4. Discussion
This study highlighted the significant changes in the riparian zone, induced by pressure indicators, under different land uses within the TGDR in China. The results also revealed the extensive distribution pattern for wide-ranging RHIs and pressure indicators in different geographical locations (rural, rural–urban transitional, and urban areas). These significant deteriorations are highlighted by Arif et al. (2021) within the TGDR. However, the indicators all followed different patterns as the situations changed, similar to results obtained in other land-use change studies (Rodrigues et al., 2018; Wohl, 2017). Geographical locations and humans played a central role in riparian zone changes (Ferreira et al., 2005; Perry et al., 2012). Our results showed that rural–urban transitional areas offered relatively better riparian zone conditions than urban and rural areas (Figure 5). Similar results were obtained for all the health indicators, as shown by their relatively high score percentages. Rural–urban transitional areas are distinctive locations, as they exhibit the characteristics of both rural and urban regions, containing vegetation similar to rural sites but being managed as urban sites. Although the administrators of the TGDR paid more attention to urban areas than rural–urban transitional and rural regions, urban riparian zones were more disturbed due to development and concrete-based activities, which explains why the pressure impact was highest in urban regions. A particular area can hold a specific stress milieu that alters environmental circumstances depending on the prevailing land-use system (Zheng et al., 2021b; Johansen et al., 2007). Southwest China has a mountainous topography (Sang et al., 2019; Wang et al., 2016; Yang & Li, 2016), and riparian areas in this region are highly variable due to environmental changes caused by the contingent pressure indicators. As a result of the building of the TGDR, the whole riparian zone structure in the reservoir was modified (Zheng et al., 2021a). Geographical circumstances are responsible for the riparian vegetation at the time of the study being different from past natural cover (Wang et al., 2016; Yang et al., 2014). Similar results from other parts of the globe, such as Brazil (Mello et al., 2018), have led to recommendations to conserve riparian zones with unique characteristics, such as the TGDR.
The Kruskal-Wallis tests distinguished sharp contrasts between the different transects of geographical locations and RHIs. The test results confirmed the extent to which geographical locations significantly changed with respect to habitat, plant cover, regeneration, erosion, and exotic characteristics of the reservoir, as well as in the development of riparian vegetation, as demonstrated by Zema et al. (2018) using a different statistical approach. Dam construction can cause nearby land to change and, consequently, alter river features (Bombino et al., 2014). These morphological changes modify the river width and circumference, which change the river hydrology (Shieh et al., 2007). Multiple factors influence the dynamics of pressure indicators, and hydro-morphological changes are key reasons to boost their impact (Bombino et al., 2008; Galia et al., 2016; Hooke & Mant, 2000). As riparian health conditions respond to these variations when exposed to changing environments and land use, these fluctuations can influence the plant community in terms of structure and development (Steiger et al., 2005). However, the impact on riparian zones varies with transect location and magnitude of the fluctuation (Bombino et al., 2009). Although these transects were located in different geographical sites, their features were, to some extent, identical. The study sites were very similar in terms of vegetation indicators, structure, and the extent of riparian vegetation. The differences within a riparian structure can be attributed to the unique environmental characteristics of individual land uses. The interactions between geographical location and RHIs were always significant for rural and rural–urban transitional site indicators (H, PC, R, Er, Ex, and C). However, significant effects were detected on stress indicators among transect locations across the entire TGDR area. Zema et al. (2018) similarly found significant changes in riparian zone conditions, considering the environment and land-use variation, as expected within a dam territory, and multivariate statistical analysis confirmed similar results in buffer zones in Italy (Zema et al., 2015). These techniques have been widely employed and appear to be valid in ecological studies (Kazakis et al., 2018; Zema et al., 2015).
PCA was conducted for RHIs and pressure indicators, as these were correlated in an earlier study (Arif et al., 2021). The selected RHIs displayed more than 65.24% variance in the data set, while the pressure indicators accounted for 70.90% of the total variance (Figures 6). PCA indicated the extent to which the original or essential indicators modified variance during data configuration. The results of these loadings confirmed the close relationships between riparian health and stress groups by considering each group individually. Three critical components for each indexing, as shown by PCA, were defined for riparian health and pressure indicators. These factors expressed strong values, ranging from low to high (-0.624 – 0.963) from all indicators (positive and negative loads), as shown in Figures 6. The rotated component matrix values for riparian health and pressure showed that all factor groups fell within similar values. PCA displayed distinct similarity (positive values) and contrasts (negative values) from indexing and subsets (for habitats, plant cover, regeneration, erosion, and exotic parameters). Moreover, most of the pressure characteristics functioned as expected. This process demonstrates that PCA can aggregate key indicators without the need for a broader set of indicators (McIntosh, 2015; Zema et al., 2018).
By contrast, the loads on PCA were found to be insignificant (relative to evident loadings for characteristics of the riparian zone) among several individual indicators within different geographical locations in the TGDR. Regarding contribution and effectiveness, these indicators presumably played a role in transect locations regarding the geographic areas within the TGDR. The history of selective indicators can be traced from similar studies conducted on Australian riparian zones (Johansen et al., 2007; Johansen et al., 2008). Vegetation cover is a significant indicator of riparian health conditions (Oliveira et al., 2016; Wang et al., 2013) compared to other indicators. It is multilayered and may assist other indicators in performing their roles as a safety filter (Ding et al., 2021). However, the understory cover of riparian vegetation seemed more highly influenced than other vegetation characteristics and subsequently disappeared in areas of hydro-fluctuation. All RHIs, however, were affected by flow regulation. They displayed more drastic changes in the upper sections and a parallel decrease in the lower parts of the dam (Zema et al., 2018), indicating that the riparian health of different sites mostly depends on their geographical locations. As a result, this boosted the impact of erosion and increased the invasion of exotic species in the TGDR; similar observations have been recorded near other dams (Jian et al., 2018; Merritt & Cooper, 2000).
Pearson correlation equations applied to riparian pressure indicators are considered independent variables (x ), while those applied to RHIs are considered dependent variables (y ) in different geographical locations. These equation ranges are shown in Table 2 and Figure 7. The equations in Table 2 contained high descriptive levels (r = -0.731 – 0.529) and showed regression strength for the different transect areas (rural, rural–urban transitional, and urban). Following a more in-depth analysis of each of the regression strengths, the Pearson correlation values showed that RHIs influenced changes in the riparian zone condition, the impact of the pressure indicators modified riparian health, and riparian health characteristics varied according to the transect location in the TGDR (Supplementary A, Tables A.1–A.5). As expected, in our study the extent of riparian health response to the effects of postoperative pressure adjustment through environmental and land-use changes highlighted the hierarchical relationships among different transects (Mantyka-Pringle et al., 2014). We discovered that even under similar conditions, pressure indicators in these riparian areas might react in reverse. The pressure distribution pattern mainly depended on the geographical locations and their structural designs (Catford et al., 2013; Nilsson et al., 2013; Singer et al., 2013). In our study, this pressure increased the riparian condition fragmentation and altered the distribution pattern (latitude and longitude) in the TGDR area, resulting in different Pearson correlation strengths (Figure 7). Transects from urban areas displayed the highest correlation between RHIs, whereas these associations were relatively low in rural–urban transitional regions. Correlation strengths for pressure indicators were generally highest in urban transects and lowest in rural transects. Associations between pressure indicators and RHIs revealed that the highest correlation strengths were from urban areas, while the most moderate associations were from rural–urban transition areas. Our study can help assess the impact of pressure indicators that operate in complex patterns since pressure indicators are causing changes in riparian health, influenced by environmental and land-use changes in the TGDR. Our findings were consistent with those from several other studies (Petts & Gurnell, 2005; Rood, 2006; Tealdi, 2011), although not with those from certain specific riparian zones (Johansen et al., 2007; Johansen et al., 2008).
We found multiple scenarios (for parameters for habitat, plant cover, regeneration, erosion, and exotics) where the average pressure impacted on different land-uses. These results confirmed that riparian area conditions are not always affected by the same pressure indicators (Lamb et al., 2003), which is why urban areas showed significant structural changes and even displayed a strong reaction to pressure indicators, resulting in rapid changes in riparian conditions (Dempsey et al., 2017). Occasionally, pressure indicators did not significantly affect riparian conditions, resulting in steady and healthy conditions (Lamb et al., 2003). If a pressure indicator in a given area does not cause a significant long-term change in buffer areas, resulting in a stable state, the riparian conditions will not alter significantly. The present study showed that dependence on riparian area conditions strongly correlated with pressure on the indicators (Table 2). These indicators can cause the buffer zone to deform, causing structural changes and lowering the efficiency that maintains the riparian zone condition (Biswas & Mallik, 2010; Mallik et al., 2011).
The AHC method exposed dissimilarities between different land uses. According to published literature (Bombino et al., 2019), all geographical areas that retained parallel sites were clustered in the same set. It was evident that humans contributed to altering the environmental and land-use scenarios, which ultimately caused significant changes in riparian zones. Our findings agree with those from studies investigating other regions with characteristics comparable to the TGDR region (Rodrigues et al., 2018; Wohl, 2017).
Overall, our study attempted to eliminate traditional analytical methods based on field observations and instead assessed the deteriorated impact of the dam on habitats, vegetation, regeneration, erosion, and exotic indicators in different environmental and land-use conditions under the influence of pressure indicators. Pearson correlation provided a quantitative assessment of the riparian zone’s ecological responses to the pressure indicators of big dams. These correlation strengths confirmed that certain steps are required to develop management strategies that will minimize the impacts of the dam on riparian conditions, under the influence of environmental and land uses, as highlighted by others (Aguiar et al., 2016; Henry & Amoros, 1996; Mantyka-Pringle et al., 2014). We examined the changes in riparian zones under the influence of land-use changes and observed the effect of pressure indicators on rural, rural–urban transitional, and urban areas within the TGDR. This vast reservoir is a combination of multiple main streams and linked tributaries. Stream bodies emerge from different parts of China and have dissimilar topographies, resulting in variation in the riparian zones of these water bodies. Future research should investigate indicators that influence diverse water bodies in the TGDR; since the uses of these water bodies are different, pressure impacts could vary as well. Ultimately, riparian zones may experience an unequal degree of pressure. Therefore, future work should focus on changes in the riparian conditions of large dams and water bodies such as the TGDR. The results will provide comprehensive information to the reservoir administrator, thereby helping them implement functional changes commensurate with the condition of their respective water body.