Geospatial segmentation of high-resolution photovoltaic production maps for Switzerland

Abstract

Until last year, most of Switzerland’s photovoltaic (PV) installations were built on roof tops. But the amount added is not enough to reach the country’s energy transition goals. With the adjustments of September 2023, the government incentivizes large-scale, free-standing photovoltaic installations. It is now essential to identify the best installation locations and to accurately estimate their production potential. Past studies have assessed different landcover classes, but much of the efforts have gone into separating out zones that are not suitable for PV plants; for technical, economical and also legislative reasons. All along, the underlying radiation data that was used to compute the local energy yield remained at a spatial resolution > 1 km. Given the complex terrain of the southern half of the country, this resolution is not high enough to capture the local variability in production potential. Our study introduces a new methodology to derive solar irradiance at a very high resolution of 25 m. Satellite data is combined with high resolution terrain information to compute accurate horizons and to account for local shading effects. These base radiation maps are then converted into potential electricity production from a PV panels. A comparison of the production from a typically chosen panel tilt with the production that can be achieved when the tilt is locally optimized based on the high-resolution radiation maps underlines the value of our new method. In a first application, this data set was used to estimate the lumped production potential of two major landcover classes in Switzerland: agricultural land and water surfaces, each of them divided into two subclasses. The geospatial segmentation was based on land use maps and the total available area within each class was calculated. Comparing the results to the production potential from Swiss roofs shows that these newly incentivized installation areas have a much higher production potential than the conventional roofs; both, in an absolute sense of total potential production (roofs: 120 TWh/a, agricultural: 2,250 TWh/a, water: 210 TWh/a), and in a relative sense of energy yield per installed capacity, especially in winter (roofs <50kWh/m2, agricultural >100kWh/m2, water ≈100kWh/m2).