An extended hazard interaction matrix for exploring multi-hazard complexity in data-scarce regions: An application to Kerala, India

Desai, Anisha; Barendrecht, Marlies; Jalayer, Fatemeh; Taylor, Faith E.

doi:10.5194/egusphere-2026-2999

Preprints

https://doi.org/10.5194/egusphere-2026-2999

Preprints

19 Jun 2026

| 19 Jun 2026

Status: this preprint is open for discussion and under review for Natural Hazards and Earth System Sciences (NHESS).

An extended hazard interaction matrix for exploring multi-hazard complexity in data-scarce regions: An application to Kerala, India

Anisha Desai, Marlies Barendrecht, Fatemeh Jalayer, and Faith E. Taylor

Abstract. This paper extends the analysis of multi-hazard interrelationships beyond the primary focus on cascading and amplifying mechanisms, through the development of an evidence-based database in a data-scarce context. The methodology is applied to Kerala, India. To support interpretation, an adapted hazard interaction matrix was developed that extends existing frameworks by (i) incorporating a broader range of interaction mechanisms beyond traditional cascading and amplifying effects, and (ii) enabling representation of three-way hazard interactions, advancing beyond conventional pairwise depictions. The matrix was further enhanced to capture seasonal variation in interaction potential throughout the year. Drawing on academic literature, grey literature, and media sources, the database captures evidence for both well-documented and underreported hazards and their interactions, whether historically observed or theoretically possible. The final database contains evidence of 22 distinct hazard types across six hazard groups and 137 potential hazard interrelationships. Results indicate that, while cascading and disposition-alteration mechanisms dominate the interrelationships observed in Kerala, accounting for additional interaction mechanisms increases the number of identified interrelationships by 36 %. This suggests that restricting analyses to a limited subset of interaction types may not fully capture the region's multi-hazard complexity. The matrix was further enhanced to capture seasonal variation in interaction potential throughout the year. Incorporating seasonality reveals distinct temporal windows of elevated interaction potential shaped by monsoon rainfall and temperature variability. When applying seasonal filters, the number of potential interrelationships identified was reduced by approximately 6%. This study demonstrates that interaction-focused, seasonally informed frameworks can reveal multi-hazard dynamics that may otherwise be overlooked when analysing only a subset of hazard types and interaction mechanisms

Received: 25 May 2026 – Discussion started: 19 Jun 2026

Competing interests: At least one of the (co-)authors is a member of the editorial board of Natural Hazards and Earth System Sciences.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 6382 KB)

Supplement (107 KB)

Download & links

Anisha Desai, Marlies Barendrecht, Fatemeh Jalayer, and Faith E. Taylor

Status: open (until 31 Jul 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-2999', Anonymous Referee #1, 06 Jul 2026 reply
I would like to thank the authors for a very interesting manuscript; I very much enjoyed reading it. I agree with the authors that detailed local case studies such as this one are novel and very valuable, and that they help address the global imbalance in multi-hazard research towards data-rich regions. I also really like the visualizations; in particular Figures 2, 7 and 10 are a very nice way of conveying a lot of information, and I can see the seasonal matrix being a useful communication tool. I do have some general comments that I would like to exchange thoughts on with the authors, mostly around positioning the extended matrix within recent classification work and the interpretation of some of the headline numbers.
General comments
The authors could reflect on the recently published TST (Type-Space-Time) classification of hazard interrelations by Wenzel et al. (2026, Natural Hazards).
For example, Figure 2 of the manuscript covers what Wenzel et al. call the type dimension (and the seasonal matrix covers the temporal dimension), but the spatial dimension of interactions is currently absent: the matrix aggregates over all of Kerala. Could the authors reflect on how spatial (non-)overlap within the study region affects which of the interrelationships are actually plausible, similar to how they use seasonal (non-)overlap to filter out interactions?

It would also nhelp to briefly position the six mechanisms from De Angeli et al. (2022) relative to the TST types also given the authors’ own point (Section 1.1.1) that different terminology hinders comparison across studies.

Related to this: the manuscript uses “interaction”, “interrelationship” and “interaction mechanism” somewhat interchangeably. Given that the introduction (rightly) flags the lack of standardised terminology as a gap, it would be good to define these terms explicitly early on and use them consistently.

I would be curious to read a slightly more in-depth reflection on the intended use of the matrix in practice. Section 5.4 positions it as an exploratory tool for scenario development and stakeholder engagement, and engagement with KSDMA is mentioned as future work; can the authors say a bit more about how they envisage a practitioner actually using Figures 7, 10 and 11 (which are rich but also dense, see also below minor comment), and whether for example an interactive/digital version of the matrix is foreseen?

How are mechanisms without temporal ordering (parallel, coincident triggering, cyclic triggering) placed in the asymmetric hazard-one/hazard-two matrix, and how are they counted in the link counts of Figure 6? Are these counted once or twice (i.e. in both the upper and lower triangle)? This matters for the interpretation of the stated percentages (32% and 50% of all links).

Minor comments
Abstract: the sentence “The matrix was further enhanced to capture seasonal variation in interaction potential throughout the year” appears twice (around L. 6–7 and L. 13–14).

176: the sentence “More broadly, Kerala provides a useful case study for exploring how interaction-focused” is incomplete.

175: “structure inference”: maybe “structured inference”?

145: the formatting of the population density (860 persons km-2) appears to be negative; please check the superscript.

There are several confusing cross-references left in the text: “Table 3.1” (L. 223 and the caption of Figure 3) presumably should be “Table 1”, and “Figure 6.1” (L. 463) and “Table 6.1” (L. 466) presumably should be Figure 3 and Table 1?

Abstract L. 15: “reduced by approximately 6%”: in the text this is 9 of 137 interrelationships (≈6.6%) showing no seasonal overlap; it would help to state in the abstract what the 6% refers to (interactions ruled out for temporally coincident mechanisms only, not removed altogether).

Similarly, for the 36% in the abstract, it would help to state the baseline explicitly (i.e. relative to the 101 cascading + disposition-alteration interrelationships).

Figure 7 is impressive but very dense; maybe check legibility

Please check reference Bureau, 2025 (?)

Suggested literature
Wenzel, T., van Westen, C., Sunil, M., Pantaleoni Reluy, N., Marr, P., Glade, T., Bell, R. (2026): Towards a practical multi-hazard interrelation classification: implications for assessing their impacts. Natural Hazards 122, 82. https://doi.org/10.1007/s11069-025-07901-0

Reply
Citation: https://doi.org/10.5194/egusphere-2026-2999-RC1

Anisha Desai, Marlies Barendrecht, Fatemeh Jalayer, and Faith E. Taylor

Supplement

https://doi.org/10.5194/egusphere-2026-2999-supplement

Anisha Desai, Marlies Barendrecht, Fatemeh Jalayer, and Faith E. Taylor

Viewed

Total article views: 9 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
3	6	0	9	0	1	0

HTML: 3
PDF: 6
XML: 0
Total: 9
Supplement: 0
BibTeX: 1
EndNote: 0

Views and downloads (calculated since 19 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	3	6	0	9

Cumulative views and downloads (calculated since 19 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	3	6	0	9

Viewed (geographical distribution)

Total article views: 3 (including HTML, PDF, and XML) Thereof 3 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 13 Jul 2026

Short summary

Natural hazards rarely occur in isolation. We studied Kerala, India, to see how hazards can affect one another and how these influences change over time. Using published studies, reports, news and social media, we identified 22 hazard types and 137 possible links between them. Looking beyond the usual ways hazards are connected revealed 36% more relationships, while adding seasonal filters narrowed these by 6%. This study helps to highlight multi-hazard dynamics that may otherwise be overlooked.


Total:	0
HTML:	0
PDF:	0
XML:	0