Satellite images reveal rapid development of global water-based photovoltaic over the past 20 years

Abstract

Water-based photovoltaics (WPV) have emerged as a promising solution to land-use conflicts associated with solar photovoltaic systems. Accurate monitoring of the spatiotemporal distribution of WPV is essential for evaluating its development potential, environmental impacts, and informing policy decisions. Satellite remote sensing data offer a feasible approach for WPV mapping and monitoring. However, conventional image classification and deep learning methods often limited by sample size requirements, computational costs, and technical complexity, which hinder their widespread applicability. To address these challenges, this study proposes a novel index, the normalized difference photovoltaic index (NDPI), for WPV detection. We generated a global WPV map for the year 2023 using Sentinel-2 MSI imagery and NDPI. Additionally, by integrating NDPI with Landsat time series data, we determined the installation dates of WPV systems and evaluated their development trends from 2000 to 2023. Our results show that: (i) The NDPI demonstrated excellent performance in WPV detection, with overall accuracy for spatial location and installation dates of WPV was 0.935 and 0.927, respectively, and Kappa coefficients of 0.870 and 0.921. (ii) Global WPV coverage in 2023 reached 589.17 km2, with Asia being the primary contributor, accounting for over 97 %. China emerged as the leading country, with a WPV area of 472.92 km2, significantly exceeding other nations (< 50 km2). (iii) WPV experienced significant growth from 2000 to 2023, particularly after 2015. The increase in WPV area (434.57 km2) from 2015 to 2023 was nearly three times the total area covered in the previous 15 years. The proposed NDPI provides a universal approach for global WPV spatiotemporal monitoring and the update of basic information. It also provides potential for assessing the environmental impacts of WPV across its operational lifecycle.