Lecture 14
Climate change
Natural disasters

Esteban Moro

Network Science Institute | Northeastern University

NETS 7983 Computational Urban Science

2025-04-15

Welcome!

This week:

Climate change and Natural Disasters in urban areas from the perspective of Computational Urban Science

You are here

Data -> Methods -> Models -> Applications.

Aims

  • Understand the use of Computational Urban Science on the impact of climate change and natural disasters
  • Present some studies and applications of CUS in climate change and natural disasters
  • Discuss the potential of CUS to inform policy decisions, improve disaster response, and enhance community resilience

Contents

  • Climate change
    • CUS in Climate change
    • Effect of cities and urbanization in climate change
    • Effect of climate change on human behavior
    • Climate change communication and misinformation.
    • Climate change adaptation and resilience.
  • Natural disasters
    • CUS in Natural disasters
    • Population displacement and evacuation modeling
    • Longer-term recovery and resilience
    • Inference of damages to built environment
    • Social networks and information dynamics
  • Conclusions
  • Further reading
  • References

Climate change

CUS in Climate change

Large-scale behavioral data is an invaluable resource for understanding the impact of climate change on human mobility and social dynamics [1]. By analyzing large-scale datasets can enable system-level climate actions and policies.

  • Gain insights into how people respond to and adapt to climate change.
  • Identify the factors that influence these patterns, such as population density, infrastructure, and socio-economic status.
  • Explore the role of social networks in shaping human behavior in response to climate change.
  • Investigate the impact of climate change on social interactions and community resilience.
  • Increase trust and minimize missinformation in climate change communication.

CUS in Climate change

Computational Social Science enabled behavioural climate action, from [2]

CUS in Climate Change

The applications can be categorized into 4 main areas:

  • Effect of cities and urbanization in climate change
  • Effect of climate change on human behavior.
  • Climate change communication and misinformation.
  • Climate change adaptation and resilience.

Effect of cities and urbanization in climate change

In [3] Manoli et al found that Urban Heat Islands are mainly explained by climate and population in those urban areas.

Effect of cities and urbanization in climate change

In [4], they found that urbanization exacerbates the probability of extreme droughts: by 2050, 50-70% of urban regions would consistently suffer droughts due to urban expansion

Effect of climate change on human behavior

Analyzing tens of millons of posts on social media in different cities, Baylis et al [5] found that colder weather is associated with more negative sentiment, while warmer weather is associated with more positive sentiment.

Climate change communication and misinformation.

In a very recent paper, [6], they studied how to persuade climate skeptics using surveys. They found that exposure to factual climate change news increase concern about climate change, challenging the belief that they are unmovable.

Climate change adaptation and resilience.

People change their behavior to adapt for climate change. For example, [7] found that outdoor activity is severly reduced by extreme heat, and it is delayed in time within the day

Natural disasters

CUS in Natural disasters

Large-scale behavioral data is an invaluable resource for understanding the impact of natural disasters on human mobility and social dynamics. By analyzing data from mobile devices, we can

  • Gain insights into how people respond to and recover from such events.
  • Understand the patterns of human mobility before, during, and after disasters.
  • Identify the factors that influence these patterns, such as population density, infrastructure, and socio-economic status.
  • Explore the role of social networks in shaping human behavior during disasters.
  • Investigate the impact of disasters on social interactions and community resilience.
  • Analyze the long-term effects of disasters on human mobility and social dynamics.

CUS in Natural disasters

The applications can be categorized into 4 main areas:

  • Population displacement and evacuation modeling
  • Longer-term recovery and resilience
  • Inference of damages to built environment
  • Social networks and community dynamics

Population displacement and evacuation modeling

Evacuation and displacement estimation using mobile phone data, social media, etc. can be used to:

  • Estimate the number of people displaced by a disaster
  • Identify the locations where people are likely to evacuate
  • Understand the factors that influence evacuation behavior

To:

  • Quick identification of post-disaster needs,
  • Planning of emergency supply distribution networks,
  • Pre-positioning of resources and supplies

Population displacement and evacuation modeling

Example: Lu et al. [8], used CDR to study predictability of displacement mobility patterns after the Haiti Earthquake in 2010. 1.9 million mobile phone users: estimated 23% of population in the capital displaced. But displacement destinations were highly correlated with pre-disaster mobility patterns (mobility predictability).

Population displacement and evacuation modeling

Wang et al. [9] found that those correlations extend to many other disasters, including hurricanes, thundersorms, typhoons, wildfire

Population displacement and evacuation modeling

A general framework to study displacement and evacuation patterns is to detect mobility anomalies and universalities in post-disaster movements. For example, Yabe et al [10], found some relationships between seismic intensity levels and evacuation rates. However, distances traveled are not correlated with the intensity of the earthquake.

|

Population displacement and evacuation modeling

Despite that universality, not all people are equally affected by evacuation: Deng et al [11] using LBS data found that race and wealth strongly impact evacuation patterns, with disadvantaged minorities being less likely to evacuate during hurricane Harvey.

Longer-term recovery and resilience

One of the advantages of large-scale data in CUS is the ability to track movements of users over a long period of time. This allows us to study the long-term effects of disasters on human mobility and social dynamics like:

  • Migration patterns after disasters
  • Community resilience and recovery
  • Business recovery
  • Economic impact of disasters

This allows:

  • Understanding the factors that influence recovery
  • Identifying the communities that are most vulnerable to disasters
  • Longer term infrastructure planning
  • Evaluating the effectiveness of disaster response and recovery efforts

Longer-term recovery and resilience

Example: use causal techniques to estimate the impact of disasters on business using human mobility data [12]

Longer-term recovery and resilience

Example: use causal techniques to estimate the impact of disasters on business using human mobility data [12] (using BSTS modeling)

Longer-term recovery and resilience

Using credit card transactions in Mexico to investigate the resilience of communities to hurricanes [13]

From BBVA AI Factory

Longer-term recovery and resilience

Many studies studied potential precursors of resilience to natural disasters using large-scale data to be able to identify what factors are correlated with resilience. For example, many papers [14] [15] [16] found that social capital is a key factor in determining the resilience of communities to disasters.

  • Households with denser personal networks experience quicker recovery.

Inference of damages to built environment

Another important application of CUS in natural disasters is the inference of damages to the built environment. This could help to:

  • Identify the areas that are most affected by disasters
  • Understand the factors that influence damage to buildings and infrastructure
  • Evaluate the effectiveness of disaster response and recovery efforts
  • Plan for future disasters, relief programs, insurance claims

Inference of damages to built environment

For example, in [17] they found that Twitter activity and sentiment around hurricane Sandy landing was correlated with per-capita economic damage.

Social networks and information dynamics

Another important application of CUS in natural disasters is the study of social networks and information dynamics [18]. This could help to:

  • Understand how information spreads during disasters
  • Control and understand missinformation
  • Identify how people use that information
  • Design better communication strategies for disaster response

Social networks and information dynamics

For example, in [19] they characterize and study the propagation of fake information about hurricane Sandy and who were propagating them

Conclusions

  • CUS is a powerful tool to study climate change and natural disasters
  • The availability of longtudinal data allows us to study the precursors, monitor the instantaneous and study long-term effects of disasters or climate change
  • The applications of CUS in climate change and natural disasters are diverse and can be used to inform policy decisions, improve disaster response, and enhance community resilience.

Further reading

  • Climate change:
    • Large-scale behavioural data are key to climate policy [1]
    • Magnitude of urban heat islands largely explained by climate and population [3]
    • Weather impacts expressed sentiment [5]
    • Growing polarization around climate change on social media [20]
  • Natural disasters
    • Mobile phone location data for disasters: A review from natural hazards and epidemics [21]
    • Rapid assessment of disaster damage using social media activity [17]
    • The Role of Social Networks and Information Sources on Hurricane Evacuation Decision Making [16]
    • Building Resilience: Social Capital in Post-Disaster Recovery [15]

References

CUS 2025, ©SUNLab group socialurban.net/CUS

[1]
M. A. Jenny and C. Betsch, “Large-scale behavioural data are key to climate policy,” Nature Human Behaviour, vol. 6, no. 11, pp. 1444–1447, Nov. 2022, doi: 10.1038/s41562-022-01479-4.
[2]
R. Debnath, S. van der Linden, R. M. Alvarez, and B. K. Sovacool, “Facilitating system-level behavioural climate action using computational social science,” Nature Human Behaviour, vol. 7, no. 2, pp. 155–156, Feb. 2023, doi: 10.1038/s41562-023-01527-7.
[3]
G. Manoli et al., “Magnitude of urban heat islands largely explained by climate and population,” Nature, vol. 573, no. 7772, pp. 55–60, Sep. 2019, doi: 10.1038/s41586-019-1512-9.
[4]
S. Huang et al., “Widespread global exacerbation of extreme drought induced by urbanization,” Nature Cities, vol. 1, no. 9, pp. 597–609, Sep. 2024, doi: 10.1038/s44284-024-00102-z.
[5]
P. Baylis et al., “Weather impacts expressed sentiment,” PLOS ONE, vol. 13, no. 4, p. e0195750, Apr. 2018, doi: 10.1371/journal.pone.0195750.
[6]
A. Tohidi, S. Balietti, S. Fraiberger, and A. Balietti, “Divergence between predicted and actual perception of climate information,” PNAS Nexus, vol. 4, no. 3, p. pgaf084, Mar. 2025, doi: 10.1093/pnasnexus/pgaf084.
[7]
Y. Fan, J. Wang, N. Obradovich, and S. Zheng, “Intraday adaptation to extreme temperatures in outdoor activity,” Scientific Reports, vol. 13, no. 1, p. 473, Jan. 2023, doi: 10.1038/s41598-022-26928-y.
[8]
X. Lu, L. Bengtsson, and P. Holme, “Predictability of population displacement after the 2010 Haiti earthquake,” Proceedings of the National Academy of Sciences, vol. 109, no. 29, pp. 11576–11581, Jul. 2012, doi: 10.1073/pnas.1203882109.
[9]
Q. Wang and J. E. Taylor, “Patterns and Limitations of Urban Human Mobility Resilience under the Influence of Multiple Types of Natural Disaster,” PLOS ONE, vol. 11, no. 1, p. e0147299, Jan. 2016, doi: 10.1371/journal.pone.0147299.
[10]
T. Yabe, Y. Sekimoto, K. Tsubouchi, and S. Ikemoto, “Cross-comparative analysis of evacuation behavior after earthquakes using mobile phone data,” PLOS ONE, vol. 14, no. 2, p. e0211375, Feb. 2019, doi: 10.1371/journal.pone.0211375.
[11]
H. Deng et al., “High-resolution human mobility data reveal race and wealth disparities in disaster evacuation patterns,” Humanities and Social Sciences Communications, vol. 8, no. 1, pp. 1–8, Jun. 2021, doi: 10.1057/s41599-021-00824-8.
[12]
T. Yabe, Y. Zhang, and S. V. Ukkusuri, “Quantifying the economic impact of disasters on businesses using human mobility data: A Bayesian causal inference approach,” EPJ Data Science, vol. 9, no. 1, pp. 1–20, Dec. 2020, doi: 10.1140/epjds/s13688-020-00255-6.
[13]
E. A. Martinez et al., “Measuring Economic Resilience to Natural Disasters with Big Economic Transaction Data,” arXiv.org. Sep. 2016. Accessed: Apr. 08, 2025. [Online]. Available: https://arxiv.org/abs/1609.09340v1
[14]
B. Völker, “Disaster recovery via social capital,” Nature Sustainability, vol. 5, no. 2, pp. 96–97, Feb. 2022, doi: 10.1038/s41893-021-00820-5.
[15]
D. P. Aldrich, Building Resilience: Social Capital in Post-Disaster Recovery. University of Chicago Press, 2012.
[16]
A. M. Sadri et al., “The role of social capital, personal networks, and emergency responders in post-disaster recovery and resilience: A study of rural communities in Indiana,” Natural Hazards, vol. 90, no. 3, pp. 1377–1406, Feb. 2018, doi: 10.1007/s11069-017-3103-0.
[17]
Y. Kryvasheyeu et al., “Rapid assessment of disaster damage using social media activity,” Science Advances, vol. 2, no. 3, p. e1500779, Mar. 2016, doi: 10.1126/sciadv.1500779.
[18]
A. M. Sadri, S. V. Ukkusuri, and H. Gladwin, “The Role of Social Networks and Information Sources on Hurricane Evacuation Decision Making,” Natural Hazards Review, vol. 18, no. 3, p. 04017005, Aug. 2017, doi: 10.1061/(ASCE)NH.1527-6996.0000244.
[19]
A. Gupta, H. Lamba, P. Kumaraguru, and A. Joshi, “Faking Sandy: Characterizing and identifying fake images on Twitter during Hurricane Sandy,” in Proceedings of the 22nd International Conference on World Wide Web, in WWW ’13 Companion. New York, NY, USA: Association for Computing Machinery, May 2013, pp. 729–736. doi: 10.1145/2487788.2488033.
[20]
M. Falkenberg et al., “Growing polarization around climate change on social media,” Nature Climate Change, vol. 12, no. 12, pp. 1114–1121, Dec. 2022, doi: 10.1038/s41558-022-01527-x.
[21]
T. Yabe, N. K. Jones, P. S. C. Rao, M. C. Gonzalez, and S. V. Ukkusuri, “Mobile phone location data for disasters: A review from natural hazards and epidemics,” Computers, Environment and Urban Systems, vol. 94, p. 101777, 2022.