In 2019, record-breaking deforestation and fires in the Amazon Forest took place and brought more international attention to climate issues in Brazil. This could be connected to the increase in the number of publications from this year onwards. In fact, recent studies demonstrate the dynamics of deforestation-related fires and how the Amazon region is switching from a sink to a source of carbon to the atmosphere, an enormous challenge to be faced (Liu et al., 2020; Reddington et al., 2019; Gatti et al., 2021).
3.1) Climate change
3.1.1) Keywords analysis – Climate Change
Table 1 shows the top ten keywords and their number of occurrences for “Climate Change”.
Table 1
Top ten keywords in Climate Change publications by the University of São Paulo and number of times used.
Keyword*
|
Occurrences
|
impacts
|
159
|
Brazil
|
146
|
temperature
|
128
|
biodiversity
|
108
|
precipitation
|
106
|
Amazon
|
104
|
climate
|
99
|
forest
|
93
|
variability
|
91
|
carbon dioxide
|
89
|
Sources: Web of Science and Scielo Citation Index
* we omitted the keyword “climate change”
“Impacts”, “Brazil” and temperature” are the top three keywords, with 159, 146 and 128 occurrences, respectively. This indicates an expressive concern on climate change impacts on Brazil, particularly with temperature increase. This paints a broader picture, from which other more focused topics will branch from. For example, “Impacts” encompasses a wide array of contexts; there can be impacts on health, ecosystems, infrastructure, food production, precipitation regimes, and also urban/social impacts, such as financial losses and displacements (Benevolenza and DeRigne, 2019). So, while this keyword is very general, it represents attempts to address the main challenges of climate change in the Brazilian territory. Other keywords show concern with ecological issues, such as “biodiversity”, “forest”, and “amazon” (Paiva et al., 2020). The other most popular keywords seem to be concerned with atmospheric science studies, such as “precipitation”, “variability”, “carbon dioxide” and “climate” – however, “precipitation” and “climate” can also easily be linked to “impacts” or other topics.
To better clarify the relationship between keywords and identify the main research areas, we present the clustering of keywords for Climate Change in Fig. 2. We were able to identify four main clusters of investigation within climate change: 1) impacts on forest and plant development (green in Fig. 2), 2) land use and ecology (red in Fig. 2), 3) adaptation/governance (purple in Fig. 2), and 4) climate/atmospheric studies (blue in Fig. 2). The size of the circle is associated to the number of repetitions of each keyword individually, line thickness indicates how often they appear together, forming clusters of the same colours.
Cluster 1 (green in Fig. 2) is more focused on plant and vegetation from a perspective of the plant sciences, and possibly addressing biophysical processes (“photosynthesis”, “growth”) in a more specialized manner than cluster 2, such as analysis of the impacts on the productivity of plant species under scenarios of climate change in (Marin and Nassif, 2013; Grandis et al., 2021). Cluster 2 (orange/red in Fig. 2) shows keywords more connected to ecology and geography (“land use”), thus bringing a more systemic/interdisciplinary view of biodiversity and landscape, such as the coupling economic assessment with crop production and pollination services (Borges et al, 2020). These research fields represented by clusters 1 and 2 are both important quite complementary. Keywords such as “amazon” and “forest” are part of both fields and located at the edge between these clusters. “Land use”, a popular topic in urban studies, is part of cluster 2 but much closer to cluster 3 (purple in Fig. 2), which makes sense, since this cluster is more related to the human dimensions of climate change, (“adaptation”, “management”) (Serrao-Neumann, Di Giulio and Choy, 2020).
The main keyword “climate change” is part of cluster 3 – this implies that the keywords used in cluster 3 appear more often with climate change than the keywords in the other clusters. Cluster 3 is related to the social, political and institutional aspects of climate change. It is reasonable to suppose that the environmental problems related to climate change faced by the Brazilian society in the past few years (such as human-induced fires in forests) might have propelled a greater number of studies aimed at better understanding these aspects. Also, there are good examples of climate governance in other spheres of government, such as municipal climate plans, which are aligned with this scientific production (Carvalho et al., 2020). Finally, cluster 4 (blue in Fig. 2) is related to climatic and atmospheric studies which are more concerned with understanding the natural sciences basis of climate change (Braz et al, 2021). These studies are vital for building the climate science knowledge, as they form the basis upon which all the other more impacts-oriented studies are developed. Using tools such as modelling, these studies strive to make climate science as robust as possible (“model”, “variability”) for the understanding of climate change, usually for larger spatial scales than other clusters (“South America”).
Figure 3 shows the temporal evolution of the keywords and clusters related to climate change database. The blue cluster showed a slight change: “growth” and “responses” were left out in the latter years, while “patterns” and “evolution” were included, suggesting that the atmospheric sciences studies were possibly more concerned with the temporal evolution of the climate system attributes. The green cluster, concerned with ecological studies, initially showed a more continental scale focus (“south america”) while later, such studies were more focused on the Amazon region and its connection to climate (“amazon”, “rainforest”). Most of the ecological keywords such as “biomass” and “rainforest”, initially present in other clusters, moved to the green cluster in the later years, indicating that the amazon region has been receiving an increased and interdisciplinary attention in climate change research (Gatti et al, 2021). Outdated terms such as “global warming” were excluded from the later graph in the red cluster, while the keywords “drought”, “carbon” and “carbon dioxide” kept on, demonstrating a steady focus on the carbon balance and its association to drought episodes. Recent studies were more focused on plant sciences (“photosynthesis”, “plants”), demonstrating an important trend (Grandis et al., 2021). Finally, the purple cluster has become more attuned to interdisciplinary studies with the addition of “impacts”, “adaptation” “land use” and “ecosystem services”, which is quite relevant for successful climate adaptation strategies (Hobday et al., 2016). The change in the top 10 keywords is available as Supplementary Information 2 (climate change) and 3 (climate adaptation).
3.1.2) Origin of authorships – Climate change
The region of origin of authors is presented in Table 2 (international) and Fig. 4 (national). In Table 2, clusters were identified in VOS Viewer according to the most common partnerships in the papers, showing which countries in our database usually perform research together.
Table 2
Clustering according to the country of origin of co-authorships, considering “Climate Change”
Cluster 1 (18 items)
|
Cluster 2 (17 items)
|
Cluster 3 (11 items)
|
Cluster 4 (6 items)
|
Cluster 5 (2 items)
|
Argentina
|
Austria
|
Australia
|
Belgium
|
Brazil
|
Bolivia
|
Czech Republic
|
Canada
|
India
|
USA
|
Chile
|
Denmark
|
England
|
Israel
|
|
Colombia
|
Estonia
|
Ireland
|
New Zealand
|
|
Costa Rica
|
Finland
|
Japan
|
South Africa
|
|
Ecuador
|
France
|
China
|
Turkey
|
|
Ghana
|
Germany
|
Philippines
|
|
Kenya
|
Greece
|
South Korea
|
|
Malaysia
|
Italy
|
Taiwan
|
|
Mexico
|
Norway
|
Thailand
|
|
Netherlands
|
Poland
|
Vietnam
|
|
Panama
|
Portugal
|
|
Peru
|
Romania
|
|
Scotland
|
Russia
|
|
Singapore
|
Spain
|
|
Uruguay
|
Sweden
|
|
Venezuela
|
Switzerland
|
|
Wales
|
|
Sources: Web of Science and Scielo Citation Index
Although there are slight variations, cluster 1 is generally related to Latin American countries and cluster 2 is dominated by European countries. Asian countries prevail in cluster 3, cluster 4 has mixed “outliers” without a prevailing region, and cluster 5 is composed of only Brazil and USA.
Cluster 1 is the largest cluster, which demonstrates the relevant presence of many Latin American authorships in the scientific production in climate change from USP, investigating issues related to conditions experimented in Latin America. However, Brazil’s strongest association is with the USA. Perhaps partnerships originating from USP make it easier for scientific cooperation with this country. Cluster 2 is entirely made of European countries and the second largest cluster. It is interesting to note that countries were grouped by continent, so, we conclude that Brazilian research on climate change presents a high level of regional sectorization, which makes sense due to the fact that closer regions may experience similar impacts and therefore similar challenges when confronting climate change (IPCC, 2021; Moore et al., 2021).
Figure 4 shows the regional distribution of authorships among Brazilian regions. Research on climate change is dominated by institutions from the Southeast and Northeast regions, which are the most populous regions in Brazil. While concerns in the Southeast are usually related to the increase in extreme precipitation events (particularly due to the large urban population and its vulnerability), the Northeast is the driest region in Brazil, for which desertification has already been identified as potential impact (Marengo et al., 2020). Anyway, these regions concentrate more than 60% of the authorships in Brazil, but there are, in fact, institutions from all Brazilian regions, which demonstrates that USP maintains research partnerships across the country and thus can serve as a proxy for research carried out countrywide. In fact, our analysis revealed that the twelve institutions with most partnerships in this database were Brazilian. Apart from the Amazon Forest, which has been a constant focus of research, there is a need to improve the representativeness of the less densely populated regions in Brazil.
3.2) Climate adaptation
We had to carefully screen the papers in this database to understand whether their research was focused on biophysical/biological adaptation or climate adaptation in the form of societal responses, which is the focus of this study. Therefore, we classified the climate adaptation studies in four categories of relevance, as shown in Table 3, to encompass different instances and definitions in which “climate adaptation” may appear (Berrang-Ford, Pearce and Ford, 2015.). Category 1 is the closest to our definition of climate adaptation, with studies focused deliberately on climate adaptation actions, policies and case studies. In category 2, we grouped studies closely related to adaptation, such as vulnerability and infrastructure. Category 3 was assigned for studies in which there is a potential link to our definition of climate adaptation, and category 4 for those clearly focused on biological adaptation.
Table 3
Classification of papers according to relevance for the climate adaptation scope used in this study
Relevance to climate adaptation
|
Description
|
N. of papers
|
Percentage
|
1
|
Investigates a past, present or future issue in climate change adaptation within the scope of this study
|
57
|
37%
|
2
|
Focus on topics closely related to climate adaptation such as impacts, vulnerability or infrastructure
|
24
|
16%
|
3
|
There is a potential link to the adopted scope of climate adaptation
|
15
|
10%
|
4
|
Biological/genetic adaptation or other non-related topics
|
58
|
37%
|
Sources: Web of Science and Scielo Citation Index
Nearly a third of the papers (37%) were classified as category 4, and so, not included in the climate adaptation analysis (Figs. 5–7, Tables 4 and 5). However, the same percentage of papers (37%) were classified as the highest relevance (category 1), that is, studies explicitly concerned with climate adaptation actions and policies. A smaller percentage of studies were classified in the relevance 2 (16%), which are studies closely related to climate adaptation such as infrastructure, health, vulnerability, or deal with climate change impacts from an adaptation perspective. An even smaller percentage of studies (10%) were not clear in their link to our definition of climate adaptation, but still hold some potential and therefore were classified as relevance 3. Given the smaller amount of adaptation papers retrieved compared to the broader climate change literature, we decided to keep relevance 3 studies in our analysis.
3.2.1) Keywords analysis – Climate Adaptation
The same procedures (top ten keywords, cluster analysis, temporal evolution, and origin of authorships) were used to characterize scientific production on climate. Table 4 indicates that the same two keywords from the climate change analysis (section 3.1.1) – “impacts” and “Brazil” – were also the most popular in climate adaptation. However, the remaining keywords show a much closer resemblance to what was found in cluster 3 from the cluster analysis of the “climate change” string documents (purple in Fig. 2). Thus, the most frequent keywords for climate adaptation are related to sustainability, management, and policy-related studies, such as “cities”, “governance” and “sustainability”. “Vulnerability” stands out as the 3rd most used keyword, which is clearly associated with the relevance 2 studies from Table 3.
Table 4
Top ten keywords in climate adaptation publications by the University of São Paulo and number of times used:
Keyword
|
Occurrences
|
Impacts
|
20
|
Brazil
|
18
|
Vulnerability
|
14
|
Cities
|
9
|
Governance
|
9
|
Management
|
8
|
Temperature
|
8
|
Sustainability
|
7
|
Climate
|
6
|
Fisheries
|
6
|
Sources: Web of Science and Scielo Citation Index
Figure 5 shows the identified clusters for climate adaptation and the most frequently used keywords within them. Three main clusters were identified: 1) adaptation actions and policies; 2) urban environment, vulnerability and health and 3) food and coastal impacts (the keyword “Brazil” stood out from all clusters, as discussed below). Cluster 1, the largest (red in Fig. 5), is related to adaptation actions and policies, namely, strategies to improve climate adaptation and enhance societal and environmental resilience and quality of life. Management topics within this cluster can be identified by the keywords “policy”, “management”, “sustainability” and “challenges”, by the proposition of policies, or their evaluation (Sotto et al., 2019). Some of the most popular strategies or focus areas for climate adaptation can be identified by the keywords “ecosystem-based adaptation”, “water”, “adaptative capacity” and “agriculture” (Bustamante et al., 2019). This can be related to the relevance 1 studies from Table 3. Cluster 2 (red in Fig. 5), however, is more aligned with the urban environment, vulnerability, and health, thus, more aligned with relevance 2 studies in Table 3. The health concerns are evident from the keywords “mortality” and “vulnerability”, although the latter can be used in many different contexts, such as with “climate” and “cities”. The keyword “climate change” is part of this cluster, showing that, in our database for climate adaptation, the overarching topic of climate change is more strongly related to the impacts and vulnerabilities than the institutional approach revealed in cluster 1 (red in Fig. 5).
The keywords “mitigation” and “ecosystem services” show cluster 2 studies also try to include strategies characterized by co-benefits (both for adaptation and mitigation, for instance) in the attenuation of climatic impacts, for example, via green infrastructure, urban greenery, etc. (Bustamante et al., 2019). Cluster 3 (blue in Fig. 5) is smaller and focused on coastal impacts, possibly due to the threats posed by sea level rise, ocean acidification and impacts on marine or coastal ecosystems, shown by the keywords “marine ecosystems” and “coastal” (Popova et al, 2016). There is an evident concern with food production on coastal environments (“food security”, “fisheries” (Hobday et al., 2016)). This cluster also shows the keywords “impacts” and “adaptation”, which demonstrate that even though there is a wide umbrella of scientific production concerning climate adaptation, in Brazil it is strongly linked with coastal impacts. It is vital to study such present and future threats, as 25% of Brazil’s population lies on the coast, with many important cities of economic relevance, and the fact that coastal/ocean impacts are forecasted by many climate change scenarios (IPCC, 2021).
Still, there was less evidence of scientific production concerning climate adaptation for environments such as smaller cities, rural communities and certain biomes such as the Amazon Forest, including indigenous populations, which demonstrates a possible research gap in this area. It is also worth noting that “adaptation” was also a central keyword, strongly tied to all clusters (noticeable by its central position and line thickness to keywords of all clusters). The keyword “Brazil” (yellow in Fig. 5) was not strongly associated with any of the clusters in particular, but strongly connected with the main keywords “adaptation” and “climate change”. This suggests that all clusters produce literature relevant to Brazil.
The temporal evolution of the keywords for climate adaptation (Fig. 6) was slightly harder to perform than for climate change (Fig. 3), due to the smaller number of documents and their uneven yearly distribution. Anyway, a few aspects can be highlighted, such as a steady tendency to study climatic adaptation for coastal locations in Brazil, such as in the work of Popova et al (2016). Actually, the blue cluster in both periods suggest that, beyond understanding the nature of impacts (such as “fisheries” and “marine ecosystems” in the period 2007–2109), there has been an increased interest in understanding the change and evolution of impacts (such as in the keyword “change impacts” in 2020–2021). The evolution of keywords in the purple cluster suggests a recent trend in the institutional aspects of climate change adaptation, with the addition of keywords such as “policy” and “challenges”, and interestingly enough, “mitigation”, showing that there might be a trend in studies which align adaptation and mitigation, which is a promising path forward (Lampis et al., 2021). “Water” and “food security” were not as present in the most recent studies (2020–2021), but “vulnerability” remained throughout the entire period, which is extremely important for adaptation across the Brazilian territory (Pinho, Marengo and Smith, 2015). “Governance” as a separate keyword in the latter years indicates that not only this topic is important (since it was already present from 2007–2019) but also evenly distributed through all climate change adaptation clusters in 2020–2021, which is quite relevant, since the successful implementation of climate adaptation policies is heavily influenced by the interchange between different sectors of society (Pinho, Marengo and Smith, 2015; Serrao-Neumann, Di Giulio and Choy, 2020, Tagliari et al., 2023).
3.2.2) Origin of authorships – Climate adaptation
Table 5 shows the distribution of authorships in the international scale, and Fig. 7, in the national scale, for climate adaptation. In the international scale, Brazil’s scientific partnerships in climate adaptation is dominated by developed countries. There seems to be no pattern of regional division, as it was observed for the broader climate change literature (Table 2). The only Asian countries are in Cluster 1 – Japan and China – and the remaining countries in this cluster are European. Cluster 3 has the only two Latin American countries apart from Brazil (Argentina and Mexico) – together with France and Germany. Cluster 2 is composed of different countries from different continents. Overall, the profile is rather different than the broader climate change literature, with a much smaller range of countries, no regional division and dominated by developed countries, and with an absence of many Latin American and Asian countries present in the climate change research.
Table 5
Clustering according to the country of origin of co-authorships, considering climate adaptation:
Cluster 1 (8 items)
|
Cluster 2 (7 items)
|
Cluster 3 (11 items)
|
Canada
|
Australia
|
Argentina
|
China
|
Brazil
|
France
|
Ireland
|
England
|
Germany
|
Italy
|
India
|
Mexico
|
Japan
|
Netherlands
|
|
Spain
|
South Africa
|
|
Sweden
|
USA
|
|
Switzerland
|
|
|
Sources: Web of Science and Scielo Citation Index
Concerning the national scale, our analysis indicates that 10 out the 14 institutions with the largest number of co-authorships in climate adaptation are Brazilian, which indicates that USP’s production is somewhat representative of a national scale research and knowledge production. However, nearly two thirds (62%) are dominated by the Southeast and South regions, which are the most developed regions in Brazil. The rest presents low percentages such as the North with 14%, Northeast, 9% and Centre-west, 5%. As shown in our results, climate adaptation is a relatively newer concept compared to the broader climate change field in Brazil (Figs. 1, 4 and 8), therefore, it is natural that the richest regions will dominate the scientific production in a new trend. However, it is necessary that institutions from more rural and less developed regions in Brazil to adopt climate adaptation as a priority for research, considering they are home to Brazil’s endangered biomes, such as the Amazon Forest, the Caatinga (semi-arid zone) and the Cerrado (Brazilian savannah), as well as indigenous and other vulnerable population groups (Pinho, Marengo and Smith, 2015).
There are a few limitations to consider when interpreting the results of this work. Firstly, the institutional distinction: this project is part of a larger project evaluating USP’s dissemination of scientific knowledge, but it can provide a trustable proxy for the most recent research trends in Brazil. Another point to consider is related to the keywords and indexation processes. Articles relevant to climate change adaptation may have not been captured by our analysis because they did not use ‘climate change’ in their indexing (title, abstract and keywords), particularly if coming from more specialized research fields (which usually don’t employ interdisciplinary approaches or keywords). Finally, using more databases could encompass a wider range of scientific research, which could be performed in future works.