the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Jan 2026
Spatiotemporal reorganization of global earthquake disaster impacts within coupled human-Earth systems
Abstract. Understanding how earthquake disaster impacts reorganize across space and time is essential for interpreting seismic hazards within coupled human-earth systems. Using global disaster records from EM-DAT spanning 1980–2024, this study examines multi-scale spatiotemporal patterns of earthquake disaster impacts and their socio-environmental associations at global and national scales. Temporal analyses show a pronounced decoupling between seismic occurrence and disaster consequences: while earthquake frequency, exposed population, and cumulative economic losses increased overall, mortality rates declined markedly after the early 2000s. Spatial analyses reveal strong heterogeneity across continents, countries, and major tectonic plates. Asia accounts for a substantial share of global earthquake occurrences, affected populations, fatalities, and economic losses, yet national-level impacts vary considerably even under comparable tectonic settings. Standard deviation ellipse and centroid analyses further indicate an eastward to southeastward migration of the global earthquake disaster centroid over time, accompanied by relatively stable orientation and a modest contraction in spatial dispersion. To explore factors associated with national differences in fatalities, a Geographical Detector model is applied using cumulative fatalities as the dependent variable and a set of natural, climatic, socioeconomic, governance, infrastructure, and health-related variables as explanatory factors. Results show that population density and development- and governance-related indicators exhibit relatively high explanatory power, while interactions among factors generally strengthen spatial associations through bilinear or nonlinear enhancement. Overall, the findings suggest that global disparities in earthquake disaster impacts reflect the spatial co-configuration of hazard exposure, development conditions, and institutional capacity, contributing to a system-level understanding of how seismic disaster impacts evolve within coupled human-Earth systems.
- Preprint
(1493 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 05 Mar 2026)
-
CC1: 'Comment on egusphere-2025-6499', Yongji Wang, 14 Jan 2026
reply
-
AC1: 'Reply on CC1', Yanjun Ye, 31 Jan 2026
reply
We sincerely thank the commenter for this very positive and thoughtful evaluation of our study. We greatly appreciate the recognition of the risk-oriented perspective and of the emphasis placed on human systems in shaping earthquake outcomes.
Our aim was to move beyond a purely hazard-based interpretation and to provide system-level evidence that disaster impacts emerge from the interaction between seismic activity and societal conditions. We are grateful for the acknowledgement of the multi-method framework, which was designed to integrate temporal trends, spatial patterns, and explanatory factors in a coherent way.
We believe that this type of integrative perspective is essential for advancing both scientific understanding and practical approaches to disaster risk reduction. Thank you again for engaging with our work and for contributing to this open and constructive scientific discussion.
Citation: https://doi.org/10.5194/egusphere-2025-6499-AC1
-
AC1: 'Reply on CC1', Yanjun Ye, 31 Jan 2026
reply
-
CC2: 'Comment on egusphere-2025-6499', Yannan Wang, 14 Jan 2026
reply
Very interesting paper. It provides new method and discussion about the spatiotemporal reorganization of global earthquake disaster. I hope it can be published in the journal. Based on my own background, I feel a little puzzled about the strong strain belt. Does it correspond to both the active margin and passive margin?
Citation: https://doi.org/10.5194/egusphere-2025-6499-CC2 -
AC2: 'Reply on CC2', Yanjun Ye, 31 Jan 2026
reply
We sincerely thank the commenter for the positive evaluation of our study and for the encouraging remarks regarding its potential contribution.
Regarding the question on the “strong strain belt”, in this study the term is used in a broad, global-scale sense to describe regions characterized by concentrated tectonic deformation and high seismic activity. It primarily corresponds to active plate boundaries, including both convergent margins (such as subduction zones and continental collision zones) and major transform fault systems. Passive margins, which are generally characterized by low tectonic strain and limited seismicity, are not included in what we refer to as strong strain belts in this analysis.
We appreciate this opportunity to clarify the terminology and will consider further improving the description in the manuscript to avoid potential ambiguity.
Thank you again for your interest in and support of our work.
Citation: https://doi.org/10.5194/egusphere-2025-6499-AC2
-
AC2: 'Reply on CC2', Yanjun Ye, 31 Jan 2026
reply
-
RC1: 'Comment on egusphere-2025-6499', Anonymous Referee #1, 30 Jan 2026
reply
The paper by Zhang et al investigates changes in earthquake impacts in the 1980-2024 time interval. The EM-DAT us used as data source and spatial and temporal changes are analyzed using metrics associated with cumulative or normalized deaths, affected people and economic losses.
The topic of the contribution is relevant, given the importance of seismic events as sources of damage. The analysis integrating coupled human-Earth systems is interesting, however I have found some flaws that may affect the results and interpretations. These deal with data completeness, both in terms of investigated time interval and data compilation. The authors briefly mention the limitations of their approach in section 5.2.5, while I would like to see a broader discussion on the limitations.
Some of the metrics used to infer the conclusions are strongly affected by a few large-scale events. For this reason, I think that some extra effort should be devoted to data analysis and cross-validation (e.g., leave one out or moving average windows approaches). I also encourage the authors in providing some thoughts on the causal relations between the drivers and metrics.
Below I list my major comments, which I hope may be useful in the revision stage.
Title: I find “earthquake disaster impacts” redundant; is the word disaster needed? Consider changing throughout the text to “earthquake impacts”
Data completeness. I acknowledge EM-DAT is probably the best database to run this kind of analysis, but I encourage the authors in explicitly mention eventual issues on data completeness. Three out of the 8 metrics are normalized over the number of events, making the database completeness a critical factor for the reliability of the obtained results.
Time interval. Is the time interval 1980-2024 enough to address the research question? In many regions of the world, recurrence intervals of large earthquakes are much longer than the considered time interval; thus, the analyses may be flawed by an incomplete sampling. Indeed, all panels in Figure 1 show prominent peaks, related to specific earthquakes (e.g., Sumatra 2004, Wenchuan 2008, Haiti 2010, Tohoku 2011, Turkey-Syria 2023), which strongly affect the average values and thus the authors’ interpretation.
In line with the previous comment, I encourage the authors to exploring statistically how a single event affects the overall trend. For instance, 2023 is characterized by a peak in the number of affected people and fatalities. If the analyses are run on the 1980-200 interval, how temporal trends would be affected?
Geometric centroid calculation. At line 144, you mention that the weight W is set to 1. Have you considered the possibility to assign a weight based on the metrics (e.g., affected people, deaths, mortality rates)? The events included in the database have huge differences in terms of losses, as can be observed in figure 1. Assigning an equal weight obscures such differences, possibly driving to misleading conclusions. In other words, the equal weight basically means that the centroid migration is related to the rate of seismicity (of the earthquakes included in the database). For such an analysis, a more complete seismic catalogue, regardless of the earthquake impact, may be more suitable.
Lines 183-184. How have you assessed that events in 1980s are 2020s are equally recorded in the EM-DAT? Can record incompleteness, especially in the first part of the time interval, bias the trend?
Section 4.1.2. If you considered the intervals 1980-1994; 1995-2010; 2011-2024, they are 15,16 and 14 years long, respectively. This is quite a gross overlooking, the intervals should be of equal length, otherwise the conclusions regarding the cumulative numbers are biased.
Figure 2. I have the impression that if you adopt the intervals 1995-2009 and 2010-2024 the peaks in the second period would be much reduced. Indeed the Haiti earthquake (2010) has a strong influence on the number of fatalities (either cumulative or per event) and mortality rates.
Section 4.2.2. Have you considered the possibility of running the analyses on economic losses metrics by normalizing the data on the countries’ GDP?
Figure 5. it is quote hard to geographically locate the panels. Consider adding the country boundaries.
Section 4.2.3. you should mention the limitations of this analysis, regarding for instance the relatively short time interval analyzed with respect to recurrence intervals on many faults around the globe.
Table 3. add the references for the data sources, to facilitate the reproducibility of your analyses.
Section 4.3. Maybe I missed something, but are the driving factors considered over the 1980-2024 interval? Or did you use the most recent available data? I suppose that the variation of the driving factors over time is spatially heterogeneous.
Lines 307-312. Same concept repeated twice.
Citation: https://doi.org/10.5194/egusphere-2025-6499-RC1 -
AC3: 'Reply on RC1', Yanjun Ye, 31 Jan 2026
reply
We sincerely thank the anonymous referee for the careful reading of our manuscript and for the detailed and constructive comments. We appreciate the positive evaluation of the relevance of the topic and of the coupled human–Earth system perspective adopted in this study.
The issues raised regarding data completeness, temporal representativeness, and the influence of extreme events are highly relevant for global-scale analyses of earthquake impacts. We acknowledge that EM-DAT, while currently the most comprehensive global disaster database available, is affected by spatial and temporal heterogeneities in reporting, particularly in the earlier part of the study period. These aspects require careful consideration when interpreting long-term trends and deserve explicit discussion.
We also agree that some of the metrics considered may be sensitive to a small number of high-impact earthquake events. The potential influence of individual events on aggregated temporal patterns needs to be carefully considered in the interpretation of results and appropriately addressed in the analysis and discussion.
With respect to the geometric centroid analysis, we appreciate the comment on the use of equal weights. In this study, the centroid migration was intended to illustrate the spatial reorganization of reported earthquake impacts at the global scale. At the same time, we agree that alternative weighting schemes based on impact-related metrics can provide a useful complementary perspective, and this point will be more clearly discussed.
We further thank the referee for the specific suggestions regarding temporal binning, normalization of economic losses, figure readability, documentation of data sources, and the treatment of driving factors. These comments are very helpful and will be carefully considered to improve the clarity, robustness, and transparency of the manuscript.
Overall, we are grateful for these insightful remarks, which will help us strengthen both the analysis and the interpretation in the revision stage.
Citation: https://doi.org/10.5194/egusphere-2025-6499-AC3 -
AC4: 'Reply on RC1', Yanjun Ye, 31 Jan 2026
reply
We sincerely thank the anonymous referee for the careful reading of our manuscript and for the detailed and constructive comments. We appreciate the positive evaluation of the relevance of the topic and of the coupled human–Earth system perspective adopted in this study.
The issues raised regarding data completeness, temporal representativeness, and the influence of extreme events are highly relevant for global-scale analyses of earthquake impacts. We acknowledge that EM-DAT, while currently the most comprehensive global disaster database available, is affected by spatial and temporal heterogeneities in reporting, particularly in the earlier part of the study period. These aspects require careful consideration when interpreting long-term trends and deserve explicit discussion.
We also agree that some of the metrics considered may be sensitive to a small number of high-impact earthquake events. The potential influence of individual events on aggregated temporal patterns needs to be carefully considered in the interpretation of results and appropriately addressed in the analysis and discussion.
With respect to the geometric centroid analysis, we appreciate the comment on the use of equal weights. In this study, the centroid migration was intended to illustrate the spatial reorganization of reported earthquake impacts at the global scale. At the same time, we agree that alternative weighting schemes based on impact-related metrics can provide a useful complementary perspective, and this point will be more clearly discussed.
We further thank the referee for the specific suggestions regarding temporal binning, normalization of economic losses, figure readability, documentation of data sources, and the treatment of driving factors. These comments are very helpful and will be carefully considered to improve the clarity, robustness, and transparency of the manuscript.
Overall, we are grateful for these insightful remarks, which will help us strengthen both the analysis and the interpretation in the revision stage.
Citation: https://doi.org/10.5194/egusphere-2025-6499-AC4
-
AC3: 'Reply on RC1', Yanjun Ye, 31 Jan 2026
reply
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 132 | 43 | 27 | 202 | 13 | 8 |
- HTML: 132
- PDF: 43
- XML: 27
- Total: 202
- BibTeX: 13
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This is an exceptionally rigorous and insightful study. By demonstrating a clear decoupling between earthquake frequency and mortality after the 2000s, it powerfully shifts the narrative from hazard to risk, emphasizing the critical role of human systems. The innovative multi-method approach—spanning temporal trends, spatial centroid migration, and factor detection—provides compelling, system-level evidence that disaster outcomes are shaped by the complex interplay of exposure, development, and governance. The findings are both academically significant and immediately relevant for global risk reduction strategies.