Abstract
This article combines two neglected elements within the history of online news: public service news sites and weather reporting, and it does so by utilising web archives, which – surprisingly – do not figure very prominently in journalism history. The two elements have – in isolation and in combination – at least in Denmark, become increasingly important as the online news sections of the two public service institutions Denmark’s Radio (DR) and TV2 consistently are among the most visited news sites and since reporting on the weather has gained in prominence and more recently, at least on DR TV, has become increasingly educational in its linking to issues of climate change. This article focusses on online news and conducts a historical analysis of the weather reporting on DR.dk from 2005 to 2022. The analysis seeks to balance the coding of journalistic texts with considerations of the online form of journalism, which here broadly means reading the webpage as a text. A key focus in the analysis is how meteorological data have been woven into cultural and social narratives, some of which are linked to climate change.
Keywords
Introduction
In Home Territories, Morley (2000) discusses how, in broadcast history, ‘[e]ven the weather was nationalised and how this meant national limits [being] clearly demarcated’ (p. 106). An important part of this was the daily UK shipping forecast whose ‘mesmeric voice and tireless rhythms are buried deep in the public consciousness’ (Chandler cited in Morley, 2000: 106) while in Sweden ‘the names of the coastal observation posts . . . were read like a magical chant, as outposts encircling the nation’ (Lofgren cited in Morley, 2000: 106). If you are a Dane of a certain age, you may in similar ways recall the sound of, and atmosphere linked to, names such as Doggerbanke (Dogger Bank) and Fladen (Fladen Ground), North-Sea fishing locations whose forecasts were on national radio.
The forecasts within general journalistic outlets, and certainly in public service media, have remained focussed on the weather as a daily ritual formation of ‘collective identity’, which was (and perhaps still is) possible as the weather ‘is not controversial’ (Berland, 1994: 99). It is indeed noticeable that as weather forecasts have become increasingly less important for most of us professionally, the weather has become woven into a much broader range of cultural narratives. On TV and also in print, weather reports have thus in many places become more comprehensive and significant. In The Weather Obsession, Zion (2017: 3) points to a 2012 PEW study in the US showing that ‘the weather was the most followed topic in local news’.
But the character of weather and its reporting is obviously not uniform across cultures; this is brought out very clearly in a comparison between modes of address in Greek and English TV weather reporting in the mid-1990s (Sifianou and Tzanne, 1996). Based on a range of linguistic characteristics (e.g. levels of formality and authority), the analysis shows that the Greek form is generally rather distant in its authoritative presentation of ‘raw weather data’ whereas the English ‘forecasts construct a relationship of closeness and familiarity where presenters identify themselves with the audience and what the weather means for everyday life’ (e.g. traffic, golf and gardening). This is done through a vernacular – ‘a friendly chat about weather conditions rather than an official report’ (2) – in which the host is sharing (and inviting) experiences of the weather. It is, however, as we will discuss in more detail, precisely when raw meteorological data are mediated and interpreted within specific social and cultural contexts that ‘news about the weather’ (Sifianou and Tzanne,1996: 3) become news in a broader journalistic sense.
The DR online weather reporting in focus here has developed on the URL http://www.dr.dk/nyheder/vejret since around the turn of the century. The webpage has consisted of a mixture of visualised forecast data and news stories with varying degrees of relations to forecast data and with shifting ties to the weather coverage on TV. Yet, almost from the beginning, the page seems much closer to the form and style of the English TV forecasts than to the Greek ones described above. While all mediations of data are cultural (as are data in themselves), there is after all a continuum running from minimal presentations in which data are more or less left to speak for themselves to presentations in which the data form the background of a range of news narratives about experiences linked to vacations, hay fever, sunbathing and gardening; most weather coverage mixes content from different positions on this continuum. As such, weather forecasts exhibit in a very stripped down fashion one of the key tensions within climate change journalism, namely that between scientifically based forecasts and their meaning for daily life. But, while it is very easy to imagine what your day tomorrow will be like depending on the weather, it is much more difficult to envision what summer in your location will be like in a generation. Yet, while most mediated weather forecasts have a timeframe of 5–7 days, we are seeing a shift in which longer time frames – both backwards and forward in time – inform the coverage of the weather and thus how this coverage starts taking on additional meaning. One element pointing in this direction is the coverage of unusual weather and extreme weather events and their increased frequency – both aspects that may involve more explicit discussions of climate change.
Given that most of the relatively little research on weather reporting is focussed on TV, we turn our attention to online weather pages in order to understand how a specific type of content has evolved along with the rise of the internet and how specific characteristics of online communication plays into this. By focussing on shifting balances between weather data and culture will seek to contribute to broader discussions about how projections and reporting relating to climate change may be made culturally and politically relevant at a more intimate level for those for whom this is not yet the case. Our research question is: How has meteorological data been mediated in the weather reporting on the weather page of Denmark’s public service institution DR from 2005 to 2002? Below, we will explain our choice of case study in more detail. Before that, however, we turn to the research context of this study.
Research context
The weather, like music, mediates between our physical and social bodies. Its rhythms and irregularities, and the rituals we construct around them, shape what it means to be part of the social, both within a particular time and space, and across to other times and places, as we imagine or remember them (Berland, 1994: 899). We are related to it, we do affect it, and to that extent we are responsible for its well-being. All we need now is a way to write that on to the maps, in all the diverse ways that such inscription occurs (Berland, 1994: 112).
The two sentences above are respectively the first and last of Berland’s 1994 article ‘On reading “the weather”’ and as such both grounds and sets the direction of this article. In terms of grounding, we share Berland’s (1994) argument that the various representations of the weather it reveal themselves to ‘to students of discourse and society as a product of current regimes of knowledge’ (p. 112). As far as direction is concerned, we are similarly interested in following how climate change may introduce a new cultural dynamic in our understanding of (severe) changes in the weather and how this may be written on weather maps, which we translate into a concern with how meteorological, journalistic and everyday knowledge about the weather have been negotiated on DR’s webpage in the decades following Berland’s rather prophetic article from 1994.
At least in ‘the modern industrialized world’, writes (Rayner, 2003: 286) ‘our experience of nature is . . . is invariably mediated by science’. Thus, for a number of years ‘information about weather [has been less] derived from personal observation, but from mass media communications based (presumably) on observations using scientific instruments and analysed with the aid of computers’ (Rayner, 2003: 280). As such we may, in the words of Berland (1994: 111), see
weather forecasts . . .[as] a ritualized celebration of the technological sophistication of satellite surveillance, space-based photography, and computerized image data analysis.
Yet, while scientific terms have entered our everyday vocabulary, there is a broad difference between science-based and experienced-based perceptions of the weather and this is partly because of the role of memory, which as Zion (2017: 10) argues ‘generates meaning, not statistics’. There is, says Berland (1994: 112),
something called ‘nature’ [which] makes a special unsanctioned appearance beneath and between the cracks of modern science, [and] . . . “residual” forms of knowledge and belief about the weather - whether of a classical, theistic or pantheistic nature - are thriving in the loquacious everyday world of popular culture.
The meaning of the weather is thus to be found in mediations between everyday lore and/or ethnometeorology and complex scientific modellings as these feed into weather forecasts, which ‘presents weather in increasingly technological terms’ (Zion, 2017: 10), which – as we shall see – increasingly is mediated and accompanied by vernacular expressions partly emanating from users of TV or online news.
The media crop up in studies within the fields of geography (e.g. Hulme and Burgess, 2019; Keeling, 2010) and anthropology (e.g. Golinski, 2003). Within the field of media studies there has been some engagement with the weather most of which is of a historical nature: Mergen (2008) depicts the development of broadcast weather reporting in the US, while Zion (2017) focusses on the developments across media Australia (and to some extent the UK and the US). Both books outline the interesting and changing relationships between national meteorological agencies and the media and Zion outlines how the Australian Bureau of Meteorology is now a media organisation in its own right while its information is permeating a range of internet platforms and apps.
Over the approximately two-decade lifespan of journalism studies there has been remarkably little research on weather reporting. This is probably partly because this type of coverage is not deemed important for processes of democracy. Given an increasing interest in the issue of climate change there has, however, recently been a growing interest in the coverage of extreme weather events – for a recent study and overview see Strauss et al. (2022). While offering important insights into how weather events and climate are or are not connected by journalists, most of this research is focussed on news journalism and not weather coverage. An important tendency is thus for weather issues becoming news. Zion (2017) has a chapter entitled ‘The weather becomes the news and with the same title, a well-known Danish meteorologist and TV weather host published a book in 2021 subtitled’ ‘20 years with climate changes’ (Theilgaard, 2021). What our study shows, however, is that the newsworthiness of weather is not only linked to climate change.
We are mainly interested in understanding the development of day-to-day coverage of the weather and what this might tell us about the possibilities of incorporating data and news addressing issues beyond the immediate weather. We are thus more interested in the ritual aspects of weather communication than in the transmission of specific information. As such, we are – while not focussing on climate communication as such – in line with Donald (2022: 33), who argues for ‘[s]tudying climate communication as a community ritual’. Following Carey (1989) – who introduced the ritual/transmission distinction in journalism studies (although not in relation to climate issues) – this important as it focusses our attention on continuity, which is a significant challenge with regard to climate change communication (Adam, 2021; Moser, 2016; Ytterstad and Bødker, 2022).
Finally, Durham Peters’ The Marvelous Clouds (2015) should be mentioned as it grapples with media and the natural world in a more fundamental way than a concern with technologically-enabled human communication. For Peters, the digital asks us to revisit and reformulate a more fundamental idea of media, namely as broader infrastructures that have meaning but do not speak directly. From this perspective, the sky and the weather are ‘being altered by media, understood as infrastructures of data and control’ (Peters, 2015: 2); and this is also what Berland sees emerging in 1994, when she talks about the necessity of learning to relate to the weather as something reflecting human choices. This is, however, as we shall see, not what lies behind the renaming of the DR weather page from ‘the weather to’ ‘our weather in 2017’.
At some level, weather reporting as a kind of data journalism:
some domains with a prospective orientation—like weather forecasting, election campaigns, sales and marketing, crime prevention, or economic modelling—have been profoundly changed by the unprecedented computing power (Pentzold and Fechner, 2021: 716).
That is indeed true as data has made more accurate forecasts possible while also allowing for more elaborate and detailed visualisations. The interesting thing is, however, that the development that we outline and discuss below seems to have moved somewhat away from weather reporting as explicitly data-based and towards a format in which the ritual and identity-related aspect have been foregrounded. This does obviously not mean an absence of data but rather a development through which the available data form the background of culturally relevant news reporting.
One could argue that the increasing awareness of the human impact on present and future climates and thus the weather potentially makes weather coverage an extended form of data journalism in the sense that this is not only about forecasting the weather but also about predicting trajectories based on the calculated impact of policies and lifestyles. So, when Zion (2017) argues the main difference between conventional weather forecasting and climate-change forecasting is the timespan, he is only partly right, as future trajectories must, as noted, factor in predictions of human activities and their accumulated effects. This indeed makes the communication of climate forecasts challenging.
Methodology
Choice of case
While the coverage of the weather on TV by the two Danish public service institutions, the Danish Broadcasting Company DR and TV2, has developed and become more prominent over the years, the websites of these institutions are consistently amongst the most visited in Denmark: in March 2023, the sites TV2.dk and DR.dk were respectively ranked number three and four – after the sites of the two major Danish tabloids (Danish Online Index, 2023). In addition, for DR, which is the public service institution chosen as our case study, TV weather reporting has become increasingly explanatory in its linking weather phenomena to climate change. In Denmark, climate change is a relatively less polarising topic compared to, for example, the US., and this makes Denmark a suitable supplement to existing research on weather and climate reporting, which mostly stems from an Anglo-American context.
Public service media (PSM) institutions (ideally) target all members of the public, and a national community (Denmark’s Radio (DR) Public Service Asessment, 2021; Sehl, 2020); and 68% of Danes agree that public service is important for society (Newman et al., 2023: 44). In terms of news, 78% list online as a source of news (Newman et al., 2023: 71), 83% trust DR news and DR online news has a weekly reach of 40% (Newman et al., 2023: 71). In terms of access, 48% state they get their news directly from the source (Newman et al., 2023: 12) rather than ‘via side-door routes such as social media, search or mobile aggregators and this runs counter to the development in most other parts of the world’ (Newman et al., 2023: 10).
Based in the above, it would be fair to say that the weather reporting of DR (TV and online) is the weather of ‘the community of residents in Denmark, and draws on shared culture and narratives, collective stock of accumulated knowledge, and normative assumptions of what is important, preferable, and undesirable vis-à-vis the weather’. We are specifically interested in how this plays out online as, firstly, most research on weather reporting has been focussed on TV, and, secondly, that a look at online formats and text may give us a different insight into how meteorological data are journalistically mediated and utilised in various collective narratives.
Data and analytical approach
We analyse the evolution of the webpage DR.dk/nyheder/vejret. Archived versions of the webpage were extracted from the internet archive via the Wayback Machine in the period from 2005 to 2022. The Wayback Machine is an interface to the internet archive, which was founded by the nonprofit organisation The Internet Archive. Surprisingly, web archives do not figure very prominently in research on the history of journalism (Bødker, 2018; Weber, 2018), which might be due to a range of challenges that vary according to the specific archive and its unique methods for harvesting data. Indeed, an internet archive is not ideal for collecting and analysing textual data exhaustively and comprehensively, as the archived web page is a unique version since the actual act of finding, collecting and preserving changes the webpage that once was on the live web (Brügger, 2008: 156). During the archiving process, moreover, the content of a webpage may change, which has two consequences. The first is that we cannot be certain that everything published is present. The second consequence is that we might get something in our archive that was different from what appeared in the past, as an archived version of a webpage can be a combination of elements from two or more versions that were never present at the same time (Brügger, 2005: 23). In addition, data harvests done by internet archive crawlers are often not consistent from year to year, which can result in differences in the amount of data archived, which in turn, might create challenges for the exhaustiveness and the representativeness of a study. Nevertheless, as our purpose is to track the historical development of the webpage including both its visual appearance and its textual themes, an internet archive is, in our case, necessary.
First, we extracted all archived versions of the webpage DR.dk/nyheder/vejret from 2005 to 2022, which resulted in 1722 unique data elements each of which consists of a screenshot and a text file with all the text from the page. It is worth noting here that we distinguish between website and webpage, where the former is a collection of many different pages (e.g. DR.dk), the latter is a specific page within a collection of many others on a specific site (see Brügger, 2018). By scrolling through the 1722 screenshots, we identified five distinct periods in the development of the webpage. Given our interest in the journalistic mediation of weather data, we distinguished the periods by looking for shifts in the general graphical layout of the page (colours, fonts, columns and menus), the amount and position of visualised meteorological data, the role of weather hosts, reoccurring interactive elements and the type, amount and positioning of weather stories and news.
By going through the screen shots, we iteratively developed a codebook consisting of eight coding categories and classification rules with the aim of capturing the various degrees of journalistic mediation of data as well as variation in specific types of mediation. 1 To validate the coding categories, we calculated intercoder reliability based on a test coding of a smaller sample, that is, 67 weather report elements conducted by two coders. We obtained a Krippendorff’s alpha of 0.619, and a Cohen’s Kappa of 0.626, which indicates an intercoder agreement ranging from acceptable to substantial. From the 1722 data items, we constructed a sample of an artificial summer week for each of the five identified periods. We first established the number of data items within each period from the three Danish Summer months (June, July and August). We then divided this number by seven to get the interval between days in the artificial week. We then selected the first Monday, skipped the interval number and took the first Tuesday after that, and so forth. Two coders conducted the coding analysis of in total 35 days (seven from each of the five periods), which amounted to 658 elements. Looking at the same season throughout, we would, we hoped, be more likely to identify changes.
To balance the coding of journalistic headlines with considerations of the online form of journalism we considered the webpage as a text in itself (see Bødker and Brügger, 2018). First, we manually analysed the visual appearance of the webpage throughout the entire period from 2005 to 2022 (n = 1722) to identify and diachronically track the appearance and placement of its main elements and shifts in graphic layout. The underlying premise of the visual analysis is that it is possible to deduct the level of importance given to specific elements by their physical placements on the webpage. For example, following conventional insights from print journalism, the elements that are evaluated as most important are placed at the top of the page and follow a hierarchical order of importance, where the least important objects are placed at the end of the page. In addition, as the reading direction is from left to right in North Germanic languages, the elements appearing on the left side of the page (followed by the centre of the page) are similarly taken as signifying comparably greater importance, than elements appearing on the right. As we shall see in the following findings section, we operationalise visual elements as visual objects that transcends mere text by containing a graphic and/or interactive element to complement the text. For example, these include national and local maps, satellite images, weather reports as illustrated with conventional icons, audience’s personal weather photos, photo galleries and interactive elements. Lastly, to obtain a greater insight into the qualities of the news content, we conducted a qualitative coding analysis of all textual heading of elements published within the five constructed news weeks (n = 658). The coding categories will be outlined in detail in the following analytical section.
Analysis – From meteorological data and information to leisurely commonality
The overall analysis starts with a mainly descriptive outlining in chronological order of the five constructed periods. Each of the five sections will contain an overall description and interpretation of the visual elements on the webpage, insights from the coding of weather stories and include discussions of selected illustrative examples of news headlines. Each of the sections will concentrate on significant changes from the preceding period(s) and this means that the first section is somewhat longer as it sets the stage, part of which is to describe the coding categories. After going through the five periods, we will address some overall themes that emerge by drawing together some of the key developments of the entire period from 2005 to 2022.
Period 1 (2005–2010): Data and education
Visually, the webpage is characterised by, first, a grey and blue-toned background, which later shifts towards green, blue and white, which has the effect of a clearer demarcation of the various elements appearing on the page. 2 The page is divided into three columns. On the centre of the page, appears the traditional 24-hour weather forecast written in neutral language and as such comes across as more or less unmodified and/or unprocessed meteorological data. The visualisation accompanying the meteorological description consists of the conventional weather forecast symbols, for example, suns and clouds. Next to the forecast is a black-and-white satellite photo. In the middle column there is an element with audience photos and, also, a bit later, an element encouraging audiences to send in ‘your weather photo’. This feature not only remains but gains in significance through the whole studied period. In terms of audience relations, it is notable that in 2007, a new element appeared on the page: a mobile weather studio in the form of a boat to bring the forecast out among its audiences (in the summer period). At the bottom right-hand corner is a link to the latest TV forecast.
At the top of the page located to the left of the main forecast, is a photo of the TV weather hosts (‘the weather’s hosts’) – a feature that stays throughout the entire period studied, that is until 2022. This clearly is carried over from and relates to the TV forecasts where – as quoted above – hosts ‘construct a relationship of closeness and familiarity [and] identify themselves with the audience’. Online, the perceived need for a personified mediator between forecast data and culturally relevant content is signalled by the consistently prominent positioning of the hosts whose online presence frames the content. What we are presented with at this page is thus not only forecast idata but information that has been filtered by someone you (presumably) can relate to.
With regards to the characteristics of the textual content on the page, our constructed news week for the first period comprised 138 web elements – where an element is a ‘coherent semiotic entity’ (Brügger, 2018: 33). Out of this content, the category Forecast Data was the most dominant with 43%. In the Forecasts Data category, textual elements are distinguished by neutral data description and contain linguistic expressions from the field of meteorology. Furthermore, textual elements in the Forecasts Data category are information heavy compared to the other categories we coded for and this category contains content with relatively raw and unprocessed data in relation to which the journalistic mediation is minimal.
The second and third most dominant categories are Forecast Data and Purpose and Educational Content each taking up 21% of all content respectively. Forecast Data and Purpose comprise content where data is contextualised by reference to its utility for the audience. For example, pollen levels are contextualised by reference to allergies, UV levels by reference to presence in the sun, snow by reference to skiing and wind by reference to sailing. Textual elements in this category contain specific data that is journalistically mediated, however minimally, according to an instrumental logic with the audiences as the subject. User Generated Content takes up 8% of all content and comprises predominantly audiences’ weather photos.
Textual elements in the Educational Content category are characterised by recurring teaching formats such as ‘The Hurricane School’, ‘The Weather School’, ‘Valuable Knowledge regarding the Weather’ and ‘Quiz’. This represents a top-down approach to communicating information on the weather. Rather than more Socratic or pedagogical ways of teaching, information is mediated as a one-way exercise instead of a hermeneutical one. It is assumed that audiences need to be ‘taken (back) to school, and the authoritative expertise of the DR weather department is underscored. Lifestyle themes were present in 5% of all the content. Textual elements in this category commonly take the form of weather stories that are centred around leisurely activities oftentimes involving nature, for example, swimming at the beach, hiking in the autumn-coloured forest or cooking recipes for specific seasons.
In the sample for the first period, Data as News was the smallest category with only 2%. This category comprises news stories based on specific data, most commonly forecasts data as news. Moreover, this category also includes articles that refer to unusual weather phenomena, and so captures the process by which data gets interpreted in relation to past weather The elements containing Weather-related Stories are broader nature and weather stories that are not tied to the immediate forecast data, and Climate Change issues were absent from the sample we analysed from the first period. Taken together, the categories Lifestyle, Data as News, Weather-related News and Climate Change, represents the content where the journalistic mediation of data are most prominent.
Period 2 (2011–2012): The rise of the user
In the second period most of the characteristics from period 1 are carried over. A significant difference is, however, a shift to a much whiter and more airy layout. And the element with the weather host is now renamed ‘your weather host right now and contains only one person (with name and photo) at the top left part of the webpage’. This is a marked shift from period 1 where there were four hosts depicted and as such underlines the importance given to personifying the weather for the audience. Another small feature worth mentioning here is the appearance of a small Facebook logo and the name ‘The DR weather right under the photo of the host’. As such, this clearly ties the general entity – the DR weather – to a specific person.
Another noticeable change is the move to only one column at the top of which there are various data-based forecast visualisations and this is also where the latest televised forecast is now embedded. Below this there are three stories spanning the width of the page and below that there are four columns with smaller stories. As such, the webpage has become clearer in its editorial choices. It is also notable that the picture accompanying the main story (July 4, 2011) is a user photo. This signals the growing space and prominence given to user impressions of the weather, which is drawn on to help situate the broader weather patterns in contexts probably more relatable to a majority of the audience.
If we turn to the content of the elements, the most significant change is that the category Forecast Data has been relegated to second place and now comprises 24% of the total of 89 coded elements. The most prominent category is now User-generated Content, which makes up 33%. Another significant development is that the categories Forecast Data and Purpose and Educational Content, which each took up 21% of the content in period 1 now have been reduced to approximately 10% together. Simultaneously with this, we see a rise in the category Data as News, which goes from 2% to 8%. An example of such a news story from Saturday August 18, 2012, is headlined ‘Humid Heat with the subtitle ‘While 30 degrees only will be reached in some locations, it is still going to be a hot experience’. The most significant change, however, as already hinted at, is the significant rise in user generated content, which goes from 8% to 33% of the coded elements.
Period 3 (2013–2015): Weather (data) as news
The most visible change in this period is that the background has turned green and you can vaguely see leaves of grass in the background. As a new feature you can, at the very top of the page, click on one of five Danish cities and below you can search for your own location. This is a step towards meeting more specific needs of the audience. What is also worth noting here is that the weather host (now in the top right-hand corner) has a quote listed. On Wednesday August 28 in 2013 the host, Søren Jacobsen, is quoted as saying: ‘Did you also find it foggy this morning?’ Then take a look at the video that I have attached. If this (also) tickles your curiosity, then watch Our Weather this evening. Following the point made above about personifying the weather through the host and making a connection with the audience it is here noticeable that the host is now directly addressing the audience – ‘Did you also find it foggy this morning’.
The significant rise in period 2 of Data as News is now followed up visually as there are now two main columns where the one on the left is called ‘Nyheder fra Vejret’ (‘News from the Weather’) and the right one contains forecasts and where the top one is an embedded TV forecast (underneath which there is a satellite photo headlined ‘see where it is raining’). News about or from the weather here gains as status on par with forecasts. Another thing worth mentioning is that viewer/user photos are still given more space.
Looking at the distribution of the elements on the page the most significant change is that the category Data as News goes from 8% to 23%. This brings it, as also visible in the layout as pointed out above, on the same level as the category Forecast Data.
Period 4 (2016–2018): Even more news
In terms of the visual aspects the page returns to a mainly white layout with some stories highlighted in a light green. It is also worth noting that the embedded video disappears and below the five cities at the top there is now a major story, which is an overall interpretation of the forecast, for example, ‘Fog is teasing on Sunday’. The page now looks more like a news page with a clear prioritising of content and a major news story at the top. In addition to that there is now a link to an Instagram account. While there are still audience photos, the visibility of the audience is now also seen in the appearance of a list of most shared articles. Another feature that is worth mentioning here is the appearance of links to what the Danish Meteorological Institute and the main public service competitor TV2 write about the weather.
Given these additions the number of elements on the page goes from 93 in period 3 to 116 in period 4. In terms of the distribution of the elements, the most significant changes are still another rise in the category of Data as News, which goes from 23% to 29% and that the category Weather-related Stories goes from 9% to 15%. There is thus now a significant number of stories about living with the weather and nature, stories which are not tied to any immediate meteorological data. It is also noticeable that educational content now only makes up 3%.
Period 5 (2019–2022): A broadening of contexts: Nature and climate
In period 5, the last period in our study, the webpage now consists of one column, where a major weather news story is placed at the top centre. Right below appears a box where users can type in their geographical location, to receive a specified weather report completely tailored to their individual context. Conventional visual symbols of the weather forecast, for example, skies and suns, are again appearing on the webpage but in more detailed and visually appealing versions compared to the first period. Also, in contrast to the first period, these weather symbols are not placed in the top centre of the page but are parsed in-between weather news stories. Audiences’ personal weather photos are still a dominant feature and are now gathered in photo galleries appearing in several places on the page. The page still contains a list of links to TV2, and the Danish Meteorological Institute with the headline ‘What Others Write About the Weather’. Around 2020–2021, we identified a shift towards a more conventional news page layout and a significant increase in the number of weather news stories, most of which are accompanied by photos.
Our artificial news week from the last period comprised 201 elements. Out of those, Forecast Data comprised 8%, which is a significant decrease compared to the first period of our sample where 43% of the elements were classified as such. Similarly, the category Forecast Data and Purpose has decreased to 3% compared to 21% in the first period. So has the previously rather prominent category, Educational Content with 3% compared to the initial 21%. Importantly, the categories that represent the most significant degree of journalistic mediation have all increased during the last period: Data as News (from 2% to 37%), Weather-related Stories (from 0% to 27%), Lifestyle (from 5% to 9%) and Climate-change Issues (from 0% to 2%). The Figure 1 below gives an overview of the development.
The distribution of elements in the five periods.
Discussion
Based on the overall developments identified through the changes in the layout and content over the five periods, we here draw together some of the observations. What is noticeable is – as the chapter subtitle suggests – that we see a broad development away from data and education; in fact, the webpage in 2023 no longer has any visualised data at all. This constructs a culturally embedded commonality that is somewhat different from the one identified by Morley (2000) with which we began this article. What we see is a much more deliberate construction, which has many of the characteristics identified by Sifianou and Tzanne (1996) quoted above. A key element here is what they call the ‘element of “weather personification”, which “adds to the personalised character of. . .” weather forecasts where weather is reported not in scientific terms but as views of an individual’ (Sifianou and Tzanne, 1996: 363). What is different, however, is that the online weather functions as a meeting ground between a personalised host and partly individualised users in the form of audience photos. One could say that this somehow attempts to mirror everyday colloquial exchanges about the weather and as such becomes a phatic and ritual exchange rather than mainly one of information.
While many of the elements of the page are undergirded by meteorological data, these are pushed into the background. Related to this it is notable that this happens despite a growing amount of available data as well the growing possibilities of visualisation online. As such, the page can only to some degree be described as data journalism. What Berland wrote in 1994 is thus only partly true: ‘Today, weather forecasts offer us a ritualised celebration of the technological sophistication of satellite surveillance, space-based photography, and computerised image data analysis’ (Berland, 1994: 111). While ‘the increasing Improvements in satellite measurements and observational data’ are contributing ‘to forecasting accuracy for even the most isolated of locations’ (Zion, 2017: 106), what seems celebrated on the page is rather the possibility of engaging with the weather mainly as a backdrop to leisure activities – for example, barbecuing on the terrace or going to the beach. While this is linked to the fact that fewer and fewer among the audience are interested in the weather for professional reasons – and those who are find more specific information elsewhere – it is also a celebration of affluence and thus also a celebration of a national identity where not all but many can afford to think about the weather as mainly linked to pleasure.
This movement away from forecasting and data in the direction of pleasure and news makes the page ill-suited for the communication of issues linked to climate change and there are in fact very few stories linked to that. One example is a story from June 12, 2022, headlined ‘Temperatures in Southern Europe break records and climate change is blamed’ and on the same day, but closer to home, ‘Climate changes can mean raised water levels more than 100 times a year’. Such stories do, however, sit somewhat uneasily with the broader colloquial frame and the page, following from the points made above, thus faces the same challenge with regard to communicating about climate change that each of us face in our more day-to-day encounters with others about the weather.
Yet, if we follow Callison (2014: 1), there is a growing need to develop ‘vernaculars through which we understand and articulate our worlds and the nuanced and pluralistic understandings of climate change’ (emphasis in the original) and she finds evidence of that through her case studies. On the weather page we have studied there is, however, no signs of such a vernacular. Given the developments described above it rather seems that what has developed is a vernacular of affluent leisure shielding itself from uncomfortable facts about climate change. This is, of course, also related to the fact that Denmark, via its geographical location and wealth, remains somewhat removed from severe impacts of climate change. Regardless, however, it does not seem that the online public service weather has found a way, as Berland (1994: 112) writes, to ‘write on to the [weather] maps’ that we are ‘responsible for its well-being’; the term ‘our weather’ often used on DR’s page does, however, not convey a sense of responsibility but is rather meant to signal a convivial sense of being in this weather together – something we share over and above other differences. What Zion (2017: 94) calls the ‘weather/climate disconnect’ is thus not overcome as this requires connecting ‘impersonal statistics’ with ‘personal experience[s]’.
It is noticeable that the televised weather forecast on DR has been much more advanced with regard to situating the present weather in relation to long-term trends; and a recent article in The Guardian interviewing a range of TV weather hosts in the UK also points in that direction. ‘There’s now’, the article says, a greater ‘responsibility to educate’ and one of the hosts adds: ‘We have to communicate climate in almost every forecast because it is our responsibility to give people perspective’ (Wollaston, 2023: n.p). What we have seen on the Danish broadcaster’s page is, however, a steady movement away from education and not towards it; and this is indeed a pity since some knowledge – general and/or experienced – is important for engaging in exchanges about climate change as one of the respondents to a Danish study says: ‘I don’t know much about the subject so wouldn’t speak about it’ (Bruhn Jensen, 2017: 447). While there has been some criticism of approaching climate change journalism in relation to a perceived lack of knowledge (e.g. Gunster, 2017) the weather may have an advantage as it is very linked to experience, identity and community. Why the communication on TV in this regard is rather different is difficult to say. One aspect could be that the TV forecast follows directly from the news and thus establishes a more serious frame, which is linked to the flow of TV where elements are tied together temporarily. This is not the case for content online, which people seek out more directly and for a variety of reasons. Another reason linked to this is that TV for most of the studied period appears as the primary medium, which is marked by embedded videos of the TV forecasts.
Conclusion
This article constitutes a first step towards understanding the development of online weather reporting in the first two decades of the 21st century; and contrary to expectations, we saw little evidence of a more towards a more data-heavy reporting and thus no evolution of an online language attempting to breach the weather/climate disconnect. There can be several reasons for this, one of them being a wish – or need – to personalise the weather in order to create connections to the day-to-day experiences of the audience in order to attract users to the site. Another aspect is the broader development of the various elements making up the web sphere of the weather, which means that the need for up-to-date and local weather data increasingly may be served by apps, which may have pushed the online weather page more towards culture and community.
As far as the comparison with TV is concerned, this would require an analysis of the characteristics of weather reporting on DR TV and thus also the videos embedded on the weather page (which was not possible to do in the archived material we worked with). In relation to such analyses, it would be very relevant to conduct interviews with the team behind the weather page. It is also important to note here that the data used could be expanded in various ways to include more days and seasons as well as by moving beyond headlines to conduct both quantitative and qualitative readings of the full articles as well as of the embedded videos (as mentioned above). Regardless, however, it is our hope that this article can be a productive steppingstone for further studies of linkages of weather and climate across an increasingly diverse complex media landscape. This could also lead to broader comparative studies of how various (national) cultures construct relations to the weather.
