The growth of Residential rooftop solar (RRS) in some western countries has predominantly been driven by individual or market behaviour and has been extensively studied. However, the development landscape of RRS in China differs, and its driving mechanisms remain unclear. To address this research gap, we investigate the spatial distribution pattern and driving factors of RRS growth using city-level data on RRS installation. Employing the Geogr. The growth of Residential rooftop solar (RRS) in some western countries has predominantly been driven by individual or market behaviour and has been extensively studied. However, the development landscape of RRS in China differs, and its driving mechanisms remain unclear. To address this research gap, we investigate the spatial distribution pattern and driving factors of RRS growth using city-level data on RRS installation. Employing the Geographical Detector Model, we calculate indicators to identify the contributions of various socio-economic factors to RRS growth and the strength of their interactions. Our key findings include: 1) significant spatial heterogeneity in RRS growth across regions with different natural and socio-economic characteristics, which impact RRS growth in two patterns: one-way and inverted U-shaped; 2) although solar radiation abundancy is important, air pollution and certain socio-economic factors appear more influential, be it comparing between the eastern and western China, or north and south; 3) the significance of fiscal subsidies has diminished, but benchmark electricity prices (BEP) could serve as a useful alternative; 4) substantial synergistic effects exist between different factors, with environmental and demographic factors displaying particularly strong synergies with others, suggesting that they are essential considerations for future RRS planning. Our findings contribute to a comprehensive understanding of RRS development in China and hold critical implications for future policy design.••Spatial distribution and driving factors of RRS growth are explored using GDM.••Spatial heterogeneity exists among regions with different socio-economic features.••Air pollution and some socio-economic factors are more important in RRS promotion.••BEP can be an alternative incentive tool for fiscal subsidy to promote RRS growth.••Residential rooftop solarSpatial heterogeneityDriving factorsSynergistic effectResidential rooftop solar (RRS) for electricity generation is essential in the new power system and vital during the low-carbon green energy transformation, which is being adopted globally (Moore and Bullard, 2021). In recent years, China's RRS has been expanding rapidly, with the annual growth rate ranking first in the world. However, RRS growth is spatially uneven due to significant inter-regional differences in China. Therefore, the driving mechanism shaping the current RRS development remains a question.The adoption of RRS can be affected by various factors, such as cost (Kaufmann et al., 2021), incentive policy (Briguglio and Formosa, 2017), home ownership (Briguglio and Formosa, 2017), peer effect (Mundaca and Samahita, 2020, Rai et al., 2016), place attachment and environmental attitudes (Abreu et al., 2019, Corbett et al., 2022), income (O Shaughnessy et al., 2021), race and ethnicity (Sunter et al., 2019), and so on. After reviewing 173 studies on the adoption behaviour of RRS system, Alipour et al. (2020) identified 333 predictors and classified them into three dimensions: individual, social, and informational. A comparison of the frequency of predictor usage indicated some most popular predictors. Although factors promoting RRS have been thoroughly studied, these conclusions are primarily drawn from Western countries and focus on individual decision-making behaviour, while the story in China might. 2.1. Geographical detector modelAs proposed by Wang et al. (Wang et al., 2016), Geographical Detector Model (GDM) is a widely used technique for spatial stratified heterogeneity analysis (Song et al., 2020). The basic idea of GDM is that a risk would exhibit a spatial distribution similar to that of a factor if the factor attribute leads to the risk (Wang and Hu, 2012). Originally, it identifies the spatial relationships between human diseases (risks) and disease determinants (factors) (Wang et al., 2010). Later, GDM is applied to study other “risks” such as minimum mortality temperature (Yin et al., 2019), quality of rural life (Fang et al., 2020), air pollution (Ding et al., 2019), and so on. Through GDM analysis, we can identify the distinct mechanisms across different strata and recognize the determinants of the observed process.Since RRS growth involves many spatially related factors such as geographic resources and regional economic characteristics, forming many different “strata” overlapping with the distribution of RRS growth, GDM can be a useful tool to analyze the similarity of RRS growth and these strata, which can better disclose the spatial correlations between RRS growth and possible factors. Compare to econometric models without considering the spatial characteristics, GDM has stronger explanatory power in finding driving factors of RRS growth. Moreover, GDM is technically free by multicollinearity a.
