the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Sep 2024
Derivation of Moisture-Driven Landslide Thresholds for Northeastern Regions of the Indian Himalayas
Abstract. Landslides pose a significant threat in the Northeastern Himalayas, driven by monsoonal rains and exacerbated by rapid urbanization. This research establishes moisture (primarily rainfall) thresholds that can cause landslides in Northeastern Himalayas 'hotspots' based on 490 rain-driven landslides catalogued between 2006 and 2019. Coupling the innovative Regularized Expectation-Maximization approach with non-crossing quantile regression, we reveal critical insights into antecedent moisture conditions and their role in shallow to deep landslide genesis. Our derived moisture threshold for the Northeastern Himalayan region, E (mm) = −11.10 + 0.62 D (hour), for 24 < D < 1440 hr, fits within global bounds for both deep and shallow landslides. The spatial analysis demonstrates significant heterogeneity, with Guwahati (located at 26.14° N, 91.74° E in Assam) and Shillong (located at 25.58° N, 91.89° E in Meghalaya) requiring higher cumulative rainfall for landslide triggers compared to Aizwal (located at 23.73° N, 92.72° E in Mizoram). Our analysis shows that environmental controls, e.g., elevation, slope, land use types, CND, and rock types, play significant roles in shaping rainfall thresholds to trigger landslides. The insights from this research offer effective landslide risk management strategies and advance the predictive capabilities of Landslide Early Warning Systems with broader implications for climate resilience and disaster preparedness.
- Preprint
(1937 KB) - Metadata XML
-
Supplement
(1255 KB) - BibTeX
- EndNote
Status: open (until 25 Dec 2024)
-
RC1: 'Comment on nhess-2024-152', Anonymous Referee #1, 10 Oct 2024
reply
Dear Authors,
The work proposes an approach for defining rainfall thresholds for “moisture-driven landslide” (MDL) forecasting in the Himalayas. The authors calculated the soil moisture content empirically, by multiplying the antecedent cumulative rainfall by a decay constant 𝑘=0.9. This approach is not physically rigorous, and while it may be acceptable for defining empirical thresholds, the limitations were not sufficiently clarified.
The definition of MDL provided in the introduction is too limited. I suggest adding details on the types of landslides considered (whether deep or shallow) and the type of movement (whether slow or fast landslides, debris flows, or slides). Additionally, two landslide catalogs were used, but only MDLs were collected from them. Since the definition of MDL is unclear to me, do these catalogs specifically contain this type of landslide, or were other similar types grouped together?
Finally, for a clearer understanding of the study's methodology, I suggest briefly mentioning in the introduction how you intend to extrapolate soil moisture from the antecedent rainfall. This missing information might lead to think that the antecedent rainfalls were directly used without deriving the soil moisture.
It is frequently stated in the manuscript that ED thresholds were purposely used to account for antecedent rainfall, something that ID thresholds do not allow. I’m not sure about that, as ED thresholds, like ID thresholds, are event-based, and they depend on how the duration is defined. For D=15 days, an E is calculated over the 15 days, and the same applies to I, with the difference that I is distributed over the entire duration (mm/day in this case). I suggest to provide a different justification for using ED threshold instead of ID.
I found paragraph 3.2, which relates to the methodology for defining rainy days, quite confusing. It was decided to define a day as rainy when at least 10% of the seasonal rainfall occurs. I think this method has high limitations, because, as highlighted in the text, 10% of the monsoon season's rainfall can be quite high (it was calculated to be about 13mm). In my opinion, this threshold is too high, as daily rainfall of 10mm would not be considered as a rainy event, which means that 30mm over three days would not be considered. In my opinion, they could still have an impact in the context of MDLs and could help refine the rainfall thresholds. I suggest providing further justification for this method and highlighting its limitations.
The methodology for defining the lag-time, particularly in identifying the trade-off between correlation and scattering bias of triggering vs antecedent rainfall, is not clear to me and also needs further clarification.
In the discussion and conclusion sections, you frequently refer to landslide susceptibility ("This study provides crucial insights into the landslide susceptibility of the NEH region"). However, this term is out of context, as susceptibility refers to spatial predisposition, while rainfall thresholds are used for temporal forecasting (in the case of this study, spatialized within 30km around each rain gauge). I would suggest you revise this part.
Best regards,
Anonymous Reviewer
Citation: https://doi.org/10.5194/nhess-2024-152-RC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 130 | 30 | 179 | 339 | 16 | 5 | 6 |
- HTML: 130
- PDF: 30
- XML: 179
- Total: 339
- Supplement: 16
- BibTeX: 5
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1