Topic modelling of the wetland condition assessment literature reveals trends, key gaps, and opportunities for combining different assessment techniques

Wetlands, crucial for global biodiversity and ecosystem services, face escalating degradation due to human activities. Despite the growing urgency to track and improve wetland condition, challenges persist in developing adaptable and comprehensive assessment methods. This study employed advanced topic modelling techniques to analyze the wetland condition assessment literature, revealing trends, methods, and research gaps. Topic modelling was applied to 1,969 articles from Web of Science, using Latent Dirichlet Allocation (LDA) to identify topics that represent the key ideas based on the co-occurrence pattern of words from abstracts, titles, and keywords. Our analysis identified diverse topics such as ‘habitat for biodiversity’ ‘climate change’, and ‘scenario modelling’. We assessed the similarity and popularity of these topics over time, revealing dynamic trends in wetland condition assessment research. A research gap analysis was also undertaken to identify pairs of topics separated in both thematic content and co-occurrence within articles. The findings highlight the evolution of research interests, with some topics gaining prominence while others decline over time. For instance, ‘remote sensing’, ‘contaminant accumulation’, ‘wetland degradation’, ‘ecosystem services’ and ‘climate change’ emerged as hot topics, reflecting the increasing emphasis on the development of new technologies and the concern about wetland change due to anthropogenic impacts. In contrast, topics like ‘biological and ecological integrity’, ‘forested wetlands’, and ‘wetland boundary delineation’ showed declining prominence over time, indicating a potential shift in research focus. We identified a gap between the ‘remote sensing’ topic and most other topics, showing the importance of combining, for example, water quality and biodiversity data with remote sensing analyses for wetland condition assessment. The topic of ‘contaminant accumulation’ was also distant from other topics, bringing attention to the need to interconnect this topic with other approaches to wetland condition assessment. Our results highlight the need for policies that prioritize the integration of advanced technologies, such as remote sensing, with traditional ecological indicators to enable more holistic wetland assessments. These findings also emphasize the importance of region-specific conservation strategies to address challenges such as pollution and land-use changes, while fostering global efforts to harmonize monitoring frameworks and enhance the resilience of wetlands to climate change. This innovative approach demonstrates the potential of automated text analysis in rapidly synthesizing and comprehending vast collections of scientific literature, offering a holistic overview of wetland condition assessment research worldwide.