A research group led by scientists from China’s Tianjin Normal University has used satellite images to reveal the rapid development of global water-based PV (WPV) over the past 20 years.

WPV comprises both floating PV systems and photovoltaic plants deployed on fixed structures in shallow waters.

For the task, the team has developed an index-based remote sensing method, which they called the normalized difference photovoltaic index (NDPI). Index-based methods are mathematical formulas that identify land features based on spectral indices, which are calculated using different bands of light captured by satellites.

“Compared to traditional thresholding and machine learning approaches, index-based methods are based on the universal features of target objects, offering greater adaptability to variations in environmental conditions,” the academics said. “Additionally, index-based methods provide faster and more flexible solutions for WPV mapping and information updates, as they directly respond to the spectral characteristics of the target objects. Moreover, WPV arrays, typically located on water surfaces, exhibit lower susceptibility to environmental disturbances compared to terrestrial PV power plants.”

To create their NDPI method, the researchers analyzed 5,000 sites globally from Google Earth. The samples were divided into five land cover types, namely bare soil, impervious surface, vegetation, water, and PV. Based on the analysis of the images, solar panels reflect more shortwave infrared (SWIR1) light than most surfaces and reflect less visible light and near-infrared (NIR) than land or water.

Following these discoveries, the team was able to design formulas that detect solar panels on water. The method was then verified over 1,000 WPV, and accuracy (OA) for spatial distribution and installation date of 0.935 and 0.927, respectively, was achieved. “Overall, the installation dates of WPV showed high accuracy, with user’s accuracy (UA), producer’s accuracy (PA), and F-score values exceeding 0.9 for all years except 2012 and 2013,” they said.

As the method was verified, the scientists tested it on 2023 images from the Sentinel-2 MSI satellite sensor operated by the European Space Agency (ESA). Using 2000-2023 images from NASA’s Landsat satellite, the team has tested the NDPI for installation date trends.

“The global WPV area reached 589.14 km2 in 2023. Notably, the growth of global WPV accelerated significantly after 2015, with an annual growth rate of 56.06 km2 yr-1, nearly ten times that of the 2000–2015 period (5.98 km2 yr-1),” the results showed. “WPV development has been particularly rapid in Asia, driven primarily by China. In 2023, the WPV area in China reached 472.92 km2, accounting for more than 80% of the global total.”

The analysis revealed that Asia is the primary contributor to global WPV, accounting for over 97% of the WPV area, followed by North America with 1.569%, while all other continents had a combined contribution of about 1%. While China was the dominant contributor, with a WPV area of 472.92 km2, Thailand and Vietnam emerged as key areas with areas exceeding 20 km2 in each.

“In China, policy support has notably accelerated WPV development, with the WPV area increasing from 104.89 km2 in 2016 to 472.92 km2 by 2023,” they noted. “Apart from China, the WPV area in Vietnam also increased from 6.17 km2 in 2017 to 24.24 km2 in 2023, and the WPV area in Thailand nearly tripled from the previous year to 2023. In Europe and North America, WPV areas also continue to show an upward trend, with a significant acceleration in growth between 2005 and 2009.”

Concluding their research, the team said it estimates the global installed capacity of WPV to be 21.18 GW, with projections suggesting it could reach 35.64 GW by 2030.

Its findings were presented in “Satellite images reveal rapid development of global water-based photovoltaic over the past 20 years,” published in the International Journal of Applied Earth Observation and Geoinformation.

Scientists from China’s Tianjin Normal University, Henan University of Technology, Advanced Laser Technology Laboratory of Anhui Province, Fujian Agricultural and Forestry University, Zhejiang University, and Finland’s National Survey of Finland have participated to the research project.

This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: editors@pv-magazine.com.