Climate-denying rumor propagation in a coupled socio-climate model: Impact on average global temperature

rumor quickly and assist in reducing the temperature. This portrays the importance of group propagation as shown in [49]. The univariate sensitivity plot attained by varying the average number of group members (m) displays that rumor propagation is quicker when the average number of group members is higher (S7 Fig). The baseline value for m for the model is 70. At lower m values, rumor propa- gation is considerably slow. It also leads to a higher temperature and CO2 emissions at lower m values which means that rumor propagation in groups with fewer members is less capable of spreading the rumor. Having more members in a group elevates the speed of the rumor propagation, making more individuals to decide at a time. The univariate analysis for the parameter �c, which is the average number of contact for each individual, shows that higher �c (baseline value 27) does not aid much in decreasing the tem- perature or the rumor propagation (S8 Fig) whereas lowering it slows down the rumor propa- gation leading to an elevated temperature level. Even though having more individual connections does not help in reducing the temperature, having fewer connections can cer- tainly increase the temperature anomaly by making the information spread more localized. S9 Fig is the univariate analysis for the parameter �d, which is the average rate of transmitting mes- sages per group. As the �d value is decreased, the rumor propagation slows down, and the tem- perature shoots up. There is a considerable variation in the temperature anomaly as �d

decreases, implying that lowering the number of messages transmitted in the group can increase the temperature level, as it slows the rumor transmission among the population and making fewer individuals think about climate change. The parameter distribution for each model parameter is calculated using Approximate Bayesian Computation is in S10–S16 Figs. The parameter plane by varying k and �d illustrates that the temperature anomaly decreases as the average network degree and the messages transmitted per group increase (Fig 5). The hypothesis was that reducing the average degree of the network will not much affect the rumor propagation and temperature anomaly when the average number of messages transmitted in the groups is high. This hypothesis is tested using parameter planes by changing the parame- ters simultaneously, and the result supports our hypothesis. The dotted lines in Fig 5 represent Fig 5. Global temperature anomaly is unaffected by low network degree at a high number of group messages. Parameter plane showing the changes in (a) rejector population and (b) temperature anomaly due to the variation in the average number of messages transmitted per group (�d) and the average degree of the network (k). The dashed lines show equal �d ∗ k values and the gray lines show equal size of rejector population and equal temperature anomaly. https://doi.org/10.1371/journal.pone.0317338.g005 PLOS ONE Climate-denying rumor propagation in a coupled socio-climate model: Impact on average global temperature

PLOS ONE | https://doi.org/10.1371/journal.pone.0317338 January 16, 2025 11 / 18