3.3 The Response of Stress and Riparian Health Indexing and Sub-Indexing
Pearson correlation revealed the inter-indicator relationships between stress indicators and RHIs, allowing us to understand riparian zone changes within the TGDR under land-use variations. Although PCA provided information on the key indicators responsible for the total variation in riparian zones with significant factor groups, Pearson correlation was performed for all RHIs and stress indicators, and valuable information was obtained independently for the key stress indicators (7) and RHIs (14). Pearson correlation relationships are described below.
With regards to the stress indicators, associations were continuously significant at p < 0.01 (|r | ≤ 0.933), and both positive (r ≤ 0.933) and negative (r ≤ ̵̶ 0.584) correlations were found between pressure indicators (Table 2). The highest correlations were generally observed in urban transects, whereas lower correlations were detected in rural transects (Supplementary A, Tables A.1–A.5).
The associations between RHIs within different geographical locations were widely significant at p < 0.01 or p< 0.05 (|r | ≤ 0.989), except for regeneration. Positive correlations were found in most of the situations for H (r ≤ 0.624), PC (r ≤ 0.989), R (r ≤ 0.455), Er (r ≤ 0.963), and Ex (r ≤ 0.935) (Table 2). The transects from urban areas showed the highest correlation, whereas those from rural–urban transitional areas showed a lower correlation. Stronger relationships were formed from PC, Ex, and Er indicators, whereas a relatively lower correlation was observed from those of H and R (Supplementary A, Tables A.1–A.5).
During the last phase of the Pearson correlation, the correlations between pressure and condition indicators for the different geographical locations were determined. These correlations were significant both atp < 0.01 and p < 0.05 (r = -0.731 – 0.529) between the indicators of pressure and the indicators of H, PC, R, Er, and Ex (Table 2). The highest comparative correlation strength was generally observed in urban transects, whereas the lowest strength was primarily found in rural–urban transitional transects. The results showed that pressure indicators correlated with habitat (r = 0.207 – 0.624), plant cover (r = -0.658 – 0.436), regeneration (r = -0.693 – 0.377), erosion (r = -0.731 – 0.583), and exotics (r = 0.168 – 0.529) (Supplementary A, Tables A.1–A.5).
[Table 2 to be inserted about here]
Heat maps were developed separately for stress and riparian health indices as well as sub-indices to indicate their correlation strength via their colors (Figure 7). Results showed that indices and sub-indices exposed unique features for different land-uses. The correlations were relatively positive in rural areas, where conditions correlated with plant cover (0.852**), habitat (0.646**), erosion (0.597**), and regeneration (0.574**). Comparably, correlation strengths were moderate (both positive and negative) in rural–urban transitional areas. Exotics correlated negatively with regeneration (-0.483**), plant cover (-0.469**), erosion (-0.332**), and condition (-0.273*). In contrast, condition showed a relatively positive correlation with plant cover (0.781**), erosion (0.587**), and habitat (0.468**). Urban zones displayed a markedly strong positive (in condition and plant cover) and negative (in exotic and pressure) correlation among other indices and sub-indices.
[Figure 7 to be inserted about here]