the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Jun 2025
Leveraging Social Media for Geoscience Communication: Insights from the British Geological Survey's Multi-Hazard and Resilience Campaigns
Abstract. Social media offers a unique avenue for scientific communication; however, it remains underutilised by many scientific organisations. This study examines the social media strategy of the British Geological Survey (BGS), the UK’s leading geoscience organisation, to assess its effectiveness in engaging the public with research on Multi-Hazard and Resilience. We investigate two key research questions; 1. how effectively does BGS engage the public through its social media efforts, and 2. what challenges does BGS face in using these social media platforms to enhance public understanding?
Scientific organisations often rely on the deficit model of communication, characterised by a one-way transfer of knowledge. Yet, emerging studies suggest that a dialogue-based approach, tailored to different social media platforms and formats, may foster better public engagement. This paper provides a framework for assessing social media activity that can be applied to scientific organisations worldwide.
To address research question 1, we conduct content and sentiment analysis on BGS social media posts – including X, Facebook, LinkedIn, Instagram, YouTube, and BlueSky – from May 2023 to March 2024. A systematic codebook is developed to categorise descriptive and interpretive variables for any social media output. To answer research question 2, we conduct semi-structured interviews with five BGS employees who manage departmental social media accounts to understand their attitudes towards social media engagement.
Our findings suggest several actionable strategies, such as streamlining communication across platforms, maximising the reach of ‘Multi-Hazard and Resilience’ themes, increasing video content output, and better incorporating public feedback. Although focused on BGS, our mixed-methods approach and methodology offer a valuable template for other scientific organisations seeking to enhance their online presence and science communication efforts. This study highlights BGS’s successful establishment of a multi-platform online presence, showcasing a range of content formats that effectively engage audiences.
Competing interests: The authors declare that there are no competing financial interests. However, it should be noted one of the authors for this paper is the chief executive editor of Geoscience Communication.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(1160 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-1963', Anonymous Referee #1, 07 Nov 2025
-
AC1: 'Reply on RC1', Eleanor Dunn, 14 Dec 2025
Overall
An insightful piece of work looking at how effectively does BGS engage the public through its social media efforts and associated challenges. This topic is particularly important given the increasing frequency and severity of natural hazards, which heightens the need for organisations like the BGS to communicate clearly, responsively, and accessibly with the public. However, if the goal is to establish a framework for understanding how social media can be a meaningful tool for scientific organisations, the paper would benefit from a stronger theoretical grounding, clearer methodological detail, and more robust engagement with existing literature to ensure its wider applicability and impact.
- The authors wish to thank reviewer 1 for their engagement with the text and for their valuable insights and recommendations.
- The revised manuscript has been re-written to align with the recommendations of the reviewer. More context and clarity has been provided when discussing the literature and methodology,
Some of the paragraphs are noticeably short and could be merged to improve readability and flow
- Some of the shorter paragraphs have been merged, most noticeably, subsections of section 2.2 have been combined to remove extra detail and strengthen the flow of the text.
It would have been valuable to see more empirical analysis. There is quite a lot already done with social media analysis.
- A clear limit exists to what further empirical analysis can be extracted from the dataset. The paper already reports descriptive patterns, cross-platform comparisons, sentiment distributions, and Chi-square tests. These analyses are the standard and appropriate tests for this type of content-analysis dataset. Any additional statistical tests would add volume without improving interpretive strength or validity.
Introduction
Provides a clear and succinct overview of the British Geological Survey, successfully establishing why its social media activity is a relevant subject of study. The purpose of the paper is also stated clearly.
- The authors wish to thank reviewer 1 for this comment.
Page 4 – “This paper is designed to act as a framework that can be applied to national and international scientific organisations to allow them to understand and assess their social media activity going forwards.” – remove the ‘s’ from ‘forwards’
- This correction has been made [Section 1, paragraph 8]
What are the previous works in this area, what were their outputs and how do they reinforce the need for this work. I don’t see strong support from the literature to back up the need for this work, which I do think is important; it just hasn’t been fully established here.
- The authors agree that the original introduction did not sufficiently situate the study within the existing research. In response, we have added a paragraph [Section 1, paragraph 5] summarising the relevant literature on government and scientific organisations’ use of social media for crisis communication, public engagement, and science communication. This paragraph highlights key findings including the prevalence of one-way communication, the challenges researchers face in online engagement, and the uneven effectiveness of institutional communication during hazards, and clearly identifies the gap that no previous study has examined BGS’ social media activity or the organisational barriers researchers encounter. These additions strengthen the justification for the present study while keeping the introduction concise.
Social media and science communication
The section effectively outlines the three dominant models of science communication, but the argument would be strengthened by a deeper examination of how these models diverge in practice and what that means for the study’s analytical approach. For the most part, it is lacking theoretical depth.
- We have strengthened the theoretical framing by adding a concise explanation of how the deficit, dialogue, and participation models diverge in practise and how these distinctions translate into observable communication behaviours [Section 2, paragraph 2]. This addition clarifies how each model predicts different interaction patterns on social media, and it explicitly links these theoretical differences to our analytical research, explaining why one-way engagement metrics are an appropriate first step for situating BGS activity along the dissemination-participation spectrum.
1 Geoscience communication: Not too sure why this section is need as it is repeated with the exception of the journal reference, for which I don’t see the immediate relevance.
- Following an almost identical comment from reviewer 2, this section has been deleted.
2 Social media and scientific organisations: Not see the connection to the broader literature, and lacks critical depth.
- We have strengthened Section 2.2 by adding a concise framing paragraph [Section 2.2, paragraph 9] that explicitly connects scientific organisations’ social media practises to the broader literature on institutional communication models, resource constraints, and the gap between theoretical ideals of engagement and real-world practise. This addition clarifies how previous research on governmental and scientific agencies informs our analytical approach and explains why examining BGS’s online behaviour is theoretically relevant. The revised section now situates BGS within established patterns identified in the literature and enhances the conceptual depth of the argument.
3 Social media and BGS: Good description overall.
- The authors wish to thank reviewer 1 for this comment.
Methodology
Details are missing; for example, how were the data collected from the different platforms -APIs, manually, and how were they classified into groups -e.g., sentiments. If specific computer programming packages were used in this case, stating what they are would be helping to understand what was done and some of the associate limitations.
- The authors have included more detail here [Section 3.1, paragraph 3], including information about how the social media posts were downloaded.
‘50 random social media posts are selected and analysed at the beginning of the data collection period and then analysed a second time, four weeks later, to ensure correct content and sentiment identification, before the remaining posts were categorised’ – This is assuming that there is little to no change in content properties and sentiment after the following 4 weeks. I’m not sure why this assumption is made or if this wording itself is not what the authors have intended.
- We did not assume stability in content properties or sentiment over the four-week period. The second coding round was used solely as an internal reliability check on our classification decisions, which is a standard approach when independent coders are not available. The purpose was to verify that our theme and sentiment labels were applied consistently, not to track changes in engagement metrics or post characteristics. We have revised the wording to make this explicit [Section 3.1, paragraph 3] and to avoid any implication that temporal stability of the posts formed part of the analytical logic.
‘To identify the correct research theme for content analysis, the social media posts are viewed at face value – as in it is assumed the viewer does not click on any attached information to understand the posts’ theme further.’ – Please state explicitly what this means. An example would help. Also, if this process is being done manually, this would be a lot of data to process for one person to process, or even two for that matter. Would there not be potential for misclassification by researchers at some point. What corrective measures were used to account for this?
- We now state explicitly that ‘face value’ coding means classifying each post only on what a typical user sees without opening links or attached materials [Section 3.1, paragraph 4, for example coding an X post from its visible text rather than the full report it links to. Because manual coding carries a risk of misclassification, we introduced a corrective step by recoding a random sample of 50 posts four weeks later to check that theme and sentiment labels were applied consistently. The revised wording makes clear that this second round was a reliability check rather than an assumption about temporal stability.
Were spam/bot like posts removed? How? This is not clear.
- Spam comments were kept in the analyses as a way to observe if any research discipline received an unusually high number of spam interactions in comparison to other research themes. This is now clarified [Section 3.1, paragraph 3].
Results
Figure 2 – the explanation is fair but the figure looks noisy to the point where I question whether it adds value being here. Another option may be to use stacked figures for each platform, or may be a table.
- Figure 2 has now been edited to address comments made by both reviewer 1 and 2. The colour scheme has been improved and the legend has increased in size.
Page 18 – ‘In Table 4, all p-values are must smaller than 5’ – Change ‘0.5’ to ‘0.05’
- This correction has been made [Section 4.1.2, paragraph 2].
Generally informative and great seeing the use of Chi-test.
- The authors wish to thank reviewer 1 for this comment.
Missed opportunity
- The authors noted that the phrase ‘missed opportunity’ is not linked to any specific section or issue, so it is not possible to address it in a meaningful or constructive way. The manuscript has been strengthened in all areas where clear guidance was provided, and without further detail this comment does not warrant additional changes.
Discussion and Conclusion
Overall good logical suggestions based on the results.
- The authors wish to thank reviewer 1 for this comment.
Good links to existing works.
- The authors wish to thank reviewer 1 for this comment.
Citation: https://doi.org/10.5194/egusphere-2025-1963-AC1
-
AC1: 'Reply on RC1', Eleanor Dunn, 14 Dec 2025
-
RC2: 'Comment on egusphere-2025-1963', Anonymous Referee #2, 18 Nov 2025
Please find attached comments to support a major revision of this manuscript.
-
AC2: 'Reply on RC2', Eleanor Dunn, 14 Dec 2025
This paper explores how social media can be used in geoscience communication, particularly asking how effectively the British Geological Survey engage the public on multi-hazard / resilience focused work, and the challenges that BGS staff encounter when using social media platforms to enhance public understanding of this theme. The paper draws on existing methodologies to deliver these research objectives, combining analysis of social media post engagement with (a small number of) semi-structured interviews. Below are a set of major comments that I believe would significantly strengthen the paper, and minor comments that would improve overall readability and clarity.
- The authors wish to thank reviewer 2 for their engagement with the text and for their valuable insights and recommendations. The revised manuscript has been updated to align with the major and minor recommendations of the reviewer.
Major Comments
Independence (e.g., Line 59 – 62)
It feels odd that the paper suggests that it is an independent perspective on BGS’ social media output and gives an objective overview of social media activity and its effectiveness because “the lead author is not a BGS employee" when another author (involved in the analysis and authorship, as set out in the author contributions) is a BGS employee. The independence of a paper is not solely shaped by the lead author’s distance from the organisation it seeks to study. While not suggesting the authors (even if having connections with BGS) are unable to look objectively at the data – I would challenge the idea that the paper gives ‘an independent perspective’ and would strongly encourage more transparency about the measures taken to ensure objectivity and any implications of positionality of authors that may shape the recommendations and conclusions being made.
- We have reworded this paragraph to remove any implication that independence rests on the lead author’s employment status [Section 1, paragraph 2]. The revised text explains transparently that while one co-author is a BGS employee, they did not conduct the analysis or draft the manuscript, and their role was limited to advising on data access and discussing analytical options. This makes the division of labour clear without overstating independence, and the Author Contributions section has been updated accordingly.
Structure/Superfluous Paragraphs (e.g., Section 2.1)
Overall, the paper would benefit from a careful proof-read and edit, to improve the flow of the narrative. There are some paragraphs that feel superfluous to the core focus of the paper (e.g., Section 2.1 - Why is this needed, giving a history of the journal you are seeking to publish in?), and some of the general background on X ownership and features (e.g., Lines 134-141). There may be other places where you can tighten up the flow and content also.
- We have undertaken a full proof-read and removed or merged paragraphs that did not advance the argument. Section 2.1 has been folded into the preceding material, and the platform descriptions in Section 2.3 have been streamlined into a single integrated section. These changes improve narrative flow without altering the substance of the analysis.
Contribution beyond BGS (and engagement with wider literature)
There are some sections where more engagement with the published literature would help the reader to understand how this work builds on others, what it is adding in terms of originality, and an evidenced explanation of how this work is of relevance beyond BGS. For example, you make very brief reference to your work being useful beyond BGS – but it’s not clear how (i.e., what you have added to the methods and approaches set out in the wider literature).
- We have strengthened the discussion [Section 5, paragraph 1] by adding concise paragraphs that situate the study within the wider literature and explain its relevance beyond BGS. These additions clarify how our approach extends existing methods and how the findings can inform similar organisational analyses in other scientific settings.
Data and Rationale for Assertions/Conclusions (e.g., Line 428-430)
Various recommendations and assertions are made, but the data/evidence to support these is not always set out. Some specific examples are given below, but I think the article would benefit from a careful check throughout to ensure there is a clear link between each recommendation and the data (results), reflecting on the broader literature where relevant.
- Alongside each recommendation, the authors have supplied a clear link to the results to demonstrate how the authors came to each conclusion [Section 5.1.1, paragraph 1; section 5.1.2, paragraph 1; section 5.1.3, paragraph 1], with references to various literature.
Need for nuanced reflections
Different social media platforms have different audiences, and the article is lacking any significant engagement with what typical audiences may be and how that may impact on the way that posts are engaged with on different social media. For example, there may be substantially more engagement with BGS posts on Linked In by the wider professional geoscience community than on Facebook. I don’t know if this is the case or if data exists to explore this – but I was surprised there wasn’t more nuanced reflection on audiences in Sections 2, 4, and 5.
- Within Section 2.2 [Section 2.2, paragraph 9], the authors have now included a paragraph which details the different demographics of each platform and how this could influence the results. The authors have followed this up in Section 5 [Section 5.4, paragraph 3], providing a link to how these differing demographics could have influenced the results.
Minor Comments/Suggestions
Line 58 (and elsewhere). Here you note that the accounts are run on a ‘voluntary’ basis. It would be helpful to know what this means in practice, and have it defined. It may not be explicitly in their job description, but is something vaguer about communicating their science to diverse audiences described (for example)? Do people note this as an example of their communication / operational support in promotion applications / annual reviews, in which case it goes beyond a voluntary task?
- BGS researchers do not receive any financial compensation for being involved in running social media accounts, however it is encouraged throughout the organisation. The authors do not feel comfortable stating that any involvement in running social media accounts may assist with promotion/bonuses as this may differ depending on job role/seniority. The authors have added an additional sentence to reflect this information [Section 1, paragraph 1].
Line 79-80. You note that the interviews (five) are used “to understand what barriers individuals may may face when using social media to promote their research and how these barriers can be tackled”. This currently reads as if the perspectives of five individuals with tightly constrained selection criteria are being used to address a much broader question about barriers to promoting research. While those five individuals will have valid perspectives, it feels like that question needs to be explored with a bigger and more diverse audience to harness informative and more generally applicable recommendations.
- This sentence has been changed [Section 1, paragraph 6] to reflect the fact that responses only represent a possible subset of all opinions within BGS.
Line 106. Insert the word ‘about’ between communication and research.
- This sentence has been removed to address another comment by reviewer 2.
Line 107-108. What is the evidence for the claim in this sentence (that there is an increasing number of scientists who want to facilitate a valuable exchange....)?
- This paragraph has now been removed to address another comment by reviewer 2.
Line 112-116. A claim is made about governmental organisations - this is then supported with papers from ~ a decade ago. Are there any more recent studies that explore whether this is still an issue. If not, can you add critical reflection on whether these sources still mirror the reality of 2025.
- The authors have added in three more recent references to support the argument [Section 2.1, paragraph 2].
Line 157-158. I suggest adjusting the language from ‘Facebook is... leading the way’ to ‘Facebook is host to substantive amounts of fake news...’. The former implies an intended objective, and I don’t think this is substantiated in your manuscript.
- This correction has been made [Section 2.2, paragraph 4].
Line 200-204. In your current structure, this sits under Section 2.3.5, but these lines are not specific to that section. Delete or merge with Section 3.
- This correction has been made [Section 2.2, paragraph 2].
Table 1. Make clear in the caption that the accounts run by BGS Comms are just for the Multi-Hazard and Resilience Theme.
- This correction has been made [Table 1, caption].
Line 234. Is a laughing emoji or reaction always positive? Are there not situations where one could be laughing out of sarcasm or ridicule, or because they found something bizarre?
- A clarifying sentence has been added to address this comment [Section 3.1, paragraph 3]. The true intention of a laughing emoji is impossible to verify on its own unless you know the user’s intent and therefore we took reactions at face value. However, we were able to interpret potential sarcasm/nuance when analysing comments based on contextual evidence.
Figure 1. This figure includes an acronym (LANDSLIP) of a very specific project which jars with the rest of the figure, which is more generic. Perhaps note in the caption that the arrows and annotations are exemplars (i.e., this is not a comprehensive overview of links).
- This correction has been made [Figure 1, caption].
Line 273. Please include the ethical review reference number / code (if given) and clarify in the text that ethical approval was obtained prior to commencing the work.
- This correction has been made [Section 3.2, paragraph 2].
Table 3. For question 7, here and throughout, some nuanced discussion about what is and isn’t in the job description may be helpful. See the earlier comment on Line 58.
- The authors have included an additional line [Section 1, paragraph 1] to reflect the fact that running the social media account is not included within the job description.
Figure 2. The resolution could be improved, legend made larger, lines made less feint (very hard to see for the yellows and oranges, particularly).
- Figure 2 has been improved by increasing the size of the legend and increasing the width and opacity of the lines to improve visibility. The authors have also changed the colour map to utilise more vibrant colours which account for colour blindness.
Line 300. I’m not sure if bias is an appropriate word here, it implies a deliberate or inadvertent favouring – but there may be deliberate and fair reasons why certain departments are better represented over the relatively short time frame of the study.
- This correction has been made [Section 4.1.1, paragraph 3].
Line 303. I don’t think ‘popular area’ is an appropriate phrase here – can you think again about this sentence.
- This correction has been made [Section 4.1.1, paragraph 3].
Line 304. Change “% on” to “% of posts on”.
- This correction has been made [Section 4.1.1, paragraph 3].
Figure 4. It is not easy for a reader to engage with this Figure. The legend is too small, the numbers on chart itself are too small and difficult to read.
- The authors have changed the orientation of Figure 4 and increased the font size on the labels, legend, and titles.
Figure 5. It is not easy for a reader to engage with this Figure. The blue / teal colours are all very similar. In lines 309-310 you note that the likes are normalised, but it’s not mentioned in the figure caption (nor clear if and how this has been done).
- The authors have changed the colour map and verified the colours with the app ‘Color Oracle’ to ensure the Figure adheres to colour blindness tests. The figure caption has also been edited to explain how normalisation is calculated.
Line 337-338. Here a recommendation is made to increase posts, and to streamline content best suited to different platforms. This seems out of place in the results, not well evidenced (how did you get from result to recommendation) and again, what role does audience have on the nature of engagement?
- The authors have removed this sentence from the results section [Section 4.1.2, paragraph 1].
Line 344. Remove word ‘must’.
- This correction has been made [Section 4.1.2, paragraph 2].
Table 5. What is NSFW?
- This correction has been made [Table 5]
Line 374. Unclear if you mean the interviewee’s enjoyment managing the account or the enjoyment of users following the account?
- The authors have clarified this sentence [Section 4.2.2, paragraph 1] to reflect the interviewees enjoying X less, rather than the users following the account.
Line 389. Again, do you think the suggestions should be dotted in the results and discussion, or grouped together in the latter? With this example, please add a little more detail… “it is suggested [by who?] that support [what support?] be provided… will enable the continued success of the account” [what is the evidence for ongoing success – given you use the word continued – and that the suggestion will enable this?].
- This sentence has been re-written to provide clarification [Section 4.2.3, paragraph 3]. However, the authors have left the sentence in the results section as the suggestion was made during the interviews and therefore falls under the results section rather than the discussion section.
Line 396. Change ‘many’ to ‘may’
- This correction has been made [Section 4.2.4, paragraph 1]
Line 402-403. I’m not sure this ‘clear interest’ is evidenced by the previous bullet points (it seems to contrast with bullet #2, Line 399).
- This sentence has been clarified [Section 4.2.4, paragraph 3] to reflect the fact that an interest was shown by 4/5 interviewees rather than every interviewee.
Section 5.1.1. (a) It’s not clear why this is being posed as a problem and needing corrective action. (b) On the recommendation of having an ‘on call’ roll - there is huge diversity in the group described, and it is not clear how a landslide expert could offer advice/support on questions/comments regarding geomagnetism, for example.
- To address this, the authors have added some clarifying sentences [Section 5.1.1, paragraph 1] which provide details on why unequal post distributions is a problem and why it warrants corrective action. The authors have also clarified that the recommendation refers to having one on-call person per Multi-Hazard and Resilience group.
Line 423. I’m unclear what ‘underutilised market’ in this context means – a market for what purpose?
- This sentence has been clarified [Section 5.1.2, paragraph 1].
Line 427. Change ‘regarding’ to ‘compared to’ (or something similar).
- This correction has been made [Section 5.1.2, paragraph 1].
Line 428-430. What is the evidence for this assertion (re. BlueSky)?
- The authors have added two references to support this assertion [Section 5.1.2, paragraph 1].
Line 429. Change ‘consisting’ to ‘consistent’.
- This correction has been made [Section 5.1.2, paragraph 1].
Line 436. What is the evidence for this assertion (re. videos)?
- This sentence has been clarified [Section 5.1.3, paragraph 1].
Line 449. Check use of word ‘present’ – this doesn’t read well.
- This correction has been made [Section 5.1.4, paragraph 1].
Line 485. How many questions/comments were unanswered? Out of how many? The evidence for including this suggestion could be strengthened.
- Thank you for this comment. A systematic audit of response rates to social media questions was outside the scope of this study, and no quantitative count of unanswered comments was undertaken. We have revised the text [Section 5.2.5, paragraph 1] to remove the implication of measured non-response and to frame the recommendation as a precautionary practice informed by existing literature rather than a quantified finding from the dataset.
Line 494-495. Update to ‘and thus, there is a chance for discrepancies to exist between…’ and add (for example) at the end of the sentence.
- This correction has been made [Section 5.3, paragraph 1].
Lines 505-507. This sentence is not clear and may need to be rewritten. Can you also clarify if the beneficiaries of these insights are those within or beyond BGS, and if beyond the evidence for that / limitations.
- The authors have re-written this sentence [Section 5.3, paragraph 3]. A clarifying sentence has been added at the end of 5.4 to discuss who the beneficiaries of the research are.
Line 517. This sentence seems to imply that you are presenting a methodology that is new – and can be applied elsewhere. Have these approaches not been applied before? Can you be explicit in this section about what you are offering that is novel in terms of methodology, vs what existing methods you are applying.
- The authors have added a clarifying sentence [Section 5.4, paragraph 3].
Line 535. Change ‘where’ to ‘were’.
- This correction has been made [Section 6, paragraph 3].
Ethical Statement. Can you add in the ethical approval reference number if you have it.
- Thank you for this comment. The British Geological Survey does not operate a formal human research ethics committee and therefore no ethical approval reference number exists. Ethical approval was granted internally by BGS prior to data collection. The ethical statement has been revised to clarify this.
Checks throughout:
- Both X (formerly Twitter) and just Twitter are used – please be consistent: The authors have removed the references to Twitter (excluding a sentence mentioning its origins).
- Capitalisation of some terms (Likes, Shares, Comments) is not aways consistently applied: The consistency has been improved.
- This is a multi-authored piece, so use of ‘I’ needs to be removed (e.g., line 362, maybe elsewhere): The use of ‘I’ has been removed.
- The piece needs an additional proof-read prior to resubmission, with some incorrect phrasing / missing words used (examples given above): An additional proof-read has been carried out.
Citation: https://doi.org/10.5194/egusphere-2025-1963-AC2
-
AC2: 'Reply on RC2', Eleanor Dunn, 14 Dec 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 814 | 138 | 38 | 990 | 45 | 56 |
- HTML: 814
- PDF: 138
- XML: 38
- Total: 990
- BibTeX: 45
- EndNote: 56
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Overall:
1 Introduction
2 Social media and science communication
Methodology
Results
Discussion and Conclusion