction
For the purposes of understanding the impacts of solar variability on the electricity network, estimates of hourly aggregate solar power generation for a region are often required. However, the availability of solar production data for the UK remains limited, and studies tend to rely on historical meteorological (met) station data, satellite data products, mesoscale model output or reanalysis datasets. Their respective resolution varies from point measurements to typically 0.05 degrees to around 50 km for reanalysis. In addition, the availability of data on installation distribution is generally on a postcode district level, representing areas which range between 0.01 and 3,500 km2. The extent to which the resource and the locational data sources are able to capture the spatial and temporal variability in the resource is a consideration for the resulting credibility of any analyses based on them. This paper examines the importance of this variability on the characteristics of aggregated time series of solar production.

Data and Methods
Satellite data products offer a convenient and homogeneous gridded dataset from which to examine the effect of scale on the characteristics of the resource. Here, CMSAF hourly data has been processed to provide multi-year hourly time-series estimates of photovoltaic panel output across Great Britain. As well as gridded information, this has been processed for postcode districts based on their centroid for all 2880 postcode districts.
The analysis compares the inherent variability between estimates at a postcode level and also variability within postcode districts. This is intended to shed light on ‘how big’ an area over which data is aggregated can become before potentially important information is lost.

Results and conclusions
While the analysis is currently at a preliminary stage it is demonstrated that there is variability across regions and within postcode districts. The extent to which lower resolution reanalysis data is sufficient to represent aggregate power generation in a sub-region is to be determined.