The move is partly thanks to UN efforts, according to Carlo Giupponi of Ca' Foscari University, Italy, and Claudio Biscaro of Johannes Kepler Universität in Austria, who analysed developments by combining data mining of bibliographic databases with a standard literature review. The pair published their work in Environmental Research Reviews, a new section in Environmental Research Letters (ERL).

"I have always been fascinated by the role that language plays in our life and in particular by the barriers we find when we have to use terms with contested or ambiguous definitions," Giupponi told environmentalresearchweb. "This is evident in science, when the same generic terms – vulnerability, resilience, risk – are used in different but related fields with reference to contrasting definitions."

To find out more, Giupponi teamed up with Biscaro, who is an expert in machine learning. Together, the pair employed techniques such as network analysis to assess how collaborations within and across research communities have determined the evolution of concepts and definitions. They analysed 3757 papers from the ISI Web of Science database in this "very lively" research environment – 800 on disaster risk reduction, 2639 on climate change adaptation and 155 that crossed both fields.

"We demonstrated the growing interactions of two distinct research communities and how the increasing collaborations and consequent terminological convergence were catalysed by international organisations, namely the IPCC [Intergovernmental Panel on Climate Change] for climate-change adaptation and UNISDR [United Nations International Strategy on Disaster Reduction] for disaster-risk reduction," said Giupponi.

Concentration and focus

The literature on climate-change adaptation, the researchers found, tended to cover ecological issues, policy and governance, food security, assessment methods, resilience and climate science. Disaster-risk research was generally more concentrated on hazards and disasters and, to some extent, indicators, indexes and maps. Typically, the papers spanning both topics focused on policy and governance, resilience, and hazards and disasters.

Disaster-risk literature developed first and it was rare for disaster-risk-reduction specialists to publish on climate-change adaptation, the study showed. But around 10–15% of researchers who published mainly on climate-change adaptation also wrote papers for the disaster-risk sector. Since the early 1990s, the proportion of authors publishing in only one research stream has decreased from 0.88 to 0.80, indicating intensifying relationships between the two communities.

Over time, the disaster-risk literature showed more interest in data analysis to produce vulnerability indices and maps, and less interest in hazards and disasters. The make-up of topics in the climate-change adaptation publications stayed roughly the same.

In 2014, working-group II of the IPCC's Fifth Assessment Report focused on "Impacts, Adaptation and Vulnerability". Reflecting "progress in science", the group included a much more vague definition of vulnerability than the 2007 Fourth Assessment Report, which considered the concept as the degree of susceptibility of a system to the adverse effects of climate change. In 2009, the UNISDR, said Giupponi and Biscaro, which doesn't have a specific focus on climate change, defined vulnerability according to susceptibility to the damaging effects of a hazard, and talked about community as well as acknowledging that vulnerability may arise from different combinations of "physical, social, economic and environmental factors". Others feel that the term vulnerability can denigrate large parts of the world and prefer alternatives such as resilience.

Authors involved in the IPCC's 2012 Special Report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation (SREX) were distributed between climate-change adaptation and disaster-risk reduction, showing "the convergence of key authors of the two communities in a joint effort under the auspices of the IPCC".

Giupponi and Biscaro reckon that without adequate consideration of the documents and reports produced by UN institutions, it is impossible to understand the evolution of scientific literature; the UN documents are key references for definitions of terms but also catalyse the scientific literature. SREX and the Fifth Assessment Report helped to reconcile contrasting definitions, the researchers said, by defining vulnerability concisely in a way close to the disaster-risk tradition, making vulnerability one of the elements of risk. By including outcome and context vulnerability as extra definitions, the reports also highlighted the impossibility of adopting a single definition.

Machine versus convention

As well as network analysis, Giupponi and Biscaro carried out a conventional literature review, including papers that didn't fit the search terms for the automated assessment, for example those from fields such as sustainability science. This review also showed that the climate-change-adaptation community has picked up on concepts typical of disaster-risk publications, such as the notion of risk, its assessment and management.

"We believe we demonstrated how traditional literature review can benefit by the integration of machine-learning techniques," said Giupponi. "For example, the latter allows the classification of topics and the calculation of quantitative indicators and their evolution over time, while only literature review can allow us to consolidate the identification of seminal papers, understand the meaning of quantitative indicators, and eventually discuss results and derive conclusive statements."

Indeed, the literature review revealed one of the surprises for the researchers – the role played by international organisations. This was not detectable "without accurate literature review going beyond bibliographic databases to examine documents that are usually defined as grey literature". The pair were also impressed by the abilities of machine-learning techniques to "support our reflections through effective graphical representations of the dynamics of literature evolution".

So what's next? "Imagining that all future works on vulnerability will conform to the same glossary is out of consideration, and it would also mean that we would exclude the possibility of any further conceptual development," said Giupponi. "What we recommend instead is more research efforts focused on providing solutions for the – possibly quantitative – assessment of the two main interpretations of vulnerability as an end-point and as a starting point. Algorithmic approaches will substantially reduce the ambiguity of the concepts, but also they will contribute to consolidate the interpretation of related terms, which are at least as ambiguous as vulnerability and which quite likely will appear as related variables, for example exposure, sensitivity and susceptibility."

Giupponi and Biscaro plan to continue studying the main concepts at the basis of global change and sustainability sciences, apply their approach to institutional networks, and look at the interplay between other scientific communities and international organisations, such as biodiversity studies and the recently established Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES).

One thing is clear – vulnerability isn't going away anytime soon. "The ever greater relevance of global change in our lives will increase even more the interest of scholars for assessing the vulnerability of socio-ecosystems," said Giupponi. "The complexity of those systems will limit the possibility of providing policy/decision makers with the robust modelling approaches needed to accurately simulate future evolutions and what-if scenarios soon, but that should be within the ambitions of scholars in the field in the coming years."

Related links

Related stories