Rapid mapping and spatial analysis on the distribution of photovoltaic power stations with Sentinel-1&2 images in Chinese coastal provinces

Abstract

Photovoltaic (PV) solar energy generation attracts considerable attention to archive carbon neutrality goals worldwide. Geospatial data describing the PV system based on satellite images are critical for PV deployment. However, it remains challenging to obtain relative data in coastal zones due to frequent cloud cover. There have also been relatively few studies on spatial analysis of PV development in terms of solar resources and the energy demands. Therefore, this study took Chinese coastal provinces as example, and proposed a random forest classification method with information on morphology, optics, and SAR to address cloud pollution. The results obtained were analyzed with geographical and socioeconomic factors to determine its development status. The results have shown that (1) by October 2022, the total PV area in Chinese coastal provinces reached 837.3 km2, with an overall accuracy of 96.9 % and a Kappa coefficient of 0.91. Water-based PV represents 33.7 % which tended to be distributed near the shoreline. (2) Bounded by the Huai River, the PV density in the north was 9.0 km2 per 104 km2, which is more than twice that in the south. The development of water-based PV is a key reason for the high PV construction density in coastal areas. (3) PV distribution was slightly mismatched with solar resource and power demand, especially in Liaoning and Guangdong. Liaoning has relatively high potential for PV development. The study findings can play a key role in evaluation of PV power generation, development potential, and ecological environmental impacts.

Keywords

• Photovoltaic power generation
• Remote sensing
• Carbon neutrality
• Random forest classification
• Synthetic aperture radar
• Spatial analysis