2.1 Study area
The Wannan Mountains, locating at the south of Anhui Province, China, are composed mainly of the Jiuhua Mountain, Huangshan Mountain, and Tianmu Mountain, with the highest elevations of 1,342 m; 1,841 m; and 1,787 m, respectively. Due to the influence of the subtropical monsoon climate, this area is characterized by asymmetric seasonal temperature and precipitation distributions. Annual air temperature ranges from –2.1 °C (January) to 27.5 °C (July) with the mean of approximately 17.8 °C. Annual rainfall is 1,100–2,500 mm/year, with a mean of approximately 1,900 mm/year, of which more than 60% of the rainfall occurs from May to August. The upland headwater streams in this area are highly complex and abundant, these streams are confluent to the Xin’an, Chang, Qiupu, and Qingyi rivers, among which the Xin’an River belongs to the Qiantang watershed, and the other three rivers belong to the Yangtze River.
Fifty-three low-head dams in the first-order headwater streams of the Wannan Mountains were surveyed, of which 21, 9, 4, and 19 low-head dams were located in the Xin’an, Chang, Qiupu, and Qingyi rivers, respectively (Figure 1). Overall, low-head dams (< 5 m in dam height) were built in the tributaries for agricultural irrigation and water supply in this region (Yan et al., 2013; Li et al., 2021). For more details about the dam size, see the results of Li et al., 2021. Given the constraints of time and lack of historical data, biotic homogenization can be studied effectively as a spatial process (reviewed in Dar & Reshi, 2014). The spatial patterns may also provide tentative evidence for changes in current fish assemblages (Liu et al., 2019). As fish dispersal is restricted to via-watercourse routes in the river ecosystems, low-head dams can affect stream fish assemblages by changing habitat conditions and by blocking upstream movement (Porto et al., 1999; Yan et al., 2013). Compared with the free-flowing segments below dams, choosing the references above dams usually ignore the dams blocking fish movement, which may underestimate the ecological effects of low-head dams on current fish assemblages (Li et al., 2021). Therefore, two types of sampling sites were set for each low-head dam: the impoundments created by low-head dams (i.e., the treatment sites) and the free-flowing segments below low-head dams (i.e., the reference sites). In addition, the effects of low-head dams on geomorphological footprints are often less than 2 km (Fencl et al., 2015), and the free-flowing segments were ensured at least 2 km from the dams.