With the intensification of global climate change and the increasingly frequent occurrence of extreme weather, flood disasters have recently become an important problem that must be confronted in urban development and construction. China has rapidly urbanized since 2000, with an urbanization rate reaching 60 %, similar to that of middle- and high-income countries (Liu and Yu, 2019). This rapid urbanization has significantly changed the type and composition of underlying urban surfaces. The areas of impervious surfaces have increased, negatively affecting natural hydrological processes and weakening the flood control and waterlogging resistance of cities, resulting in increasingly frequent urban flooding disasters (Xia et al., 2017). Therefore, the ability to improve the urban stormwater regulation capacity and strengthen flood prevention and control capabilities is an important aspect of urban construction. China officially launched the pilot sponge city construction project in 2015. The construction of an urban "sponge" is the main mechanism for controlling runoff in sponge cities (Ministry of Housing and Urban-Rural Development (MOHURD), 2014). Green roofs can improve the generation, accumulation, and discharge of rainwater runoff, which is one of the main measures of source control.
Many studies have summarized the efficacy of green roofs at reducing rainwater runoff and alleviating urban waterlogging. In addition, green roofs can generate ecological and economic benefits by improving the heat island effect, energy consumption, water quality, and air quality (Berndtsson, 2010; Deng et al., 2018; Vijayaraghavan, 2016). Common urban water management facilities usually require a large urban area and the transformation of existing urban infrastructures (Czemiel Berndtsson, 2010). These facilities can also conflict with and restrict urban development planning. Therefore, implementing them in highly developed and dense urban areas is often difficult and expensive. However, green roofs, given their position on a building, lead to a strong spatial advantage in highly dense urban areas. The roofs of buildings account for 40–50 % of the total urban underlying impervious surface area, thereby representing a significant proportion of the area that can be utilized for runoff control measures (Palla et al., 2009). Therefore, reasonably transforming building roofs into green roofs can help to effectively utilize urban spaces and improve the urban stormwater regulation capacity.
Now widely implemented and popular, modern green roofs have been incorporated into urban construction planning in many countries. Modern green roofs mostly originated in Germany, with the installation of the first green roofs in the early 20th century. The green area in Germany has increased by 13.5 million m²/y and the proportion of green roofs on new buildings has reached 14 % (Oberndorfer et al., 2007). Subsequently, the United States, Canada, Australia, and other western countries, as well as Japan and Singapore, among other Asian countries, have adopted green roofs. In Portland (Oregon, United States), green roofs are required in all new buildings; consequently, the green roof coverage rate has reached 70 %. In Toronto, Canada, all new development projects with a building area of > 2,000 m² must implement green roofs for 20–60 % of the roofs. In Tokyo, Japan, all new buildings and building expansions with a land area > 1,000 m² must install green roofs (Chen, 2013; Vijayaraghavan, 2016). Therefore, green roofs have become a common aspect of new buildings.
With the popularization of ecological, green, and sustainable development, there has been an increase in research on green roofs. At present, research on green roofs has been mostly based in the United States and European Union (Blank et al., 2013). Although green roof research in China is fairly recent and there are no comprehensive studies, many scholars have explored the hydrological and economic effects of green roofs in Chinese cities. Based on the rainwater runoff retention capabilities of green roofs, many studies have used fixed-point monitoring based on green roof green roofs to confirm their feasibility in China. Several studies have investigated the optimal green roof configuration scheme, considering components such as the drainage layer materials, planting matrix, vegetation type, roof structure, and other elements (Hu et al., 2020; Li et al., 2019; Shen et al., 2020; Zhang et al., 2019; Zhai et al., 2015). However, most studies have focused on the configuration and composition of green roofs, without a direct comparison between green roofs and surface runoff, including their impact on urban flooding. Previous studies have used small-scale laboratory roofs, with instrument monitoring data typically used to evaluate their performance. Therefore, there is a lack of follow-up studies that further simulate the performance of large green roofs in a real urban environment. Although many studies have confirmed the functions of green roofs, research results cannot be directly applied to actual urban planning and construction projects (Liu et al., 2015). Thus, considering the constraints of time, space, and resources, computer models should be used to simulate the performance of green roofs (Babaei et al., 2018).
The construction of a sponge city is costly. According to the notice issued by the MOHURD (2015) of the People’s Republic of China, the yearly subsidy standard for pilot sponge cities is 600 million yuan for municipalities directly under the central government, 500 million yuan for provincial capital cities, and 400 million yuan for other cities. To better promote the effectiveness of sponge cities, we must establish a quantitative relationship between green roofs and surface runoff via a rainfall-runoff model to provide the basis for future decision-making and planning (Liu et al., 2014). Regional characteristics have a significant influence on green roofs; their performances vary according to geographical, climate, and hydrological conditions (Lee et al., 2012; Vijayaraghavan, 2016). Many scholars have emphasized the importance of green roof localization, reporting that roofs should be adapted to local conditions rather than following the same construction schemes used in other regions (Li et al., 2017; Ma et al., 2020; Williams et al., 2010).
Kunming is a rare plateau city; the effects that its special geographical climate and hydrological conditions have on the functionality of green roofs remain unknown. Therefore, this study took the Beichen District of Kunming as an example, used the Storm Water Management Model (SWMM) to simulate and predict the runoff reduction capacity of green roofs, and analyzed their impact on urban flooding based on five indicators: runoff volume reduction, runoff peak, peak delay time, number of ponding points, and longest ponding duration. In addition, the entropy weight-technique for order preference by similarity to an ideal solution (TOPSIS) method was used for Multi-Standard decision-making to select the optimal green roof construction scheme in the study area.