Revealing the Related Hereditarian Classification Patterns for Water-Soil Biogenic Elements and Riparian Buffer Characteristics: An Interpretable Stepwise Cluster Analysis Method

Abstract

Implementing zoned management for riparian buffer zones to control pollution is essential. Due to the inherent dynamic variability in pollution monitoring, the critical biogenic elements (C, N and P) to map the macro-scale pollution patterns within permissible buffer widths through the lens of stable background cause as well matter. In this study, an interpretable stepwise cluster analysis model was proposed for the biogenic element classification mapped by riparian buffer characteristics. The model could robustly identify the prior information from the buffer zones in multi-scale characteristics and effectively mitigate classification errors and bolster the credibility of the classification results. Pearl River Basin in China with 60 water-soil monitoring stations was taken as the study area for deeply searching landscape pollutants control schemes in riparian buffer zones. As data collection advances that will dynamically update classification value estimates, a more adaptive strategy to riparian buffer zone management can be achieved. The landscape of structure, connectivity and climate were leveraged in riparian buffers seek to reveal the hereditarian relationships behind the partitioning of the biogenic elements. The results indicated that the biogenic element classification could be mapped by riparian buffer characteristics with the optimal partition level of α in the interpretable stepwise cluster analysis model, and the results are valuable for supporting zoning adjustment of controllable characteristics on the existing riparian buffer management strategies.

Keywords: riparian buffer zones, landscape pollution management, carbon-nitrogen-phosphorus, interpretable clustering analysis