Introduction

The international scientific consensus on the responsibility of human beings to act on climate change (CC) has meant that the attention of researchers is not solely focused on the issue of global warming itself, but that especial attention is also paid on the socio-cultural aspects of CC, including the practices and discourses that surround this issue. The construction and circulation of discourses in contemporary societies is developed through mass media, which explains the strong interest generated in the last few decades in the study of these forms of communication that transmit content regarding CC.

Media attention on CC in English-language newspapers went from 700 references in the western media in 1988 to 7,500 in 2006. In the United Kingdom, it quadrupled between 2003 and 2006 due to the success of Al Gore's documentary "An Inconvenient Truth" and the impact of the Stern Report (^{Boykoff and Roberts, 2007: 6}). ^{Schmidt, Ivanova and Schäfer (2013)} demonstrate that CC is a relevant issue in 37 media outlets from 27 countries that were included in their study, representing an average of 0.62% of all news items published between 1997 and 2009. This is compared to media attention for other science-related issues such as stem cell research or human genome, which received less coverage in countries such as Germany, France and the United States over the same period.

This general interest has also been demonstrated in the wide range of scientific literature available on how information related to CC has been covered. Studies have demonstrated that media attention does not develop in a linear manner, but it fluctuates because of spikes in attention and specific events (^{Boykoff, 2011}). This is particularly the case during the international Conferences of the Parties (COP) for the United Nations Framework Convention on Climate Change (UNFCCC), more commonly known as the United Nations Climate Change Conferences (^{Brossard, Shanahan and McComas, 2004}; ^{Anderson, 2009}; ^{Sampei and Aoyagi-Ussui, 2009}; ^{Painter, 2010}; ^{Liu, Lindquist and Vedlitz, 2011}; ^{Schmidt et al., 2013}; ^{Kunelius and Eide, 2012}).

The majority of the studies have been focused on the media coverage in the United States and Europe. If the studies did include other regions, they were generally limited to english-speaking countries. This is why ^{Anderson (2009: 176)} states that future research could explore the reasons for different levels of coverage in different countries, their different emphases and why the views of developing countries are rarely aired. Similarly, it is important to consider that there is a need to increase the studies centered on online media given the enormous growth of the news websites and the Internet as a source of news, and the parallel decline of printed news media.

Due to all of these reasons, this study aims to be a starting point for the analysis of online media coverage of CC in Spanish-speaking countries with a specific focus on the annual COP's. Following ^{Kunelius and Eide (2012)}, "the summits force different kinds of actors and forms of knowledge into a compressed time and space, providing an opportunity for researching climate change politics, social climate, political theory in general and climate journalism in particular" (2012: 268). They understand that COP offers a framework for the trans-national representation of two areas: the politics of CC and climate journalism. For the political actors it is clear that economic interests, technological knowledge, international alliances and the ownership of resources, such as carbon or oil, determine their position in the negotiations.

This study builds on previous studies by ^{Pinuel et al. (2012)} and ^{Suárez, Arcila and Pinuel (2013)} and describes the way in which online media in Spanish-Speaking countries are presenting news on CC during COP16 in Cancun and COP17 in Durban, held in December 2010 and December 2011, respectively.

Media coverage of climate Change

The case of CC has especial relevance for social studies. Official reports from international agencies (^{Pachauri and Reisinger, 2007}) state that the consequences of CC could be mitigated with a change in social behavior, which includes the adoption of socio-economic practices that are environmentally friendly. Therefore, a process to raise social awareness about the issue is necessary. The media have an important role to play in this area.

^{Boykoff and Roberts (2007: 4)} highlight the first references to global warming in printed media in the 1930's. However, it was not until the 1980's when the issue was highlighted in the media. ^{McComas and Shanahan (1999)}, in their analysis of the media coverage of CC in The New York Times and The Washington Post from 1980 to 1995, identified a peak of interest from these two newspapers in 1988, when the articles had a more pressing and occasionally a desperate tone.

As ^{Ungar (1992)} concluded, in order to be part of the media agenda it is necessary for CC to have an impact on the real world: "Environmental problems like CC must 'piggyback on dramatic real-world events" (1992: 483). Even if CC was in the media decades before, it was the 1988 heatwave across the United States that awoke interest in the issue. That year also saw the formation of the Intergovernmental Panel on Climate Change (IPCC), established by the United Nations Environment Program (UNEP) and the World Meteorological Organization (WMO). ^{Jaspal and Nerlich (2014: 2)} also state that 1988 "can therefore be seen as a crest of a wave of social, political and media recognition of climate change as a global problem in need of global solutions". CC had transcended the field of science and entered into the socio-political sphere, resulting in an increase in media coverage in the United States (^{Ungar, 1992}; ^{Trumbo, 1996}), the United Kingdom (^{Carvalho and Burguess, 2005}) and Germany (^{Weingart, Engels and Pansegrau, 2000}).

The same trend has appeared in other countries. The Project Global Media Map on Climate Change at the University of Hamburg, which measures media attention for CC over a 15-year period in 23 countries from across the World, shows that the coverage of CC has slowly increased as from 1996, and in particular after 2004 in western countries. Spikes in coverage were noted when the IPCC IV Report was published in 2007 and during the Copenhagen Conference in 2009 (^{Schmidt et al., 2013}).

^{Cottle (2009: 506)} also recognized 2007 as "a transformative moment in the news career of climate change". CC finally came of age as a "global crisis" demanding responses from all the world's nations. Cottle focuses on the use of a visual environmental rhetoric. The western news media's spectacular visualization of CC through dramatic and symbolic scenes collected from around the world has undoubtedly helped to establish its status as a widely recognized global crisis.

The evolution of the number of references to CC in the mass media can also be demonstrated with a Spanish-language example, that of the El País newspaper in Spain. From the first appearance of this term in 1976 until 2009, there were a total 3,547 references over 34 years (^{Mercado, 2012a: 443-444}). In the first ten years of this period there were only 15 mentions of the term "climate change". In 1988 there were four articles that mentioned CC, which highlights an enormous difference in the media coverage of this phenomenon compared to that of the United States due to the different socio-political, economic and journalistic contexts of the two countries. The Earth Summit held in Rio de Janeiro in 1992 resulted in the number of mentions in the El País newspaper reaching 62 that year. Up to 1995 there still were a low number of references about CC (96). The negotiations prior to and during COP3 in Kyoto in 1997, generated hundreds of mentions of the term CC in El País. The reduction of emissions was the main focus of El País until 2007, when CC reached its maximum level of media coverage in the newspaper with 819 mentions.

Theoretical Framework

This investigation analyses the journalistic coverage of CC from the theoretical perspective of framing, which is understood as the process by which news organizations and journalists feature, emphasize, and/or select certain events, issues, or sources to cover over others (^{Nisbet, 2009}). The framing theory implies that information content not only sets the public agenda, but also implicitly forms a way of thinking about certain issues through news frames.

^{Entman (1993)} explained that frames in the news can be examined and identified by the presence or absence of certain keywords, stock phrases or particular sentences, stereotyped images, and sources. ^{Nisbet (2009)} highlights the importance of sources in the news agenda, which are defined as the featured voices, actors or groups in news coverage, such as government officials or environmentalists.

According to ^{Entman (1993: 52)} framing is to ''select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation''. In this sense, frames can be understood as the way in which information is presented by: i) the source quoted; ii) the position of the discourse; and, iii) the recommendation, this is to say, the message of its conclusion.

Method

The goal of this study is examining the treatment given by Spanish-language digital media to the news coverage of CC during the Cancun (2010) and Durban (2012) conferences. To achieve this objective, content analysis was carried out on a sample of news published in Spanish-language media websites, following the established criteria for this quantitative technique (^{Krippendorff, 2004}). Several methods and approaches have been used in CC media coverage research, but a recent overview of the literature in the field has shown that this variety makes "comparison and analysis in one coherent framework much more difficult" (^{Schäfer and Schlichting, 2014: 145}). To address this difficulty and contribute to the vast body of research in the area, this study has chosen the quantitative approach, which is still predominant in the investigation of media coverage of CC (^{Schäfer and Schlichting, 2014}) and of science topics (Schäfer, 2010).

In relation to the treatment of the content and based on the framing theory, it is possible to focus attention on very different elements, ranging from the selection of sources included in the content to the position of the discourse used to interpret the events covered by the articles. This study has placed attention on both aspects; in the first place, the source (who speaks?) in both their selection and position in the coverage represents an initial element of treatment by the media. On the other hand, the way in which information is framed by the position of the discourse (if the information expresses conformity with the issue or clashes with it) and the message of its conclusion (if it proposes solutions or highlights impasses) also affect how readers perceive this issue. These variables, based on the work by ^{Pinuel et al. (2012)}, allow for an analysis of the way in which discourses on CC are presented by online media. Analysis of these variables also facilitates an understanding of how online media presents information and how controversy surrounding this issue is communicated to citizens.

Findings

Once data were collected, the analysis of the results was carried out using SPSS (Statistical Package for the Social Sciences) software. The central trends of the variables studied were extracted (measures of frequencies, averages and percentages). This was followed by a non-parametric test known as the Chi-squared test of independence (X²⁾, with the goal of identifying the existence or not of a statistically significant association between the categorical variables.

It was found that the media coverage of CC during the Cancun and Durban conferences by news websites demonstrated similar trends (RQ). However, there was an important reduction in the number of news items published in Spanish-language digital media on the issue of CC from 2010 to 2011, dropping from 626 news items to 263.

As indicated in Table 1, if the total sample (N= 889) is examined, it can be observed that the subjects who are used as sources in articles referring to CC (RQ1) are primarily politicians (36.8%), followed by witnesses (20.9%), experts (14.6%), and social organizations (11.7%). These trends were consistent between 2010 and 2011.

Within the news it was found a total of 605 names used as sources and mentioned 1,882 times. Only 34 persons were mentioned more than 10 times and of this group just 4 persons were scientists: Mario Molina, Nobel Prize in Chemistry for his role in elucidating the threat to the Earth's ozone layer of chlorofluorocarbon gases (n =12); Rajandra Pachauri, Chairman of IPCC (n=13); Michel Jarraud (n=14) and Jeremiah Lengoasa, Secretary-General and Deputy Secretary-General of WMO of the World Meteorological Organization (WMO) (n=10). Additionally, one ecologist was incorporated as a source: Gustavo Ampugnani, Latin American Policy Coordinator at Greenpeace (n=10). UN high-level representatives were also mentioned. Specifically, Christiana Figueres, Executive Secretary of UNFCCC (n=52), and Ban Ki-moon, UN Secretary-General of (n=46).

Mexican President Felipe Calderon (n = 94) was the most mentioned source in our study (RQ2). We can notice that Mexican media had a more complete coverage of COP16 due to the proximity. Calderon appeared more times than Patricia Espinosa (n = 54), who was his Minister of Foreign Affairs. The second most quoted person was Evo Morales (n=79), President of Bolivia, who during COP16 called for a deal to "save the planet" and sentenced that any agreement must include the "rights of Mother Earth". Thus, the speech of Pablo Solon Romero, Ambassador of Bolivia to the UN and negotiator in Cancun, was covered 32 times by the media. Politicians of other Latin American countries were also significantly quoted in news: Ecuador (President Rafael Correa, n= 38; María Fernanda Espinosa, n = 29); Venezuela (President Hugo Chavez, n= 35; Claudia Salerno, n= 12); Brazil (President Luiz Inacio Lula da Silva, n= 26); Argentina (President Cristina Fernandez, n= 23) or Colombia (President Juan Manuel Santos, n=19).

Regarding the frames produced by the position of the discourse (RQ3.1), the findings demonstrated that in general the majority of the information did not commit to a particular position (49.27%), although an important portion of these news items expressed confrontation (39.37%) within their content. As it is noticed in Table 2, only a small percentage expressed conformity (11.36%). Similarly to the previous category, there were no significant differences between the samples from 2010 and 2011.

Although the position of the discourse provides clues about the attributes and focus of the articles, the frames of the proposals of the articles' conclusions were also analyzed (RQ3.2). This category records whether or not a solution was proposed within the article for the problems generated by CC. Table 3 demonstrates that, in general terms, 40% of the articles did not propose a problem or solution in the news item (a category that reduced considerably from 2010 to 2011), as there was no clear position provided in the general conclusion of the article. However, the study found that almost 40% of other news items clearly proposed a solution, which means that they included proposals based on improving interactions with the environment and/or within the same social system (^{Pinuel et al., 2012}). This sub-category remained constant between 2010 and 2011. Finally, 19.57% of the news articles presented an impasse, which means that even though they proposed a problem they did not offer any type of solution. The percentage of articles with impasses increased from 2010 to 2011.

Almost a half (47%) of the sample of news items studied was from news wire services (RQ4), mainly international agencies such as EFE (36.1%), AFP (20.49%), AP (16.34%), and to a lesser extent regional and national news agencies such as Notimex (6.34%). These wires were published without any editing. In the case of the images, videos and podcasts that were included in the publication of the news item, a large number of these were not attributed to any source (Table 4). Even when AP or AFP has nowadays services in Spanish, EFE (based in Spain) was the first international news wire service exclusively in Spanish with a more complete coverage of Latin American issues, and thus its role in the Spanish-Language media is very relevant.

Politicians were the most quoted sources in the news from wire services (57.4%) and in the news written by media reporters (51.3%). There were not differences in the position presented in the news from wire services and those from reporters: news items that do not commit to any particular position are 47% in those from wire services and 52% in those from reporters; news that express confrontation are 39% in the case of wire services and 38% in the case of reporters; and news that express conformity are 13% in wires and 9% in media reporters' news.

In relation to the proposal of conclusion, in 20% of the news from wire services and from media reporters there is no solution to the problem (caused by CC). Against one could imagine, proposals for solutions are presented in wires to a greater extent (43.2%), compared to news written by media reporters (36.3%). Furthermore, the percentage of news items in which the final paragraphs do not propose a problem or solution was higher in those written by media reporters (43%) than those from wire services (37.6%).

With the goal of finding associations between the inclusion of a source and the news frame (position of the discourse and proposal of conclusion) a series of statistical tests (X²⁾ were carried out (RQ5).

When victims [X² (2, N = 889) = 26,383, p <.01] (|3.8| > 2.58) or activists [X² (2, N = 889) = 19,055, p <.01] (|2.6| > 2.58) appear, there is a trend for the news item to express confrontation. Similarly, when one of the sources is a politician there is a tendency for the article to commit to a particular position [X² (2, N = 889) = 21,460, p <.05] (|−2.2| < -1.96), but the analysis does not reveal if it is either conformity or confrontation. Clearly, witnesses were less likely to commit to a particular position [X² (2, N = 889) = 28,830, p <.01] (|3.1| > 2.58) and scientists less likely to express conformity [X² (2, N = 889) = 8,994, p <.05] (|−2.2| < -1.96). According to data, the presence of scientists [X² (2, N = 889) = 36,213, p <.01] (|3.4| > 2.58) and politicians [X² (2, N = 889) = 14,126, p <.01] (|1.9| > 1.96) is associated with a greater possibility that a solution is proposed in the news item, but when witnesses [X² (2, N = 889) = 14,126, p <.01] (|3.1| > 2.58), activists [X² (2, N = 889) = 9,188, p <.05] (|1.7| > 1.96) and victims [X² (2, N = 889) = 9,144, p <.05] (|2.0| > 1.96) appear the trend is to not propose a solution.

Additionally, we found that the relationship between the origin of the news item (news wire service, reporter, mixed) and the participation of a politician in the article was significant (X² (2, N = 889) = 7,840, p <.05*/Fisher's exact=0.019 < 0.05), but it was not possible to find which factor within the variables strengthened the relationship. It was also found that when a witness was the source there was a lower tendency that the origin of the article was mixed (X² (2, N = 889) = 12,644, p <.01/Fisher's exact=0.001 < 0.01) (|-2.1| < -1.96). The data analysis did not show any association between the origin of the news item and the position of the discourse or the proposal of conclusion.

Discussion

The significant decline (by 65%) in the number of news items between 2010 and 2011 is consistent with the findings of ^{Boykoff and Mansfield (2012)} in their analysis of the European media: almost 600 news items referred to COP16 in Cancun while just over 400 covered COP17 in Durban. This trend did not occur in the media in the United States, where coverage slightly increased over the same period. This can be linked to the dependence of the news websites included in this study on the international news agencies based in Europe (near 70 % of the total) where the worsening of the economic situation in 2011 increased interest in economic or work-related news, to the detriment of other issues such as CC. Moreover, as our data revealed, the news items with frames of no solution proposed (impasses) increased from 2010 to 2011, which we infer may also be related to the change of the economic context in those years.

Likewise, the context in which media content is created and transmitted is of especial importance. In the case of news items on CC that were broadcast by television networks in Spain during the Cancun and Durban conferences, ^{Gaitan and Pinuel (2013)} demonstrated a relationship between the economic crisis and the discourse on CC presented to audiences. Building on this research, the data is confirmed by ^{Aguila (2013)} in a study that shows how the coverage of these events by television networks dropped between COP in Cancun and following year's meeting in Durban, precisely because the economic crisis the country faced in the same period increased. This demonstrates how the constrictions of the economic system noticeably affect the development of discourses and content with topics of general interest.

In Latin America there is an even more marked decline between the coverage of the two conferences as that the first (COP16) was held in Mexico. The large Latin American media organizations covered this international conference, sending special correspondents that filed daily reports on the negotiations. These were discussions that involved important leaders from these countries and the declarations from the Conference had a resounding impact on media coverage from the region. As it has been previously observed in relation to the coverage of COP's, politicians are the actors who are most quoted as they are the main participants in the negotiations, much more so than experts or scientists.

Similar studies in countries across the region that were not specifically focused on an analysis of COP's also concluded that politicians were the most quoted sources in the media coverage of CC. In a study of 1,628 articles on CC published in 2003, 2005 and 2007 in Chile (^{Dotson et al., 2012}), the politicians represented 57% of the sources quoted in the El Mercurio and 68% in La Nación, while experts were quoted in 22% and 21% of the articles on CC, respectively.

Supporting the contributions of ^{Liu, Lindquist and Vedlitz (2011: 407)}, ^{Schmidt et al. (2013)} affirm that the coverage of these conferences is part of the media relations approach to CC, describing what is at stake in terms of the consequences of CC and utilizing the high profile of political actors who participate in these international negotiations. During the summits, the politics of CC are reduced to being a purely international component (^{Brossard et al., 2004}).

In the initial reports on global warming when the issue first appeared in the public agenda, scientists were the main sources of information, but later on politicians and interest groups (industry and environmental) came into picture (^{Trumbo, 1996}). ^{Wilkins (1993)} highlights that in 1990, the majority of sources cited in news items on global warming shifted from scientists to politicians and interest groups. The change from a discussion dominated by science to a discussion led by political and industry figures might have contributed to the confusion surrounding CC that exists in the United States media (^{Williams, 2001}).

^{Painter (2010)} confirms that in the media coverage of COP15 in Copenhagen scientists were hardly quoted, with the scientific perspective being secondary to politicians and industry leaders. In the 427 articles analyzed, almost 80% of the news items had less than 10% of their content specifically relating to CC science. Only 9% of the texts dedicated more than a half of their content to scientific aspects with 22 of these 34 articles published in western media outlets (Australia, Italy, UK, USA) that directly quoted scientific organizations like IPCC and NGOs. Kunelius and Eide also highlight the limited presence of politicians in Copenhagen, despite the unprecedented number of scientists who were present and the abundance of scientific information that was available. "This implies that scientists were not actively included (by journalism) in the political bargaining" (2012: 275). "Official sources are more likely to present mainstream and status-quo views and partake in conflict scripts among institutional actors, limiting thus the scope of possible concerns and solutions presented in the press" (^{Zamith et al., 2013: 350}).

Addressing RQ1, in the Spanish-speaking Online Media our data confirm that politicians are the main sources of news. Only 4 scientists declare (previously mentioned: Mario Molina, Rajandra Pachauri, Michel Jarraud and Jeremiah Lengoasa). Molina in COP16, as he is Mexican, and the others because WMO presented in COP17 its Climate Annual Report, which stated that thirteen of the warmest years recorded have occurred within the last decade and a half. Very significant was the appearance of just one ecologist. Our study also found that witnesses were important sources (20.9%) during the coverage of COP's, even more quoted than scientists (14.6%).

Furthermore, the media treatment of COP16 indicates a nationalistic domestication (^{Kunelius and Eide, 2012}) by Latin American media at two levels (RQ2). In one level the domestication refers to Mexico as the COP16 host country: there is a high percentage of Mexican politicians used as sources in the news analyzed in this study. Besides Mexican president and the Mexican organizer of COP16, there were important quotes by Marcelo Ebrard Casaubón, head of the Government of Federal District. The most quoted scientist (M. Molina) is also Mexican. Hence, in Cancun, the large presence of Mexican politicians seems to present the nation as a global actor.

At another level, we can observe this domestication as a region. The Presidents of Bolivia, Ecuador, Venezuela and Brazil (as well as the Ecuadorian Minister of Foreign Affairs) were more quoted than international names such as Barak Obama, Connie Hedegaard (European Commissioner for Climate Action) and Todd Stern (United States Special Envoy for Climate Change). In COP16 in Cancun, the media preferred to include high political leaders (as presidents) instead of technical actors, like Juan Elvira Quesada, Mexican Secretary of Environmental Issues. Nonetheless, in COP17 in Durban, Maite Nkoana-Mashabane, President of COP and Secretary of International Relations and Cooperation of South Africa, was even more quoted than President Jacob Zuma.

RQ3.1 and RQ3.3 address CC news frames. These frames were divided into those that did not commit to any particular position (news items concerning major declarations on the issue) and others in which the journalist included confrontation and highlighted the opposition to the policy or actions that are the subject of the article.

Almost a half of the sample of the news items studied came from international news wire services (RQ4), coinciding with the results from the research by ^{Takahashi and Meisner (2013)}, which examines the role of news agencies in media treatment of CC in Peru. This eliminates the capacity of online media to consult local actors. ^{Liu et al. (2011)} emphasize the importance of local and regional commitments made by media organizations regarding CC. Therefore the inclusion of a high percentage of news items from international agencies shows the lack of a connection between the global problem and concerns and action at a local level. However, our data shows no significant differences of treatment in items from news agencies and those written by media reporters. In both cases politicians (and their agendas) are extremely relevant, instead of agents of recognition such as ecologists and experts. Our findings suggest that Spanish-Speaking Online Media do not take advantage of COP's to bring CC topics to media readers closer.

RQ5 suggested that some actors of CC (victims or activists) are more likely to have a confrontational position that may be understood as a position of protest with current policies. This trend is confirmed when we observed that scientists (the first to alert about CC causes and consequences) were less likely to express conformity. This way, there is a relationship between the speaker and the way their speech is presented in the news item.

This last finding has theoretical and practical implications. In the first place, it is clear that our results confirm framing theoretical framework and suggest it is a useful approach to understand media coverage. The incorporation or not of a source changes the way news items are presented in media, thus the framing process begins when the journalist has to choose who speaks about CC. On the other hand, this fact might orientate specialized journalistic practice. Those who give the information significantly affect news writing, so if journalists are aware of this effect, the final news frame can be less sensitive to the source. This can also change the influence of public understanding and political decision-making in society (^{Hansen, 2011}).

Conclusions¹

The data included in this study can be used to improve the understanding of how CC is covered by online media in Spanish-speaking countries. We found that the media treatment focuses on the political debate and there is a significant association between quoted sources and the news frame. Data also showed that a significant body of news came from news wire agencies and that sources were especially related to countries' media, which implies a nationalistic domestication. Thus, our study reveals that foreign coverage trends are similar in the Hispanic region, and this empirical evidence aims to reduce the gap in CC media coverage research in non English-speaking countries and developing nations. Moreover, the link between the quoted source and CC news frames supports the framing theoretical approach, which serves as an adequate framework for news coverage analysis.

This study has its limitations in terms of sample size and representativeness, and is an initial approximation for research on CC in Latin America and Spain. This research area can continue to be advanced, above all by incorporating an analysis of qualitative discourses in the search for a more in-depth approach to the problems surrounding CC in the media, as an instrument to raise social awareness and the first step towards taking action. Future research may also incorporate a comparative analysis to detect any significant difference in the way each country of the region covers CC.