If you want to get “technical” about it, dentistry, is about 13,000 years old. However, while the fillings Neolithic peoples performed with only rocks and tar are impressive (and they are, jeez just imagine!), one would hardly make the case they form the basis of modern dental theory—which like modern chemistry, medicine, and many other fields, find their roots in the 18th and 19th centuries.
Ecology is not entirely different. Though Aristotle’s text “History of Animals” is one of the earliest biological texts, it wasn’t until the 18th century and later that works from Humboldt, Wallace, Linnaeus, Darwin, Vernandsky and many others formed the basis for the field of ecology as we know it—the literal definition of which we still bicker about. And it wasn’t until the 20th century, with the founding of the British Ecological Society (BES) in 1913 and the Ecological Society of America (ESA) in 1915, that ecology hit the big time.
Thanks to the combination of modern data science text mining methods and the nearly 100 year record of publications from ESA journals, authors Ji Yoon Kim, Gea-Jae Joo, and Yuno Do, have yielded a fascinating narrative of how ecological research has evolved over the past century. In their recent article, “Through 100 years of Ecological Society of America publications: development of ecological research topics and scientific collaborations,” published in Ecosphere, Kim and co-authors Joo and Do use modern text mining methods to scour the abstracts and titles of 22,179 publications from ESA journals, including Ecology (founded in 1920), Ecological Monographs (1931), Ecological Applications (1971), Frontiers in Ecology and the Environment (2003), and Ecosphere (2010). Their findings not only describe how the field has changed, but how it has been shaped by history and innovation.
THE BIG PICTURE
Using a network analysis approach called a word network map, the results show that ecological research over the preceding decade roughly clusters into four groups (the number in parentheses corresponds to the number of occurrences in a given abstracts/keywords for an article):
1) Population Ecology cluster (3578) – Including the keywords survival (1409), predator (1301), behavior (1118), mortality (1100), probability (879), selection (818), and movement (546). This cluster also included connections to the terms population dynamics, competition, predation, and behavior. The terms trait, stage, and risk have made recent ground in this cluster as well.
2) Conservation and Landscape Ecology (1976) included the terms management (1314), landscape (1337), estimate (1134), biodiversity (768), framework (718), and simulation (529). The terms restoration, ecosystem service, policy, scenario, and network, were all more recent.
3) Community and Forest Ecology cluster including the terms forest (2282), diversity (1466), tree (1301), disturbance (1175), plot (1050), composition (1045), and climate (919). Climate change, fire regime, community composition, functional group, and invasion were more recent terms.
4) Ecosystem Ecology* cluster, including ecosystem (2188 occurrences), biomass (1420), production (1222), soil (1170), treatment (1152), concentration (1018), productivity (933), nitrogen (862), and carbon (724). Species richness, flux, food web, and trophic cascade were more recent terms in the ecosystem ecology cluster.
These results are not actually that surprising. Given the recent 100 year anniversaries of both the BES and ESA, there has been some research and thought into how the members of each society view the “big ideas” of ecology. William Reiners of the University of Wyoming has headed up two such heady articles, in 2015 and 2017. In the 2017 Reiners et al. paper, survey respondents identified concepts such as scale, ecosystem, disturbance, population, community, etc. as the most vital to ecology–virtually mirroring the patterns and trends found in Kim and co-authors’ analyses.
We are, if anything, a persistent bunch.
But as Kim and co-authors point out, these patterns were firmly established from the onset. The initial composition of the ESA was a bit biased toward forest and plant ecology with self-identified research focus, here as a percentage of total membership, biased towards those disciplines: plant ecology 28.7%, animal ecology 28.0%, forestry 14.0%, entomology 12.7%, marine ecology 4.6%, agriculture 3.9%, and others 8.1%.
THE CHANGING LANDSCAPE
Perhaps it is my own bias, but the most fascinating aspect of this research, is the change. Some terms and research foci come and go. But taken as a whole, these trends tell the story of us. Not just ecologists, but all of us.
In the 1920s, there is a strong focus on the concept of “succession,” how the species structure of a system changes over time. Given the strong European and North American centering of the ESA and BES, this make perfect sense. 1910 to 1920 represents the peak of deforestation in the United States, with agricultural land-use at its maximum. Prior to the Green Revolution, and before commercial fertilizer became readily available, large swaths of land were needed to grow food for the growing nation. However, fields soon began to fall fallow for varies reason (e.g. continued industrialization, urbanization, failing economies, disasters, you name it), and forests began to regrow in many regions of the Eastern and Midwestern US. As forests regrew, it made sense to study them and understand them.
In the 1960s, the use of the term “disturbance” increases dramatically. During this time period, perhaps best punctuated by the publication of Rachel Carson’s Silent Spring in 1962, there is a growing awareness of the damage we have done to the earth system. In the post World War II era, there is growing recognition of the effects of pollution, disease, and negligence. During this time, we see the rise of the environmental movement, culminating in the first Earth Day in 1970.
Technology also drives research. The terms carbon and nitrogen begin to rise in the 1950s, likely a result of the maturation of the nuclear age. Sub-disciplines of ecology, such as stable isotope ecology, became more prominent in part due to the proliferation of knowledge, but also the more ready access to technology such as mass spectrometry. Understanding how isotopes “worked,” and being able to measure them accurately (and reliably!) resulted in a whole new addition to an ecologist’s tool kit.
Ultimately, as the authors point out, this study shows a measure of the degree of “interest” in a topic, rather than a true study of the development of ideas. Text mining approaches can only give an impression of change over time, rather than wholly interpret that change. Each corner of the ecology universe is focused on the next “big thing.” Whether it is using traits over species, new statistical tools (Bayesian takes an uptick in the late 90s by the way), or new technologies such as LiDAR (light detection and ranging) that simply didn’t exist during the last century, we’re always moving forward. And while some would may consider the slew of contemplative pieces that have arisen in the past couple of years that celebrate our various centennials as naval-gazing, this self-assessment is warranted. If anything, applying text mining and other data science techniques to past publications could be further expanded. Actually linking ecological concepts through time to search terms, would greatly increase the utility of this work. That allow us to move past “impressions of change” to an actual quantification of how concepts have shifted. Perhaps something like sentiment analysis or other computational linguistics tools could provide better insight. Science is constant reevaluation and reinvention.
*Ironically, in the 2017 article, “100 Articles Every Ecologist Should Read” from Courchamp and Bradshaw published in Nature Ecology and Evolution, many central, key manuscripts focused on these concepts were omitted. Whichone could interpret as giving weight to the argument presented here that ecosystem ecology and its related concepts are more “new school” ecology. It is of note that the the Kim et al. research outlined above differs strongly from that of Courchamp and Bradshaw, as Courchamp and Bradshaw curated survey data from the editorial boards of journals that were more explicitly focused on conservation and evolutionary ecology.