Toward a cohesive understanding of ecological complexity

Abstract

Ecological systems are quintessentially complex systems. Understanding and being able to predict phenomena typical of complex systems is, therefore, critical to progress in ecology and conservation amidst escalating global environmental change. However, myriad definitions of complexity and excessive reliance on conventional scientific approaches hamper conceptual advances and synthesis. Ecological complexity may be better understood by following the solid theoretical basis of complex system science (CSS). We review features of ecological systems described within CSS and conduct bibliometric and text mining analyses to characterize articles that refer to ecological complexity. Our analyses demonstrate that the study of complexity in ecology is a highly heterogeneous, global endeavor that is only weakly related to CSS. Current research trends are typically organized around basic theory, scaling, and macroecology. We leverage our review and the generalities identified in our analyses to suggest a more coherent and cohesive way forward in the study of complexity in ecology.

Fig. 1.. Analytical roadmap.

Summary illustrating the stepwise process of data collection and analyses in this study. (A) Preliminary assessment of the literature done through a search on Web of Science. (B) Examination of the papers, books, and book chapters as well as (C) the standardized literature search and full-text extraction to search for the (D) 23 features identified on (B) in the full text of articles retrieved in (C). (E) Analyses on complexity and control articles in the search for generalities in the field of ecological complexity. TS, topic; WC, Web of Science categories; TI, title; AK, author keywords.

Fig. 2.. The study of ecological complexity in space and time.

(A) Global network of collaborations including all authors from the articles that referred to “ecological complexity” in their title or keywords (n = 172). Points represent researchers’ affiliation addresses, and lines indicate collaboration between authors. (B) Cumulative production (from 1970 to 2021) between articles mentioning “complexity” in their titles and abstracts including all scientific fields (gray line) and, separately, for ecology and environmental sciences, as approximated by the search term “ecological complexity” (red line).

Fig. 3.. Comparison between control and complexity articles.

Comparison between control (gray) and complexity (red) groups considering the features retrieved by the systematic mapping (listed in Table 1). The control group includes articles randomly selected from the ecological literature, and the complexity group includes articles explicitly referring to “ecological complexity” in their title or keywords. Note that six articles (control = 4, complexity = 2) did not include any of the features described in Table 1 and were excluded from the analysis. (A) The richness of features of each article and (B) the exponential of the Shannon entropy calculated on relative frequency of feature usage were significantly higher in the complexity articles. (C) Study uniqueness (i.e., the distance from each article to its group median) was smaller in complexity articles, indicating that these were typically more similar among themselves. (D) The relationship between study uniqueness and feature richness shows that articles mentioning fewer features were on average more distant from their group mean, suggesting that these features were rarely mentioned by other articles. In (A) to (C), the data distributions are depicted with a kernel density plot with a dot representing the median value, and a box-and-whisker plot with outliers representing the minimum, Q1, median, Q3, and maximum with the length of 1.5 × the interquartile range.

Fig. 4.. Connections among complexity features in ecology.

This unipartite network shows the projection of a bipartite network linking complexity features (Table 1) based on their co-occurrence in the “complexity” group of articles. Features (nodes of the network) are shown, with more red color indicating that features are more significantly associated with the complexity articles based on indicator species analysis. Co-occurrence strength (edges) is represented by the sum of the edge weights of the adjacent edges of the node.

Fig. 5.. Seminal literature and the topic clusters in the ecological complexity literature.

Weighted co-citation network for the top 100 co-cited articles in the complexity articles. The colors reflect co-citation clusters: foundational complexity theory [(18); in blue]; scaling, hierarchies, and cross-scale dynamics [(70); in gold]; and macroecological theory and large-scale systems [(2); in pink]. Two additional clusters [(43, 151); in gray] count 10 or less articles and emerged from the use of “ecological complexity” in a more specific context [e.g., pest control in agriculture (151)].