The Contribution of Extreme Event Communication to Climate Change Mitigation: Outrage and Blame Discourse in Twitter Conversation on Severe Fires

Abstract

Risk communication from the perspective of Extreme Event Attribution (EEA), which assesses the extent to which climate change influences various extreme weather events, has significant potential for climate change communication due to its ability to make the phenomenon more relatable to citizens. This study examines the digital conversation generated following the wave of wildfires in Spain in 2022, which was declared the worst year of the 21st century in terms of hectares burned. By using the Social Network Analysis (SNA) methodology, 145,081 tweets were analyzed to construct a mention network, capturing the digital clusters formed around this discussion and highlighting the predominant tones in the debate. The findings reveal that the conversation predominantly adopted a tone of outrage and assigned responsibility. This research study offers a renewed perspective on risk communication, highlighting significant challenges faced by environmental activism on social media and underscoring the need to improve communication strategies to increase awareness and mobilization around climate change.

1. Introduction

Climate change is affecting the frequency, duration, and intensity of various extreme weather events, such as heatwaves, droughts, heavy rainfall, and fire weather conditions. The risks associated with extreme events involve three main factors: natural hazards, exposure, and vulnerability (IPCC, 2023). Natural hazards are phenomena that can cause harm or damage. Exposure refers to assets at risk, and vulnerability reflects how much those assets can be affected. Climate change influences natural hazards by changing their frequency, duration, and intensity, though these effects differ among various weather types.

Risk communication is currently framed within a context of growing public concern (Eurobarometer, 2019), although there is a tendency to believe that the impacts of climate change will affect “future generations” and “poor countries” (Meira et al., 2013). In this context, communicating risks that are not fully known or imminent is a complex task, yet essential for the social construction of the concept of risk. This social construction is key to the formulation of strategies and public policies, as “each society shares visions of what it considers acceptable (or unacceptable) risks” (Heras, 2021, p. 21). With the goal of engaging citizens in climate change adaptation efforts, the scientific community has demonstrated the importance of communicating the climate crisis by balancing the highlighting of risks with causes and solutions (Fernández-Reyes & Jiménez, 2019; García, 2021).

The literature on natural hazards is extensive and encompasses a wide range of studies that explore how individuals perceive, interpret, and respond to risks associated with these events (Browne & Peek, 2014; Cutter et al., 2000; Mileti & Sorensen, 1990). This body of literature provides a robust framework for understanding the dynamics between risk communication and public response. In fact, it has been shown that public risk perception directly influences the adoption of protective measures (Lindell & Hwang, 2008). Key factors shaping this perception include the trustworthiness and credibility of the source (Scheer et al., 2014) and the efficiency and reach of communication across diverse platforms (Fakhruddin et al., 2020). These elements are particularly relevant in the context of climate-induced natural hazards. It has also been pointed out that integrating traditional media with social networks significantly enhances message reach and community engagement (Fakhruddin et al., 2020).

This study focuses on the wave of wildfires in Spain in 2022, which was considered the worst year of the 21st century in terms of hectares burned—over 300,000—in forest fires in Spain (EFFIS, 2022). This wave coincides with concerns that climate change is increasing the incidence of wildfires, given recent outbreaks of fires worldwide (Jones et al., 2022). Wildfires exemplify the interaction between rapid-onset and slow-onset hazards. While the ignition of a wildfire is a rapid-onset phenomenon, the conditions that exacerbate it are directly linked to climate change, a slow-onset hazard. These combined disasters pose unique challenges (Staupe-Delgado, 2019), as efforts often focus solely on immediate responses, neglecting the structural actions needed to address the underlying causes.

Also, in 2022, the Sixth Assessment Report of the IPCC (2022) provided one of the most comprehensive scientific descriptions to date of the impacts of climate change. Furthermore, global social movements such as Extinction Rebellion call for immediate action to address the global risks arising from the environmental crisis; this demand is expressed on their website and through their protest actions (Extinction Rebellion, 2023, May 22).

Based on this context, this research study specifically examines how the public interacts through social media following an extreme event, emphasizing its potential to increase public awareness and foster a more meaningful dialogue on this issue. To this end, this study examines the main characteristics of the digital conversation on Twitter, now called X, about the wave of wildfires in Spain in 2022 (RQ1), as well as the main clusters formed around the conversation (RQ2). Based on this information, the degree of attribution to climate change within the digital conversation and its clusters (RQ3) is also analyzed.

1.1. Risk Communication Through Extreme Event Attribution (EEA)

Within the field of risk communication, Extreme Event Attribution (EEA) is increasingly attracting the attention of the scientific community. These studies analyze the extent to which climate change influences various extreme weather events. The link between extreme events and current human activity is based on the fact that we have altered the planet’s general conditions, and to confirm this, these studies are conducted in two steps: first, detection—for example, identifying clear trends in heatwaves and their influence on the generation of wildfires; and second, attribution—for example, demonstrating that human-induced warming has triggered changes in heatwave trends. EEA allows scientists and communicators to move beyond general statements about the relationship between weather and climate change, providing numerical estimates of how climate change has impacted the frequency and/or intensity of specific extreme weather events (Betts, 2021).

In line with this, there is growing interest in how extreme weather events can drive understanding and concern about the risks posed by the climate crisis (Cutler et al., 2020; Howe et al., 2019; Painter & Hassol, 2020). Joshua Ettinger et al.’s (2023) study defines extreme weather events as “teachable moments”. Based on Susan A. Flocke et al.’s (2012) concept, teachable moments are personal experiences that can influence individual risk perceptions and decision-making by making an abstract threat more personally salient. In this sense, the ability of these phenomena to bring climate change closer to citizens is emphasized, considering the communicative potential of linking it to individual health (Maibach et al., 2021).

Similarly, attention has been focused on how wildfires specifically can contribute to this cause (Hopke, 2020). Research has shown that during wildfire outbreaks, there is increased coverage of climate change, as well as a greater link between these events and climate change, referred to as climate issues. However, this is not always the case, and most studies highlight the need for greater contextualization of these events with climate change, or in other words, the need for more frequent mention of climate change in the coverage of extreme weather events (Burgess et al., 2020; Cordner & Schwartz, 2019; Painter et al., 2020).

Beyond merely mentioning climate change, attention must be given to the communicative approach used to address these phenomena, as well as the influence that different approaches can have on their contextualization. Some studies have noted a prevalence of the “blaming game” surrounding wildfires, which ultimately silences a potentially meaningful debate on the impact of climate change on weather conditions (Anderson et al., 2018). This phenomenon can be better understood through the lens of attribution theory (Fiske & Taylor, 1991) and theories of blame (Malle et al., 2014), which examine how individuals interpret and assign causes to events they experience, considering internal factors (personal attributions) or external factors (situational attributions), and how these attributions influence judgments of responsibility, emotions such as guilt, and subsequent behaviors. These perspectives are particularly relevant in the analysis of digital discourse, where attributing responsibility to specific actors (such as governments, corporations, or individuals) can shape public narratives and affect both the level of engagement and societal demands for action.

At a more general level, communicating attribution has been shown to evoke negative reactions from politically conservative audiences in the United States, exacerbating distrust in solutions to address extreme weather risks (Hai & Perlman, 2022). At the same time, others have been critical of media coverage of extreme events that neglects to mention climate change (Painter & Hassol, 2020).

Although the debate is ongoing, the potential of extreme events for climate communication is evident. In fact, some studies propose a support tool for risk communication professionals to use in evaluating risk messages (Jit et al., 2023). Beyond journalism, the influence that this risk perspective can have on digital conversations on social media is a particularly interesting area.

1.2. Social Media Activism and Extreme Weather Events

The use of social media as a tool to assess public opinion on various issues, including climate change, is well established (Cody et al., 2015; Fownes et al., 2018; UN Global Pulse, 2015). The combination of information dissemination and emotional elements on social media is particularly effective for engaging citizens regarding climate change (Painter et al., 2018) and for the creation and consolidation of online clusters (Segerberg & Bennett, 2011). While some studies highlight the versatility and adaptability of social media as a strength in climate change communication (León & Erviti, 2021), others point out weaknesses such as the predominance of traditional media as sources of information (Cook et al., 2018), misinformation circulating on these platforms (Treen et al., 2020), and the polarization of users (Cook et al., 2018).

Although dialogue-based methods to engage the public in climate change have been less studied compared with one-way communication strategies (Badullovich, 2022), climate dialogue can foster “knowledge, trust, and agency by linking individuals and clusters to local experiences and scientific information, enhancing understanding of how others engage and take action” (Kelly et al., 2020, p. 418). These statements underline the importance of considering digital conversation, also noting that it can facilitate political action by increasing participants’ willingness to act (Stier et al., 2017). The digital conversation about climate change often revolves around whether it is real or whether there is scientific consensus on the issue (Veltri & Atanasova, 2015). Other popular topics include perceived impacts of climate change (Pathak et al., 2017) and the consequences of extreme weather events (Newman, 2016).

Within the broad realm of social media, each platform serves distinct communicative purposes. This study focuses on the microblogging platform Twitter due to its close relationship with the use of hashtags and hashtag movements (Larrondo-Ureta & Orbegozo-Terradillos, 2021), as well as the academic interest it has attracted over the past decade (Lazer & Radford, 2017).

Various studies have documented how tweets about climate change reference extreme events such as fires, heavy rains, droughts, or floods (Anderson, 2020), with greater attention given to those of larger magnitude (Sisco et al., 2017). Additionally, an increase in tweets about climate change has been observed following extreme events (Cody et al., 2015). A decrease in the number of skeptical tweets about climate change when the conversation addresses extreme weather events has also been recorded (An et al., 2014). According to Jennifer R. Fownes et al. (2018), these data indicate that a portion of the public perceives a strong link between extreme weather events and climate change. In this sense, the most significant aspect is that these extreme events can act as catalysts for the conversation about climate change, marking critical moments that can influence public opinion.

2. Materials and Methods

This study proposes a qualitative methodological approach, utilizing Social Network Analysis (SNA) techniques within the Big Data paradigm. Specifically, the Twitter platform (San Francisco, CA, USA), now called X, was selected as the object of study due to its representativeness as a spontaneous and real-time social action scenario (Campos-Domínguez, 2017), as well as its proven ability to generate visibility, mobilization, and influence in public opinion regarding social movements (Bruns et al., 2015).

A sample was formed through hashtags related to the various wildfires that occurred in Spain in 2022. The sampling process involved two phases, utilizing snowball sampling. First, hashtags related to the major fires of 2022 were identified based on their severity, understood in this case as the number of hectares burned. While there are various mechanisms to quantify the damage caused by a wildfire, in this particular case, we have focused on the number of hectares burned because it is precisely this extent that has made this phenomenon so catastrophic. Once these hashtags were located, those that had generated the highest number of tweets were selected, thus forming the sample for the first phase. In this initial sample, an exploration was conducted in which new informants were identified—specifically, new hashtags related to the wave of fires that, despite not covering the fires with the highest number of hectares burned, had generated a significant number of tweets. Hashtags were collected through the Twitter Academic API retrospectively and subjected to an exploratory and inductive analysis, as is common in the field of Social Network Analysis (SNA). Consequently, the sample consisted of 145,081 tweets, extracted from the following list of hashtags: #IFLosacio [Forest Fire Losacio], #IFCebreros [Forest Fire Cebreros], #Zamorasequema [Zamora drought], #IFLadrillar [Forest Fire Ladrillar], #IFAteca [Forest Fire Ateca], #IFMonsagro [Forest Fire Monsagro], #IIFF [Forest Fires], #IFSierraDeLaCulebra [Forest Fire Sierra de la Culebra], #EspañaCalcinada [Spain Burned], #BRIF [Forest Fire Reinforcement Brigades], #IFVegalatrave [Forest Fire Vegalatrave], and #IFCasasDeMiravete [Forest Fire Casas de Miravete].

Based on the analysis of the identified hashtags, as well as the accounts from which they originated, a mention network was created by using Gephi software version 0.9.2 (Paris, France). In this network, each conversation participant is represented by a node, and each interaction (retweet, response, or mention) is represented by an edge or connection. This network was interpreted as a directed network (where edges have a defined direction, i.e., an origin and an end) and weighted (edges have an associated value or weight).

These types of networks allow for the detection of user clusters through community detection algorithms. In this case, the Louvain Multilevel algorithm (University of Louvain, Belgium) (Blondel et al., 2008) was applied, based on the optimization of modularity. It is worth noting the qualitative nature of the clusters, despite the algorithm assigning them numerical values. The network as a whole was visually represented through a graph, and each of its main clusters was described in relation to the following elements:

• Number of nodes or actors contained in each cluster and their percentage out of the total actors of the total graph. Each cluster was assigned a name based on its main theme and tone.
• Predominant themes in each cluster.
• The time period of conversation for each cluster.
• The leaders of each cluster in terms of Indegree, i.e., mentions received by other users.

Additionally, to detect the degree of attribution to climate change within this digital conversation regarding the wave of fires, three textual filters corresponding to the terms ‘climate change’, ‘climate crisis’, and ‘climate emergency’ were applied to the sample.

3. Results

3.1. General Characteristics

The captured conversation consisted of a total of 41,705 actors—represented by nodes—whose interactions were represented by a network graph (Figure A1). In this graph, a total of 153,717 unique interactions were synthesized, with the average number of interactions per node being 5.765. After applying the Louvain Multilevel algorithm (Blondel et al., 2008) for community detection, a total of 143 clusters were identified. It is important to note that the numbers of the clusters were randomly assigned by the algorithm and, therefore, should not be interpreted ordinally.

Among all the identified clusters, those containing at least 5% of the nodes were selected for in-depth analysis, with the analysis being conducted on a total of eight clusters. Through the network graph available in Appendix A (Figure A1) we can observe how the digital conversation developed in a centralized scenario, in which large clusters form the central core of the interaction exchange.

In the analysis of conversations in various clusters regarding the fires in Spain, a notable variation in the focus and duration of the discussions is observed (

Table 1. Conversation times.

Cluster 10	Cluster 1	Cluster 6	Cluster 19
Cluster 37	Cluster 22	Cluster 127	Cluster 2

Source: own creation.

). For example, in Cluster 10, the conversation was temporarily centered on the first days of the Losacio fire in Zamora, while Cluster 1 shows a similar pattern concerning the Monsagro fire in Salamanca. Similarly, in Cluster 6, discussions are concentrated on the first days of the fire in Cebreros, Ávila. However, Cluster 19 and Cluster 2 present a more sustained conversation over time, indicating prolonged and constant interest in the topics discussed. On the other hand, Cluster 37 is characterized by concentrating discussions at the peaks of the fire in Casas de Miravete, Cáceres, and Cluster 22 shows a notable peak in conversations coinciding with the third day of the fire declaration in Losacio, Zamora. Finally, Cluster 127 stands out for focusing its conversation on mentions of a specific tweet. These patterns indicate that the dynamics of the conversation can vary considerably depending on the cluster and the specific event.

3.2. Clusters and Topics

In

Table 2. Explored clusters.

Cluster	Assigned Name	Number of Users	Percentage of Users
10	Search for impact and attribution of responsibilities	8.685	20.82%
1	Denunciation of firefighter conditions and political responsibility attribution	6.036	14.48%
6	Praise for the performance of forest firefighters	3.347	8.03%
19	Contextualization of the severity of fires in the Iberian Peninsula	3.086	7.41%
37	Last minute information	2.667	6.41%
22	Attribution of political responsibilities	2.546	6.1%
127	Neighborly solidarity	2.183	5.23%
2	Narrative of BRIF work	2.163	5.19%

Source: own creation.

, the explored clusters, along with their detailed characteristics, are listed.

There was also significant variability among the most viralized tweets from each cluster (see Appendix A:

Table A1. Top 5 tweets of each cluster.

CLUSTER	TOP 5 TWEETS (ORIGINAL + TRANSLATION)	RTS
10: Search for impact and attribution of responsibilities	RT @memes_zamora Esta es la superficie total que se ha quemado en la provincia de Zamora desde que comenzaron los incendios hace un mes. Quizás así se entienda mejor la dimensión de lo que está ocurriendo. #IFLosacio #ZamoraSeQuema #EspañaCalcinada [This is the total area that has burned in the province of Zamora since the fires started a month ago. Perhaps this helps better understand the magnitude of what is happening.]	1.912
	RT @opiniondezamora ÚLTIMA HORA | La Junta de Castilla y León es denunciada por Greenpeace por su actuación en el incendio de la Sierra de la Culebra. #IFSierraDeLaCulebra #IIFF [BREAKING NEWS | Junta of Castilla y León has been denounced by Greenpeace for its actions in the Sierra de la Culebra fire.]	549
	RT @Juan13Navarro Me envía un bombero este vídeo del #IFLosacio desde Tábara (Zamora). El panorama es desolador y me cuenta él que muchos están disponibles pero que la Junta de Castilla y León no los llama. Se trata de una zona cercana a la parte de la sierra de la Culebra que no ardió en junio. [A firefighter sent me this video of the Losacio fire from Tábara (Zamora). The scene is devastating, and he tells me that many firefighters are available, but Junta of Castilla y León does not call them. This is an area near the part of the Sierra de la Culebra that did not burn in June.]	419
	RT @yolsclemente Para que os hagáis una idea, esto es lo que se ha quemado al noroeste de Zamora en el último mes (seguramente más de 45.000 hectáreas con perímetros aproximados hasta ahora). Más de un 4% de la provincia entre el #IfSierradelaCulebra #IfFigueruela y #IFLosacio [So you get an idea, this is what has burned in the northwest of Zamora in the past month (likely over 45,000 hectares with approximate perimeters to date). More than 4% of the province among the Sierra de la Culebra, Figueruela, and Losacio fires.]	391
	RT @opiniondezamora VÍDEO Losacio revive su peor pesadilla. El incendio vuelve a nivel 2 y desalojan a la población. #IFLosacio #IFVegalatrave [VIDEO Losacio relives its worst nightmare. The fire returns to level 2, and the population is evacuated.]	292
1: Denunciation of firefighter conditions and political responsibility attribution	RT @BBFFLuchaCyl Estamos viendo como saltan las costuras de @naturalezacyl @jcyl #INFOCAL hasta el punto en que han CONTRATADO un periodista con su cámar para hacer PROPAGANDA contra lo que estamos SACANDO A LA LUZ. #IFMonsagro Los #BomberosForestales de #CyL YA NO CALLAMOS! #QUIÑONESDIMISION [We are seeing how the seams of INFOCAL are coming apart to the point that they have HIRED a journalist with a camera to do PROPAGANDA against what we are BRINGING TO LIGHT. The forest firefighters of CyL WILL NO LONGER REMAIN SILENT!]	711
	RT @BBFFLuchaCyl Esta es la realidad de los #IIFF en #CyL #IFNavafria -Personal deshidratado—Intoxicado—Abandonado por @jcyl @naturalezacyl—Maltratado Estos señores que hoy están ingresados el 30 de septiembre estarán en el PARO! #INFOCAL DIGNO YA! #QUIÑONESDIMISION [This is the reality of the wildfires in CyL—Dehydrated personnel—Intoxicated—Abandoned by Junta of Castilla y León]	700
	RT @BBFFLuchaCyl Esta es la realidad de los #IIFF en #CyL #IFNavafria -Personal deshidratado-Intoxicado—Abandonado por @jcyl @naturalezacyl—Maltratado Estos señores que hoy están ingresados el 30 de septiembre estarán en el PARO! #INFOCAL DIGNO YA! #QUIÑONESDIMISION [This is the reality of the wildfires in CyL—Dehydrated personnel—Intoxicated—Abandoned by Junta of Castilla y León]	427
	RT @BBFFLuchaCyl Una imagen vale más que mil palabras. @naturalezacyl @jcyl @alferma1 @PopularesCyL DESMENTID esto si tenéis el valor. #IFLosacio #QUIÑONESDIMISION #ARRANZDIMISION [A picture is worth a thousand words. Deny this if you dare. #LosacioFire]	423
	RT @BBFFLuchaCyl Respondiendo al nuevo gabinete de prensa de @naturalezacyl donde intentan lavar sus vergüenzas y tapar nuestras desgracias. Todo este material de primeros auxilios y avituallamiento es gracias a los vecinos del #IFMaillo #IFMonsagro. Os vamos a plantar cara siempre! [Responding to the new press office of the environmental department where they try to cover their shame and hide our misfortunes.]	407
6: Praise for the performance of forest firefighters	RT @AlertaZeta Bonito y entrañable gesto de los bomberos forestales que dan de beber a un corzo deshidratado por el incendio #IFLosacio. Gracias Vídeo de @INFOCYL [Beautiful and heartwarming gesture from the forest firefighters who gave water to a dehydrated roe deer during the Losacio fire. Thank you.]	501
	RT @Pivei99 Los verdaderos héroes sin capa y no los del trajecito en los despachos Resiste #Cebreros #IFCebreros [The real heroes without capes, not those in suits sitting in offices.]	195
	RT @bomberosdeavila Así llegaban ayer nuestras dotaciones a la defensa de Hoyo de Pinares por la Av-502 #IFCebreros [Yesterday, our crews arrived to defend Hoyo de Pinares via Av-502.]	112
	RT @InfuesAvila #IFCebreros Ávila Espectacular descarga del Grupo 43 @EjercitoAire Gracias @BomberoForestaI #IF @INFOCYL @InfuesAvila [Spectacular water drop by Group 43. Thank you.]	97
	RT @InfuesAvila ¿Quereis ver fotos de como están los montes públicos, aquí en Ávila por ejemplo? Si conseguimos más de 100 RT. Lo tendreis, un buen álbum… Y si superamos +1000 RT, el antes y después de un incendio como en #IFCebreros… #BomberosForestales [Would you like to see pictures of how public forests look, here in Ávila for example? If we reach more than 100 RTs, you’ll have a great album… And if we get over 1000 RTs, you’ll see the before and after of a fire like in Cebreros…]	78
19: Contextualization of the severity of fires in the iberian peninsula	RT @IncendiosES Situación dramática en toda la Península Ibérica a causa de los #IIFF. #Galicia, #Cáceres, #Málaga, #Segovia, #Zamora, #Portugal… Así recogen los satélites de @CopernicusEMS los puntos calientes a lo largo de todo el territorio [Dramatic situation across the Iberian Peninsula due to wildfires. Galicia, Cáceres, Málaga, Segovia, Zamora, Portugal… Copernicus satellites capture hot spots throughout the territory.]	166
	RT @Storm_Malaga Como si de un volcán se tratase, así de espectacular ha captado el satélite #Sentinel2 el #IFLasMestas, #IFMonsagro, #IFLadrillar, #Cáceres, #Salamanca, hoy 13 de julio. [As if it were a volcano, this is how spectacular the satellite Sentinel2 captured the fires in Las Mestas, Monsagro, Ladrillar, Cáceres, Salamanca.]	94
	RT @Storm_Malaga Estremecedor. El #IFLosacio: 32.071 ha (rojo) e #IFSierradelaCulebra: 29.670 ha (azul), se sitúan respectivamente como el 2° y 3° más extensos jamás registrados en #España. Ambos #IIFF han llegado a fusionarse. ¡61.741 ha! Territorio equivalente a #Antequera. Mamma mía… [Staggering. The Losacio fire, 32,071 ha (red), and Sierra de la Culebra fire, 29,670 ha (blue), respectively the 2nd and 3rd largest ever recorded in Spain. Both fires have merged.]	91
	RT @i_ameztoy Impresiona la vista nocturna de los incendios vistos desde el espacio. #IFLadrillar #IFLasMestas #IFMonsagro #IFCandelario #IIFF Cáceres—Salamanca SNPP/VIIRS Julio 2022 #EO #Wildfires [The night view of the fires from space is impressive. #Wildfires]	56
	RT @Plan_INFOCA Hoy celebramos el #WorldRangerDay, el #DíaMundialDeLosGuardasForestales. En #Andalucía, realizan sus funciones los celadores forestales y los @AAMM_And. Estos últimos, ejercen como directores de extinción de #IIFF en su ataque inicial. Para todos: ¡Muchas felicidades compañeros! [Today we celebrate World Ranger Day. In Andalusia, forest rangers and environmental agents perform extinction leadership tasks in the initial attack on wildfires.]	53
37: Last minute information	RT @saratuit Monfragüe fue declarado Parque Nacional en el año 2007. Por favor, medios de comunicación, mencionadlo como se merece. Se está quemando la joya de Extremadura y uno de los espacios naturales más importantes de España. Nos duele el alma. #IFMonfragüe #IFCasasDeMiravete [Monfragüe was declared a National Park in 2007. Please, media outlets, give it the recognition it deserves. The jewel of Extremadura and one of the most important natural areas in Spain is burning. Our hearts ache. #IFMonfragüe #IFCasasDeMiravete]	167
	RT @UMEgob Esta imagen es una muestra de a que se enfrentan los servicios de extinción en los incendios forestales de estos días. Mucha fuerza compañeros #StopIncendios #JuntosSomosMásFuertes #IFMonsagro en Morasverdes [This image shows what firefighting services are facing during these forest fires in recent days. Stay strong, colleagues #StopFires #TogetherWeAreStronger #IFMonsagro in Morasverdes]	137
	RT @Blog_ElCanchero ¡Sentido común! Lección magistral de un vecino de @CasasdeMiravete a los que se dedican a hablar durante horas sobre el campo, pero no dicen nada con un lenguaje impostado y amparados en la sostenibilidad ambiental, transición ecológica, etc. ¡Grande Alejandro! #IFCasasdeMiravete [Common sense! A masterclass from a resident of @CasasdeMiravete for those who talk endlessly about the countryside without saying anything meaningful, using pretentious language and hiding behind terms like environmental sustainability, ecological transition, etc. Well done, Alejandro! #IFCasasdeMiravete]	126
	RT @UMEgob Desde la #UME, con el corazón encogido, lamentamos el fallecimiento de brigadista mientras participaba en las tareas de extinción del #IFLosacio Nuestro pésame a los familiares, amigos y compañeros D.E.P. [From the #UME, with heavy hearts, we regret the death of a firefighter who was participating in the firefighting efforts at #IFLosacio Our condolences to family, friends, and colleagues. Rest in peace.]	87
	RT @JaVi_MoRen0 Declarado un nuevo incendio forestal en Casas de Miravete, en Cáceres. Muy cerca de Monfragüe. #IFCasasDeMiravete #Extremadura [A new forest fire has been declared in Casas de Miravete, Cáceres, very close to Monfragüe. #IFCasasDeMiravete #Extremadura]	82
22: Attribution of political responsibilities	RT @luciap_06 No veréis a ningún influencer, a ningún equipo de fútbol ni a nadie pidiendo solidaridad como en el volcán de la palma Zamora no le importa a nadie a excepción de los zamoranos #IFLosacio #ZamoraNoSeCalla #ManuecoDimision #QuiñonesDimision [You won’t see any influencers, any football teams, or anyone asking for solidarity like they did with the La Palma volcano. Zamora doesn’t matter to anyone except for the people of Zamora. #IFLosacio #ZamoraNoSeCalla #ManuecoResign #QuiñonesResign]	709
	RT @JesusCintora «El operativo de incendios en Castilla y León es patético, anacrónico e ilegal». Los sindicatos de los bomberos forestales denuncian las condiciones «precarias» del sector y apuestan por un trabajo previo en el monte durante todo el año #IFLosacio [“The firefighting operation in Castilla y León is pathetic, outdated, and illegal”. The unions of forest firefighters denounce the “precarious” conditions in the sector and advocate for year-round preventive work in the mountains. #IFLosacio]	543
	RT @DavidAloGar El silencio de todos y cada uno de los miembros de la Junta de Castilla y León -presidente, vicepresidente y los 10 consejeros- en una jornada con dos muertos, más 30 pueblos desalojados y casi 10.000 personas evacuadas es sencillamente atronador. #IFLosacio #IFCebreros [The silence from each and every member of the Junta de Castilla y León—president, vice-president, and the 10 councilors—on a day with two deaths, over 30 villages evacuated, and nearly 10,000 people displaced is simply deafening. #IFLosacio #IFCebreros]	23
	RT @Yolanda_Diaz_ Con mucho pesar lamentamos el fallecimiento de un brigadista en #IFLosacio en la extinción del incendio. Los equipos de emergencias merecen los mejores medios y el máximo reconocimiento. La seguridad en el trabajo es esencial y hay que seguir reforzándola para evitar accidentes. [With great sadness, we mourn the death of a firefighter at #IFLosacio while fighting the fire. Emergency teams deserve the best resources and the highest recognition. Workplace safety is essential and must continue to be strengthened to prevent accidents.]	256
	RT @Losacio Ahí os dejo una entrevista que me hizo @PabloIglesias para dar visibilidad y no dejar en el olvido a Losacio y agradecer a la gente que ayudó en #IFLosacio [Here is an interview I did with @PabloIglesias to bring visibility and ensure Losacio is not forgotten, and to thank everyone who helped during #IFLosacio]	190
127: Neighborly solidarity	RT @BomberoForestaI Desde aquí quiero agradecer a todos los vecinos de todos los pueblos que se vuelcan de una manera increíble con los más afectados y sobre todo con nosotros, con todo el mundo de la extinción, GRACIAS, DE VERDAD. #IFCebreros #IFLosacio [From here, I want to thank all the residents of every town who go above and beyond for those most affected and especially for us, everyone involved in firefighting. THANK YOU, TRULY. #IFCebreros #IFLosacio]	954
	RT @BomberoForestaI Como bombero forestal estas imágenes me destrozan el alma como a todos los que aman la naturaleza y su entorno. Son de hoy del #IFCebreros. Muchas gracias a todos los medios que hoy han dado todo por salvar cada hectárea y vida. Orgulloso de esta profesión, de esta forma de vida. [As a forest firefighter, these images break my heart just as they do for everyone who loves nature and their surroundings. These are from today at #IFCebreros. Many thanks to everyone who gave their all today to save every hectare and every life. Proud of this profession, of this way of life.]	214
	RT @BomberoForestaI #IFCasasDeMiravete (Cáceres). Así cruza el fuego la N-V… [#IFCasasDeMiravete (Cáceres). This is how the fire crosses the N-V highway…]	198
	RT @BomberoForestaI Así se juegan la vida los compañeros de los Romeos, ¡Muchísimo cuidado por favor! #OACEL #IFLosacio [This is how our colleagues from the Romeos risk their lives. Please, be extremely careful! #OACEL #IFLosacio]	138
	RT @BomberoForestaI Iniciativa increíble de mi amiga Sonia. Se agradece difusión. Los bomberos forestales también necesitamos ayuda psicológica. #IFCebreros #IFLosacio #IFAteca @IncendiosES @AT_Brif [An incredible initiative by my friend Sonia. Please share. Forest firefighters also need psychological support. #IFCebreros #IFLosacio #IFAteca @IncendiosES @AT_Brif]	131
2: Narrative of BRIF work	RT @BomberoForestaI Iniciativa increíble de mi amiga Sonia. Se agradece difusión. Los bomberos forestales también necesitamos ayuda psicológica. #IFCebreros #IFLosacio #IFAteca @IncendiosES @AT_Brif [An incredible initiative by my friend Sonia. Please share. Forest firefighters also need psychological support. #IFCebreros #IFLosacio #IFAteca @IncendiosES @AT_Brif]	396
	RT @AT_Brif 23:03 #IFCasasDeMiravete #Cáceres Impresionante ritmo de trabajo @BRIF_PINO RD con sus motosierras + herramienta manual construyendo una línea para evitar que el fuego avance hacia el Parque Nacional de Monfragüe. GRANDES Somos #BRIF de @mitecogob [23:03 #IFCasasDeMiravete #Cáceres Impressive work pace by @BRIF_PINO RD using chainsaws and hand tools to build a line to prevent the fire from advancing towards Monfragüe National Park. HEROES We are #BRIF from @mitecogob]	91
	RT @AT_Brif 23:20 #IFAteca #Zaragoza Los compañeros de @BrifDaroca RD se disponen a defender las cercanías del municipio de Ateca. Somos #BRIF de @mitecogob [23:20 #IFAteca #Zaragoza The team from @BrifDaroca RD is preparing to defend the outskirts of the town of Ateca. We are #BRIF from @mitecogob]	82
	RT @BomberoForestaI Desde aquí quiero agradecer a todos los vecinos de todos los pueblos que se vuelcan de una manera increíble con los más afectados y sobre todo con nosotros, con todo el mundo de la extinción, GRACIAS, DE VERDAD. #IFCebreros #IFLosacio [From here, I want to thank all the residents of every town who go above and beyond for those most affected and especially for us, everyone involved in firefighting. THANK YOU, TRULY. #IFCebreros #IFLosacio]	80
	RT @EMIF_IIFF #IFMonsagro #IFLadrillar Duro trabajo el realizado ayer por los @EMIF_IIFF de @PLANINFOEX y @naturalezacyl para consolidar el flanco izquierdo en el límite de provincia entre Cáceres y Salamanca, junto a efectivos de @BRIF_PINO, @briftabuyo, @UMEgob, @PLANINFOEX, @naturalezacyl [#IFMonsagro #IFLadrillar Hard work yesterday by the @EMIF_IIFF team from @PLANINFOEX and @naturalezacyl to secure the left flank at the provincial border between Cáceres and Salamanca, alongside personnel from @BRIF_PINO, @briftabuyo, @UMEgob, @PLANINFOEX, and @naturalezacyl]	67

Source: own creation.

).

Cluster 10 is the largest, comprising 8685 actors or nodes, 20.82% of the total. This cluster is led—with leaders being those users who have received the most mentions from the rest of the cluster—by media and journalists with a strong geographical component, all of them related to the affected region. Temporally, the conversation concentrated on the first days of the Losacio fire (Zamora). The conversation is centered, on one hand, on contextualizing the dimension and severity of the fire, and on the other hand, criticizing Junta de Castilla y León for its management. Media silence suffered by the region is also denounced. The fire is compared to the eruption of the La Palma volcano to highlight differences in media coverage, and the burned area is shown in terms of the surface area of other regions to graphically convey the dimension of the event. The death of a firefighter during extinguishing efforts also plays a prominent role in the conversation. Criticisms of organizations such as Junta de Castilla y León or the central government, the media, the passivity of citizens, or even the silence of certain influencers denote an intention to attribute responsibility for disaster management, as well as a search for media coverage of the phenomenon.

Cluster 1 is the second largest, with a total of 6036 actors, 14.48% of the total. Among the most mentioned users are Junta de Castilla y León, its president Alfonso Fernández Mañueco, and his party, the Partido Popular. Temporally, the conversation is concentrated on the first days of the Monsagro fire (Salamanca). In this case, the conversation focuses essentially on denouncing the working conditions of firefighters: long working hours, insufficient and poor-quality food, lack of resources, safety for firefighters, etc. It is condemned that the fire danger level was not raised in time, and Junta de Castilla y León is appealed to as responsible for the disaster. The hashtag #Quiñonesdimision stands out in relation to Juan Carlos Suárez-Quiñones y Fernández, the environmental councilor of Junta de Castilla y León.

On the other hand, Cluster 6, with 3347 actors representing 8.03% of the total, is led by firefighter communication accounts, highlighting @AlertaZeta and the air force. In this cluster, the conversation, centered on the first days of the fire in Cebreros, Ávila, includes a viral tweet showing forest firefighters giving water to a dehydrated roe deer, symbolizing the dedication of the firefighters. Criticisms of the authorities for the lack of resources are also present, with mentions of food provided by neighbors due to resource shortages.

Cluster 19, with 3086 actors, equivalent to 7.41% of the total, mentions the Forest Fire Prevention and Extinction Service, forest education accounts, and the Copernicus satellite. This cluster presents a sustained conversation over time, focusing on the severity of the fires in the Iberian Peninsula during the summer of 2022, using graphic resources to illustrate the situation.

Similarly, Cluster 37, with 2667 actors, constituting 6.41% of the total, highlights mentions of @UMEgob, the forest firefighters of Extremadura (@PLANINFOEX), Junta de Extremadura, and the Ministry of Defense. The conversation in this cluster focuses on the peaks of the fire in Casas de Miravete, Cáceres, sharing updated information about its severity and the evacuated clusters. Among the highlighted stories is the death of a brigadier in the Losacio fire.

Cluster 22, with 2546 actors, representing 6.1% of the total, shows diversity among its leaders, including football players, journalists, and leaders. The conversation reaches a peak on the third day of the Losacio fire, once again highlighting the death of a brigadier and criticizing the management of Junta de Castilla y León. A tweet from journalist Jesús Cintora highlights the criticism of fire management, demanding greater media coverage.

Cluster 127 comprises 2183 actors, 5.23% of the total. A diversity of actors receiving mentions is observed: official entities, journalists, NGOs, or influencers. However, a centralized leadership is noted in the @BomberoForestal account, for the publication of a highly salient tweet (5074 mentions) in which a video is shared about how residents of the affected areas organize to prepare food for those affected. In this line, the conversation is essentially focused on mentions of this tweet.

Finally, Cluster 2 comprises 2163 actors, 5.19% of the total. In this case, a cluster led by a diversity of actors is found, among which various Forest Fire Reinforcement Brigades (BRIF) accounts stand out. The contents of the cluster circulated over a more extended period, with the conversation focusing on describing different tasks and actions of the forest brigades. Again, a tone of gratitude for their work is observed, around which the most mentioned tweet (3969 mentions) stands out, sharing a private initiative for psychological support for the teams that worked in the fires.

3.3. Attribution to Climate Change

Regarding the attribution of the fires to climate change in the different clusters analyzed, very limited mention of this phenomenon is observed.

Table 3. Attribution to climate change in clusters.

Cluster	Attribution to Climate Change
10	2.12%
1	0.21%
6	0.23%
19	0.67%
37	0.11%
22	0.15%
127	0.12%
2	0.09%

Source: own creation.

illustrates the presence of the terms ‘climate change’, ‘climate emergency’, or ‘climate crisis’ in relation to the total number of tweets in each cluster. This table reveals that the inclusion of these terms barely exceeds 2% only in the largest cluster, Cluster 10. This indicates that despite the magnitude of the fires and their devastating effects, the conversation around climate change has not been predominant in the cluster discussions.

It is also interesting to note that even in Cluster 10, where the mention of climate change is relatively higher, this topic has not become a central axis of the discussion. This could be due to several factors, including a lack of clear and accessible information about the relationship between fires and climate change or a greater focus on immediate criticisms of fire management by local and regional authorities.

4. Conclusions and Discussion

The analysis of the digital conversation on Twitter regarding the wildfires that affected the Iberian Peninsula during the summer of 2022 reveals, firstly, a discourse characterized by public outrage and the assignment of responsibility for the fires (RQ1). This indignation reflects protests about the lack of resources and ineffective management in combating forest fires, with the blame being primarily directed at relevant authorities and political leaders. In this context, the study confirms the dominance of the responsibility attribution discourse in the communication surrounding extreme events (Anderson et al., 2018) and underscores the need for a critical perspective on the detrimental effects of this discourse on the potential for a substantive climate debate. Ultimately, the critical tone dominated the conversation, although there remains space for other narratives that, while less prominent, have introduced positive tones, such as commendations for the work of firefighters and recognition of the solidarity among those affected.

With respect to the digital clusters formed around this conversation (RQ2), a diversity of predominant themes are evident within each cluster, ranging from the denunciation of firefighters’ working conditions to criticisms of the relevant authorities. However, all clusters are characterized by a critical and indignant tone. The digital conversation evolved within a centralized framework, wherein large clusters constituted the central hub of interaction. Additionally, the permeable nature of the various clusters is noteworthy, with common narratives often circulating across clusters—for instance, the story of the firefighter who died while extinguishing the Losacio fire. In general, no clusters were observed to be particularly isolated from others. This interaction between clusters can be seen as a positive aspect, facilitating the exchange of opinions and fostering climate discussions among users.

The conversation in each cluster also varied according to its temporal dynamics, typically exhibiting a pattern of conversational peaks driven by information and interaction, resembling what could be described as event-based digital activism. This contrasts with the scientific cluster’s call for continuous communication about climate change. The Association of Environmental Information Journalists (APIA) has called on the media to commit to environmental reporting and its continuity (García, 2021). In this regard, it is important to highlight the potential of more sustained activism on social media, moving beyond the event-driven news cycle. The capabilities of social media offer a unique opportunity to foster this continuous communication.

On the other hand, the limited mention of terms such as ‘climate change’, ‘climate crisis’, or ‘climate emergency’ in the tweets (RQ3) suggests activist actions that fail to fully contextualize the phenomenon. Given the existing trend to simplify the complexity of climate change by focusing on climate countermovements and the literal denial of the phenomenon (Almiron & Moreno-Cabezudo, 2022), it is crucial to present these events in direct and comprehensible terms for the audience. While political ideology remains the dominant factor in interpreting these events, perceptions of damage and proximity can be leveraged to improve communication and enhance climate change awareness in affected clusters (Zanocco et al., 2018). However, the limited connection to climate change observed in this study contrasts with journalists’ tendency to link wildfires more frequently to climate change than other extreme events, such as hurricanes (Molder & Calice, 2023).

The limited attribution of wildfires to climate change observed in these digital conversations reflects broader challenges identified in the literature, such as the normalization of extreme events due to repeated exposure (Moore et al., 2019) or the difficulty of attributing specific events to climate change due to statistical models and margins of error (Stott et al., 2016). Addressing these challenges requires communication strategies designed to bridge the gap between scientific complexity and public understanding in real time.

It has been observed that extreme weather events, such as wildfire outbreaks, act as a catalyst for public digital discussions about the consequences of climate change. This opportunity for environmental communication inherently brings the challenge of framing these phenomena within the broader context of climate reality and their anthropogenic origins. There is a recognized need for improved public literacy regarding current weather risks and the long-term implications of climate change to foster informed decision making and support for mitigation efforts (Fleischhut et al., 2020). Identifying effective communication formats remains a significant challenge for traditional media, public administrations, and environmental non-governmental organizations, all of which play a critical role in raising awareness about the climate emergency.

Within the context of indignation evidenced in this study, it is particularly important to highlight alternatives for effective environmental communication. In this regard, proposals for climate change communication that focus on extreme events represent a promising area of exploration. Among these strategies are narratives centered on hope and accessible solutions to reduce emotional rejection (Merkel et al., 2020) or the promotion of messages that reinforce a sense of agency and personal self-efficacy (Ockwell et al., 2009). Such alternatives, especially when combined with personalized messaging and the use of personal and local stories (León et al., 2022; Merkel et al., 2020), hold significant potential for improving public engagement.

This research study is not without limitations. A primary limitation of social media research is that tweets cannot be treated as surveys to infer the state of public opinion or attitudes (Lin et al., 2013). However, by analyzing tweets, public beliefs with the potential to influence individuals, media, and policymakers can be discerned (Freelon et al., 2016). It is also important to acknowledge that characteristics such as technological usage or socio-economic homogeneity may limit the generalizability of the results. Nevertheless, the exploratory nature of the study prioritizes the identification of key narratives within the digital conversation, without aiming to represent the general public. In this respect, Social Network Analysis (SNA) techniques, applied within the Big Data paradigm, offer a valuable tool for deepening our understanding of digital public opinion.