Visual digital intermediaries and global climate communication: Is climate change still a distant problem on YouTube?

PLOS ONE OPEN ACCESS Citation: Segerberg A, Magnani M (2025) Visual digital intermediaries and global climate communication: Is climate change still a distant problem on YouTube?. PLoS ONE 20(4): e0318338. https://doi.org/10.1371/journal.pone. 0318338 Editor: Andrea Tomassi, Uninettuno International Telematic University, ITALY Received: March 11, 2024 Accepted: January 13, 2025 Published: April 8, 2025 Copyright: © 2025 Segerberg, Magnani. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data availability statement: All files needed to replicate the results of the study are available from the Open Science Framework, URL: https://osf.io/gmb7n/ Funding: AS: Swedish Research Council grant 2015-01835. https://vr.se AS, MM: Swedish Research Council for Health, Working Life and Welfare (FORTE) grant 2021-01646. https://forte.se, under the CHANSE ERA-NET Co-fund programme, which has received funding from the European Union’s Horizon RESEARCH ARTICLE Visual digital intermediaries and global climate communication: Is climate change still a distant problem on YouTube? Alexandra Segerberg1∗, Matteo Magnani 2 1 Department of Government, Uppsala University, Uppsala, Sweden„ 2 InfoLab, Department of Information Technology, Uppsala University, Uppsala, Sweden ∗ alexandra.segerberg@statsvet.uu.se Abstract This article addresses the role of digital intermediaries in visual climate communication,

and specifically their contribution to the persistence of a ‘green ghetto’ of traditional com- municators and repertoires online. We argue for a comparative sensibility: global plat- forms convey global issues to global audiences, yet the same platform may distribute conditions of visibility for compelling communication unevenly around the world. The study analyses how a major global visual platform, YouTube (Search), articulates climate change in 232 countries in their official languages. It combines API research, channel coding and computational image analysis to assess the processing and presentation of top-ranked results with respect to their diversity and proximity to local context. The find- ings show that YouTube Search establishes visibility winners who typically sustain the classic visual repertoire of climate change as a distant problem, and that Global North sources dominate irrespective of region. However, there are notable exceptions to these patterns. 1 Introduction It is notoriously challenging to engage citizens on climate change, an issue that is often per- ceived as abstract and remote in time and place. Strategic recommendations therefore encour- age making climate narratives relevant by bringing them close to people’s experience and spheres of concern [1–4]. The recommendations are well known, but observers still note an ingrained ‘green ghetto’ [[2] ch. 6] in communication on climate change, in which a delim- ited circle of much the same (Global North) speakers, stories, and styles dominate and rein-

force the notion of climate change as a distant problem. The primary problem with the green ghetto from the perspective of citizen engagement is not in the first instance what is said about climate change, but how it is said, by whom, and the relation to the audience. There have been expectations that digital media would dissolve the green ghetto, by sup- porting fresh voices from beyond the circle of ‘usual suspects’ who might bring repertoires that resonate with local audiences [5–7]. After all, contemporary complex media ecologies PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 1/ 23

PLOS ONE Visual digital intermediaries and global climate communication soften the grip of editorial gatekeeping, make it easier to produce custom content, enable cit- 2020 Research and Innovation Programme under Grant Agreement 101004509. AS, MM: Swedish Research Council grant 2021-02769. https://vr.se The funders did not play any role in the study design, data collection and analysis, decision to publish, and preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. izens to seek out issues they care about in interaction with their peers, and amplify visual con- tent, which is known to engage. Empirically, however, it is not clear whether the green ghetto has lost traction online. The initial expectations did not account for the digital platforms that have such a crucial position in this communication landscape [8]. Moreover, studies of online climate communication have tended to focus on text-centred social media such as Twitter, as opposed to the swath of platforms that primarily centre on visual content (e.g. pictures, memes, and videos) such as YouTube, Instagram, or TikTok [6]. In this paper, we are interested in the role of visual platforms in global climate commu- nication and, more specifically, how they contribute to the persistence (or dissolution) of a green ghetto in public discourse online. We approach search engines and social media as dig- ital intermediaries [9,10] that mediate information flows, attention, and connection between climate communicators and audiences.

We focus on rankings (e.g. ranked presentations of content or contributors) as an area in which platform intermediation is overt. Two aspects of rankings are significant: the con- tent that platforms present at the top and the processes that produce these results. With the former, platforms articulate the scope and character of issues, and audiences easily assume that platform priorities in organising information reflect the issue’s core. With the latter, platforms mediate the processes through which issues, actors, and content become prominent, setting conditions for the broader (in)visibility of fresh voices and local con- nections. We conceive of digital intermediaries in ranking processes as encompassing agencies on several sides. On the platform side, platforms filter, hierarchise, and display content as ranked results, shaped by its politics, policies, algorithms, and affordances. On other sides, core platform doings embed in broader ranking assemblages [11]. These also invoke the (re)action of heterogeneous others such as content producers and user audiences. We argue that a comparative sensibility is crucial to advance understanding of online climate discourse and the role of digital intermediaries within it. Previous work on climate journalism in traditional media observes that the dearth of research attention to Global South contexts is problematic, not least since the societal challenges of climate change call for transnational interventions that require substantial public support [12–15]. In a parallel argu-

ment, we propose that a more differentiated knowledge base is important also in the case of online climate communication. To this end, it is crucial to consider not only the differences between digital intermediaries [16], but also within them. Although online climate commu- nication is typically studied in a single context (usually US/English), it is not obvious that the operation of a digital intermediary in one context applies more broadly. Indeed, there is rea- son to expect that ranking processes on global platforms play out differently across contexts. Both the platform itself and other actors involved in the broader ranking assemblages may do things differently across contexts. This motivates two sets of questions about digital intermediaries in global climate com- munication, and their role in perpetuating (or dissolving) a green ghetto in public discourse online: • How do global digital platforms present climate change to differently situated global publics? More specifically, to what extent are fresh voices and local connection present in top-ranked results? • Do the conditions of visibility for fresh voices and local connection in top-ranked content distribute evenly across contexts? PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 2/ 23

PLOS ONE Visual digital intermediaries and global climate communication We address these questions with empirical focus on a major global visual platform that combines both search and social network affordances: YouTube. Pursuing an approach of quantitative description [17] and large-scale observation [11], the study examines how YouTube Search articulates ‘climate change’ to audiences searching for information on the issue around the world. It combines API research, qualitative channel coding, and compu- tational image analysis to analyse and assess the top-ranked search results returned for ‘cli- mate change’ for 232 countries and 85 official languages with a comparative perspective on the underlying processes and results. In order to bring the digital intermediary into focus, we delimit this study to the specifically visual dimension of the platform and videos (i.e. not considering audio or user comments), and assess the presence of a particular set of classic distancing visual themes: arctic landscapes and polar bears. The results show that top-ranked results on YouTube Search largely perpetuate a green ghetto, with some modification. ‘Usual suspects’ such as professional media continue to figure prominently, now joined by some fresh voices in the form of user-led and digitally native pop- ular science channels. Still, both old and new climate communicators continue to reference classic distancing arctic themes. The results also show that the conditions for local connec- tion are constrained. This is in particular the case for regions that share a global language.

Top results for queries in global languages return much the same videos for geographically disparate API search settings (e.g. France, Canada and Cameroon), and channels from one country dominate (e.g. in the case of French, channels based in France). Closer examination reveals distinct profiles in how major language clusters articulate the scope and character of the issue. The study draws on research in the distinct fields of visual climate communication and algorithm audit approaches in platform studies. In the following sections, we therefore first outline the relevant debates in these two fields and how they inform our research questions, before discussing the case of YouTube Search as a global visual platform in the context of online climate communication. We then proceed to the research design and analysis. 2 Visual climate communication and digital platforms 2.1 Towards fresh voices and local connection in visual climate communication? Visual representation, as found in photos, videos, gifs, and memes, is considered a powerful component in climate change communication. Images engage attention and make informa- tion more memorable and credible [18–20]. They shape emotional connection to issues and policy [21,22], and underpin a sense of intimacy and direct connection [23]. What is more, under the right conditions, they can transcend linguistic and national borders [24,25]. This means that visual communication can be key to bringing climate issues close to home. Conversely, visual content can also bolster the impression that the issues of climate change

are abstract and remote in time and place. Ironically, a recurring element in visual climate dis- course is the distancing frame [[26] p. 16], which is broadly assumed to be problematic for engaging people with climate change. The distancing frame reinforces the idea that climate issues are geographically, temporally, and psychologically remote from everyday life. Typi- cal examples include the classic visual themes of political or celebrity elites, polar bears, and empty (arctic) landscapes that became prominent and entrenched at the turn of this century in Global North civil society advocacy campaigns [27], stock footage infrastructures [28], legacy popular science media and climate journalism [26,29–32]. The evidence for the under- lying psychological theory is mixed, and audience perception studies diverge on exactly when and how climate images are experienced as distancing [33–35]. However, a range of studies PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 3/ 23

PLOS ONE Visual digital intermediaries and global climate communication suggest that many find the classic repertoires distancing, and some audiences more so than others. There is experimental support for abstract images (e.g. empty landscapes) being more likely to create a perception of climate change as spatially and temporally remote [36]. More- over, observational studies find that audiences in the Global North interpret the classic dis- tancing visual themes associated with mass media climate reporting as salient but demotivat- ing [37–39]. For some, the iconic visual themes have become worn clichés that associate with cynicism and fatigue [37,39,40]. In addition, while images travel, audiences also interpret the same image differently. For indigenous communities and audiences outside the Global North, the themes can be far removed from both familiar visual repertoires and lived experience [41]. Accordingly, recommendations for compelling visual climate communication echo the general call for communicators to prioritise speakers and content with relevance to the audi- ence at hand. The recommendations for visual content emphasise relatable proximity in actual depictions: the advice is to show ordinary or ‘real’ people in authentic (not staged) situations and scenes with local connection [1,40,42]. More generally, recommendations emphasise the value of deploying fresh visual narratives (not one-size-fits-all clichés), and being sensitive to ways in which visual themes, references, and repertoires may be so culturally loaded that they

are not readily relatable outside that context [1,4,24,40]. Testing such principles in the digi- tal sphere, an analysis of 380 images in Twitter ‘top tweets’ confirmed that user engagement in comments, likes and retweets was higher for climate change messages that contain images with real people and include a local connection [3]. In principle, such findings suggest that the digital context may offer an organic path away from the green ghetto of usual suspects and themes. It is easy to assume that the ability to source one’s own visuals would make it easier, and perhaps more likely, to tell stories with local themes [7]. It would follow that, if a variety of communicators use their own sources to tell stories, if images with ordinary people and local connection are popular, and if we factor in how visuals increase the spread of messages [43] and how imitation features in networked visual publicity [44], then online climate discourse would become more varied and usual sus- pects and timeworn distancing themes less dominating. Yet, this is where digital interme- diaries complicate the picture. What climate communicators post is one thing, the visibility afforded them is another. 2.2 Diversity and proximity in online rankings An important part of what platforms do in the world is present issues to audiences. Research in platform studies emphasises the persuasive power inherent in how search engines and social media present actors and content. User audiences stay with default settings [9], and

may not consider who or what is not included. They respond to signals of salience and urgency in engagement metrics and trend tickers [45–47]. They pay attention to, and ascribe authority to, posts positioned at the top [48,49]. Rankings — for which platforms filter, hier- archise, and display actors and content — become pivotal in suggesting the shape, scope, and relevance of issues. Algorithm audit research helps investigate this aspect of what rankings do. Such work sys- tematically probes algorithmic workings in the light of societal implications [50,51], for exam- ple with respect to how rankings amplify gender and race bias [52], misinformation [53], memory politics [54], and political information bias during election periods [55]. Two strands in this literature are pertinent for assessing the persistence of a green ghetto in top rankings. The first centres on the (lack of) diversity at the top, which has bearing on the extent to which fresh voices gain prominence. This work suggests issues of composition PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 4/ 23

PLOS ONE Visual digital intermediaries and global climate communication (which actors) and concentration (their distribution) can be crucial. A number of considera- tions may play into making results relevant, including freshness, user satisfaction, and intro- ducing fresh content by serendipity (while perhaps balancing with sponsorship). The evidence challenges common assumptions about the constraints on diversity in typical ranking logics [56]. However, studies also present instances in which particular platform ranking practices privilege large or already popular sources while under-representing others [57], concentrate attention to a select number of sources [58], or skew content due to differences in what high- and low-ranked sources present [59]. The second strand examines the relevance of top-ranked content for differently situated audiences in terms of proximity, which has bearing on issues of local connection. Impor- tantly, this work illuminates systematic variations in the ‘localness’ of top-ranked content (beyond personalisation). Platforms roll out features and moderation technologies at differ- ent pace across locations or in keeping with local regulation [60]. The same arrangements may also produce distinct results in different contexts. In one study, for example, Google Search returned city information with different localness of sources depending on how developed publishing and scientific industries were in the respective regions [61]. In a different take on local connection, another study found that object recognition systems of the kind that

underpin image search performed poorly in recognising household items such as spices from non-Western and low-income communities [62]. These two strands inform our approach to the question of how platforms present climate change to differently situated publics, and specifically the degree to which top-ranked content includes fresh voices and local connection. We specify our first question as follows: (RQ1) What characterises top-ranked results for climate change with respect to diversity and proxim- ity on YouTube Search — and does this differ across region/language search contexts? 2.3 The conditions of visibility across contexts Another important part of what digital intermediaries do hinges on the processes that pro- duce top ranks. If we want to understand the role of digital intermediaries in climate discourse in light of the green ghetto, we need to consider not only whether fresh voices and local con- nection appear in the top results, but also what the conditions of visibility are for this to occur. Here, the workings of the digital intermediary go beyond ‘the’ platform or ’the’ algorithm to also encompass actors such as the users that post content and the audiences that engage with it. Research on how algorithms intervene in the world emphasise the socio-technical dimen- sion [63]. A strong case is made for acknowledging the distributed agencies that converge as assemblages in ranking processes [11], what we here refer to as ‘ranking assemblages’. On this view, ranking processes are embedded in, and are an entanglement of, platform politics [64]

and use cultures [65]. Intensifying the multi-sided tangle, interaction and cues run in sev- eral directions. For example, between algorithms premised on user engagement (among other things), actors that post content and anticipate such curation (among other things), and audi- ences that discover content to like. Consequently, in specific assemblages, some outcomes become more likely and others less so, which circumscribes the conditions of visibility in that context. Taking this starting point seriously implies that it makes sense to consider ranking assem- blages and their conditions of visibility in a comparative perspective. Given that not only plat- forms but also other actors imprint ranking assemblages, we can expect differences across contexts in several tangents. On the one hand, as noted, platform-side doings may in fact vary PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 5/ 23

PLOS ONE Visual digital intermediaries and global climate communication across contexts — or, a platform doing the same thing may integrate locally to different effect. On the other hand, what content contributors and audiences do also plays in. The intuition about varied patterns is supported by emerging research. An exploratory cross-platform study identified distinct visual vernaculars associated with climate change across platforms [16]. Related work (not on climate issues) suggests that there is also reason to expect differences within the same platform. In the case of YouTube, a handful of studies show that ranking cul- tures differ across issues (e.g. have a stable or ‘newsy’ morphology [11]), and take different shape around the same issue across language/region contexts [66,67]. Such suggestive find- ings motivate the question whether the conditions for dissolving a green ghetto are the same everywhere on a platform such as YouTube. We specify our second question as follows: (RQ2) What characterises ranking assemblages underpinning top-ranked results on climate change on YouTube Search — and does this differ across region/language search contexts? The multi-sided interactions in ranking assemblages are inherently difficult to pick apart. Nevertheless, we argue that they are still accessible to comparative analysis. In this study, we follow the proposal to address the tangled agencies through API-based large-scale observation [11,66]. This, like most other approaches to algorithm auditing, leverages outside observa-

tion to handle the challenges of access, transparency, comprehensibility and randomisation in dealing with algorithms [50]. We recognise that other approaches could be more power- ful for focusing particular relations (e.g. agent-based audit to focus user-platform dynamics [68]). However, an advantage of API observation for our purposes is that it makes it possi- ble to systematically probe the multi-sided interactions in how they unfold over time, and thereby lends traction to comparative analysis. Since the API for a platform such as YouTube Search runs as close as possible to a baseline search, the API approach is also useful when the analytical focus is the platform-level process rather than what specific users actually see [11,69]. 3 Climate change on YouTube search We focus on YouTube, a video sharing and social media platform that is receiving increas- ing attention for its pivotal position in global online climate discourse. The U.S.-based plat- form is a global destination for climate news, science, entertainment, and education con- tent that spans the spectrum from professional to amateur [70,71]. It is localised in over 100 countries and 80 languages [72]. A sizeable audience is in the U.S., which is the most stud- ied. By 2022, however, YouTube was twice as big in India, and had major markets in countries such as Indonesia, Brazil, Russia, and Japan, followed by others such as Mexico and Germany [73]. The platform offers several popular ways to discover, navigate and consume content,

including Search, following recommendations for related content via the UpNext listing, or subscribing to particular channels. Search is a central affordance on the platform. The default setting ranks results by ‘rele- vance’, displaying thumbnail image, title, views and date uploaded. YouTube prioritises three key considerations in connecting users with content in Search: the relevance of keyword match in titles, thumbnails, video description and content; viewer satisfaction in aggregate signals of engagement such as watch time, likes, and comments; and quality. For quality, the platform surfaces authoritative sources and demotes borderline content, especially for news, politics, science, and medical information. Personal search and watch history contribute to results if a user is logged in [74,75]. As noted, ranking assemblage differences have been observed within and between other issues on YouTube Search [11,67]. PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 6/ 23

PLOS ONE Visual digital intermediaries and global climate communication YouTube has been accused of being both too lax and too censoring of actors and content (in different contexts) [76]. If global search results on climate change follow the patterns from English/US studies on mainstream climate search terms, we would expect them to be rela- tively conservative: they would be primarily characterised by ‘usual suspects’ and classic visual themes. Previous research on US/English samples finds that news media channels are promi- nent, and that they dominate among those showing classic distancing themes such as the ‘dis- aster impact’ frame [69,70]. We can expect continuities across regions on this point. Climate change is precisely the type of issue for which YouTube elevates authoritative sources such as news media, and research on global climate journalism shows that while there are region- specific trends, there is continuity across countries in common news frames, including the distancing frame [12,14]. Regarding the user-led channels, studies of YouTube Search indicate that they also appear in top results, and that they tend to be more popular [69,77]. However, it is less clear what to expect from them with respect to local connection in content. A qualita- tive study with focus on ‘gain’ and ‘loss’ frames observed that user-led climate channels tend to re-inscribe legacy media repertoires into their videos [78], but the study on the disaster impact frame did not note user-led channels using this frame in top results [69]. 4 Method

This study pursues an approach of quantitative description [17] and large-scale observation [11,66] to examine what YouTube Search presents in top rankings and the underpinning pro- cesses with respect to climate change. For our first question, about platform presentation, we consider the channel actors and visual content in top ranked results. For the second question, on the underpinning processes, we consider the platform, channel, and audience sides of the ranking assemblages. 4.1 Delimitations As we detail in the following sub-sections, we make several key delimitations in order to focus platform-level processes in this context. Chief among these include the following: • We analyse only the visual dimension of YouTube and its videos, and do not consider other aspects such as the audiovisual elements and user commentary. • We look at API results, not at what specific users would see when searching (e.g. based on their search history or other criteria that YouTube might be using but not corresponding to specific API settings). • We focus on a keyword that has been shown to have a mainstream profile on YouTube search, ‘climate change,’ [70], since we expect green ghetto discourse to characterise main- stream communication. For this reason, in this study we have not used other keywords from the literature, such as ‘climate hacking’ or ‘chemtrails’, better suited to the study of specific aspects within climate change and/or specific groups of producers and audiences. Practically, using a general keyword also simplifies the translation to other languages,

which is an important factor given the world-wide scope of our research. • We examine only one kind of distancing theme (in its arctic rendition), which is selected to give us leverage on the comparative analysis of relevance in terms of freshness and local- ness. • We approach the image data as still images, and do not consider their sequence since we are only looking at which videos show this type of content. • Finally, although we are working with a sizeable dataset, our observations are by design based on small sample sizes, since we focus on the videos that the platform ranks at the very PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 7/ 23

PLOS ONE Visual digital intermediaries and global climate communication top across contexts. Our results are consistent across data collections performed at different times, which allows us to make observations regarding the presence or absence of specific themes and actors in the search results. However, we do not use our data to support statis- tically significant statements such as the fact that the proportion of videos showing arctic themes for one country is higher or lower than for another. 4.2 Data collection The data for this study was collected from the YouTube Search API in May 2021. The collec- tion retrieved search results for all countries that can be specified in the API in their main official languages, based on videos published up until December 2020. We decided to use all languages that are listed as official for at least one of the country codes that can be specified in the YouTube API in order to be as complete as possible. We achieved satisfactory coverage of 232 countries and 85 languages, despite not being able to find translations for a few languages (which we document in the data released with the study). The accounts used to retrieve the data are set to not use personalisation. We used the following API settings: • We set order by ‘relevance’, the default search setting. • The search term or query (‘q’) was ‘climate change’ or its equivalent in each respective language. • Searching for all official languages meant that for some countries we made multiple distinct

queries. For example, in the case of Canada, we searched for both Canada/English/‘climate change’ and Canada/French/‘changement climatique’. • We used the RegionCode parameter to pass country information to the API. This parame- ter only retrieves videos that can be viewed (i.e. are not blocked) in the specified country. • We did not use the location and locationRadius parameters, which constrain the results to channels located in the specified radius, since they do not seem to be used in the default search settings. The data consists in the ranked results of 276 queries, which cover 232 countries and 85 languages. Our self-imposed limit was 400 results per search. We here only report find- ings based on those region/language pairs with 400 results. 253 queries returned the full 400 results, covering 219 countries and 62 languages. We focus the analysis particularly on top-20 rankings [48,49]. 4.3 Data preprocessing: Channel coding and image annotation To prepare the channel data, we coded all channels with videos in top-20 rankings by actor type, based on channel description. The coding scheme of 17 categories builds on previ- ous work [69,79], to include actors that are conventionally prominent or studied (e.g. legacy media, science, and civil society organisations), traditional actors that have conventionally been less prominent (e.g. large businesses, political parties, public agencies, religious organisa- tions), and actor types created for this context (e.g. user-led channels, online popular science

and education actors). The inter-coder reliability between two coders on a random sample of 10% of a larger data set of 1631 channels was Cohen’s kappa .81 (see supplementary mate- rial: coding schemes, channel categories). While the agreement was high in general, there was some disagreement on the code Popular science group, which we note here because this category became important in the analysis. This disagreement was analysed and resolved by discussion, as it had to do with miscommunication regarding the focus on non-institutional PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 8/ 23

PLOS ONE Visual digital intermediaries and global climate communication groups dedicated to popular science communication (as opposed to individuals, public intel- lectuals, or school classes). For the visual analysis, we used a method centred on computational image analysis to assess the presence of a distancing visual theme that has been strongly represented in con- ventional climate communication such as mainstream media reporting. We sought a means of analysing local connection in visual repertoires as part of assessing a potential break in convention and move towards relatable climate communication. Because of the difficulty in achieving robust large-scale comparative coding of ‘local’ themes, we focused solely on the opposite to local, i.e. the presence of a classic distancing theme. We selected the distancing theme in its arctic rendition: polar bears, icebergs, and ice cliffs in water. Building on the liter- ature outlined earlier, we assume that many find the arctic themes not relatable. This, because these classic themes are geographically remote (for most), devoid of people, and specific to a particular visual repertoire tradition. Moreover, the use of the classic arctic themes suggests thematic continuity rather than innovation. To prepare the visual data for computational image analysis, we segmented the videos into shots (contiguous frames from an individual camera) and selected the central frame from each shot, resulting in about 150 000 images for a set of top-20 results. Following a procedure

reported in greater detail in the Supplementary material, to identify polar bears, we classified the images using a combination of pre-trained and custom deep neural networks, all with a single corresponding label. We also added polar bear images that we discovered in the Google Vision API search for ice images, and manually eliminated false positives. For icebergs and cliffs, we used Google Vision API to identify frames with ice (and related labels) and manually coded the resulting data according to a visual code book (see supplementary material: coding schemes, visual themes). The inter-coder reliability between two coders on this material was Krippendorf ’s alpha .84. The high reliability score can partly be explained by the decision to only consider unequivocal depictions of the themes. For example, pictures coded as icebergs must show an iceberg in a prominent position, with the ice clearly protruding from the water (the visual codebook and examples of positive and negative images are provided in the Sup- plementary material). We note that this approach yields a conservative assessment of the pres- ence of the themes: allowing broader definitions of these themes would have led to an even higher amount of content being coded as icebergs and cliffs. 4.4 Assessing diversity and proximity in top-ranked results We draw on the channel coding and image analysis just outlined to assess the platform pre- sentation in terms of diversity and proximity (RQ1). To assess diversity, we analyse both the

concentration and composition of channels. We consider concentration of channel types using entropy and classification error as measures to examine whether the platform presents a broad or narrow set of voices in top-ranked results. We analyse the composition of channel types to examine the diversification of voice beyond traditional climate communicators (see supplementary material: measures). To assess proximity in top-ranked results, we analyse the local connection of contributing actors (i.e. channels) and visual themes in the video content. To analyse localness of channels, we refer to channel metadata about source location. For visual themes, as discussed earlier, we focus on classic arctic distancing themes as preeminent non-local motifs. We analyse the pro- portional prevalence of these specific themes both in top-20 videos and across lower-ranked results. PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 9/ 23

PLOS ONE Visual digital intermediaries and global climate communication 4.5 Examining ranking assemblages To examine the conditions of visibility in the underlying ranking assemblages across con- texts (RQ2), we undertake systematic large-scale observation [11,66] with focus on platform-, user-, and channel-side factors. As noted earlier, it is difficult to disentangle the various agencies that contribute to rank- ing assemblages. Nevertheless, we argue that it is possible to systematically probe the multi- sided contributions in a comparative perspective. The systematic analysis pivots on a logic of potential position. In brief, we assume a space of possibilities for how platform, channel, or user (i.e. audience) agency could in principle influence the ranked content, then locate the actual position in the top-ranked results. We focus on the potential of a particular side to con- tribute to diversifying the ranking in ways relevant to dissolving the green ghetto on a plat- form such as YouTube. To dissolve the green ghetto, platforms need to surface fresh channels and themes, but it is also the case that channels need to bring fresh content and audiences need to engage with it. For the platform side, we analyse the extent to which the same query yields different results across time and place, with an eye to the visibility given to particular types of channels and the conventional arctic visual themes. We assume that the platform in principle could actively diversify results for the same query at different times and could return locally relevant results

at different locations. For the temporal analysis, we execute the same queries at different times, from a few seconds to three months apart, then compute the overlap between top- ranked results as the number of common top-20 videos divided by 20 (see supplementary material: diversity, time). We only requested videos published until December 2020 (several months prior to the data collection) in order to avoid contaminating the results with recent videos not available to previous searches. To assess localness in the platform context, we take all available region/language pairs query results, and examine location indirectly, through lan- guage (expecting there to be a correlation between channels and videos with metadata from regions where that language is spoken), and directly, by analysing the effect of the regionCode parameter, which returns only videos that can be viewed in the specified country (videos can be blocked in a specific region either by the channel or by the platform). For the channel side, we assess if diversification of (channel) voice brings diversification of visual content by examining whether all channel types in the same context have the same probability of showing the arctic visual themes. Similarly, to probe the role of user audiences, we consider diversification in channel preferences. We assume that if audiences strongly pre- fer a channel, this would favour its visibility in the top (even if channels with little engage- ment also appear). We take engagement in views, comments and likes as proxy for the broader

set of user popularity signals that platform ranking algorithms may draw on. The YouTube data only provides global metrics for each video, but we can still compare relative preferences across contexts. 5 Findings 5.1 Diversity and proximity in top-ranked results We first consider how YouTube articulates climate change to global audiences: what charac- terises the top-ranked results on YouTube Search with respect to diversity and proximity, and does it differ across contexts (RQ1)? The findings show that on YouTube, the issue is still typ- ically a distant problem delivered by a constrained circle. In general, the issue is defined by a delimited set of channel types, who do in fact expand beyond the usual suspects but still sus- tain a traditional distancing visual repertoire. This said, it depends on where you look. Closer PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 10/ 23

PLOS ONE Visual digital intermediaries and global climate communication examination reveals that language groups articulate the scope and character of the issue with notably distinct profiles. The aggregate analysis focusing on the top-20 results suggests that YouTube supports ‘vis- ibility winners’ [66] on climate change. A concentrated set of channel types dominate top results, and their videos are consistently more likely to top search results across different regions. As Table 1 shows, media organisations are the most prevalent, with a higher frac- tion of videos in the top results. Their videos are highly amplified — the API returns the same video in several regions’ rankings — but not the most so. Actors fresh to conventional climate communication are also highly amplified, and audi- ence engagement with these channels is higher than for professional media channels. User-led channels are relatively prevalent at the top, highly amplified, and (often) highly engaged with. Videos posted by digitally native popular science groups and online education actors are the most amplified, although with limited prevalence in top results. For example, the same Ger- man popular science channel appears across several global regions. These popular science and education channel types are also the most engaged with. By contrast, other traditional climate communicators have little visibility in the aggregate results. Authoritative channels such as (inter-)governmental and science organisations, and potential stakeholders, such as political

actors, civil society, and religious organisations, are neither prevalent, amplified, nor engaged with. Nevertheless, the iconic themes of the distancing frame continue to be part of the climate narrative. One fifth of the search results show classic arctic themes (and recall that we are using a conservative coding for these themes). While fresh actors from beyond the conven- tional climate communication circle do appear in the top ranked content, they do not notice- ably reduce the prevalence of the arctic theme. If anything, fresh climate communicators sup- port its persistence. Table 1 shows the extent to which videos in the respective channel cat- egories include the themes. Professional media maintain arctic content, and online educa- tion actors match this. Yet, the other newly prominent actors, user-led channels and popu- lar science groups, surpass these levels. By contrast, traditional communicators such as sci- ence organisations, civil society, and (inter-)governmental organisations show little of these themes, but as just noted, are also not particularly visible themselves. Beyond the aggregate, there is significant variation across contexts. Language is the main differentiating factor, as opposed to geographical location on the globe or categorisations such Table 1. Composition and content of the combined search results from all country-language pairs, considering the top-20 results (videos) for each pair. #res: number of search results. %res: fraction of search results. #cha: number of distinct channels. #vid: number

of distinct videos. amp: fraction of search results divided by fraction of videos minus 1. #vie: median number of views per video. #com: median number of comments per video. #lik: median number of likes per video. %PB: fraction of videos showing polar bears. %Ice: fraction of videos showing icebergs or ice cliffs in water. %PBI: fraction of videos showing polar bears, icebergs or ice cliffs in water. Only categories with at least 2% search results. Category #res %res #cha #vid amp #vie #com #lik %PB %Ice %PBI All Categories 5060 1.00 1067 1369 0.00 2785 34 2 0.08 0.18 0.20 Prof. media org. 1932 0.38 310 436 0.20 3402 39 3 0.08 0.20 0.22 User-led channel 965 0.19 183 199 0.31 13778 192 32 0.13 0.27 0.29 Pop. sci. groups 424 0.08 43 55 1.09 28369 542 63 0.15 0.31 0.31 No description 397 0.08 199 222 -0.52 706 10 0 0.07 0.13 0.14 (Inter)gov. org. 317 0.06 80 131 -0.35 1253 10 0 0.02 0.13 0.14 Civil society 245 0.05 74 93 -0.29 841 10 0 0.06 0.11 0.13 Online education 235 0.05 41 48 0.32 8516 117 8 0.17 0.12 0.21 Business 119 0.02 32 35 -0.08 3595 23 3 0.09 0.14 0.17 Science org. 110 0.02 44 52 -0.43 1240 10 0 0.06 0.13 0.15 https://doi.org/10.1371/journal.pone.0318338.t001 PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 11/ 23

PLOS ONE Visual digital intermediaries and global climate communication as Global North/South. Recall that the data is collected through the Search API with no per- sonalisation: individual YouTube users may see different results, which is not what we assess in this study. The top results evince different profiles, and channel types behave differently across contexts. We here therefore present the results according to language groups. The English-language queries, as the largest cluster, follow the aggregate profile. As Table 2 shows, the top-ranked results concentrate to a small number of channels. ‘Climate change’ is primarily defined by professional media and user-led channels, and to some extent also popu- lar science groups. The top-20 channels tend not to be local: channels from only 14 countries are represented in the 79 English-language searches, and we note a clear dominance from the US and UK. Almost one third of the top-ranked videos include arctic distancing themes. Beyond English, the results in the language groups span a spectrum of profile types. Top results in French and Japanese distribute relatively evenly across channel types. At the other extreme, Arabic and German render climate change as an issue highly defined by one type of channel. In these examples, the channel type in focus is professional media, but this is not always the case in concentrated profiles. For example, in other (smaller) languages with concentrated profiles, (e.g. Turkmen, Zulu) user-led channels dominate (see supplementary material: top-ranked results).

Among the globally spoken languages, sources based in Europe often dominate. In the French cluster, the top-ranked videos are almost all posted by French (and some Belgian and Swiss) channels. This means that top-ranked results, which have a relatively high prevalence of arctic content (38%), are not local to other regions that get these results, such as, Canada, Cameroon, and Haiti. In the Spanish language cluster, which has levels of arctic content more akin to English (27%), channel sources spread between Spain and Latin America. Neverthe- less, Spain contributes the most channels. This is twice that of Mexico, despite Mexico being the biggest YouTube market in the cluster. By contrast, large markets with their own language have a larger portion of local chan- nels at the top. Japanese, German, Hindi and Indonesian queries (almost) exclusively present sources with local connection. Russia is an exception, with only half the videos stemming from Russian channels and the others from a range of eight countries. There is also a different pattern for Hindi and India/English: the former strongly features online education channels and channels with local connection, the latter aligns with the English cluster pattern of skew- ing towards media, user-led, and popular science channels, and only three are local. Still, in all of these languages — large markets but small YouTube languages — the arctic themes persist in up to one third of the videos (from Indonesian at 20%, German 22%, Japanese and Hindi at

25%, to Russian at 29%). For the Arabic cluster, in which most top channels are regional news media or localised versions of global broadcasters such as BBC, Deutsche Welle, and France 24, arctic themed content sits at 30%. The real standout is Portuguese. This is the only global language cluster in which reference to climate change as a distant problem via the arctic theme is virtually absent. Only one video, an animated infomercial posted by a No Description channel, depicts an arctic theme. This is also the only cluster dominated by channels local to a single Global South country, Brazil. The issue is articulated through a highly distributed set of channel types that inverts the oth- erwise typical profile: (inter)governmental and civil society channels are prominent, explicitly user-led channels less so, and professional media presence is limited to one single video. 5.2 Conditions of visibility: Ranking assemblages across contexts We now dig into the conditions of visibility underpinning the top-ranked results and how they contribute to sustaining or dissolving the green ghetto: What characterises ranking PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 12/ 23

PLOS ONE Visual digital intermediaries and global climate communication Table 2. Composition and content of search results for selected languages. Channel concentration is computed using the concentration measure based on entropy (supplementary material: measures), with higher values indicating higher concentration of the results towards a smaller set of categories (and thus lower diversity). Arabic Channel concentration: .41. Content: .08 (PB) .30 (PBI). Lang Category #vid #cha #vie #com #lik %PB %Ice %PBI ar Prof. media org. 27 15 3367 41 4 0.07 0.22 0.26 ar Pop. sci. groups 5 4 443622 7840 804 0.00 0.40 0.40 ar User-led ch. 3 3 4658 114 22 0.00 0.00 0.00 ar (Inter)gov. org. 2 2 5916 58 2 0.50 0.00 0.50 ar Dig. media prod. 1 1 204555 3239 109 0.00 1.00 1.00 ar Online education 1 1 17544 664 39 0.00 0.00 0.00 German Channel concentration: .39. Content: .15 (PB) .22 (PBI). Lang Category #vid #cha #vie #com #lik %PB %Ice %PBI de Prof. media org. 20 15 189056 3667 1206 0.15 0.20 0.20 de No description 2 2 221063 5089 2768 0.00 0.00 0.00 de User-led ch. 2 2 15856 279 132 0.00 0.50 0.50 de Civil society 1 1 5413 210 5 1.00 1.00 1.00 de Dig. media prod. 1 1 285431 4017 916 0.00 0.00 0.00 de Music 1 1 418060 712 26 1.00 0.00 1.00 de Online education 1 1 226329 2553 986 0.00 0.00 0.00 de Pop. sci. groups 1 1 295842 9558 2457 1.00 1.00 1.00 de Religious org. 1 1 285420 2273 30 0.00 0.00 0.00 de Science org. 1 1 5222 174 38 0.00 0.00 0.00 English Channel concentration: .34. Content: .10 (PB) .29 (PBI). Lang Category #vid #cha #vie #com #lik %PB %Ice %PBI

en Prof. media org. 43 32 47974 785 329 0.05 0.30 0.30 en User-led ch. 25 24 89116 4689 540 0.08 0.32 0.32 en Extr. weather ch. 5 2 42125 686 131 0.00 0.00 0.00 en Pop. sci. groups 5 3 4574151 243961 22929 0.00 0.20 0.20 en (Inter)gov. org. 3 2 4000 119 9 0.00 0.00 0.00 en Business 2 2 122354 1238 34 1.00 1.00 1.00 en No description 2 2 923957 14904 2843 0.00 0.00 0.00 en Online education 2 2 299392 3138 98 0.50 0.50 0.50 en Talk platform 2 2 685784 14126 2388 0.00 0.00 0.00 en Civil society 1 1 806 21 0 0.00 0.00 0.00 en Science org. 1 1 21123 211 119 0.00 0.00 0.00 French Channel concentration: .13. Content: .14 (PB) .38 (PBI). Lang Category #vid #cha #vie #com #lik %PB %Ice %PBI fr Prof. media org. 12 10 40314 452 44 0.00 0.33 0.33 fr No description 6 5 6420 124 18 0.50 0.50 0.67 fr User-led ch. 5 5 56007 2683 475 0.00 0.00 0.00 fr Pop. sci. groups 4 3 143014 1041 12 0.00 0.25 0.25 fr (Inter)gov. org. 3 3 26626 3 0 0.00 0.33 0.33 fr Civil society 3 2 38318 393 19 0.33 0.33 0.33 fr Politician, party, … 2 2 14362 2117 300 0.00 0.00 0.00 fr ch. discontinued 1 1 11793 182 130 1.00 1.00 1.00 fr Science org. 1 1 19748 238 160 0.00 1.00 1.00 fr Talk platform 1 1 13214 241 113 0.00 0.00 0.00 Japanese Channel concentration: .05. Content: .15 (PB) .25 (PBI). Lang Category #vid #cha #vie #com #lik %PB %Ice %PBI ja User-led ch. 5 4 22099 783 58 0.00 0.00 0.00 ja Prof. media org. 4 4 15662 194 42 0.25 0.75 0.75 ja Business 3 3 9485 103 4 0.00 0.00 0.00 (Continued) PLOS ONE https://doi.org/10.1371/journal.pone.0318338 April 8, 2025 13/ 23

PLOS ONE Visual digital intermediaries and global climate communication Table 2. (Continued) Lang Category #vid #cha #vie #com #lik %PB %Ice %PBI ja Science org. 3 1 43567 534 0 0.33 0.00 0.33 ja (Inter)gov. org. 2 2 8007 62 6 0.00 0.00 0.00 ja Pop. sci. groups 2 2 73106 2133 227 0.50 0.50 0.50 ja Civil society 1 1 1224 118 3 0.00 0.00 0.00 Portuguese Channel concentration: .11. Content: .00 (PB) .02 (PBI). Lang Category #vid #cha #vie #com #lik %PB %Ice %PBI pt (Inter)gov. org. 9 9 2289 133 1 0.00 0.00 0.00 pt Civil society 7 6 841 22 3 0.00 0.00 0.00 pt No description 4 4 1262 14 0 0.00 0.25 0.25 pt Pop. sci. groups 4 3 548 114 0 0.00 0.00 0.00 pt Online education 3 3 37395 1315 0 0.00 0.00 0.00 pt User-led ch. 3 3 13778 3401 135 0.00 0.00 0.00 pt Prof. media org. 1 1 2829 60 0 0.00 0.00 0.00 pt Religious org. 1 1 11263 301 34 0.00 0.00 0.00 pt Talk platform 1 1 1360 87 0 0.00 0.00 0.00 https://doi.org/10.1371/journal.pone.0318338.t002 assemblages underpinning the top-ranked results for ‘climate change’ on YouTube Search, and does this differ across region/language search contexts (RQ2)? Our findings suggest that the ranking assemblages, and the conditions of visibility that they entail, are fairly constant. On this mainstream climate keyword, ranking assemblages operate in largely similar ways across contexts, and this in turn translates into different prospects for fresh and locally rel- evant climate communication in particular contexts. We probe the platform, channel and audience sides of the ranking assemblage, and find that agencies on the respective sides