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.