2.3.2. Estimation of urbanization and climate change contribution
In this study, we evaluate the contribution of climate change and urbanization based on the “trajectory” method (Feng, H. et al. 2014, Feng and Huihui 2016). At local scales, climate change and urbanization jointly influence the precipitation trends at some certain grids. Their potential contributions at a certain UA can be calculated as follows.
UE = Ru-Rr , (1)
UC =Ssub ×(Ru -Rr)/Ru×100% , (2)
CC = (Rr/Ru)×100% , (3)
where Ru and Rr are the precipitation trends in urban and rural areas, respectively. The urbanization effect (UE) is defined as the trend difference between the urban and the rural series in Eq. 1. Ssub is the proportion of urban grids to total grids. If the rural and the urban grids have the same number in the contribution analysis at regional scale, the parameter Ssub can still be remained for the sake of equation generality. The relative urbanization contribution (UC) in percentage is defined as the fraction of the overall trends of the extreme precipitation events attributed to the UE in Eq. 2. The relative climate change contribution (CC) is defined as the ratio of the precipitation trend in the rural areas to that in the urban areas in Eq. 3 (Gu, Zhang et al. 2019).
2.3.4. Extreme precipitation index
We adopt the precipitation data from April 1 to July 1 in this study, excluding the persistent heavy precipitation caused by summer monsoon. Four extreme precipitation indices are selected to represent the intensity of presummer precipitation over the LP based on the gridded data in the CMFD. The four indices include the maximum daily precipitation (Rmax), the very wet day precipitation (R95p) which is the total precipitation when daily precipitation > 95th percentile mean daily precipitation, the extremely heavy precipitation (R>10mm) which is the total precipitation when daily precipitation >10 mm and the number of light rain days (R<10mm) which is the count of days with daily precipitation < 10 mm during the presummer season.