Technological Forecasting & Social Change 200 (2024) 123197

Contents lists available at ScienceDirect

Technological Forecasting & Social Change

journal homepage: www.elsevier.com/locate/techfore

Social network platforms and climate change in China: Evidence

from TikTok
Yunpeng Sun a, *, Ruoya Jia a, Asif Razzaq b, c, Qun Bao d
a
School of Economics, Tianjin University of Commerce, Tianjin, China
b
Adnan Kassar School of Business, Lebanese American University, Beirut, Lebanon
c
Research Division, CAREC Institute Urumqi, China
d
School of Economics, Nankai University, Tianjin, China

A R T I C L E I N F O A B S T R A C T

Keywords: The actions and policies enacted by today’s youth hold profound implications for future generations, under­
Climate change news scoring their pivotal role in advocating for climate issues. Younger cohorts exhibit heightened concern regarding
TikTok climate change and are highly visible on social media platforms. Hence, this study aims to delineate the portrayal
Centrality
of climate-related news and disasters on TikTok. The research draws upon fifty TikTok accounts focused on
PageRank method
Superedge rank method
climate-related content. Employing social network analysis, PageRank, and Superedge Rank methodologies, this
Digital economy investigation evaluates how TikTok users—called TikTokers—address climate change and its impact on social
media. The assessment scrutinizes climate-related news, disasters, and the resultant networks to gauge the in­
fluence wielded by social influencers in disseminating messages. Results reveal that among the four key enti­
ties—internet influencers, government, scientists, and producers—internet influencers exert the most substantial
influence on climate change news dissemination on TikTok, while the government plays an influential role in
climate disasters. Like other social media platforms, TikTok is a valuable arena for gauging public sentiment on
critical health concerns like global warming. Nonetheless, ensuring the reliability and depth of messages shared
on TikTok necessitates the presence of credible experts who can deliver comprehensive and scientifically sound
information within the platform’s time constraints.

1. Introduction Llaguno et al., 2022). Transnational decision-making and worldwide

discussion are essential for climate change management. As a result,
While public interest in climate has always been prevalent, the recent social media platforms like Facebook, TikTok, and Instagram have
surge in scientific evidence regarding climate change has elevated it to a emerged as a necessary avenue for citizens to engage with researchers
significant societal concern (Corbett and Savarimuthu, 2022; Wei et al., and policymakers (Cody et al., 2015; Nisbet et al., 2009).
2021). This heightened attention is reflected in various indicators, such More than 88 % of Chinese use at least one social media site
as its prominence in Gallup’s rankings, online search trends, and the compared to 70 % of Americans (Pew Research Centre, 2021), demon­
prevalence of climate discussions across social media platforms, all strating a tremendous increase in popularity over the previous 15 years.
underscoring a growing apprehension about the adverse effects of These tools are accessible to people of all ages, although some de­
climate change (Rosenthal, 2022). To construct a discourse on climate mographics tend to use them more frequently than others. Therefore,
change, society values reliable information sources. The literature shows many social media sites appeal to a wide range of demographics. Five
that “prestige newspapers,” news aggregators or organizations, and hundred ninety-seven million people in China are using social media.
opinion leaders, which can include celebrities, are the most trusted Most young adults (19–30) in China use at least one kind of social media.
sources of information regarding environmental topics (Lee and The­ It includes popular sites like TikTok, YouTube, and Instagram. Because
okary, 2021). Nevertheless, social networks serve as a platform for the choices and policies of today will have enduring effects on the
exchanging and disseminating opinions, ideas, and dialogues, trans­ climate and sustainability, today’s kids play a vital part in climate
forming the 21st-century agora (Ballestar and Sainz, 2020; Martin- advocacy. Having safe places to voice discontent might prevent young

* Corresponding author.
E-mail addresses: tjwade3@126.com (Y. Sun), asifrazzaq@yahoo.com (A. Razzaq), baoqun@nankai.edu.cn (Q. Bao).

https://doi.org/10.1016/j.techfore.2023.123197
Received 1 September 2023; Received in revised form 25 December 2023; Accepted 26 December 2023
Available online 8 January 2024
0040-1625/© 2024 Elsevier Inc. All rights reserved.
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

people from isolating themselves, becoming passive, or lashing out at information (Agryzkov et al., 2012; Giustolisi et al., 2020; Wiedermann
other marginalized groups, including economic migrants (O’Brien et al., et al., 2013), as network nodes can demonstrate importance (such as
2018). Accessible climate change information available to youth from based on intrinsic relevance or external information) that is independent
numerous sources, including social media, may increase future activism of network topology.
and change (Han and Ahn, 2020). Other successful youth-led cam­ The ten-year-old Adapted PageRank Algorithm (APA) model
paigns, such as the “March for Our Lives,” have shown the power of (Agryzkov et al., 2012) is gaining popularity as a centrality measure. It
social media’s unrestricted reach for gathering information, raising modifies the PageRank centrality model by ranking the intersections
awareness, and collecting support for causes (Bandura and Cherry, (nodes) of a city’s street network (Brin and Page, 1998). The primary
2020). goal of the APA model is to score the meetings based on the presence or
TikTok is a rapidly growing platform, with over 2 billion app absence of nodes (such as stores, restaurants, or tourist attractions) and
downloads worldwide. Research into using new social media sites like the structure of the underlying nodes of the network. Like the PageRank
TikTok for distributing public health information has recently begun model, the APA model can be understood as a random walker who
(Basch et al., 2021). Certain public health agencies and groups have systematically traverses the node grid; the longer the walker spends at a
turned to social media and social media influencers to raise public particular intersection, the greater its value. One major flaw in the APA
awareness of crucial health issues. For instance, when asked what model is that the random walker has an equal chance of teleporting to
problem the world faces most, those aged 18 to 25 from 22 nations every corner in the street network, regardless of how far away it is. It
named climate change (Barbiroglio, 2019). Young people are starting to comes from the PageRank model. To reconcile this seeming inconsis­
worry more and more about climate change. Therefore, this research tency, Agryzkov et al. (2016) tweaked the APA model to make the
aims to characterize climate-change-related news and climate-change random walker more likely to hop to a nearby intersection. While their
disasters due to TikTokers in China. method is an improvement, more work is required to make the APA
The rapid development of computing has resulted in the emergence model user-friendly for networks and a complete solution to the tele­
of the network as the fourth type of media, joining print, radio, and portation problem. As a result, we suggest a centrality model that refines
television. The Network Age has arrived. The Internet is becoming an the APA model by making the random walker more likely to hop to an
essential medium for the propagation of public opinion. Internet users adjacent intersection rather than a distant one. To achieve this goal, we
can share their thoughts and ideas freely, leading to rapid social change take advantage of recent developments in non-local random walks,
thanks to the Internet’s transmission technology. It is inescapable that which allow a random walker to hop to any node in the network (not just
certain undesired things found online will harm society and the envi­ adjacent nodes) with probabilities that depend on the distance sepa­
ronment. Still, online public opinions are closely related to the various rating the nodes (Estrada et al., 2018; Riascos & Mateos, 2012 and
paradoxes and sensitive themes throughout China’s Social Trans­ 2021).
formation Period. The environmental and climatic crises have made Sheffi introduced the term “supernetwork” in 1985 (Sheffi, 1985).
TikTok a vital platform for spreading awareness (Cody et al., 2015). Few University of Massachusetts professor Nagurney defined supernetworks
studies, however, have attempted to characterize and assess the influ­ in 2002 as networks that exist on top of and beyond existing networks
ence of environmental leaders in these systems. There has been no (Nagurney, 2005). These networks have many characteristics, including
insinuation that the tools and techniques used to measure influencer layers, levels, attributes, and erratic degrees of coordination and
impact (such as the amount of reels or follows). congestion. An example of a supernetwork would be how its constituent
Existing research has mostly ignored intangibles like followers’ networks interact and influence one another. Its framework has made
perceptions of an influencer’s credibility and the quality of the infor­ studying how different networks affect one another easier. Utilizing a
mation they provide. However, the literature gives “environmental supernetwork’s many features and benefits can bring it on par with the
influencers” (a subset of “digital activists”) very little attention in terms average internet user’s knowledge, psychology, and viewpoint. It allows
of analyzing their function, traits, and techniques in comparison to other for a more accurate depiction of the growth and development of online
types of influencers. The stability and security of society are at imminent public opinion. It builds a 4-layer supernetwork of internet users, in­
risk if proper oversight and guidance are not in place. Any Internet user formation, psychology, and viewpoint to classify the online public
who can operate a computer, communicate their opinions in a specific opinion leader, and investigate the function mechanism. At the same
style, and participate in the issue addressed in a particular method may time, it may explore the environmental information that has the most
become the online opinion leader in the process of public opinion significant impact on internet users and the underlying psychological
transmission online. Lazarsfeld et al. (1944) developed the term drive that shapes public opinion.
“opinion leaders” in the 1940s. Opinion leaders have much sway over Here is how the rest of the paper is structured. In the second piece,
others and use their voices to help the general public grasp complex we’ll talk about how researchers in the past have tried to pinpoint key
societal issues. Opinion leaders quickly stir strong reactions and opinion formers. In the third section, we’ll review the fundamentals of
considerable attention when significant events occur at home and the centrality measures, page rank, and supernetwork model for
abroad. Identifying online opinion leaders is vital since doing so is a analyzing online public opinion. We also propose an algorithm based on
necessary first step towards influencing public opinion online. the model and call it Superedge-Rank. We then introduce the factors that
Networks help conceptualize many physical, biological, social, and influence the algorithm, such as the amount of information shared, the
information science systems. Therefore, the scientific study of networks likelihood of change between personality types, and the similarity of
is an important topic that may inform nearly every other scientific and keywords representing different points of view. Ultimately, we will
technological discipline (Newman, 2018). To depict a system in the propose a system for locating influential people in the digital sphere by
actual world, mathematicians often use the term “network,” which is employing the SuperedgeRank algorithm in a supernetwork. In the final
synonymous with “graph.” Centrality measures are among the most piece, we’ll examine how this algorithm was used at a recent show in
popular and extensively used metrics for understanding and evaluating China and the US.
networks because of their ability to pinpoint the most pivotal nodes in
the structure. Several centrality measures, including degree, eigen­ 2. Literature review
vector, Katz, PageRank, betweenness, and proximity, have been pro­
posed since numerous ways exist to characterize significance. At least 150,000 people die yearly because of climate change (World
Conventional centrality measurements typically consider either the Health Organization, 2021). Most climate scientists agree that
local or global network topology. There has been rising interest in increasing levels of greenhouse gases are raising global temperatures.
expanding traditional centrality metrics to encompass node-specific Climate experts warn that climate change and its health hazards will

2
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Fig. 1. Chinese climate change news and climate disaster from 1st October 2022 to 31st December 2022.

worsen without action (NASA, 2019). The effects of human activities Twitter; therefore, it was picked as the hashtag to analyze. Every video’s
such as industrialization and urbanization, which have contributed total views, comments, and favorites were logged. The presence of
significantly to global temperatures, may be permanent. More enormous certain standardized features in the videos was also tracked. Basch et al.
wildfires, more prolonged droughts, different weather patterns, rising (2022) concluded that popular opinion on public health issues like
sea levels, higher flood risks, and melting ice caps are only some of the global climate change could be better understood with the help of social
climate disruptions that have increased due to climate change (NASA, media platforms like TikTok.
2019). Wildfires destroyed an average of seven million acres annually Many members of today’s internet culture get their news not from
between 2000 and 2020. The annual acreage burned during this time the mainstream media but from influential voices (their followers) in the
was 3.3 million (Congressional Research Service, 2021). The ice in the field. The ability of opinion leaders to influence their followers because
Arctic is decreasing by 13.1 % every decade. Because of this, sea levels of their superior position, leadership, and social prestige is crucial to
have increased by an average of 3.3 mm annually since 1993 (NASA, developing communities and improving group performance (Li et al.,
2019). 2019). Identifying opinion leaders is a topic of study in several academic
The effects of rising global temperatures vary from region to region. disciplines, including those about the stock market, politics, public
One study looking at the impact of global warming on the Great Lakes opinion on the Internet, and online reputation management.
region found that longer growing seasons for crops also meant longer Context-specificity is a crucial finding from catastrophe research on
pollen and allergy seasons. Constant human influences disrupt the nat­ ICT and social media. As mentioned above, research findings on social
ural state of things, including the health hazards associated with climate media trust and government use are often contradictory (Takahashi
change (Butler, 2018; Walter et al., 2022). A healthy earth is crucial to et al., 2015; White and Fu, 2012). In the Philippines, the local govern­
health, peace, equality, and prosperity (Pew Research Centre, 2021), ment did not use social media (ibid). Still, Hurricane Sandy case studies
and pollution of the air, groundwater, and soil, as well as catastrophic show that local government authorities and traditional media were the
climate disturbances, harm local communities and businesses, including most critical nodes in social networks. China’s top-down crisis man­
agriculture and infrastructure (Ebi et al., 2006; Patz et al., 2008). The agement approach renders social media’s function in information
dissemination of knowledge about climate change to educate the public dissemination unclear.
is of the utmost importance. China’s environmental risk management, including disaster plan­
Basch et al. (2022) proposes that the actions and policies of today’s ning, is top-down. China’s top-down, reactive, Integrated Coastal Man­
youth can have far-reaching effects on future generations, making them agement policy for natural catastrophes lacks public engagement, policy
crucial to climate advocacy. Younger people are increasingly worried integration, and resources. Therefore, China’s disasters involve social
about the impacts of climate change. One hundred climate-related media. After the 2010 Yushu earthquake, citizens utilized Sino-Weibo, a
TikTok videos in English were analyzed for the study. At the time of Chinese Twitter-like microblog, to express sentiments and criticize
the research, the topic of climate change received the most mentions on official answers (He et al., 2016).

3
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

China has been seeking to increase public participation in environ­ individuals, to actively look for reliable and authoritative sources of
mental policy decisions. The 2002 Environmental Impact Assessment information on contemporary climate trends, scientific research, and
(EIA) Law demands public engagement. The 2015 Environmental Pro­ policy developments. Consequently, individuals responsible for creating
tection Law has an extensive chapter on environmental information and producing content are taking advantage of this heightened interest
transparency and participation (He et al., 2017; Zhang et al., 2016). The by providing educational and up-to-date climate-related news infor­
Chinese government has invested in online government services; how­ mation to keep their viewers or readers adequately informed.
ever, online environmental data is still biased and inaccurate (Brombal, Simultaneously, “disasters” serve as a comprehensive representation
2017; Mol, 2008). More Chinese residents, especially in cities, use of the tangible outcomes of climate change in the world. In a period
WeChat, Sina Blogs, and Sino Weibo to protest and report environmental characterized by a notable increase in occurrences of severe weather
risks and pollution (Hsu et al., 2020). Sometimes with government phenomena such as wildfires, hurricanes, and floods, these observable
agencies. The Chinese government’s “Urban Black and Odorous Water consequences need urgent consideration and evoke a profound response
Information Platform” uses WeChat to monitor water pollution (Hsu from observers. Disasters offer an unmediated perspective on the
et al., 2020). It tracks and pressures local government implementation. destructive impact that climate change may have on societies, compel­
Social media protests against government-supported industrial projects, ling observers to actively participate, demonstrate empathy, and
such as the Xiamen paraxylene (PX) factory protest, occurred in 2007 mobilize efforts towards transformative action. By employing personal
(Chin-Fu, 2013). Social media has fueled demonstrations against the tales and on-the-ground reporting, the disaster-related videos on TikTok
Liulitin trash incinerator. Thus, China is “a society in transition, in which can establish a connection between abstract deliberations concerning
public (environmental) activism is emerging and the relationship among climate matters and the pressing imperative for prompt action. The
various public and private stakeholders is increasingly tense” in envi­ combination of “Climate change news” and “disasters” functions as
ronmental policy-making (He et al., 2017). effective mechanisms for cultivating climate consciousness, stimulating
Distributing knowledge via social media and fostering citizen active participation, and driving constructive transformation on this
participation has not inevitably opened ICT. Social media and the data prominent social media platform. Social network analysis is given in
society are surveillance tools. The Chinese government monitors and appendix (see Appendix B.2).
controls social behaviors through digital data. Google and Facebook are
censored (Feng and Guo, 2013), while Chinese social media platforms
3.1. The algorithm of page rank
have active censorship (Zeng et al., 2017). Chinese internet users know
that the government censors and filters material for public control. Rural
The PageRank algorithm was developed in the 1970s by Pinski and
and urban environments may differ. In a recent survey, Beijingers valued
Narin (1976) and popularized in the 1990s to rank web pages by their
ICT gears (including social media) to increase public involvement in
‘popularity.’ It was reorganized as a search–engine ranking algorithm
environmental sustainability. Still, they recognized that its impact on
that prioritized pages with many incoming web hyperlinks or a small
decision-making had been limited. White and Fu (2012) found that rural
number of incoming associations from highly graded pages. Notably, the
Sichuan residents, after the 2008 earthquake, distrusted social media as
algorithm gives a more excellent score to highly connected pages and
a source of rumors (e.g., kidnappings) and trusted official information
pages that are allied by prominent websites. The model underlying the
from TV news stations.
algorithm’s web application assumes a web surfer who follows links
From our research of the relevant literature, we know that many
between web pages, becomes bored after a series of movements and
different hypotheses have been formulated to explain how to locate
lands on a random page. Therefore, the PageRank of a given page is
thought leaders by employing PageRank-based methods. Still, there
proportional to the likelihood that a random surfer will arrive on that
hasn’t been much study into the effectiveness of using supernetwork
page. Thus, the model can be viewed as a Markov process in which the
analysis to track down influential people in the digital space. For
states are web pages, and the connections between web pages determine
instance, Social Network Analysis zeroes focused on the online com­
the transition probabilities. The similarity between the PageRank
munity leaders (the people who participate and talk online) in forming a
calculation and the derivation of the Markov stationary distribution is,
consensus. It hasn’t considered how internet users react to new infor­
therefore, not remarkable. Specifically, the PageRank-fixing equation
mation, how people’s motivations and emotions shape what they write
reads π as:
online, etc. To effectively identify the online opinion leaders of public
events and study their involvement in the occurrence and evolution of 1− d
π= ωπ
1 + d̃ (1)
events, it is not sufficient to focus attention just on the reacting behav­ N
iors of the internet users. We must also clarify the six elements of when,
where, who, what, why, and how. Additionally, we found limited evi­ where 1 is the unitary N-dimensional vector and where the elements ω̃ ij
{ }N
dence that analyzed the association between opinions and climatic of the matrix ω̃ are given by ω = ωij i,j=1 , the weight matrix of the
conditions using Page Rank and Superedge Rank methods. Considering network observed above. The quantity d is the ‘damping factor’. It is
the above, we coupled both these methods in our study to better un­ connected to the probability of exiting a page and touching an arbitrary
derstand how to identify influential voices in the digital sphere. website. The damping factor, along with the first term on the RHS of Eq.
(1), is incorporated to certify an evolution when landing on a page
3. Materials, methods, and results without out-going links to reserve the lack of any point of the process
and to guarantee the convergence of π to a unique stationary density
Fig. 1 shows two essential topics in this period: Climate change news (Masuda et al., 2017). The PageRank is typically figured iteratively, with
and disasters. The selected account is the top 50th TikTok search ac­ an initial guess π(1) that gets rationalized by applying Eq. (4) above as:
count in this period. The details are shown in the table in the appendix.
There are four attributes: 15 Internet Influencers, 14 Government, 12 1− d
π(n+1) = ωπ(n) , π(n)n→∞
1 + d̃ → π (2)
Scientists, and 9 Producers. N
The predominant focus on “Climate change news” and “disasters” in The N-dimensional vector resulting from this algorithm represents
Fig. 1 indicates the shifting dynamics of climate discussions on TikTok, the likelihood of viewing any TikTok account i = 1,...... N, or the
where these subjects have gained significant attention due to many ‘popularity’ of that account. Our work focuses on metrics like audience
factors. The phrase “Climate change news” signifies the increasing need growth, video plays, profile visits, likes, comments, and shares of
for access to information and updates on climate change developments. climate change-related content. In this context, we may infer the net­
There is a growing trend among TikTok users, particularly younger work’s collective interest in every piece of climate-related news. Unlike

4
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 1
The Fifty Subjects PageRank Ranking.
Account Type Pagerank (after 50 iterations)

LW Internet influencers 0.050082

Wind Aid Internet influencers 0.048026
Mr. Xiaobo Internet influencers 0.047796
HN Internet influencers 0.042466
CWE Internet influencers 0.039695
WE Internet influencers 0.033571
CSP Internet influencers 0.033107
JSW Government 0.03129
SDL Internet influencers 0.02846
MV Government 0.028445
DISR Internet influencers 0.026531
Miss Guo Internet influencers 0.025482
CW Government 0.025178
NWI Government 0.023283
CGN Government 0.021151
BJW Government 0.020968
Gao Han Internet influencers 0.020866
CPN Government 0.020819
SBJ Government 0.019338
TTN Internet influencers 0.019164
UNEP Internet influencers 0.018707
HNE Government 0.017383
CMA Government 0.017363
BB Internet influencers 0.016964
LH Internet influencers 0.016352 Fig. 2. The Network of Fifty Subjects PageRank.
CCTV News Government 0.015974
WTSF Scientist 0.015959
TK Scientist 0.015875 often utilize stories from their lives to make the topic more approachable
UNCF Producer 0.015453 and advocate for sustainable behaviors. Working with other groups and
CMB Producer 0.014655
initiatives helps spread the word and raises the profile. However,
BJR Government 0.013887
MEE Government 0.013844
influencers must be responsible and knowledgeable enough to distribute
ENF Producer 0.013063 only the facts. Influencers can inform and motivate their audiences to
CT Producer 0.012928 combat climate change.
PS Scientist 0.012904 Online influencers may help spread different stories and points of
CBCDF Producer 0.012435
view about climate change. They have the potential to elevate the voices
BJPSD Producer 0.012428
WWF Producer 0.012125 of underrepresented groups and draw attention to the disproportionate
CAAS Scientist 0.012108 effect of climate change on the most disadvantaged people and com­
AHE Government 0.011549 munities. By sharing these personal accounts, influencers may draw
CS Scientist 0.011549
attention to climate change-related social justice and equality concerns.
STM Scientist 0.011503
SSS Scientist 0.011395
In addition, influencers may motivate their audience to act by increasing
GP Producer 0.011157 engagement and participation. They may mobilize their audience to
SRR Scientist 0.010609 support environmental activities by promoting climate-related cam­
ICS Scientist 0.01036 paigns, encouraging people to engage in climate strikes, and striking
CSC Scientist 0.009677
themselves. By working together, we can raise consciousness about
HN Producer 0.009336
CME Scientist 0.008412 climate change and generate a feeling of urgency about finding
ST Scientist 0.008329 solutions.
Opinion leaders must check the integrity of the data they present and
assess the credibility of their sources. Inaccurate information or
other methods (for example, Pinski and Narin (1976)), the PageRank “greenwashing” can hurt our ability to combat climate change suc­
algorithm we employ here does not consider the total number of posts cessfully. Influencers who care about their reputation should verify the
each TikToker generates. We went this route since the academic fields accuracy of their material and work with reputable sources in the sci­
we study are pretty small. The size distribution of the accounts under entific community. In conclusion, online influencers significantly impact
consideration is quite close to the mean, except for a handful of small the dissemination of information about climate change because of the
TikTok accounts that (we checked) would be over-rewarded by a awareness they spread, the sustainable habits they promote, the variety
normalized version of the Page Rank algorithm. We opted to use an of narratives they provide, and the mobilization of their followers. Their
unweighted variant of the PageRank algorithm because of this problem. influence may help people learn more about climate change and moti­
The PageRank results dated 31st December 2022 are, once again, vate them to act on an individual and societal level.
interesting. In the top position, we find in Table 1 that seven internet
influencers have higher page rank values. Fig. 2 shows that pink nodes 4. Case study: the COP 27 conference
represent internet influencers against the government, scientists, and
producers. We have used 50 TikTok accounts of internet influencers, COP is an annual international climate conference hosted by the
government, scientists, and producers to check their impact on climate United Nations. COP is an abbreviation for “Conference of the Parties,”
change disasters and news. Internet influencers have a significant which refers to the international treaty known as the United Nations
impact on climatic information. Framework Convention on Climate Change (UNFCCC). The parties to
Influencers may make a big difference regarding online coverage of the treaty have agreed to adopt voluntary measures to prevent
climate change. They can change people’s minds because of the large “dangerous anthropogenic [human-caused] interference with the
number of people they reach and the interest they generate. Influencers climate system.” Countries take turns hosting an annual meeting where

5
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 2 each line signifies a superedge SEi generated by a particular agent ai.
Environmental view, psychological view, and main opinions of the COP 27 Next, the columns characterize the ecological information type em and
conference. the psychological type pj. The columns indicate the agent’s opinion type
Example kn. Only an agent, some information about the surrounding environ­
Environmental 3 E1 The impact of climate change has
ment, a psychological type, and several keywords are included in each
subnetwork under the been becoming important superedge.
news theme
E2 COP 27 claims that climate change
4.2. Superedge degree (LSE)
has turned into a massive crisis
E3 COP 27 agrees to establish a
climate compensation fund In a supernetwork, it may be said that two superedges connected by
Psychological View 8 Psychological subnetwork: The psychological the same node are considered supersedes. According to Wang et al.
tendency can be divided into optimistic, neutral, (2010), Superedge Degree is the total number of other Superedges
and adverse, and the psychological intensity is
determined by the number of keywords with
connected to a certain Superedge through its nodes. For instance, in the
opinion tendency published by the corresponding abovementioned straightforward online public opinion supernetwork
subject. model, LSE1 = 49 (Table 4).
Opinion subnetwork 20 Positive K1: People warmly celebrate the
under the news theme view opening of COP 27.
K2: COP27 is an important 4.3. The attribute of influential degree of information dissemination
conference.
K3: COP 27 can be beneficial to The climate change subnetwork contains all the information that was
climate change. made public about a particular climate change news event during its
K4: The conference shows the
equality between men and women
inception and growth (let’s assume there are N pieces). Each piece of
in climate change action. information corresponds to one information node, ei (1 ≤ i ≤ n). The
K5: All humanity should work influential degree of information distribution, or I(ei), is defined as the
together to protect the earth. degree to which certain types of information are effective. Two indi­
K6: Governments should make a
ces—the breadth of information dissemination R(ei) and the depth of
good preparation for climate
change. information dissemination D(ei)—are used to quantify the influence of
Neutral K7: The effects of COP27 need to information dissemination.
view be deeply observed. The breadth of information dissemination: The scope of information
K8: The UN is an integral part of dissemination is determined by the total frequency of information
climate change.
Negative K9: Women do not affect climate
occurring in the public online opinion supernetwork model and the total
view change. number of superedges, or, more specifically, by the ratio of connected
K10: The COP27 Conference does superedges of ei to the total number of superedges in the model is.
not have effects on climate change.
K11: Climate change is not F(ei )
R(ei ) = (3)
essential for people. N
K12: People can do nothing about
climate change. where ‘N’ is the total number of superedges in this climate change
K13: Climate change has no news supernetwork model, and F(ei) is the number of linked superedges
relationship with humanity. of information ei.
Depth of information dissemination D(ei): This refers to the extent to
which TikTok influencers distribute information and is determined by
government representatives report on progress, establish intermediate
the overall frequency of the information piece/blogs/posts/reels in
objectives, reach agreements to share scientific and technological ad­
superedges and the number of followers in the social subnetwork that is
vancements of global benefit and negotiate policy. The meeting last year
influenced by it.
took place in Glasgow, Scotland. This year’s conference, the 27th since
the UNFCCC treaty was negotiated in 1992, is being convened in Sharm F(ei )/A(ei )
D(ei ) = (4)
El-Sheikh, Egypt. N/Na
What constitutes a dangerous level of global warming is partially a
scientific issue (what impacts are likely to occur with a given class) and where F(ei) denotes the quantity of information ei’s related superedges;
partly a value judgment (how intolerable those impacts would be). In In this climate change news supernetwork model, A(ei) stands for the
this regard, social media users participate via social media. In this paper, number of followers jointly building superedges with knowledge in so­
we are also collecting feedback via TikTok. We tracked the incident from cial subnetworks, N for the total number of superedges, and Na for the
beginning to end, cataloguing not just the several TikTok accounts total number of followers in social subnetworks.
involved but also the underlying environmental subnetwork, psycho­ According to those mentioned above, the depth and breadth of in­
logical types, and opinion network that formed the basis of the climate formation dissemination work together to determine the influential
change news story. Based on this data, we created a supernetwork model degree of information dissemination I(ei). The equation is as follows:
of the public’s sentiments in cyberspace. Throughout the event, agents
F(ei )2 × Na
made 20 evaluations, comprising three pieces of environmental data and I(ei ) = R(ei ) × D(ei ) = (5)
eight psychological categories. After COP 27 in Scotland, there were two (N)2 × A(ei )
peaks in public posting. We identified all agents throughout the event The climate change subnetwork of this well-established, essential
and the corresponding environmental data, psychological categories, climate change news model entirely consists of three information nodes.
and keywords. From these data, an online public opinion supernetwork Table 5 displays the computation results for the degree of information
model was developed (Table 2). diffusion influence.

4.1. Superedges of the COP 27 conference 4.4. Transformation relations among psychological types

Table 3 displays several examples of this model’s superedges, where The netizens that publish material or blogs to engage in the

6
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 3
Superedges oef the COP 27 conference.
Climate change news

Positive Neutral Negative

Sei Ai ei pi k1 k2 k3 k4 k5 k6 k7 k8 k9 k10 k11 k12 k13

Se1 LW e1 1 1 0 1 1 1 0 0 0 1 1 0 1 0
Se2 Wind Aid e2 2 0 1 1 0 1 0 1 0 0 0 1 0 0
Se3 Mr Xiaobo e3 3 1 1 1 1 1 0 0 1 0 0 1 1 0
Se4 HN e1 1 0 1 0 1 1 0 1 1 0 1 0 1 0
Se5 CWE e3 0 0 0 1 1 1 0 1 0 0 0 1 1 1
Se6 WE e2 1 0 1 1 1 0 1 0 1 0 1 1 0 1
Se7 CSP e3 − 1 1 0 1 0 0 1 0 1 1 1 1 0 1
Se8 JSW e2 1 0 1 1 0 1 0 1 1 1 0 0 1 0
Se9 SDL e1 3 1 1 1 0 1 0 0 1 0 0 1 0 0
Se10 MV e1 1 1 1 1 0 0 1 1 1 1 1 0 0 1
Se11 DISR e2 2 1 0 0 1 1 1 1 0 0 1 0 0 1
Se12 Miss Guo e3 − 3 1 0 0 0 0 0 1 1 1 0 1 1 1
Se13 CW e3 0 0 1 1 1 0 1 0 0 1 1 0 1 1
Se14 NWI e1 − 1 1 0 1 0 0 1 1 1 1 1 0 1 1
Se15 CGN e2 2 1 1 1 0 0 1 0 0 0 1 0 1 0
Se16 BJW e3 0 0 1 0 1 0 1 0 0 0 1 0 1 1
Se17 Gao Han e1 0 1 1 0 0 1 1 0 0 1 1 0 1 1
Se18 CPN e2 1 0 0 0 1 0 1 1 1 0 0 0 0 1
Se19 SBJ e3 − 3 0 0 0 1 1 0 0 1 1 1 1 1 1
Se20 TTN e2 0 0 1 1 0 0 1 0 0 0 1 0 1 1
Se21 UNEP e1 1 1 0 0 0 1 0 0 1 0 0 0 0 1
Se22 HNE e1 0 0 0 1 0 1 1 1 1 1 1 0 1 0
Se23 CMA e3 − 1 0 0 0 0 1 1 0 0 1 1 0 0 1
Se24 BB e2 1 0 1 0 1 0 1 1 0 0 0 0 1 1
Se25 LH e1 − 3 0 0 0 0 0 0 0 1 0 1 0 1 1
Se26 CCTV News e2 2 0 1 1 1 1 1 1 0 1 1 0 1 0
Se27 WTSF e1 − 1 0 0 1 0 1 1 0 0 1 0 1 1 1
Se28 TK e3 − 3 0 0 0 0 0 0 1 1 0 0 1 1 1
Se29 UNCF e2 2 1 0 0 1 1 1 0 0 0 0 0 1 1
Se30 CMB e3 − 3 1 0 0 0 0 0 1 0 1 1 0 1 1
Se31 BJR e2 − 1 1 1 0 1 0 0 1 0 1 0 1 1 1
Se32 MEE e1 − 3 0 1 0 0 0 0 0 0 1 1 0 1 1
Se33 ENF e2 1 1 0 1 0 0 1 0 1 0 1 0 1 0
Se34 CT e2 − 1 1 0 0 1 0 1 0 0 1 0 1 1 1
Se35 PS e3 1 1 1 0 1 1 1 0 0 1 1 1 1 0
Se36 CBCDF e3 0 1 1 1 0 1 0 1 1 1 1 1 0 1
Se37 BJPSD e1 3 1 0 1 0 1 1 0 0 0 1 0 0 0
Se38 WWF e3 1 0 1 0 1 0 1 1 1 0 0 1 1 0
Se39 CAAS e2 0 1 1 0 1 0 1 1 0 1 1 1 0 1
Se40 CS e1 3 0 1 0 1 1 1 0 1 0 1 0 0 0
Se41 AHE e1 2 0 1 0 1 1 1 1 0 0 1 0 1 0
Se42 STM e2 3 0 1 1 1 0 1 1 0 0 0 0 0 1
Se43 SSS e1 2 1 1 1 1 1 1 1 0 0 1 1 1 1
Se44 GP e3 4 1 1 1 1 0 1 0 0 0 0 0 1 0
Se45 SRR e1 0 1 1 1 0 0 0 1 1 0 1 0 1 1
Se46 ICS e1 − 1 1 1 0 1 1 0 1 0 1 1 1 1 1
Se47 CSC e3 1 1 1 1 0 1 0 1 0 1 0 1 1 0
Se48 HN e2 0 1 0 1 0 1 1 0 0 1 1 1 0 1
Se49 CME e2 4 1 1 1 1 0 1 1 1 1 0 0 0 0
Se50 ST e1 0 1 0 1 0 0 1 1 1 0 1 0 1 1

conversation at the start and during the event’s evolution are of various superedges. Taking any two superedges at random, if their psychological
psychological kinds. The psychological nodes of the online public nodes have the same tendency and psychological strength, then these
opinion supernetwork model have variable inclination and strength two superedges may be related. A relationship between two superedges
because different psychological types have distinct psychological ten­ is unlikely if their psychological preferences are opposed, and the
dencies. The psychological profiles of the agents can be determined by disparity between their psychological strengths is significant. The for­
reading their internet posts or blogs. The definition of psychological type mula pij describes the relationship between two personality types, pi,
pi is an integer with a range of +1 to _1. Pi’s positive and negative di­ and pj. Here is the formula for calculating it:
rections define psychological propensity, while Pi’s absolute value de­ { ( )/
sign pi × pj ∣pi − pj ∣, pi ∕
= pj
termines psychological strength. pij (6)
1, pi = pj
_ the negative psychological propensity, represented by pi <0, and
psychological strength, represented by|pi|; ( )
_ the positive psychological propensity, represented by pi >0, and wherein sign pi × pj is a sign function. When x ≥ 0,sign(x) = 1; when x
psychological strength, represented by |pi|; < 0, sign(x) = − 1.
_ the neutral psychological propensity, represented by pi = 0, and Among the above-established simple online public opinion super­
psychological strength represented by |pi|. network model, there are a total of 6 psychological types. Let p1 = − 3,
After classifying each node in the psychological subnetwork, we p2 = − 2,p3 = − 1,p4 = 1,p5 = 2, p6 = 3. The calculated correlations
quantify the transformation relations between node types across among the 6 psychological types of this simple model are shown in

7
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 4 calculated the similarity of keywords from two distinct supersedes using
Calculation of superedge degree. the TF-IDF method. Following is the formula for the similarity between
Se1 LW – L(Se1) = SE1 and SE2:
49
∑
m

e1 Se4,Se9,Se10,Se14,SE17,Se21,Se22,Se25,Se27,Se32,
w1j × w2j
(7)
j=1
Se37,SE40,Se41,Se43,Se45,Se46,Se50 Sim(SE1 , SE2 ) = Sim1,2 = COS θ = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
)̅
√( )(
k1 Se3,Se7,Se9~Se12,Se14,Se15,Se17,Se21,Se29~Se31, √ ∑ m ∑m
Se33~Se37,Se39,Se43~Se50
√ w 1j2 2
w 2j
k3 Se2,Se3,Se6~Se10,Se13~Se15,Se20,Se22,Se26,Se27,
j=1 j=1

Se33,Se35~Se37,Se42~Se45,Se47~Se50
k4 Se3~Se6,Se11,Se13,Se16,Se18,Se19,Se24,Se26,Se29, wherein w1j and w2j represent the keyword j’s weight in superedge SE1
Se31,Se34,Se35,Se38~Se44,Se46,Se49
and superedge SE2 , respectively. Using the preceding formula, the
k5 Se2~Se5,Se8,Se9,Se11,Se17,Se19,Se21~Se23,Se26,
Se27,Se29,Se35~Se37,Se40,Se41,Se43,Se46~Se48 calculated similarity of keywords in all superedges in a simple online
k9 Se7,Se8,Se10,Se12~Se14,Se17,Se19,Se22,Se23, opinion supernetwork model is shown in Table 7.
Se26~Se27,Se30~Se32,Se34~Se36,Se39,Se46~Se49
k10 Se4,Se6,Se7,Se10,Se11,Se13~17,Se19,Se20,Se22,Se23,
4.6. Online opinion leader identification
Se25,Se26,Se30,Se32,Se33,Se35~Se37,Se39~Se41,
Se43,Se45,Se46,Se48,Se50
k12 Se3~Se5,Se8,Se12~Se17,Se19,Se20,Se22,Se24~Se35, 4.6.1. Ranking result of SuperedgeRank
Se38,Se41,Se43~Se47,Se50 In the basic online public opinion supernetwork model described pre­
viously, the attributes of superedges are calculated using the MATLAB
programming tool and SuperedgeRank algorithm. The ranking of 17
Table 5 superedges is displayed in Table 8.
Influential Degree of Information Dissemination.
ei F(ei) A(ei) N R(ei) D(ei) I(ei)
4.6.2. Robustness

e1 18 18 50 0.36 0.24 0.0864

e2 17 17 50 0.34 0.16 0.0544 4.6.2.1. Mechanism of opinion leader identification. A single Internet
e3 15 15 50 0.3 0.36 0.108 user might show up in multiple superedges. The overall score of the
superedges associated with an agent in a social subnetwork is found after
all the superedges are ranked and calculated. The average score is then
Table 6 computed by adding each agent’s node superdegree index. The leading
Transformation Relations among Psychological Types. agent is the one with the highest score.
We utilized the SuperedgeRank technique to rank all the superedges.
pij P1 P2 P3 P4 P5 P6
Then, we identified opinion leaders using the online public opinion
P1 0.00 1.00 0.50 0.33 0.25 0.2 supernetwork model of the COP 27 Conference using the MATLAB
P2 1.00 0.00 1.00 0.50 0.33 0.25
programming tool. Table 9 shows the robust of identifying 10 leaders
P3 0.50 1.00 0.00 1.00 0.50 0.33
P4 0.33 0.50 1.00 0.00 1.00 0.5 among the 50 respondents.
p5 0.25 0.33 0.50 1.00 0.00 1.00 Table 9 displays a comprehensive hierarchy of influential individuals
P6 0.2 0.25 0.33 0.5 1.00 0.00 participating in The COP 27 meeting, specifically highlighting their
categorization and related scores of influence. The table contains a wide
range of significant entities, which have been grouped into four distinct
Table 6.
groups: Internet influencers, Government officials, Scientists, and Pro­
ducers. It is worth mentioning that the top 10 opinion leaders demon­
4.5. The similarity of keywords in different superedegs strate comparatively higher influence ratings, which range from 0.758
to 0.629. These scores indicate their substantial influence in shaping
The vector space model (VSM) is frequently employed to determine conversations and spreading information within the framework of The
object correlation coefficients. The TF-IDF method is based on the vector COP 27 conference. These individuals, frequently portrayed as influen­
space model (VSM), which considers the frequency of distinct words in tial figures on the Internet, have significant roles in moulding public
all texts and the higher resolution of this word across texts. This method dialogue and help transmit messages, potentially serve as essential
is extensively employed for calculating text similarity. In this paper, we components of the communication methods employed by The COP 27

Table 7
Similarity of keywords in different superedegs.
Column1 Se1 Se2 Se3 Se4 Se5 Se6 Se7 Se8 Se9 Se10 Se11 Se12 Se13 Se14 Se15

Se1 1.00 0.34 0.67 0.57 0.57 0.40 0.53 0.57 0.46 0.50 0.57 0.43 0.67 0.63 0.62
Se2 0.34 1.00 0.63 0.51 0.68 0.47 0.32 0.68 0.73 0.45 0.34 0.34 0.32 0.30 0.37
Se3 0.67 0.63 1.00 0.67 0.67 0.63 0.50 0.67 0.87 0.47 0.40 0.53 0.50 0.47 0.58
Se4 0.57 0.51 0.67 1.00 0.57 0.53 0.27 0.71 0.46 0.50 0.57 0.43 0.53 0.50 0.46
Se5 0.57 0.68 0.67 0.57 1.00 0.53 0.40 0.57 0.46 0.38 0.57 0.57 0.53 0.50 0.31
Se6 0.40 0.47 0.63 0.53 0.53 1.00 0.75 0.40 0.58 0.71 0.53 0.40 0.75 0.59 0.58
Se7 0.53 0.32 0.50 0.27 0.40 0.75 1.00 0.40 0.58 0.82 0.53 0.67 0.63 0.82 0.58
Se8 0.57 0.68 0.67 0.71 0.57 0.40 0.40 1.00 0.62 0.63 0.29 0.57 0.53 0.63 0.46
Se9 0.46 0.73 0.87 0.46 0.46 0.58 0.58 0.62 1.00 0.54 0.31 0.46 0.29 0.41 0.50
Se10 0.50 0.45 0.47 0.50 0.38 0.71 0.82 0.63 0.54 1.00 0.63 0.63 0.71 0.89 0.68
Se11 0.57 0.34 0.40 0.57 0.57 0.53 0.53 0.29 0.31 0.63 1.00 0.43 0.53 0.63 0.46
Se12 0.43 0.34 0.53 0.43 0.57 0.40 0.67 0.57 0.46 0.63 0.43 1.00 0.40 0.76 0.31
Se13 0.67 0.32 0.50 0.53 0.53 0.75 0.63 0.53 0.29 0.71 0.53 0.40 1.00 0.71 0.72
Se14 0.63 0.30 0.47 0.50 0.50 0.59 0.82 0.63 0.41 0.89 0.63 0.76 0.71 1.00 0.68
Se15 0.62 0.37 0.58 0.46 0.31 0.58 0.58 0.46 0.50 0.68 0.46 0.31 0.72 0.68 1.00

8
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 8 Table 9
Superedge ranks. Opinion leaders of The COP 27 conference.
Rank SuperedgeRank value Superedge number Rank Account Type Score

1 0.758 3 1 LW Internet influencers 0.758

2 0.749 7 2 Wind Aid Internet influencers 0.744
3 0.723 1 3 Mr. Xiaobo Internet influencers 0.719
4 0.712 4 4 HN Internet influencers 0.697
5 0.708 9 5 CWE Internet influencers 0.684
6 0.699 5 6 WE Internet influencers 0.665
7 0.693 10 7 JSW Government 0.655
8 0.691 6 8 SDL Internet influencers 0.648
9 0.685 2 9 MV Government 0.638
10 0.677 14 10 DISR Internet influencers 0.629
11 0.669 17 11 Miss Guo Internet influencers 0.609
12 0.653 8 12 CSP Internet influencers 0.587
13 0.649 15 13 Gao Han Internet influencers 0.557
14 0.633 12 14 CPN Government 0.549
15 0.626 11 15 SBJ Government 0.528
16 0.611 19 16 TTN Internet influencers 0.509
17 0.608 18 17 UNEP Internet influencers 0.497
18 0.597 22 18 HNE Government 0.476
19 0.59 13 19 CMA Government 0.455
20 0.581 20 20 BB Internet influencers 0.438
21 0.563 23 21 LH Internet influencers 0.419
22 0.55 30 22 CW Government 0.398
23 0.541 16 23 NWI Government 0.385
24 0.529 26 24 CGN Government 0.364
25 0.511 24 25 BJW Government 0.348
26 0.498 21 26 CCTV News Government 0.329
27 0.481 28 27 WTSF Scientist 0.318
28 0.459 31 28 TK Scientist 0.298
29 0.437 25 29 AHE Government 0.287
30 0.412 33 30 CMB Producer 0.276
31 0.385 29 31 BJR Government 0.265
32 0.366 35 32 MEE Government 0.248
33 0.337 27 33 ENF Producer 0.229
34 0.319 37 34 CT Producer 0.211
35 0.297 34 35 PS Scientist 0.197
36 0.276 32 36 CBCDF Producer 0.184
37 0.264 40 37 BJPSD Producer 0.165
38 0.239 38 38 WWF Producer 0.154
39 0.225 42 39 CAAS Scientist 0.148
40 0.212 36 40 UNCF Producer 0.134
41 0.186 45 41 CS Scientist 0.122
42 0.166 41 42 STM Scientist 0.109
43 0.14 46 43 SSS Scientist 0.964
44 0.125 39 44 GP Producer 0.837
45 0.107 49 45 SRR Scientist 0.639
46 0.095 43 46 ICS Scientist 0.432
47 0.083 50 47 CSC Scientist 0.293
48 0.054 47 48 HN Producer 0.197
49 0.038 44 49 CME Scientist 0.138
50 0.029 48 50 ST Scientist 0.093

meeting. Technologies, Lexington, KY, USA). The spatial correlation link of the
The subsequent section of the table expands its scope to embrace a government network for climate disasters was evaluated, and the asso­
wider range of opinion leaders, so indicating the active involvement of ciation matrix was developed. It can be seen that there is a meaningful
individuals from various backgrounds, including government officials, spatial network link between government news networks in China, and
scientists, and producers. Although the impact ratings may exhibit it can be detected that the network shape is consistent with typical
variability, they add to the intricate discourse around The COP 27 network structure characteristics. It is displayed that CMA, CW, NWI,
summit. This ranking offers a comprehensive viewpoint on the various CCTV News, GSW, and BJR networks are closely related to others shown
stakeholders engaged in the event, emphasizing the presence of in Fig. 3.
renowned institutions such as the United Nations Environment Pro­ Table 10 displays the primary analysis of the Government network in
gramme (UNEP) alongside relatively unfamiliar online influencers. Each China concerning the catastrophic disaster theme of climate change.
stakeholder assumes a unique role within the information ecosystem CMA, CW, NWI, CCTV News, GSW, and BJR are the most central of the
about climate-related deliberations at The COP 27 conference. six networks since they are the most established and have the most ties
to other networks dealing with disasters. The SBJ, BJW, CGN, MEE,
AHE, and MV nodes are centrally situated in disaster-related networks.
4.7. Networks of four attributes There would then be less of an impact from those disaster-themed
ideas spreading to other areas. This study examines 14 government
4.7.1. Government network under the theme of climate disaster networks, including receiving and transmitting nodes, by counting each
Fig. 3 shows the government network under the theme of climate network’s associations. In the aftermath of climate-related and natural
disaster from 1st October 2022 to 31st December 2022 using the Net­ disasters, social media has emerged as the primary means of information
draw visualization tool in UCINET software (UCINET v6.0, Analytic

9
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Fig. 3. Government network under the theme of climate disaster from 1st October 2022 to 31st December 2022.

added value of social media is worth the effort. Numerous case studies
Table 10
demonstrate how social media’s role and importance in emergencies
Centrality Analysis of Government Network under the Climate Disaster Theme.
differ across contexts. Many sources provide contradictory results, such
Account Degree Closeness centrality Betweenness centrality as comparing people’s trust in social media to their trust in more
CMA 13 1 1.055988 established, government-led sources and venues. More broadly, in
CW 13 1 1.055988 environmental policy, the Chinese government has attempted to
CGN 9 0.764706 0
strengthen the role of public engagement in decision-making. The
NWI 13 1 1.055988
CCTV News 13 1 1.055988 Environmental Impact Assessment (EIA) Law has required enhanced
JSW 13 1 1.055988 public participation since 2002. The 2015 Environmental Protection
MEE 11 0.866667 0.447222 Law includes a lengthy chapter on increasing participation and trans­
AHE 11 0.866667 0.190909 parency in environmental information (He et al., 2017; Zhang et al.,
HNE 11 0.866667 0.190909
2016).
MV 12 0.928571 0.840079
CPN 12 0.928571 0.688131
BJR 13 1 1.055988 4.7.2. The internet influencers network under the climate disaster theme
BJW 10 0.8125 0.090909 The correlation between internet influencers and climate disaster
SBJ 10 0.8125 0.215909
news themes from 1 October 2022 to 31 December 2022 using the
Netdraw visualization tool in UCINET software (UCINET v6.0, Analytic
dissemination. It allows for disseminating information from various Technologies, Lexington, KY, USA). The spatial correlation link of
sources, providing mutual help through information exchange and internet influencers on climate disasters was evaluated, and the associ­
emotional support, and maintaining contact if more conventional means ation matrix was developed. It can be seen that there is a significant
of communication become unavailable. Despite this, some wonder if the spatial network correlation among internet influencers’ networks in

Fig. 4. Internet influencers network climate disaster from 1st October 2022 to 31st December 2022.

10
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 11 one’s involvement with an issue to be focused solely on awareness. It is

Centrality analysis of internet influencers network under the climate disaster intended to entice consumers to spend hours reading through automated
theme. feeds. Users can like a single post from an eco-influencer and feel good
Account Degree Closeness centrality Betweenness centrality about themselves for taking a position against climate disaster before
UNEP 10 0.777778 7.672682
moving on to watch cat videos.
CWE 10 0.777778 5.582206 Unlike influencers, grassroots organizers advocate directly inside
Mr Xiaobo 6 0.636364 1.060714 communities, building bonds with citizens and their needs. They take
HN 7 0.666667 3.309085 accurate measures to divest institutions from fossil fuels, block the
WE 8 0.7 2.518609
establishment of toxic industrial facilities in Black and brown neigh­
Gao Han 11 0.823529 5.264098
LW 5 0.608696 1.428822 borhoods, influence local legislators on environmental laws, and so
CSP 6 0.636364 1.425752 much more. After raising awareness, they take the critical next step: to
LH 8 0.7 2.118609 effect actual change (Bjorn, 2020). The focus on individual branding and
SDL 8 0.7 1.954574 personality in influencer culture is opposed to grassroots organizing. It
BB 5 0.583333 1.092857
Miss Guo 9 0.736842 5.814348
establishes a hierarchy inside the climate movement, giving those with
Wind Aid 10 0.777778 5.481454 the most platforms more weight. The media encourages this endeavor by
DISR 7 0.666667 2.4 continuously spotlighting a few individuals as the “face of the move­
TTN 6 0.608696 0.87619 ment,” even though it is a group effort. Furthermore, these “faces” are
rarely drawn from frontline communities, whose voices should be most
heard.
Table 12
Centrality analysis of scientists’ network under the climate disaster theme. 4.7.3. Scientists’ network under the climate disaster theme
Account Degree Closeness centrality Betweenness centrality The correlation between scientists and climate disaster themes from
1 October 2022 to 31 December 2022 using the Netdraw visualization
PS 3 0.6 2
CS 2 0.428571 0
tool in UCINET software (UCINET v6.0, Analytic Technologies, Lex­
ICS 3 0.6 2 ington, KY, USA). The association matrix was created after analyzing the
STM 4 0.75 9 scientists’ spatial correlation relationship with climate disasters. It can
WTSF 3 0.8 1 be seen that there is insignificant spatial network association among
SSS 2 0.666667 0.333333
scientists’ networks in China, and it can be viewed that the network
CAAS 3 0.8 0.333333
CSC 3 0.8 0.333333 shape is less consistent than typical network structure characteristics. It
CME 3 0.8 1 is shown that only STM have extensive networks.
TK 3 0.6 2 Table 12 displays the centrality outcomes of the network of scientists
SRR 2 0.428571 0 in China for the climate issue. It demonstrates the most significant de­
ST 3 0.6 2
gree of centrality is in one scientist network only in the light of the fact
that the scientists’ network surrounding certain news subjects has grown
slowly and has fewer links. These networks are essential to all scientists’
Table 13 networks, including CSC, ICS, WTSF, and others. There is less spillover
Centrality analysis of producers’ network under the disaster theme.
in this case since the disaster topics are more beneficial and less self-
Account Degree Closeness centrality Betweenness centrality sufficient. An analysis of 12 scientists’ networks with receiving and
HN 4 0.571429 19 sending relationships considers that each network’s affiliations
WWF 2 0.533333 15 comprise the number of connections sent and received.
UNCF 3 0.444444 7 Scientists can affect climate change via popular media platforms like
GP 2 0.421053 0
TikTok. These technologies let them visually explain complicated sci­
CT 1 0.380952 0
ENF 3 0.444444 13 entific concepts. Scientists may explain climate disasters, exchange
BJPSD 1 0.32 0 research, and show their effects on society and the environment in short
CMB 1 0.32 0 videos. They may dispel climate disaster myths by offering accurate and
CBCDF 1 0.32 0
trustworthy facts. Scientists can also use TikTok to persuade viewers to
embrace sustainable practices and make eco-friendly decisions (Rosa
China, and it can be observed that the network shape is stable with et al., 2022). They help people minimize their carbon footprints and
typical network structure characteristics. It is shown that Wind Aid, Gao fight climate disasters by giving practical solutions. These actions can be
Han, CWE, and UNEP have extensive networks and are closely related to daily or lifestyle changes. TikTok users must also collaborate. Scientists
each other against other influencers, as shown in Fig. 4. can collaborate with peers, organizations, and influencers to develop
The centrality analysis of internet influencers’ networks under the content, compete, and expand their reach. Together, they can reach
climate disaster theme in China is shown in Table 11. It proves that the more people, boost engagement, and build a climate change community.
four internet influencers’ networks have the highest centrality. The Scientists on TikTok need audience involvement. They can answer
influencers’ network with those themes is developed and has more climate change topics and communicate with viewers through com­
connections with each other. These networks centre all influencers’ ments, duets, and live sessions. It builds trust, credibility, and
networks, such as TTN, DISR, BB, LW, etc. In such instances, the disaster communication.
themes are more advantageous and less self-sufficient, resulting in less
spillover. The number of associations in each network is analyzed, 4.7.4. Producer networks under the climate disaster theme
including the number of relationships sent and received, and 15 internet Using the Netdraw visualization tool in the UCINET programme
influencers’ networks with the receiving and sending connections are (UCINET v6.0, Analytic Technologies, Lexington, KY, USA), Fig. 6 de­
explored. picts the relationship between producers and the climate disaster subject
Eco-influencers at the “forefront” of the environmental movement from October 1, 2022, to December 31, 2022. The association matrix
now hoard media attention, creating a positive feedback loop of greater was developed after examining the producers’ correlation with climate
engagement and more followers. The structure of social media promotes disasters. The network form is less stable with typical network structure
features, and it is evident that there is not a sizeable geographical

11
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 14 Producers as content creators on social media platforms such as

The fifty subjects Pagerank ranking. TikTok can influence consumer behavior and indirectly affect climate
Account Type Pagerank (after 50 iterations) disasters. They can encourage their audience to make more climate-
conscious decisions by creating and promoting content emphasizing
CW Government 0.032278
NWI Government 0.032278 sustainable practices, eco-friendly products, and responsible consump­
CCTV News Government 0.032278 tion (Jiang et al., 2021). Producers can set trends, influence purchasing
CMA Government 0.030999 decisions, and nurture climate change conversations, increasing fol­
BJR Government 0.028451 lowers’ awareness and inspiring action. Collaborations with environ­
MV Government 0.02835
JSW Government 0.027737
mental organizations and activists can magnify their impact even
Gao Han Internet influencers 0.027646 further. However, producers must also consider their ecological foot­
Wind Aid Internet influencers 0.02514 print and implement sustainable production methods. Overall, TikTok
AHE Government 0.02472 producers have the potential to influence public attitudes and actions
LH Internet influencers 0.023894
regarding climate change mitigation significantly.
MEE Government 0.023269
UNEP Internet influencers 0.022013
SBJ Government 0.021904 4.7.5. The algorithm of page rank
CPN Government 0.021848 The PageRank results dated 31st December 2022 are, once again,
Miss Guo Internet influencers 0.021833 somewhat interesting. In the top position, we find in Fig. 7 that seven
HNE Government 0.020174
SDL Internet influencers 0.01994
internet influencers have higher page rank values. Fig. 7 shows that pink
HN Internet influencers 0.019906 nodes represent internet influencers against the government, scientists,
LW Internet influencers 0.019703 and producers. We have used 50 TikTok accounts of internet influencers,
WE Internet influencers 0.019583 government, scientists, and producers to check their impact on climate
BJW Government 0.019567
disasters. The government has a significant impact on climatic disasters.
CGN Government 0.019361
DISR Internet influencers 0.019348 Influencers may impact climate disaster reportage online. They can
CWE Internet influencers 0.019342 alter their views due to their big audience and curiosity. Influencers use
ENF Producer 0.019287 personal tales to make the issue more approachable and promote sus­
HN Producer 0.018939 tainable behaviors. Working with others raises awareness. Influencers
TTN Internet influencers 0.018822
must be accountable and know the truth. Influencers may inspire
SSS Scientist 0.018742
WTSF Scientist 0.018696 climate action.
CS Scientist 0.018647 Online influencers may also disseminate climate disaster narratives.
STM Scientist 0.018372 They can highlight the disproportionate impact of climate disasters on
ST Scientist 0.018176
the poor and marginalized. These personal narratives may help influ­
WWF Producer 0.018044
PS Scientist 0.017958 encers raise awareness of climate change-related social justice and eq­
TK Scientist 0.01631 uity issues. Influencers can also encourage audience involvement and
CSP Internet influencers 0.01616 action. They may mobilize their audience by endorsing climate-related
UNCF Producer 0.016132 causes, encouraging climate strikes, and striking themselves. Together,
SRR Scientist 0.016012
we can raise awareness of climate disasters and urgency about finding
Mr Xiaobo Internet influencers 0.015744
CT Producer 0.015729 answers.
CAAS Scientist 0.015729 Opinion leaders must verify their data and sources. “Greenwashing”
CMB Producer 0.015398 can hinder climate disaster efforts. Influencers who care about their
ICS Scientist 0.014845
reputation should check their heads and cooperate with recognized
CSC Scientist 0.01379
BB Internet influencers 0.01368
scientists. Finally, internet influencers disseminate climate disasters,
GP Producer 0.012612 encourage sustainable practices, offer a range of narratives, and mobi­
BJPSD Producer 0.011691 lize their followers. Their impact may help people understand climate
CME Scientist 0.010742 disasters and act individually and collectively (Table 14).
CBCDF Producer 0.008181

5. Case study (disaster: super cold wave in the United States,

network association among Chinese producers. Compared to other December 23rd-28th, 2022)
producers, it has been demonstrated that only HN, UNCF, and ENF have
substantial networks and are strongly connected to others. The winter storm dubbed Elliott by forecasters intensified into a
Table 13 displays the centrality investigation of the network of bomb cyclone near the Great Lakes on Friday, bringing high winds and
producers in China for the climate change news issue. It demonstrates blizzard conditions from the Northern Plains to western and upstate
that the most remarkable centrality is only among the two producers’ New York, along with life-threatening flooding, flash-freezing, and
networks. Producers in China have considerable sway over climate travel pandemonium as it moved. With an estimated width of 2000
change reporting thanks to their ability to shape coverage through miles, the storm has produced blizzard-like conditions and plunging
editorial decisions, story structuring, source selection, the introduction temperatures, taking out power from Texas to Maine. Forecasters in the
of possible biases or sensationalism, and collaboration with advertisers United States have warned of “potentially crippling impacts across the
(Manish et al., 2022). What the general population thinks and knows central and eastern United States.”
about climate change largely depends on their decisions. However, It is estimated that >1.5 million households were without electricity
ethical producers may raise awareness and develop educated dialogues due to the advent of the Arctic blast. After Joe Biden was briefed on
by encouraging accurate and fair reporting. Producers with a sense of Elliott at the White House, the president stated, “This is not like a snow
responsibility will see the significance of delivering thorough and day when you were a child. It is imperative.” National Weather Service
truthful coverage of climate change. They try to underline the issue’s (NWS) forecaster Bob Oravec estimated that approximately 200 million
urgency, provide several viewpoints, and underscore the scientific individuals in the 48 contiguous states were under severe weather alerts.
agreement. They help educate the public, stimulate thoughtful debate, An advisory warned that a powerful cold front would engulf the eastern
and propel China’s climate change response. United States tonight, with pervasive, dangerous cold expected to persist
across most of the eastern two-thirds of the country through the holiday

12
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 15 weekend. “Rapid temperature drops, sometimes 50 degrees or colder

Environmental view, psychological view, and main opinions of super cold wave than the previous day,” Oravec told the Associated Press. “It’s a fairly
(23rd-28th December 2022). effective system.” This event attracted the attention of many social
Example media users as well. We identified all agents throughout the event and
Environmental 3 E1 It is the largest cold alert ever.
the corresponding environmental data, psychological categories, and
subnetwork under the E2 The main flights have been keywords. From these data, an online public opinion supernetwork
disaster theme cancelled. model was developed (see Table 15).
E3 Gas production plummeted.
E4 The price of food has increased
5.1. Superedge degree (LSE)
many times.
Psychological subnetwork 8 The psychological tendency can be divided into
under the theme of positive, neutral, and negative, and the In a supernetwork, it may be said that two superedges connected by
disaster psychological intensity is determined by the the same node are considered supersedes. According to Wang et al.
number of keywords with opinion tendency (2010), Superedge Degree is the total number of other Superedges
published by the corresponding subject.
View the subnetwork 20 Positive K1: The United States should
connected to a certain Superedge through its nodes. For instance, in the
under the disaster theme view actively respond to climate abovementioned straightforward online public opinion supernetwork
disasters. model, LSE1 = 49 (Table 16).
K2: The consequences of climate
disasters should be taken
5.2. The attribute of influential degree of information dissemination
seriously.
K3: The government can do an
excellent job in post-disaster The climate change subnetwork contains all the information that was
reconstruction. made public about a particular climate change news event during its
Neutral K4: Mankind should discuss how inception and growth (let’s assume there are N pieces). Each piece of
view to prevent climate disasters.
K5: Mankind does not
information corresponds to one information node, ei (1 ≤ i ≤ n). The
understand the consequences of influential degree of information distribution, or I(ei), is defined as the
climate disasters. degree to which certain types of information are effective. Two indi­
Negative K6: Climate disasters do not ces—the breadth of information dissemination R(ei) and the depth of
view affect humanity.
information dissemination D(ei)—are used to quantify the influence of
K7: Climate disasters are
inevitable. information dissemination.
K8: Mankind is powerless against This climate disaster subnetwork is well-established; the basic
climate disasters. climate disaster model consists entirely of three information nodes.
K9: Mankind should not exploit Table 21 displays the computation results for the degree of information
nature.
K10: The consequences of
diffusion influence.
climate disasters are severe. Table 17 provides a complete assessment of the level of influence in
K11: Mankind should not disseminating information among three organizations referred to as e1,
develop the economy. e2, and e3. These entities function as repositories of knowledge, and the
K12: Climate disasters are
table utilizes many important parameters to evaluate their influence
rumors.
inside their respective networks. The table provides an analysis of the
count of persons who have been recipients of information from various
entities (F(ei)), the count of individuals who have actively accepted or
Table 16 interacted with this material (A(ei)), and the overall size of the audience
Calculation of superedge degree of super cold wave.
(N). The metrics R(ei) and D(ei) are used to assess each entity’s level of
Se1 LW – L information reach and engagement, respectively. Finally, the impact
e1 Se4,Se9,Se10,Se14,SE17,Se21,Se22,Se25,Se27,Se32, (Se1) = factor (I(ei)) is computed to offer a comprehensive viewpoint that con­
Se37,SE40,Se41,Se43,Se45,Se46,Se50 49 siders both the reach and engagement elements. The table provides a
k1 Se3,Se7,Se9~Se12,Se14,Se15,Se17,Se21,Se29~Se31, quantitative analysis of the information dissemination strategies
Se33~Se37,Se39,Se43~Se50
k3 Se2,Se3,Se6~Se10,Se13~Se15,Se20,Se22,Se26,Se27,
employed by these entities, as well as the extent of their audience reach
Se33,Se35~Se37,Se42~Se45,Se47~Se50 and level of active involvement achieved.
k4 Se3~Se6,Se11,Se13,Se16,Se18,Se19,Se24,Se26,Se29, Based on the analysis of these measures, it is apparent that the en­
Se31,Se34,Se35,Se38~Se44,Se46,Se49 tities e1, e2, and e3 possess differing levels of impact. As an illustration,
k5 Se2~Se5,Se8,Se9,Se11,Se17,Se19,Se21~Se23,Se26,
e1 achieves a 36 % reach among its audience, but e3, albeit attaining a
Se27,Se29,Se35~Se37,Se40,Se41,Se43,Se46~Se48
k9 Se7,Se8,Se10,Se12~Se14,Se17,Se19,Se22,Se23, smaller proportion of its audience (30 %), effectively engages a note­
Se26~Se27,Se30~Se32,Se34~Se36,Se39,Se46~Se49 worthy 36 % of the reached individuals. These observations demonstrate
k10 Se4,Se6,Se7,Se10,Se11,Se13~17,Se19,Se20,Se22,Se23, the mechanisms of information propagation inside these organizations,
Se25,Se26,Se30,Se32,Se33,Se35~Se37,Se39~Se41, providing a deeper understanding of the intricacies involved in the
Se43,Se45,Se46,Se48,Se50
k12 Se3~Se5,Se8,Se12~Se17,Se19,Se20,Se22,Se24~Se35,
distribution and reception of information. The present research offers
Se38,Se41,Se43~Se47,Se50 significant academic perspectives on the intricacies of information
dissemination and its many ramifications inside networks.

Table 17
5.3. Transformation relations among psychological types
Influential Degree of Information Dissemination.
The netizens that publish material or blogs to engage in the con­
ei F(ei) A(ei) N R(ei) D(ei) I(ei)
versation at the start and during the event’s evolution are of various
e1 18 18 50 0.36 0.24 0.0864 psychological kinds. The psychological nodes of the online public
e2 17 17 50 0.34 0.16 0.0544
opinion supernetwork model have variable inclination and strength
e3 15 15 50 0.3 0.36 0.108
because different psychological types have distinct psychological ten­
dencies. The psychological profiles of the agents can be determined by

13
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 18 We utilized the SuperedgeRank technique to rank all the superedges.

Transformation Relations among Psychological Types. Then, we identified opinion leaders using the online public opinion
pij P1 P2 P3 P4 P5 P6 supernetwork model of the COP 27 Conference using the MATLAB
programming tool. Table 21 shows the robust of identifying ten leaders
P1 0.00 1.00 0.50 0.33 0.25 0.2
P2 1.00 0.00 1.00 0.50 0.33 0.25 among the 50 respondents.
P3 0.50 1.00 0.00 1.00 0.50 0.33 Table 21 presents an exhaustive compilation of thought leaders
P4 0.33 0.50 1.00 0.00 1.00 0.5 throughout the timeframe linked with the “Super Cold Wave.” The
p5 0.25 0.33 0.50 1.00 0.00 1.00 classification of these influential individuals is determined by their ac­
P6 0.2 0.25 0.33 0.5 1.00 0.00
count names, account kinds, and ratings, which measure their impact or
prominence in engaging with subjects about the Super Cold Wave. The
reading their internet posts or blogs. The definition of psychological type various sorts of accounts may be broadly classified into four main cat­
pi is an integer with a range of +1 to _1. Pi’s positive and negative di­ egories: Internet influencers, Government officials, Scientists, and
rections define psychological propensity, while pi’s absolute value de­ Producers.
termines psychological strength. The ranks allocated to each opinion leader are determined based on
Among the above-established simple online public opinion super­ their respective scores, which serve as indicators of their relevance
network model, there are a total of 6 psychological types. Let p1 = − 3, within the debate around the Super Cold Wave. It is worth mentioning
p2 = − 2,p3 = − 1,p4 = 1,p5 = 2, p6 = 3. The calculated correlations that individuals who have gained prominence on the Internet referred to
among the 6 psychological types of this simple model are shown in as “Internet influencers” in the “Type” column, occupy the highest
Table 18. ranks, with LW, Wind Aid, and Mr. Xiaobo securing the first three spots.
The numerical scores allocated to each leader serve as quantitative in­
5.4. The similarity of keywords in different superedges dicators of their respective levels of influence, with higher scores
denoting a greater impact in shaping debates and spreading information
The vector space model (VSM) is frequently employed to determine of the Super Cold Wave phenomenon. The chart shown above serves as a
object correlation coefficients. The TF-IDF method is based on the VSM, great tool for identifying prominent individuals who have made signif­
which considers the frequency of distinct words in all texts and the icant contributions in addressing and disseminating information about
higher resolution of these words across texts. This method is extensively the Super Cold Wave phenomenon, notably through diverse internet
employed for calculating text similarity. In this paper, we calculated the platforms and channels.
similarity of keywords from two distinct superedges using the TF-IDF The findings of this study illustrate the broad range of opinion
method. Using the preceding formula, the computed similarity of key­ leaders engaged in conversations on the Super Cold Wave. Internet
words in all superedges in a simple online opinion supernetwork model influencers distinguish themselves by their exceptional influence rat­
is shown in Table 19. ings, underscoring their substantial impact on moulding the prevailing
discourse. Government officials, scientists, and producers contribute
significantly to the discourse, each providing distinct viewpoints and
5.5. Online opinion leader identification
valuable knowledge.
It is crucial to acknowledge that the determination of impact scores
5.5.1. Ranking result of SuperedgeRank
relies on the SuperedgeRank algorithm, which considers the intensity of
In the basic online public opinion supernetwork model described
connections and interactions within the online public opinion super­
previously, the attributes of superedges are calculated using the MAT­
network. The presented table offers useful insights into the key persons
LAB programming tool and SuperedgeRank algorithm. The ranking of
and institutions that have exerted the most substantial influence on the
17 superedges is displayed in Table 20.
discourse around the Super Cold Wave.
5.5.2. Robustness
6. Discussion and policy implications

5.5.2.1. Mechanism of opinion leader identification. A single Internet

6.1. Discussions
user might show up in multiple superedges. The overall score of the
superedges associated with an agent in a social subnetwork is found after
We analyze that the phenomena of “climate change news” and “di­
all the superedges are ranked and calculated. The average score is then
sasters” encompass separate yet interconnected dimensions under the
computed by adding each agent’s node super degree index. The leading
overarching subject of climate change as observed on the social media
agent is the one with the highest score.

Table 19
Similarity of keywords in different superedges.
Column1 Se1 Se2 Se3 Se4 Se5 Se6 Se7 Se8 Se9 Se10 Se11 Se12 Se13 Se14 Se15

Se1 1.00 0.37 0.40 0.47 0.17 0.55 0.47 0.37 0.47 0.47 0.85 0.63 0.45 0.40 0.45
Se2 0.37 1.00 0.18 0.43 0.62 0.67 0.43 0.50 0.58 0.43 0.46 0.43 0.82 0.55 0.41
Se3 0.40 0.18 1.00 0.63 0.51 0.37 0.47 0.55 0.63 0.47 0.68 0.47 0.45 0.20 0.75
Se4 0.47 0.43 0.63 1.00 0.67 0.72 0.75 0.58 0.63 0.75 0.67 0.75 0.71 0.47 0.82
Se5 0.17 0.62 0.51 0.67 1.00 0.46 0.53 0.62 0.80 0.53 0.29 0.40 0.63 0.51 0.76
Se6 0.55 0.67 0.37 0.72 0.46 1.00 0.58 0.50 0.58 0.58 0.62 0.72 0.82 0.37 0.54
Se7 0.47 0.43 0.47 0.75 0.53 0.58 1.00 0.43 0.50 0.88 0.53 0.88 0.71 0.63 0.82
Se8 0.37 0.50 0.55 0.58 0.62 0.50 0.43 1.00 0.72 0.58 0.46 0.58 0.68 0.37 0.54
Se9 0.47 0.58 0.63 0.63 0.80 0.58 0.50 0.72 1.00 0.63 0.53 0.63 0.59 0.63 0.71
Se10 0.47 0.43 0.47 0.75 0.53 0.58 0.88 0.58 0.63 1.00 0.53 0.88 0.71 0.79 0.71
Se11 0.85 0.46 0.68 0.67 0.29 0.62 0.53 0.46 0.53 0.53 1.00 0.67 0.63 0.34 0.63
Se12 0.63 0.43 0.47 0.75 0.40 0.72 0.88 0.58 0.63 0.88 0.67 1.00 0.71 0.63 0.71
Se13 0.45 0.82 0.45 0.71 0.63 0.82 0.71 0.68 0.59 0.71 0.63 0.71 1.00 0.45 0.67
Se14 0.40 0.55 0.20 0.47 0.51 0.37 0.63 0.37 0.63 0.79 0.34 0.63 0.45 1.00 0.45
Se15 0.45 0.41 0.75 0.82 0.76 0.54 0.82 0.54 0.71 0.71 0.63 0.71 0.67 0.45 1.00

14
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Table 20 Table 21
Ranking of superedge rank. Opinion leaders of the super cold wave.
Rank SuperedgeRank value Superedge number Rank Account Type Score

1 0.758 3 1 LW Internet influencers 0.758

2 0.749 7 2 Wind Aid Internet influencers 0.744
3 0.723 1 3 Mr. Xiaobo Internet influencers 0.719
4 0.712 4 4 HN Internet influencers 0.697
5 0.708 9 5 CWE Internet influencers 0.684
6 0.699 5 6 WE Internet influencers 0.665
7 0.693 10 7 JSW Government 0.655
8 0.691 6 8 SDL Internet influencers 0.648
9 0.685 2 9 MV Government 0.638
10 0.677 14 10 DISR Internet influencers 0.629
11 0.669 17 11 Miss Guo Internet influencers 0.609
12 0.653 8 12 CSP Internet influencers 0.587
13 0.649 15 13 Gao Han Internet influencers 0.557
14 0.633 12 14 CPN Government 0.549
15 0.626 11 15 SBJ Government 0.528
16 0.611 19 16 TTN Internet influencers 0.509
17 0.608 18 17 UNEP Internet influencers 0.497
18 0.597 22 18 HNE Government 0.476
19 0.59 13 19 CMA Government 0.455
20 0.581 20 20 BB Internet influencers 0.438
21 0.563 23 21 LH Internet influencers 0.419
22 0.55 30 22 CW Government 0.398
23 0.541 16 23 NWI Government 0.385
24 0.529 26 24 CGN Government 0.364
25 0.511 24 25 BJW Government 0.348
26 0.498 21 26 CCTV News Government 0.329
27 0.481 28 27 WTSF Scientist 0.318
28 0.459 31 28 TK Scientist 0.298
29 0.437 25 29 AHE Government 0.287
30 0.412 33 30 CMB Producer 0.276
31 0.385 29 31 BJR Government 0.265
32 0.366 35 32 MEE Government 0.248
33 0.337 27 33 ENF Producer 0.229
34 0.319 37 34 CT Producer 0.211
35 0.297 34 35 PS Scientist 0.197
36 0.276 32 36 CBCDF Producer 0.184
37 0.264 40 37 BJPSD Producer 0.165
38 0.239 38 38 WWF Producer 0.154
39 0.225 42 39 CAAS Scientist 0.148
40 0.212 36 40 UNCF Producer 0.134
41 0.186 45 41 CS Scientist 0.122
42 0.166 41 42 STM Scientist 0.109
43 0.14 46 43 SSS Scientist 0.964
44 0.125 39 44 GP Producer 0.837
45 0.107 49 45 SRR Scientist 0.639
46 0.095 43 46 ICS Scientist 0.432
47 0.083 50 47 CSC Scientist 0.293
48 0.054 47 48 HN Producer 0.197
49 0.038 44 49 CME Scientist 0.138
50 0.029 48 50 ST Scientist 0.093

platform TikTok. Within the “climate change news domain,” TikTok individual accounts of severe weather phenomena, such as extreme
functions as a medium for engaging in well-informed dialogues and weather occurrences and natural catastrophes, as well as the subsequent
disseminating educational material. Creators and outlets serve as plat­ consequences. Creators aim to increase public consciousness regarding
forms for disseminating up-to-date information on climate change pol­ the pressing nature of climate change by presenting the catastrophic
icy, international agreements, and scientific research. Additionally, they consequences of wildfires, hurricanes, floods, and other calamities on
provide movies that elucidate the underlying scientific principles of societies, ecosystems, and physical structures. The videos effectively
climate change, encompassing its origins, ramifications, and prospective establish a connection with viewers by presenting a personal perspective
remedies. The main objective of this information is to augment public on the hardships endured by those impacted by environmental catas­
comprehension, cultivate consciousness, and promote dialogue around trophes. These phenomena highlight the need of prompt action, the
climate change. The platform provides users with the necessary infor­ importance of being prepared for disasters, and the necessity of imple­
mation to actively participate in well-informed discussions regarding menting steps to enhance resilience to climate change. Consequently,
significant global concerns, fostering collaborative efforts towards they stimulate discussions and inspire activism in response to catastro­
adopting sustainable behaviors and policies. The expansive reach of phes associated with climate change. The dissemination of climate
TikTok enables it to serve as an inclusive platform that caters to a diverse change-related information and the portrayal of climate change-induced
audience, facilitating their ability to remain educated and up-to-date on disasters on TikTok both serve significant purposes in teaching the
advancements in the field of climate change. general public about the phenomenon. While one approach focuses on
In contrast, the content on “disasters” on TikTok tends to prioritize providing factual knowledge, the other aims to convey the harsh reality
highlighting the concrete and urgent consequences associated with and repercussions associated with climate change.
climate change. This category frequently showcases live recordings and The role of governments as opinion leaders is crucial in tackling

15
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

climate change and mitigating climate-related calamities. Policies,

legislation, and international agreements are established to reduce
climate change and enhance preparedness for climate-related calam­
ities. The aforementioned components encompass emissions reduction
objectives, renewable energy incentives, and disaster management
strategies. Governments also allocate financial resources towards the
support of climate research initiatives, as well as the provision of
assistance and recovery measures in response to climate-related in­
cidents. It is imperative for them to assume leadership roles in
addressing climate change, both inside their own country and on an
international scale, therefore serving as a model for the general populace
and other relevant parties.
As opinion leaders, scientists also play a crucial role in leading
climate change research. The provision of crucial data and analysis
serves as the foundation for our comprehension of climate change and its
associated consequences. The scope of their activity includes the sur­
veillance of climatic patterns, the execution of environmental evalua­
tions, and the anticipation of prospective calamities (Geels and Ayoub,
2023). In addition to disseminating their research within the scientific
community, scientists also disseminate their results to various external
audiences, including the general public, policymakers, and other rele­
Fig. 5. Scientists’ network under climate disaster theme from 1st October 2022 vant stakeholders. This serves the purpose of increasing public knowl­
to 31st December 2022.
edge and understanding and offering evidence-based recommendations
and guidance. They engage in partnerships with governmental entities
and various groups to formulate and implement plans to reduce the
impacts of climate change and effectively address its consequences.
Producers, encompassing both enterprises and industries, have a
substantial role in the realm of climate change and catastrophe mitiga­
tion. Organizations can implement sustainable and ecologically
conscious strategies, therefore mitigating emissions and minimizing
their overall carbon footprint. Producers have the potential to make
significant contributions towards fostering a more environmentally
sustainable economy and mitigating the impacts of climate change via
the development and implementation of eco-friendly goods and the use
of green technology. Responsible manufacturing and efficient supply
chain management practices can effectively mitigate the environmental
consequences associated with the production and distribution of prod­
ucts and services. Producers can set a precedent and influence cus­
tomers, encouraging them to adopt more sustainable behaviors.
The impact of internet influencers on the awareness of climate
change is steadily increasing. They possess a distinctive capacity to
effectively connect with and captivate a broad range of individuals via
various social media channels. Internet influencers can utilize their
widespread appeal to effectively promote public consciousness
regarding climate change and its associated impacts. They could
Fig. 6. Producers’ network under climate disaster theme from 1st October encourage environmentally conscious behaviors, endorse sustainable
2022 to 31st December 2022.
practices, and advocate for measures addressing climate change.
Through leveraging their digital presence, individuals are able to impart
knowledge and motivate their audience, so assuming a pivotal function
in galvanizing public opinion and fostering collaborative efforts towards
addressing climate change.
Internet Influencers have emerged as powerful agents of change in
the context of climate news and awareness. Their influence can be
attributed to a combination of key strategies, enabling them to shape
public opinion and attitudes towards climate change effectively
(Knupfer et al., 2023). Firstly, Internet Influencers wield the power of
large and highly engaged audiences. With millions of followers across
platforms like YouTube, Instagram, and TikTok, these influencers can
reach a vast and diverse demographic (Roshandel et al., 2023). These
audiences are more likely to interact with and share climate-related
content, creating a ripple effect of awareness and engagement (Brüg­
gemann et al., 2020).
Fig. 7. The network of fifty subjects PageRank. Secondly, what sets Internet Influencers apart is their ability to
convey complex climate-related topics in a relatable and accessible
manner. They are skilled at translating scientific jargon and intricate
environmental concepts into everyday language. This demystification of

16
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

climate science makes it easier for the general public to grasp the ur­ disaster response mechanisms. Governments should invest in robust
gency of environmental issues (Robinson et al., 2022). Additionally, early warning systems, well-coordinated disaster relief efforts, and
influencers often employ storytelling and personal anecdotes to connect efficient communication strategies during climate-related emergencies.
with their audiences connect with their audiences emotionally. This Being well-prepared for climate-induced disasters minimises their
emotional appeal can be a powerful motivator, encouraging individuals impact and fosters public trust in government leadership.
to take climate action. Internet Influencers are adept at using visual and A cornerstone of effective climate change communication is public
interactive content to capture and maintain audience attention. Videos, awareness initiatives. These initiatives should encompass both climate
infographics, live streams, and interactive elements like polls and Q&A change news and disaster-related information. In collaboration with
sessions make climate news informative and engaging (Luttrell, 2018). relevant stakeholders, governments should invest in campaigns that
The interactive nature of these platforms fosters active participation and offer accurate, accessible, and engaging content to the public. These
discussion, which are instrumental in sustaining climate conversations. campaigns can cover climate change impacts, adaptation strategies, and
Consistency and frequency in content creation also play a vital role. eco-friendly behaviors. We can drive positive change and promote sus­
Influencers maintain a regular schedule of climate-related content, tainable practices by providing the public with knowledge and action­
keeping the topic consistently in the public eye (Jin et al., 2013). This able steps.
sustained focus serves to reinforce the importance of climate change in Understanding the psychological dynamics that underlie climate
the minds of their followers. communication is crucial. It is possible to develop tailored intervention
Collaborations and partnerships are another key driver of their tactics by conducting in-depth research within psychological sub­
impact. Many influencers to collaborate with climate organizations, networks. This may involve addressing climate misinformation, coun­
experts, and fellow influencers to broaden their reach and credibility. tering climate apathy, and promoting positive behavioral changes.
These collaborations introduce their audiences to diverse perspectives, Psychological insights are instrumental in designing messages that
ensuring a more holistic understanding of climate-related issues (Dupar resonate with diverse audiences and inspire meaningful actions.
et al., 2019). An essential component of their approach is direct The proliferation of misinformation and rumors related to climate
engagement. Influencers actively interact with their followers by change necessitates swift and accurate responses. Governments and
responding to comments and questions. This two-way dialogue creates a media organizations should establish robust fact-checking mechanisms
sense of community and fosters discussions about climate change, to verify information and promptly clarify misconceptions or false nar­
further fueling awareness and action (Martin and MacDonald, 2020). ratives. Rapid responses to misinformation are essential to prevent the
Leveraging trends and challenges is yet another influencer strategy. spread of misleading content and to ensure that the public receives ac­
They often align their climate-related content with trending topics and curate and reliable information.
popular challenges on social media platforms. This strategy capitalizes The SuperedgeRank algorithm has proven to be a valuable tool for
on the broader public interest in such trends and ensures that climate identifying influential voices and emerging trends in climate communica­
change remains relevant and visible in the online space (Leung et al., tion. Utilizing data-driven approaches can enable a deeper understanding
2022). Finally, Internet influencers go beyond raising awareness. They of the evolving discourse on climate change. Analyzing the data makes it
actively advocate for concrete actions, motivating their audiences to possible to identify shifting opinion leaders, evolving narratives, and
participate in climate initiatives, projects, and environmental conser­ emerging challenges. This data-driven approach informs more effective
vation efforts. Their calls to action empower individuals to contribute to interventions and keeps climate communication strategies relevant.
addressing climate change at a personal level. In conclusion, Internet It is vital to incorporate environmental education within the
influencers have harnessed a combination of strategies and elements to educational system to empower future generations with the knowledge
convey climate news effectively and inspire their audiences to engage and awareness to address climate change. Environmental topics, climate
with the pressing issue of climate change. Their large and engaged au­ change impacts, and sustainable practices should be integrated into
diences, relatable messaging, emotional appeal, visual and interactive curricula at all levels. This educational approach equips students with
content, consistency, partnerships, direct engagement, trend utilization, the information and skills to become informed, responsible, and pro­
and advocacy collectively make them influential drivers of environ­ active participants in the fight against climate change.
mental awareness and action on a global scale. In conclusion, a comprehensive strategy to address climate change
requires a concerted effort from all stakeholders. By collaborating,
6.2. Policy implications engaging with internet influencers, enhancing disaster response, pro­
moting public awareness, conducting psychological research, fact-
To effectively address climate change challenges, fostering collabo­ checking, utilizing data-driven approaches, and emphasizing educa­
ration among diverse stakeholders is imperative. Government author­ tion, we can collectively tackle climate change and work towards a more
ities, scientists, producers, and internet influencers should engage in sustainable future. When implemented effectively, these policies
joint efforts to formulate comprehensive climate communication stra­ contribute to building a more resilient and informed society in the face
tegies. This collaboration should extend to sharing accurate, up-to-date of climate change and disaster challenges.
information and aligning their messaging to provide a unified front
against climate change. This synergistic approach enables each group to 7. Conclusion
leverage its strengths and reach, ultimately leading to more impactful
climate change communication and intervention. The effects of anthropogenic climate change are widely discussed in
Internet influencers have emerged as influential opinion leaders in scientific publications, newspaper articles, and blogs. Newspaper arti­
climate change news. Recognizing their reach and impact, governments cles may be inaccurate, while the gravity of scientific results may be too
and environmental organizations should proactively engage with these complex for the general people to comprehend. On the other hand, so­
influencers. Collaborative campaigns and partnerships can be forged to cial media is a venue where people from all walks of life may commu­
harness their extensive reach and the trust they have built with their nicate their thoughts and opinions. TikTok has become a significant
audiences. By aligning internet influencers with environmental causes resource for analyzing current events and breaking news as consumption
and providing them with credible information and resources, it becomes switches from old to new media.
possible to channel their influence towards raising climate awareness, This study employs the supernetwork analysis method in identifying
changing behaviors, and inspiring collective action. online public opinion leaders. It offers a Pagerank and SuperedgeRank
Given the pivotal role of government authorities as opinion leaders algorithm based on the superedge ranking algorithm. First, this study
in addressing climate change disasters, there is a critical need to enhance briefly discusses climate change news and disasters, followed by an

17
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

explanation of the development of an online public opinion supernet­ and interference tactics can be devised. Opposing opinion leaders, for
work model, which primarily includes three subnetworks: environ­ example, could be segregated in the social subnetwork, while positive
mental, psychological, and viewpoint subnetworks. Then, correlations opinion leaders should be safeguarded. After finding the crucial infor­
among superedges of an established online public opinion supernetwork mation in the climate change subnetwork, we may examine its time,
model are calculated, which consists of the influential degree of infor­ language, and linguistic forms to determine why it is so influential. The
mation dissemination, the likelihood of psychological transformation dominant psychological type can be identified in the psychological sub­
between different psychological types, and keyword similarity between network. Then, we can use psychological theories to analyze it, which
different viewpoints. Based on these calculations, a new SuperedgeRank may lead to improved intervention tactics. The mainstream viewpoint
method is designed to compute and rank all of the superedges of the can be determined from the perspective of subnetworks from massive
online public opinion supernetwork and the opinion leaders in the public online posts. If this is a rumor, a prompt clarification would be immensely
event. 50 Chinese TikTok accounts were used, based on internet influ­ appreciated. As a result, we can perform numerous research with varying
encers, government, scientists, and producers. Finally, the COP 27th priorities in the supernetwork using the SuperedgeRank algorithm.
meeting on climate change” and the “United States super cold wave” are
examples to demonstrate the reliability of Pagerank, the supernetwork CRediT authorship contribution statement
analysis method, and the SuperedgeRank algorithm. We conclude that
for climate change news, internet influencers play a significant role. Yunpeng Sun: Conceptualization, Data curation, Formal analysis,
Internet influencers are the prominent opinion leaders in Climate change Methodology. Ruoya Jia: Investigation, Validation, Visualization,
news, and they have more superedges than the other three attributes. Writing – review & editing. Asif Razzaq: Conceptualization, Writing –
Internet influencer has more different viewpoints than the other three review & editing. Qun Bao: Conceptualization, Data curation, Formal
attributes. Meanwhile, in climate disasters, the government has the edge. analysis, Methodology, Writing – original draft.
The government is the prominent opinion leader in Climate change di­
sasters. Government has more superedges than the other three attributes Data availability
and has more different viewpoints than the other three attributes.
Following identifying opinion leaders in an online public opinion The authors do not have permission to share data.
supernetwork, research can focus on deploying related guiding and
interference techniques. These interference techniques do not function Acknowledgement
effectively and have occasionally sparked public outrage. We can find
several types of critical nodes in the four layers using the SuperedgeRank The authors acknowledge the financial supports from the National
algorithm. Subnetworks are a type of network. Then, various guidance Social Science Fund of China (23CJY082).

Appendix A. Abbreviations, full form, and opinion leaders

Abbreviations Full Form Opinion Leaders

LW Little Weather Internet influencers

Wind Aid Wild Aid Internet influencers
Mr. Xiaobo Xiaobo Weather Internet influencers
HN Headlines Internet influencers
CWE Chinese weather enthusiasts Internet influencers
WE Weather everyday Internet influencers
CSP Cold science Internet influencers
JSW Jiangsu weather Government
SDL Science Daily Internet influencers
MV Military vision Government
DISR Director of the Institute for the Paranormal Studies Internet influencers
Miss Guo Teacher Xiao Guo Internet influencers
CW China Weather Government
NWI National Warning Issued Government
CGN Chinese Government Network Government
BJW Beijing Weather Government
Gao Han Gao Han Internet influencers
CPN China Police Network Government
SBJ Safe Beijing Government
TTN TikTok natural Internet influencers
UNEP United Nations Environment Programme Internet influencers
HNE Henan environment Government
CMA China Meteorological Administration Government
BB Big Bang Internet influencers
LH Life Help Internet influencers
CCTV News CCTV News Government
WTSF Forum of the world’s top scientists Scientist
TK True Knowledge Scientist
UNCF UNICEF Producer
CMB China Merchants Bank Producer
BJR Beijing Release Government
MEE Ministry of Ecology and Environment Government
ENF Earth Foundation Producer
CT Carbon Technology Producer
PS Panorama Scientist Scientist
CBCDF China Biodiversity Conservation and Green Development Foundation Producer
(continued on next page)

18
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

(continued )
Abbreviations Full Form Opinion Leaders

BJPSD Beijing Public Service Development Producer

WWF WWF Producer
CAAS China Academy of Aerospace Science Scientist
AHE Anhui Environment Government
CS Chinese Scientist Scientist
STM Scientific Time Machine Scientist
SSS Suyu Science Popularization Scientist
GP Greenpeace Producer
SRR Science Refutes Rumors Scientist
ICS Interviews with Chinese scientists Scientist
CSC Chinese Space Culture Scientist
HN Hello Nature Producer
CME Chinese Meteorological Expert Scientist
ST Science Travel Scientist

Appendix B

B.1. Network matrices and properties

Degree Centrality: A node’s degree of centrality in a complex network sequence can be defined as the fraction of the network’s edges that go to the
node (Wang et al., 2010). As a result, the degree centrality criteria may correspond to the links between nodes. The degree of centrality distribution
highlights important nodes and may show topological elements of government news networks. De indicates the importance of a node in the network
by counting the number of edges connecting it to other nodes. Calculating a node’s degree centrality helps pinpoint its position across the web. If a
node has the highest degree and has the most influence on the other nodes in the network, we call it the hub of the network. In a directed graph, the
degree of each node is either a point-in degree or a point-out degree, with the former being more prevalent. The formula (1) below shows how to
calculate the node’s degree centrality De(i):
De = n/N − 1 (1)

where ’n’ represents the number of intermediate nodes between the target node and the whole network, and ’N’ signifies the number of nodes in the
network.
Betweenness Centrality: A node’s betweenness centrality indicates how much power it has relative to other nodes in a network. Higher values are
associated with more confidence in a node’s ability to communicate and control its interactions with other nodes in the network. The value of ’Be’
primarily represents the extent to which a node exerts control over the connections between different nodes. The bigger the ’Be’ value, the more
control the node has over the government networks’ other nodes and how they communicate. Eq. (2) illustrates how to compute the betweenness
centrality ’Be(i)’ of a given node:
n ∑
∑ n
gjk (i)
Be = ,j ∕
=k∕
= ij < k (2)
j k
gjk

where gjk stands for the number of the shortest paths between nodes j and k and gjk (i) stands for the number of paths that connect nodes j and k through
node ‘i’, respectively.
Closeness Centrality: One may use closeness centrality to roughly calculate how far apart a particular node is from the rest of the nodes in each
network. A node’s centrality coefficient (Cl) is the inverse of the average shortest distance between that node and every other node in the network. The
clustering coefficient quantifies the degree of proximity between nodes in a network. In a network, a node’s centrality coefficient increases, and its
connectivity to other nodes improves as the distance between that node and its neighbors decreases. It is the total distance of a node from every other
node in the network, quantifying the degree to which that node is independent of the other nodes in the network. A higher score indicates that the node
has more connections with other nodes, making it more difficult to control. A node’s Cl value reflects its independence from neighboring nodes and
ability to act. An increased Cl relative to other nodes suggests that the node is less likely to affect other nodes during government network formation.
Calculating the closeness centrality "Cl(i)" of a given node ‘i’, as shown in Eq. (3), is a straightforward matter.
∑
n
Cl = dij (3)
j=1

where ‘dij’ is the undeviating distance between the two nodes, ‘i’ and ‘j,’ here (i ∕
=j).

19
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

B.2. Networks of four attributes

B.2.1. Government network under the news theme

Fig. B1. Chinese government network climate change news from 1st October, 2022 to 31st December, 2022

Fig. B1 shows the Chinese government network climate change news from 1st October 2022 to 31st December 2022, employing the Netdraw
visualization tool in UCINET software (UCINET v6.0, Analytic Technologies, Lexington, KY, USA). The spatial correlation link of government network
news was evaluated, and the association matrix was developed. There is a significant spatial network connection between government news networks
in China, and it can be seen that the network shape is persistent with typical network structure characteristics. It is shown that JSW, SBJ, NWI, CMA,
and MV news networks are closely related to others shown in Fig. B1.

Table B1
Centrality analysis of government network under the news theme.

Account Degree Closeness centrality Betweenness centrality

JSW 9 0.764706 29.5

MV 8 0.684211 20.5
CPN 4 0.52 1.5
SBJ 7 0.65 2.833333
BJW 4 0.52 12
NWI 7 0.65 2.833333
CGN 7 0.65 2.833333
CMA 6 0.619048 0
CW 7 0.65 4
CCTV News 3 0.5 0
MEE 2 0.448276 0.5
HNE 1 0.419355 0
BJR 4 0.464286 1.5
AHE 1 0.351351 0

The central analysis of the Government network under the new theme of climate change in China is depicted in Table B1. It verifies that the seven
news networks have the maximum centrality in all associations: JSW, MV, NW1, CMA, SBJ, CGN, and CW. They considered that the network with
those news themes is developed and has more connections with other news networks. These networks are central to news-related networks such as
CPR, BW, CPN, etc.
There is less spillover in such instances since those news themes have an immense advantage and a lesser capacity for self-sufficiency. Fourteen
government networks with receiving and sending dealings are examined, with the number of affiliations in each network considering the number of
relationships transmitted and received. People between 18 and 24 are most likely to access alternative news sources. This age group uses social media
three times more frequently than those over 55 to seek non-traditional sources of climate change information (Du et al., 2017). There is much
conflicting information on climate change in the media since it is a divisive subject. The effect of factual knowledge on global climate change obtained
from media sources relies on a person’s media literacy, capacity to distinguish fact from fiction, and evaluation of political claims and judgments.
Views and contributions on social media platforms may reflect broader attitudes and behaviors (Li et al., 2019). It is probably because social media
networks are cost-free, easily accessible, and have few content limitations (Du et al., 2017).

B.2.2. The internet influencers network under the news theme

Fig. B2 shows the correlation between Chinese internet influencers on climate change news themes from 1st October 2022 to 31st December 2022
using the Netdraw visualization tool in UCINET software (UCINET v6.0, Analytic Technologies, Lexington, KY, USA). The spatial link of internet
influencers on climate news was evaluated, and the association matrix was developed. There is an essential spatial network connection among internet
influencers’ networks in China, and it can be detected that the network shape is firm with typical network structure characteristics. It was revealed that
Mr. Xiaobo, HN, CSP, CWE, Wind Aid, LW, and SDL have large networks closely related to other influencers, as demonstrated in Fig. 3.
The centrality analysis of internet influencers’ networks under the climate change news theme in China is shown in Table B2. It proves that the
seven internet influencer networks have the highest centrality. They are considering that the influencers’ network with those news themes is
developed and has more connections with each other. These networks are central to all influencers’ networks, such as TTN, DISR, BB, UNFP, etc. In
such instances, the news themes are more advantageous and less self-sufficient, resulting in less spillover. It analyzes that the number of connotations
in each network contains the number of associations sent and received. Fifteen internet influencers’ networks with the receiving and sending

20
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

associations are analyzed.

Fig. B2. Chinese internet influencers network climate change news from 1st October, 2022 to 31st December, 2022.

Table B2
Centrality Analysis of Internet Influencers Network under the News Theme

Account Degree Closeness centrality Betweenness centrality

LW 14 1 2.71746
Wind Aid 14 1 2.71746
HN 14 1 2.71746
Mr Xiaobo 14 1 2.71746
CSP 13 0.933333 1.579365
CWE 13 0.933333 1.995635
WE 11 0.823529 0.361111
SDL 13 0.933333 2.013492
DISR 11 0.823529 0.571429
Gao Han 9 0.736842 0.378968
Miss Guo 12 0.875 0.882937
TTN 10 0.777778 0
LH 7 0.666667 0
UNEP 9 0.736842 0.236111
BB 8 0.7 0.111111

Regarding governing the world’s climate, China is a major player since it is the largest producer of greenhouse gases (Larsen et al., 2021). Slowly
but surely, China is becoming a leader alongside the United States in the fight against global climate change. Through their massive followings, ability
to personalize and convey stories, social media activism, education and knowledge, partnerships with NGOs and companies, and even questioning
official government narratives, Chinese internet influencers have a tremendous impact on how the public perceives climate change news. Influencers
may rapidly share information and increase awareness about climate change due to their large audiences. They make the subject more accessible and
exciting to their audience by relating personal anecdotes about their experiences.
Additionally, they use numerous social media sites to disseminate data, spark debate, and rally support for their causes. Some opinion leaders can
help dispel myths and provide accurate information about the climate because of their knowledge and experience in the field. By combining their
resources and audience, NGOs and companies may have a more significant effect through collaboration. Even though influencers periodically question
official stories, they must work within the parameters established by the state. In sum, their work helps raise consciousness, inspires people to act, and
shifts public opinion on climate change in China.

B.2.3. Scientists’ network under the news theme

Fig. B3 shows the correlation between scientists and climate change news themes from 1 October 2022 to 31 December 2022 using the Netdraw
visualization tool in UCINET software (UCINET v6.0, Analytic Technologies, Lexington, KY, USA). The association matrix was created after analyzing
the scientists’ spatial correlation relationship with climate news. It can be seen that there is a noteworthy spatial network connection among scientists’
networks in China, and it can be analyzed that the network shape is less stable than typical network structure characteristics. It is publicized that only
PS and CME have extensive networks and are closely related to each other against other scientists.

21
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Fig. B3. Chinese Scientists’ network climate change news from 1st October 2022 to 31st December, 2022

Table B3 displays the centrality investigation of the network of scientists in China for the climate change news issue. It demonstrates that the most
extraordinary centrality is among the two scientists’ networks only because the scientists’ network surrounding certain news subjects has grown
slowly and has fewer links. These networks are essential to all scientists’ networks, including CSC, STM, CS, SRR, etc. There is less spillover in this case
since the news topics are more beneficial and less self-sufficient. An analysis of 12 scientists’ networks with receiving and sending relationships
considers that each network’s affiliations comprise the number of connections sent and received.
Chinese scientists have a significant effect on the media coverage of climate change. They contribute to our understanding of climate change by
studying, producing scientific data, and suggesting policy changes. Their research adds to the sum of human knowledge and shapes media portrayals
of the world. Scientists are working to inform the public and simplify the complexities of climate change through their public outreach activities. They
work together globally and participate in assessments that help build agreement in the scientific community. The knowledge and work of Chinese
scientists have impacted how the public views climate change, which has influenced policy and media coverage of the issue.

Table B3
Centrality analysis of scientists’ network under the news theme.

Account Degree Closeness centrality Betweenness centrality

WTSF 2 0.37931 18
CME 3 0.478261 36
CS 1 0.268293 0
STM 2 0.354839 10
CSC 1 0.234043 0
SRR 2 0.297297 10
TK 1 0.333333 0
PS 4 0.478261 37
ICS 1 0.234043 0
SSS 1 0.333333 0
CAAS 2 0.37931 18
ST 2 0.297297 10

22
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

B.2.4. Producer networks under the news theme

Fig. B4. Chinese producers’ network climate change news from 1st October, 2022 to 31st December, 2022.

Using the Netdraw visualization tool in the UCINET program (UCINET v6.0, Analytic Technologies, Lexington, KY, USA), Fig. 5 depicts the
relationship between producers and the climate change news subject from October 1, 2022, to December 31, 2022. The association matrix was
developed after examining the producers’ correlation with climate news. The network form is less consistent with typical network structure features,
and it is evident that there is not a sizeable geographical network association among Chinese producers. Compared to other producers, it has been
demonstrated that only UNCF and CMB have substantial networks and are strongly connected to others.

Table B4
Centrality analysis of producers’ network under the news theme.

Account Degree Closeness centrality Betweenness centrality

UNCF 3 0.5 19
HN 1 0.347826 0
CBCDF 1 0.258065 0
BJPSD 2 0.333333 7
WWF 2 0.421053 12
ENF 1 0.296296 0
GP 2 0.470588 15
CMB 3 0.4 13
CT 1 0.296296 0

Table B4 displays the centrality outcomes of the network of producers in China for the climate change news issue. It demonstrates that the most
significant centrality is only among the two producers’ networks. Producers in China have considerable sway over climate change reporting thanks to
their ability to shape coverage through editorial decisions, story structuring, source selection, the introduction of possible biases or sensationalism,
and collaboration with advertisers. What the general population thinks and knows about climate change largely depends on their decisions. However,
ethical producers may raise awareness and develop educated dialogues by encouraging accurate and fair reporting. Producers with a sense of re­
sponsibility will see the significance of delivering thorough and truthful coverage of climate change. They try to underline the issue’s urgency, provide
several viewpoints, and underscore the scientific agreement. They help educate the public, stimulate thoughtful debate, and propel China’s climate
change response.
Also, producers may help increase the visibility of underrepresented groups in climate change coverage. Communities such as indigenous peoples
and vulnerable populations disproportionately impacted by climate change can share their tales and viewpoints. This all-encompassing approach
enables people from all walks of life to learn about climate change and its effects. It is also vital for scientists, journalists, and producers to work
together. By working together, they may verify information, put it in perspective, and assure correct reporting. The coordinated effort improves the
reliability of climate change reporting. It helps the public get a deeper comprehension of the subject. In conclusion, responsible Chinese producers
emphasize truth, balance, and inclusion despite their ability to impact climate change news through editorial decisions, framing, and source selection.
Their dedication to honest reporting and their willingness to work with scientists help spread information, stimulate thoughtful debate, and motivate
people to take action in the face of climate change.

References Ballestar, M.T., Sainz, J., 2020. A tale of two social influencers: a new method for the
evaluation of social marketing. In: Ballesta, J.F. María Teresa, D’Alessandro, S.
(Eds.), Advances in Digital Marketing and E-commerce. Springer, Berlin, pp. 80–90.
Agryzkov, T., Oliver, J.L., Tortosa, L., Vicent, J.F., 2012. An algorithm for ranking the
Bandura, A., Cherry, L., 2020. Enlisting the power of youth for climate change. Am.
nodes of an urban network based on the concept of PageRank vector. Appl. Math
Psychol. 75 (7), 945–951.
Comput. 219 (4), 2186–2193.
Barbiroglio, E., 2019. Generation Z fears climate change more than anything else. Forbes.
Agryzkov, T., Tortosa, L., Vicent, J.F., 2016. New highlights and a new centrality
https://www.forbes.com/sites/emanuelabarbiroglio/2019/12/09/generation-z-fea
measure based on the adapted PageRank algorithm for urban networks. Appl. Math
rs-climate-change-more-than-anything-else/.
Comput. 291, 14–29.

23
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

Basch, C.H., Meleo-Erwin, Z., Fera, J., Jaime, C., Basch, C.E., 2021. A global pandemic in Martin, C., MacDonald, B.H., 2020. Using interpersonal communication strategies to
the time of viral memes: COVID-19 vaccine misinformation and disinformation on encourage science conversations on social media. PloS One 15 (11), e0241972.
TikTok. Hum. Vaccin. Immunother. 17 (8), 2373–2377. Martin-Llaguno, M., Ballestar, M.T., Sainz, J., Cuerd-Mir, M., 2022. From ignorance to
Basch, C.H., Yalamanchili, B., Fera, J., 2022. # Climate change on TikTok: a content distrust: the public “discovery” of COVID-19 around international women’s day in
analysis of videos. J. Community Health 47, 163–167. Spain. Int. J. Commun. 16, 409–436.
Bjorn, Lomborg, 2020. Welfare in the 21st century: increasing development, reducing Masuda, N., Porter, M.A., Lambiotte, R., 2017. Random walks and diffusion on networks.
inequality, the impact of climate change, and the cost of climate policies. Technol. Phys. Rep. 716, 1–58.
Forecast. Soc. Chang. 156, 119981 https://doi.org/10.1016/j. Mol, A.P., 2008. Environmental Reform in the Information Age. The Contours of
techfore.2020.119981. Informational Governance. Cambridge University Press, Cambridge.
Brin, S., Page, L., 1998. The anatomy of a large-scale hypertextual web search engine. Nagurney, A., 2005. Supernetworks: an introduction to the concept and its applications
Comput. Networks ISDN Syst. 30 (1–7), 107–117. with a specific focus on knowledge supernetworks. Int. J. Knowl. Cult. Chang.
Brombal, D., 2017. Accuracy of environmental monitoring in China: exploring the Manag. 4, 1–16.
influence of institutional, political and ideological factors. Sustainability 9 (3), 324. NASA, 2019. Vital signs of the planet. https://climate.nasa.gov/causes/.
Brüggemann, M., Elgesem, D., Bienzeisler, N., Gertz, H.D., Walter, S., 2020. Mutual Newman, M., 2018. Networks. Oxford University Press, Oxford.
group polarization in the blogosphere: tracking the hoax discourse on climate Nisbet, E.K., Zelenski, J.M., Murphy, S.A., 2009. The nature relatedness scale: linking
change. Int. J. Commun. 14, 24. individuals’ connection with nature to environmental concern and behavior.
Butler, C.D., 2018. Climate change, health and existential risks to civilization: a Environ. Behav. 41 (5), 715–740.
comprehensive review (1989–2013). Int. J. Environ. Res. Public Health 15 (10), O’Brien, K., Selboe, E., Hayward, B.M., 2018. Exploring youth activism on climate
2266. change. Ecol. Soc. 23 (3), 42.
Chin-Fu, H., 2013. Citizen journalism and cyberactivism in China’s anti-PX plant in Patz, J.A., Vavrus, S.J., Uejio, C.K., McLellan, S.L., 2008. Climate change and waterborne
Xiamen, 2007–2009. China Int. J. 11 (1), 40–54. disease risk in the Great Lakes region of the US. Am. J. Prev. Med. 35 (5), 451–458.
Cody, E.M., Reagan, A.J., Mitchell, L., Dodds, P.S., Danforth, C.M., 2015. Climate change Pew Research Center, 2021. Demographics of Social Media Users and Adoption in the
sentiment on twitter: an unsolicited public opinion poll. PloS One 10 (8), e0136092. United States. Pew Research Center, Washington, DC.
Congressional Research Service, 2021. Wildfire statistics. Federation of American Pinski, G., Narin, F., 1976. Citation influence for journal aggregates of scientific
Scientists. https://fas.org/sgp/crs/misc/IF10244.pdf. publications: theory, with application to the literature of physics. Inf. Process.
Corbett, J., Savarimuthu, B.T.R., 2022. From tweets to insights: a social media analysis of Manag. 12, 297–312.
the emotion discourse of sustainable energy in the United States. Energy Res. Soc. Riascos, A.P., Mateos, J.L., 2012. Long-range navigation on complex networks using Lévy
Sci. 89, 102515. random walks. Phys. Rev. E 86 (5), 056110.
Dupar, M., McNamara, L., Pacha, M.J., 2019. Communicating Climate Change: A Riascos, A.P., Mateos, J.L., 2021. Random walks on weighted networks: a survey of local
Practitioner’s Guide. Climate and Development Knowledge Network. and non-local dynamics. J. Complex Netw. 9 (5), cnab032.
Ebi, K.L., Mills, D.M., Smith, J.B., Grambsch, A., 2006. Climate change and human health Robinson, S.A., Bouton, E., Dolan, M., Meakem, A., Messer, A., Lefond, I., Roberts, J.T.,
impacts in the United States: an update on the results of the US national assessment. 2022. A new framework for rapidly assessing national adaptation policies: an
Environ. Health Perspect. 114 (9), 1318–1324. application to small island developing states in the Atlantic and Indian oceans. Reg.
Estrada, E., Delvenne, J.C., Hatano, N., Mateos, J.L., Metzler, R., Riascos, A.P., Environ. Chang. 22 (2), 37.
Schaub, M.T., 2018. Random multi-hopper model: super-fast random walks on Rosa, Puertas, Carracedo, Patricia, Garcia-Mollá, Marta, Vega, Virginia, 2022. Analysis of
graphs. J. Complex Netw. 6 (3), 382–403. the determinants of market capitalisation: innovation, climate change policies and
Feng, G.C., Guo, S.Z., 2013. Tracing the route of China’s internet censorship: an business context. Technol. Forecast. Soc. Chang. 179, 121644 https://doi.org/
empirical study. Telematics Inform. 30 (4), 335–345. 10.1016/j.techfore.2022.121644.
Geels, Frank W., Ayoub, Martina, 2023. A socio-technical transition perspective on Rosenthal, S., 2022. Information sources, perceived personal experience, and climate
positive tipping points in climate change mitigation: analysing seven interacting change beliefs. J. Environ. Psychol. 81, 101796.
feedback loops in offshore wind and electric vehicles acceleration. Technol. Forecast. Roshandel, A., Miöen Dahlström, F., Ekström, K., 2023. Unveiling the Art of Illusion:
Soc. Chang. 193, 122639 https://doi.org/10.1016/j.techfore.2023.122639. Exploring the Fabrication of Authenticity and Trustworthiness by Social Media
Giustolisi, O., Ridolfi, L., Simone, A., 2020. Embedding the intrinsic relevance of vertices Micro-influencers to Engage their Audience: An Exploratory Multiple Case Study that
in network analysis: the case of centrality metrics. Sci. Rep. 10 (1), 3297. Aims to Analyze how Social Media Micro-influencers Fabricate Authenticity and
Han, H., Ahn, S.W., 2020. Youth mobilization to stop global climate change: narratives Trustworthiness to Engage their Audience.
and impact. Sustainability 12 (10), 4127. Sheffi, Y., 1985. Urban Transportation Networks: Equilibrium Analysis with
He, G., Mol, A.P., Lu, Y., 2016. Public protests against the Beijing–Shenyang high-speed Mathematical Programming Methods. Prentice-Hall, Upper Saddle River, NJ.
railway in China. Transp. Res. Part D: Transp. Environ. 43, 1–16. Takahashi, B., Tandoc Jr., E.C., Carmichael, C., 2015. Communicating on twitter during a
He, G., Boas, I., Mol, A.P., Lu, Y., 2017. E-participation for environmental sustainability disaster: an analysis of tweets during typhoon Haiyan in the Philippines. Comput.
in transitional urban China. Sustain. Sci. 12 (2), 187–202. Hum. Behav. 50, 392–398.
Hsu, A., Weinfurter, A., Tong, J., Xie, Y., 2020. Black and smelly waters: how citizen- Walter, Leal Filho, Wall, Tony, Mucova, Serafino Afonso Rui, Nagy, Gustavo J.,
generated transparency is addressing gaps in China’s environmental management. Balogun, Abdul-Lateef, Luetz, Johannes M., Ng, Artie W., Kovaleva, Marina,
J. Environ. Policy Plan. 22 (1), 138–153. Azam, Fardous Mohammad Safiul, Alves, Fátima, Guevara, Zeus,
Jiang, Sijian, Deng, Xiangzheng, Liu, Gang, Zhang, Fan, 2021. Climate change-induced Matandirotya, Newton R., Skouloudis, Antonis, Tzachor, Asaf, Malakar, Krishna,
economic impact assessment by parameterizing spatially heterogeneous CO2 Gandhi, Odhiambo, 2022. Deploying artificial intelligence for climate change
distribution. Technol. Forecast. Soc. Chang. 167, 120668 https://doi.org/10.1016/j. adaptation. Technol. Forecast. Soc. Chang. 180, 121662 https://doi.org/10.1016/j.
techfore.2021.120668. techfore.2022.121662.
Jin, L., Chen, Y., Wang, T., Hui, P., Vasilakos, A.V., 2013. Understanding user behavior in Wang, J.W., Rong, L.L., Deng, Q.H., Zhang, J.Y., 2010. Evolving hypernetwork model.
online social networks: a survey. IEEE Commun. Mag. 51 (9), 144–150. Eur. Phys. J. B 77, 493–498.
Knupfer, H., Neureiter, A., Matthes, J., 2023. From social media diet to public riot? Wei, Y., Gong, P., Zhang, J., Wang, L., 2021. Exploring public opinions on climate change
Engagement with “greenfluencers” and young social media users’ environmental policy in“ big data era”—a case study of the European Union emission trading
activism. Comput. Human Behav. 139, 107527. system (EU-ETS) based on twitter. Energy Policy 158, 112559.
Lazarsfeld, P.F., Berelson, B., Gaudet, H., 1944. The People’s Choice: How the Voter White, J.D., Fu, K.W., 2012. Who do you trust? Comparing people-centered
Makes up his Mind in a Presidential Campaign. Columbia University Press, New communications in disaster situations in the United States and China. J. Comp.
York, NY. Policy Anal. Res. Pract. 14 (2), 126–142.
Lee, M.T., Theokary, C., 2021. The superstar social media influencer: exploiting linguistic Wiedermann, M., Donges, J.F., Heitzig, J., Kurths, J., 2013. Node-weighted interacting
style and emotional contagion over content? J. Bus. Res. 132, 860–871. network measures improve the representation of real-world complex systems.
Leung, F.F., Gu, F.F., Palmatier, R.W., 2022. Online influencer marketing. J. Acad. Mark. Europhys. Lett. 102 (2), 28007.
Sci. 1–26. World Health Organization, 2021. Climate change. https://www.who.int/heli/risks/cli
Li, H., Zhang, M., Li, C., Li, M., 2019. Study on the spatial correlation structure and mate/climatechange/en/.
synergistic governance development of the haze emission in China. Environ. Sci. Zeng, J., Chan, C.H., Fu, K.W., 2017. How social media construct “truth” around crisis
Pollut. Res. 26, 12136–12149. events: Weibo’s rumor management strategies after the 2015 Tianjin blasts. Policy
Luttrell, R., 2018. Social Media: How to Engage, Share, and Connect. Rowman & Internet 9 (3), 297–320.
Littlefield. Zhang, L., Mol, A.P., He, G., 2016. Transparency and information disclosure in China’s
Manish, Kumar Goyal, Gupta, Anil Kumar, Jha, Srinidhi, Rakkasagi, Shivukumar, environmental governance. Curr. Opin. Environ. Sustain. 18, 17–24.
Jain, Vijay, 2022. Climate change impact on precipitation extremes over Indian
cities: non-stationary analysis. Technol. Forecast. Soc. Chang. 180, 121685 https://
Dr. Yunpeng Sun is an Associate Professor at the School of Economics, Tianjin University
doi.org/10.1016/j.techfore.2022.121685.
of Commerce, Tianjin, PR China. His area of research is green finance, carbon neutrality,
sustainability, climate change, green innovation, and financial markets. He published

24
Y. Sun et al. Technological Forecasting & Social Change 200 (2024) 123197

several papers in reputed journals such as the Journal of Environmental Management, Dr. Asif Razzaq is a senior research specialist at CAREC Institute. His area of interest
Sustainable Development, Resources Policy, and Renewable Energy. He has many highly mainly includes; digitalization, foreign direct investment, climate technology, and envi­
cited papers (top 1 %) and hot papers (top 0.1 %) in 2021, 2022 and 2023. ronmental/resources sustainability. He published many papers in reputed SSCI/SCI jour­
nals such as Technology forecasting and social change, Journal of Environmental
Management, Technology in Society, Sustainable Development, Resources Policy, and
Ruoya Jia is a master student at School of Economics, Tianjin University of Commerce. He
Environmental Research. Moreover, he has published over 50 research articles in reputed
is working in the area of climate change.
International Journals with 6000+ citations and an H-Index of 45. He is among the list of
global top 2% scientist 2023. He has many highly cited papers (top 1 %) and hot papers
Dr. Qun Bao is a Professor in Economics at the School of Economics, Nankai University, (top 0.1 %) in 2022 and 2023.
Tianjin, PR China. He published several papers in reputed SSCI/SCI journals such as
Journal of Business Research and Journal of Business Economics and Management. He has
many highly cited papers (top 1 %) and hot papers (top 0.1 %) in 2022 and 2023.

25