There is strong evidence that food systems are important contributors to climate change(Reference Steinfeld, Gerber, Wassenaar, Castel, Rosales and de Haan1–Reference McMichael, Powles, Butler and Uauy4). Actions to reduce food-related greenhouse gas emissions can have co-benefits for nutrition. The present paper describes how food and agriculture affect climate change and examines how this information has been communicated to the US general public via one medium: daily newspapers. While these messages may be greater or different in other media or other countries, US newspapers are important both because of newspapers’ policy impact in the USA and globally because of this country’s large carbon footprint. The discussion explores why food systems were under-represented in US news coverage compared with other contributors to climate change, and considers future opportunities and frames for communication. The paper does not address the expected major climate change impacts on food systems(5–Reference Tubiello, Soussana and Howden8) or health(Reference McMichael, Friel, Nyong and Corvalan9–Reference McMichael, Woodruff and Hales14), although these are a key part of what makes climate change so relevant to the public health nutrition field.
Food systems and climate change
We define food systems as everything and everyone involved in producing, processing, distributing, consuming and disposing of food, including policies, processes, products, individuals, businesses and governments. Taking a systems approach to food enables consideration of relationships between the above components and between food and other aspects of agriculture and society. A systems approach requires interdisciplinary thought and recognition that changes in one part of the system have ramifications elsewhere(Reference Sobal, Khan and Bisogni15–Reference von Bertalanffy17). There is overlap and interchange among food systems serving localities, regions and nations, and between alternative and mainstream food systems. Herein, unless otherwise stated, we are referring to food systems at the US national level.
The Intergovernmental Panel on Climate Change indicated in 2007 that agriculture contributed 13·5 % of world anthropogenic greenhouse gas emissions (CO2 equivalent), and that forestry including deforestation contributed 17·4 %(18). Much of the latter contribution is associated with food. Even in the USA, with its major industry and transportation emissions, agriculture is estimated to contribute 10 % of greenhouse gas emissions(19). While food greenhouse gas emissions go well beyond agriculture, no broad US-based estimates are known to exist.
Among food categories, food animal production is by far the top climate change contributor. The FAO estimates that livestock production alone accounts for 18 % of world anthropogenic greenhouse gas emissions – a greater contribution than transportation(Reference Steinfeld, Gerber, Wassenaar, Castel, Rosales and de Haan1).
In food systems, the greenhouse gases CH4 (especially from animal enteric fermentation and manure) and N2O (especially from fertilizers) are top anthropogenic contributors to climate change, although CO2 (especially from manufacturing, transportation and supplying energy) also plays an important role. Beyond emissions, food systems also affect the extent to which greenhouse gases are released into the atmosphere or trapped in soil and plants. A top contributor to climate change is deforestation for food production, particularly in tropical forests where some US food and animal feed imports originate. Conversely, practices like setting aside crop lands for conservation and no-till agriculture can sequester substantial amounts of CO2(Reference Steinfeld, Gerber, Wassenaar, Castel, Rosales and de Haan1, Reference Smith, Martino and Cai3, Reference McMichael, Powles, Butler and Uauy4, 19, 20). While ‘food miles’, or the distance food is transported from production to market, have received much attention, analyses suggest that transportation may account for a relatively small (though not insignificant) portion of food system impacts(Reference Weber and Matthews21–Reference Brodt25). Using food crops for fuel also affects emissions; although some evidence suggests costs outweigh benefits(Reference Fargione, Hill, Tilman, Polasky and Hawthorne26–28).
The above-described food system effects on climate change are complex, interactive and dependent on many factors. Regardless, the evidence is clear that the impact is substantial.
Individuals, business and government can reduce food-related greenhouse gas emissions. While much more research and life-cycle analysis is needed to enable prioritizing responses for the US context, some beneficial individual responses include: eating lower on the food chain and particularly reducing beef and dairy consumption; wasting less food; eating foods that are less processed, seasonal, not air-freighted or raised in greenhouses, locally produced, sustainably produced and that have long shelf-lives; avoiding packaging and plastic bags; reducing refrigeration, freezing and shopping trips; and eating less overall. McMichael et al. find that to stabilize livestock-related greenhouse gas emissions at 2005 levels, global meat consumption would need to drop to 90 g/d per person by 2050, based on expected population rises and increasing wealth and meat consumption in developing countries(Reference McMichael, Powles, Butler and Uauy4). To comply in the USA, we calculate that per capita meat consumption would need to drop by nearly two-thirds(29). While these changes can be difficult to make and sustain, cumulatively and via the effects of consumer demand on business, they could impact emissions substantially.
Beyond individuals, the magnitude of effort needed to reverse or minimize catastrophic climate change calls for focused responses from business and government. Such efforts are just beginning in the USA, far behind Europe and the UK. Food industry efforts to reduce emissions include selective and local purchasing, product labelling, reduced packaging, energy efficiency and carbon offsetting(30). Agricultural enterprises have reduced emissions including through changes in animal feeds, soil conservation and no-till farming, reduced fertilizer and pesticide use, energy efficiency, increased local distribution, improved waste management, and on-farm energy generation(Reference Paustian, Antle, Sheehan and Paul2, 30). Government incentives that could support such changes and motivate others include stimulating local and seasonal food economies, reducing governmental subsidies for animal feed grains, raising awareness about meat and dairy climate change consequences, mandating energy efficiency in vehicles, equipment and facilities used by food systems, and requiring labelling and reduced packaging. More broadly, federal and state policy makers could better incorporate food and agriculture into climate change mitigation strategies, and expand research support.
Such actions will be unlikely without public support and consumer engagement on the demand side. The US nutritional public health community can play an important role in communication and can take advantage of ‘co-benefits’ in opportunities to improve nutrition and mitigate climate change through individual and social food choices. To better shape and improve communication efforts, we need to understand more about what the US public currently hears about food and climate change.
Public awareness and concern about climate change
It is established that media coverage plays a key role in shaping which issues are considered important by the public and elected officials(Reference Wallack, Dorfman, Jernigan and Themba31, Reference Smith, Wakefield and Edsall32) and that it can be associated with behaviour change(33). Further, it is established that article framing influences how audiences perceive problems and assign problem-solving responsibility(Reference Wallack, Dorfman, Jernigan and Themba31, Reference Kim and Willis34, Reference Iyengar35). The media analysis field has moved far beyond the idea that these relationships are simple and linear(Reference Kiousis36–Reference Wammes, Breedveld, Looman and Brug39), and similarly, behavioural science has moved far beyond the idea that accurate knowledge or even intention to act necessarily leads to action(Reference Jackson40–Reference Krosnick, Holbrook, Lowe and Visser42). At minimum, however, knowledge is a ‘floor’; the public cannot be expected to take action on a problem until they know about it.
The literature indicates multiple factors driving newspapers’ agendas regarding topics warranting coverage. In addition to ‘objective’ newsworthiness, these include: journalist and editor judgement; issue promotion by interest groups; pressures to present a ‘good story’; advertiser pressures; and coverage in other sources(Reference Wallack, Dorfman, Jernigan and Themba31, Reference Smith37, Reference Clayman and Reisner43–Reference An and Bergen46). Literature on public agenda-setting suggests that issues framed as systemic (government, business) responsibilities may be seen as more important and thus more newsworthy than those framed as individual responsibilities(Reference Kim and Willis34, Reference Stone47, Reference Lawrence48).
The present study analysed US newspaper coverage of climate change and food systems’ contributions to climate change. From among the media sources, newspapers were chosen for their broad readership, policy impact and role as a source of record. The study aimed to:
1. Describe ‘food and climate change’ (FCC) coverage – time trends; newspapers; level of focus on food and agriculture; and type of article – all in the context of ‘climate change’ (CC) coverage.
2. Examine the food system components attributed as affecting climate change in FCC articles.
3. Examine how the FCC articles were framed, including whether the stated or implied responsibility for action lay with individuals, business or government.
The newspaper sample included sixteen of the top twenty US newspapers based on print circulation(49). Newspapers included were (in circulation order): New York Times, Washington Post, Chicago Tribune, Philadelphia Inquirer, Rocky Mountain News (Denver), Houston Chronicle, New York Post, Detroit Free Press, Dallas Morning News, Minneapolis Star Tribune, Boston Globe, Newark Star-Ledger, Atlanta Journal-Constitution, Arizona Republic, Long Island Newsday and San Francisco Chronicle. These papers, together, distributed an average of 10·7 million copies/d in 2006(50), with expected print readership over double this figure(51). (The other four top-twenty newspapers were unavailable via the Access World News database.)
Articles were collected for the 29-month period from September 2005 to January 2008. The ‘initial period’ of the study started three months before the premier of Al Gore’s documentary film ‘An Inconvenient Truth’ at the Sundance Film Festival, and stopped just before May 2007. To update the investigation, searches were repeated for the nine months from May 2007 to January 2008, referred to as the ‘follow-up period’.
We coded articles as ‘climate change’ (CC) if they had ‘global warming’ and/or ‘climate change’ in lead or first paragraphs. Among CC articles, we coded those that had the words ‘food’, ‘farm’ or ‘agriculture’ anywhere in their text as ‘food and climate change’ (FCC). We then excluded articles that in no way referenced food system impacts on climate change, including those focused on how climate change impacts food systems, and extraneous usages such as ‘Farmington, CT’. Articles discussing biofuels were included if they mentioned the impact on food systems. Because of the large number of CC articles, during the initial analysis a random sample was selected for closer reading (23 % after exclusions).
Table 1 describes variables examined. Two coders reviewed the articles. The lead author trained coders and evaluated consistency. Coders reviewed discordances together at the end. The mean κ statistic evaluating initial inter-reviewer concordance was 0·66 for the CC sample analysis (rising to 0·78 after excluding three subjective variables), suggesting good concordance. The κ was 0·51 for the initial FCC analysis, driven down by the low n and number of options on some variables.
*Initial period = September 2005 to April 2007; follow-up period = May 2007 to January 2008.
We performed descriptive analyses primarily involving frequencies and percentages, and, in selected cases, t tests and χ 2 analyses. We did not statistically compare the FCC with the CC articles due to sample overlap and small numbers. Because there were so many CC articles, the 658 article sample during the initial period was used for analyses involving reviewing CC articles directly. Thus, different data aggregations were used: across time (all FCC v. all CC); across newspaper (all FCC v. CC sample); key themes and responsible party (initial FCC and later FCC). Analyses were performed using the software packages STATA version 9 (StataCorp, College Station, TX, USA) and Excel (Microsoft Corp., Redmond, WA, USA).
Level of coverage
As shown in Table 2, the terms, ‘food’, ‘farm’ or ‘agriculture’ were mentioned in 109 (2·4 %) of the 4582 articles with ‘climate change’ or ‘global warming’ in their lead paragraphs from the sixteen selected top newspapers between September 2005 and January 2008. Twenty (0·44 %) were coded as substantially focused on food or agriculture contributions to climate change, forty-four (0·96 %) had short mentions and forty-five (0·98 %) had minimal mentions, as defined in Table 1. There was considerable change across time. In the first six months, a mean of 0·8 % of CC articles mentioned food contributions, while in the final six months this figure rose significantly to 2·9 % (P = 0·01, Satterthwaite’s df = 7·75). In looking at the percentage of articles coded as substantially focused on food and climate change, small numbers motivated extending the comparison to the first fourteen analysis months v. the last fourteen (leaving one month in between). In the first half of the study, one article fit the criteria (0·07 % of the total), whereas nineteen (0·63 %) did so the latter half. Comparing article type coded in the initial part of the analysis, FCC articles were more likely than CC articles to be letters to the editor (23 % v. 9 %) while CC stories were more likely than FCC articles to be news (52 % v. 42 %).
Figure 1 shows articles by month across the full study period. Monthly numbers of FCC articles ranged from zero (October and November 2005) to ten (June 2007), with a mean of 3·8 (sd 2·8). For comparison, CC articles ranged from fifty-seven (November 2005) to 305 (April 2007), with a mean of 158·8 (sd 69·8). The two followed a similar time trend, with the CC articles peaking earlier. (Note that in this and other figures, directly contrasting the numbers is inappropriate because of sample overlap.)
Figure 2 shows the initial period distribution of FCC articles by newspaper, with the same-period CC article sample for comparison. The graph reflects significant variation between newspapers but an overall similar pattern across article types. The number of FCC articles ranged from zero (Dallas Morning News, New York Post) to eleven (New York Times), with a mean of 3·3 (sd 2·9); four newspapers had only one article. In the later time period, the mean increased to 3·6 (sd 3·5), but four newspapers published zero FCC articles (Dallas Morning News, New York Post, Philadelphia Inquirer and Detroit Free Press). The highest number was still eleven (New York Times). One newspaper, the Newark Star-Ledger, substantially increased its coverage, from one to nine articles, across the two time periods.
Per newspaper CC articles ranged from five (New York Post) to 176 (New York Times), with a mean of 41·1 (sd 41·8) in the initial period. During that time, the New York Times ran over 2·3 times the number of CC articles as the Washington Post, the newspaper with the next highest number.
Content of coverage
We examined the news coverage content in terms of the food system components to which articles directly or indirectly attributed climate impacts. The highest percentage (29 %) of articles were general, such as attributing the problem to ‘food’. About 13 % referred to livestock and 12 % each to transportation and land use. Eleven per cent referred to pesticides and fertilizers, with most of these in the initial period referring to pesticides, and in the later period more often referring to fertilizers, which do have a higher impact. Nine per cent were coded as referring to the industrial nature of food production, with many specifying energy use. Finally, 15 % referred to biofuels; in the follow-up, these shifted in tone from generally positive to universally questioning or critical.
Most FCC articles directly or indirectly implied roles for individual, business and/or government in addressing the problem. Of sixty action codes overall in the initial analysis (articles could have more than one code), twenty-seven indicated individual responsibility for action, while twenty-two indicated business and eleven government responsibility. In the follow-up, these numbers were seventeen, twenty-eight and twenty-four (χ 2 = 7·22, P = 0·03), suggesting an increasing focus on the ‘big picture’.
The present analysis identified a substantial gap in US coverage of food systems’ contributions to climate change among the sixteen top newspapers studied, as well as evidence that coverage is increasing slowly. The findings can provide a basis for increased communication on these issues, while the science review in the introduction to the paper can help inform future messages.
In sum, we found that the terms ‘food’, ‘farm’ or ‘agriculture’ were mentioned in 109 (2·4 %) of the 4582 articles with ‘climate change’ or ‘global warming’ in their lead paragraphs between September 2005 and January 2008. Twenty of these (0·4 % of the total) substantially focused on the issue, with increasing coverage across time. Of the FCC articles, 29 % referred to food only in general ways. Based on the review in the introduction, the issues of livestock, land use and fertilizer might be highlighted as among the key food system factors in climate change. Respectively, these were mentioned in 13 %, 12 % and less than 11 % of FCC articles; or, in the context of CC articles, 0·48 %, 0·46 % and less than 0·41 %. Coverage varied by newspaper. The articles increasingly addressed the bigger picture of business and government responsibility compared with tips for individuals.
We suggest five possible reasons why US newspaper coverage of food-related contributions has lagged behind overall climate change coverage.
Food systems’ effects on climate change were relatively late to be highlighted by scientific, governmental and reviewing experts, particularly in the USA. Thus, the new information had to confront the barrier of pools of expert advocates, scientists and journalists communicating based on existing knowledge and expectations. The film ‘An Inconvenient Truth’, which played an important role in education about climate change in the USA, also did not mention food or agricultural contributions. Further, because CO2 is the top greenhouse gas of concern, it is possible that a ‘CO2 bias’ in awareness may have operated, whereby other greenhouse gases such as the CH4 and N2O dominant in agriculture were inadvertently discounted. It is notable that so many CC articles appeared in a single newspaper; these were often written by one journalist. Individual journalist and editor interest is also part of an issue’s idiosyncratic history and can play an important role in coverage.
There is little available life-cycle analysis data to enable quantifying and prioritizing among US food-related contributions to climate change. By contrast, in Europe, numerous research articles and non-governmental reports have described and quantified these contributions, highlighting the problem’s extent and nature (see, for example, the several dozen food life-cycle analysis reports and articles compiled at the British Food and Climate Research Network website, including only two papers about the USA)(Reference Brodt25, 52). This data gap and the further lack of lay-targeted US information on food contributions to climate change more generally make it challenging for reporters to write on the issue.
In US society, food and nutrition concerns are more commonly constructed as individual problems with individual solutions than as social problems with social solutions(Reference Kim and Willis34, Reference Lawrence48). This framing can lead editors to perceive these issues as less newsworthy or as belonging in newspaper sections such as ‘Lifestyle’. In the initial analysis, FCC articles were weighted towards individual frames, but by the later analysis, business-framed articles led, seconded by articles framing the issue in relation to government. This shift could support further increased news coverage and expand the recognized options to address the problem.
4. Advocate interest
Environmental advocacy groups have played an important role as media sources and in raising climate change’s public profile in the USA. In the past, these groups have focused primarily on sectors such as energy and transportation. While US animal welfare groups have campaigned on climate change issues(53), environmental groups are just beginning to address the issue. Environmental advocates may have feared that messages about reducing meat consumption and changing eating behaviour might undermine other climate change efforts by alienating the public, leading to prioritizing behaviours and interventions with relatively less impact, or simply by confusing people.
5. Industry responses
Sectors such as energy are commonly socially constructed as ‘big business’ and with a political- or conflict-oriented frame. Members of these industries may have inadvertently solidified this impression through efforts to foment doubt about climate change science(Reference Boykoff and Boykoff54). Accordingly, climate change advocates may gravitate towards these industries as adversaries, while the media gravitates towards the conflict. Food industries by contrast have generally laid low on climate issues in the USA. As public awareness grows, some have engaged in well-publicized efforts such as purchasing carbon credits, fashioning themselves as partners in reducing greenhouse gases rather than opponents(Reference Deutsch55). It is also possible that ubiquitous supermarket advertising in newspapers might have a subtle impact on news editorial choices(Reference An and Bergen46).
Opportunities to increase newspaper coverage
The present analysis suggests several opportunities to increase attention to food system and food animal production contributions to climate change in the USA. First, the food/climate change issue lends itself readily to some key article schemas newspaper journalists seek: investigative, consumer/health interest, events, conflict, local, telling a story. As public concern about climate change continues to grow and coverage increases, reporters also need new angles, and may thus focus on food. The public’s increasing interest in sustainable and local food production – and recognition of ancillary benefits such as taste and healthfulness – may also lead to coverage opportunities. Journalists are likely to receive more information about food system impacts, as environmental advocacy groups are increasingly recognizing this issue’s importance and animal welfare groups continue their communications campaign.
The American Public Health Association designated climate change as a theme for 2008’s Public Health Week and recognized health co-benefits of addressing food contributions to climate, designating one weekday for this topic(56).
Strengths and limitations
The present study has an unusually large sample of newspapers compared with other media analyses. For example, most other climate change media analyses we reviewed used three to seven newspapers(Reference Carvalho38, Reference Carvalho and Burgess57, Reference Boykoff and Rajan58) and one identified top wire stories in 255 papers(Reference Antilla59). The sampling and analysis were performed using methods consistent with those in the literature(Reference Smith, Cho, Gielen and Vernick60, Reference Durrant, Wakefield, McLeod, Smith and Chapman61), including using two coders, κ analysis and meeting to resolve discordance. Finally, the discussion was grounded in theory and prior literature.
The chief limitation is that it is not possible to determine how well these newspapers reflect all national newspapers, which differ in size and audience, among other factors. (None the less, smaller papers do use the same wire services, and these prominent papers have particular media agenda-setting, social and policy impact.) Coverage may also differ in other sources including television journalism, entertainment and blogs; this is a valuable area for future exploration. The searches may have missed some articles, and some of the coding was necessarily subjective. Finally, the large number of climate change articles and labour intensiveness of coding prevented us from being able to code them all so as to present consistent comparisons throughout.
Food, agriculture and land-use change are recognized by scientists as important contributors to climate change – for example, the latter two account for an estimated third of world greenhouse gas emissions, while livestock alone is estimated to contribute 18 % of anthropogenic greenhouse gas emissions. However, this information was not well disseminated to the US public in the selected major newspapers during the study period. Furthermore, even the rare articles that mention food system contributions often directed reader attention to relatively less pressing aspects of the issue, particularly in the earlier months. Given the US carbon footprint, reaching the US public with accurate information is important.
News coverage of food and climate change is clearly improving. The media follows political leader, advocacy group and expert agendas. As these groups increasingly incorporate food into their discussions of climate change, media coverage may be expected to increase further.
Greater public awareness could lead to consumer demand for food produced with a lighter environmental impact, industry actions including increased use of lower-emission agriculture and food production methods, efforts to change policy including incorporating food and agricultural policy into climate change mitigation strategies, and increased support for research.
The more we know about climate change news coverage, the more effectively we can intervene to improve message accuracy, framing and effectiveness. Public health nutrition has an important role to play. Building on co-benefits with other nutritional messages such as reducing meat consumption, and on the field’s expertise in social, environmental and biological aspects of nutrition(Reference Schwartz, Parker, Glass and Hu13, Reference Lang62, Reference Beauman, Cannon and Elmadfa63), public health nutrition can help close the information gap as well as further elaborating the science. The urgency of making rapid cuts in greenhouse gas emissions means that all avenues for reduction must be promoted. As stated in the Giessen Declaration, ‘Now is the time for the science of nutrition… to meet the challenges and opportunities faced by humankind in the twenty-first century’(64).
Conflict of interest: None declared.
Source of funding: The study was funded by the Johns Hopkins Center for a Livable Future.
Author contributions: R.A.N. initiated the study, conducted the analysis and did much of the writing. I.L.C. retrieved and coded articles and assisted with analysis and writing. K.C.S. contributed to the study design and provided guidance on media analysis methods. All authors critically reviewed and edited drafts.
Acknowledgements: The authors thank Anne Rosenthal for assistance with coding, Polly Walker, Shawn McKenzie, Robert S. Lawrence and Danielle Nierenberg for helpful feedback, and Genevieve Alelis for the title idea.