the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Jul 2025
Usage of normalized soil moisture for improving the performance of rainfall thresholds for landslides along transportation corridors
Abstract. Landslides along transportation corridors pose significant risks to infrastructure and public safety, necessitating accurate prediction and mitigation strategies. Many early warning systems for landslides are based on rainfall thresholds derived from historical data that distinguish landslide triggering from non-triggering events. However, it is widely recognized that antecedent moisture conditions have a major impact on the likelihood of a particular rainfall event leading to a landslide. We aim to improve existing rainfall thresholds for landslides along highways by incorporating antecedent soil moisture conditions. The landslide inventory was compiled using data from inclinometers at suspected landslide sites and from landslide reports following major storm events along Alabama highways. This inventory was combined with precipitation data from the National Oceanic and Atmospheric Administration (NOAA) and soil moisture data from NASA’s Soil Moisture Active Passive (SMAP) satellite. We explored the accuracy of rainfall thresholds from previous studies for forecasting landslides along the highways of Alabama. Additionally, we investigated the potential of reducing the number of non-landslide events that exceed the thresholds (false positives) by utilizing soil moisture data derived from SMAP. This study demonstrates that sites with multiple inclinometers in a landslide region produce more robust datasets compared to those with a single inclinometer, enabling more effective differentiation between landslide and non-landslide events. Furthermore, using normalized soil moisture in the development of rainfall thresholds shows potential for reducing false positives, as approximately 75 percent of the false positive cases in this study occurred when the soil moisture was at or below average conditions. Our proposed normalized soil moisture-dependent thresholds will support decision-making systems by enabling users to weigh the tradeoffs between potential false alarms and missed alarms, depending on the relative cost or risk of each for a given project. The findings will aid transportation authorities and civil engineers in making informed decisions about possible interventions or preventative maintenance in the future.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(2460 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2460 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-3419', Anonymous Referee #1, 12 Oct 2025
- AC1: 'Reply on RC1', Leila Rahimikhameneh, 24 Dec 2025
-
RC2: 'Comment on egusphere-2025-3419', Anonymous Referee #2, 29 Oct 2025
Peer Review Comments
This study introduces the concept of incorporating normalized soil moisture (NSM) into rainfall-triggered landslide threshold models to enhance early warning performance. The integration of long-term inclinometer data, NOAA precipitation records, and NASA SMAP soil moisture observations provides a valuable multidisciplinary framework. However, from the standpoint of scientific rigor and generalizability, the manuscript still presents notable limitations in terms of data representativeness, statistical validation, spatial-scale consistency, and geotechnical mechanism interpretation. Major revision is required before the manuscript can be considered for publication. The authors are encouraged to strengthen the temporal resolution, address spatial resolution discrepancies, and provide quantitative uncertainty analyses to improve the reliability and applicability of the conclusions.
-The manuscript frequently uses the first-person pronoun, which should be removed for formal scientific writing. For instance,
“We aim to improve existing rainfall thresholds for landslides along highways by incorporating antecedent soil moisture conditions.”
should be revised to
“This study aims to improve existing rainfall thresholds for landslides along highways by incorporating antecedent soil moisture conditions.”
-The use of quarterly inclinometer readings results in low temporal resolution. This makes it impossible to accurately link specific rainfall events with landslide occurrences, severely weakening the temporal correspondence between rainfall–duration (I–D) thresholds and soil moisture response, and consequently reducing statistical correlation and causal interpretability.
-The manuscript does not clarify the training background or consistency evaluation of the operators involved in inclinometer data acquisition, which could introduce subjective bias.
-The spatial resolution of the SMAP data (9 km × 9 km) is too coarse to represent local-scale topographic and soil moisture variability, limiting the applicability of the model in mountainous areas with small-scale landslides.
-The sample size and regional representativeness are limited. The dataset is small and highly localized, resulting in weak generalizability of the conclusions and making it difficult to substantiate the claim of a “transferable threshold.”
-Figure 3 fails to clearly demonstrate the relationship between rainfall, soil moisture, and landslide occurrence. The graphical evidence does not convincingly support the authors’ interpretation and should be clarified with enhanced visualization or statistical quantification.
-Geological and soil parameters were not quantitatively controlled. Although the manuscript classifies strata into three lithologic types, key physical parameters (e.g., permeability, cohesion, plasticity index) are not incorporated, which weakens the geotechnical basis of the proposed thresholds.
-The innovation is more phenomenological than mechanistic. The study focuses on statistical correlation without sufficient exploration of the underlying hydro-geotechnical processes that govern the observed trends, reducing the theoretical depth of innovation.
-The applicability boundaries of the proposed approach are not explicitly discussed. This omission reduces the methodological rigor and limits understanding of the model’s valid domain.
-The rainfall event classification is overly simplified, and the threshold selection may be too lenient, as it does not account for the independence of consecutive dry periods or short-duration, high-intensity rainfall events.
-The discussion section focuses primarily on whether prediction accuracy improved, but lacks an in-depth analysis of data sources, model structure, and scale compatibility. A more comprehensive discussion of these factors is needed; substantial revision of this section is recommended.
-Performance metrics are reported without confidence intervals or statistical significance testing. Consequently, the claimed improvements cannot be validated statistically. The authors should incorporate cross-validation or independent testing.
-The normalization procedure may obscure extreme moisture conditions. Averaging across long periods can reduce contrast between very wet and very dry states, thereby weakening the detection of extreme antecedent conditions that critically influence landslide initiation.
Citation: https://doi.org/10.5194/egusphere-2025-3419-RC2 - AC2: 'Reply on RC2', Leila Rahimikhameneh, 24 Dec 2025
-
RC3: 'Comment on egusphere-2025-3419', Yichuan Zhu, 06 Nov 2025
The manuscript presents research on improving existing rainfall thresholds for landslide prediction along highways by incorporating antecedent soil moisture conditions. The authors established an inventory containing landslide and non-landslide events, precipitation data from NOAA, and soil moisture data from NASA's SMAP. Rainfall thresholds from the literature for landslide forecasting were examined using the inventory data. Furthermore, the research proposed incorporating normalized soil moisture into the development of rainfall thresholds, which shows potential for reducing false positives in prediction. The work is highly practical and will be of interest to practitioners in landslide assessment and management. However, the overall quality of the manuscript needs improvement before it can be accepted for publication. Below are the comments on the manuscript:
Major Comments:
- In the Discussion section, the authors mention spatial resolution issues, which remain a significant concern for the current study. NASA's SMAP operates at a 9-km by 9-km resolution, while CONUS data has a spatial resolution of 28-km by 28-km. How do these resolutions align with the site-specific study presented in the manuscript?
- The soil moisture data from NASA's SMAP satellite may require calibration before incorporation into the working pipeline. Based on the reviewer's experience, systematic bias between SMAP and in-situ soil moisture monitoring can exhibit seasonal patterns. It would strengthen the manuscript if the authors could provide additional justification regarding this potential issue.
- The current work adopts a previous threshold of 5 mm to distinguish landslide from non-landslide events. While a reference is provided, it would be beneficial to include rationale for this threshold in the current manuscript. From the reviewer's perspective, whether internal movement of 5 mm should be classified as "landslide" is worth discussion. Such movement could represent only localized slope displacement rather than strain bifurcation or connection into a plastic zone. Please justify why the 5 mm threshold is effective for classifying sites as landslide locations.
- Regarding performance metrics, is it possible within the current research framework to plot a Receiver Operating Characteristic (ROC) curve and compute the Area Under the Curve (AUC)?
- The current work uses normalized soil moisture-dependent thresholds. How does this approach affect the uncertainty or sensitivity of predictions across space? Future work could include variogram or Bayesian analysis to investigate spatial uncertainty.
Minor Comments:
- Please add references for Pandas, NumPy, OS, and Matplotlib as a way to support the open-source community.
- For Figures 2 and 3, consider plotting moving averages to better illustrate seasonal or annual changes in soil moisture.
- In Figures 6 and 7, the legend shows the thresholds as shaded blocks, while the plot presents them as lines. Please make these representations consistent.
- In Figure 9, the legend notation "1.25 to<2.15" reads awkwardly. Consider using a simpler format such as "1.25–2.15."
Citation: https://doi.org/10.5194/egusphere-2025-3419-RC3 - AC3: 'Reply on RC3', Leila Rahimikhameneh, 24 Dec 2025
Peer review completion
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-3419', Anonymous Referee #1, 12 Oct 2025
General comment
The manuscript aims to optimize rainfall thresholds by using soil moisture data. The topic is of high interest as it deals with the interaction of those phenomena with road infrastructures. The idea of improving those thresholds is of high interest, too. However, the manuscript needs further improvements before its acceptance. Its structure of the manuscript is not well defined and needs careful rewriting as several sections contain sentences belonging to other parts. The following lists show the other, major and minor, issues referenced with line(s) numbers.
Major comments
- Lines 99-119 contain a summary of the materials, methods, and obtained results. I suggest that authors remove them and relocate, if appropriate, the sentences in the following sections.
- Line 142 mentions Python and several libraries; I’m wondering if it is necessary to add a citation for each one.
- Line 146: How did you choose the threshold?
- Line 148: Manual review of inclinometer data is a time-consuming and subjective process.
- Lines 150-154: Please consider Allasia et al 2020 (https://doi.org/10.3390/s20133769) in the error assessment.
- Lines 158-159: Unclear, please improve the explanation.
- Line 169: Are the stratigraphy obtained during the drilling available? This will help the classification of the sites.
- Sections 2.2 and 2.3 are too generic, and several repetitions are made in the text. Additionally, precipitation and soil moisture data are very coarse and feature different resolutions; it is not clear how the authors have faced this issue while integrating them with punctual data from inclinometers.
- Figure 5: Is the landslide outline mapped? If so, it’s better to add it to the figure to explain the behavior of the whole phenomenon more effectively.
- Lines 261-267 belong due to their content to the “Results” section.
- Lines 301-310 belong to the 2.2 subsection.
Minor comments:
- The use of the term “significant” is confusing; does it refer to the result of a statistical test or used as a synonym of "remarkable" or "noteworthy"? Please clarify.
- Lines 120-121 mention inclinometer readings; are they carried out manually? Considering the frequency, I suppose so, but it’s better to specify because Robotized (https://doi.org/10.3390/s20133769) or Automated (https://doi.org/10.3390/s23208379) approaches are also available.
- Line 140 is not necessary; please remove.
- Line 152: “+1.24 mm” à “±24 mm”.
- Figure 1: Remove the typo in the title and move the label “B” to avoid the overlay with the line.
- Figure 2 needs to be improved by adding a grid and coordinates. Moreover, authors should give, graphically, emphasis to the monitored highways.
- Lines 276-284 need to be synthesized and repetitions removed.
- Figures 6-12 are slightly blurred. Additionally, from my very personal point of view, the color choice is not ideal.
- Figure 8: Correct the label below the two plots and fix the typo in the legend of the (a) panel, i.e., “landslide” à “Landslide”.
Citation: https://doi.org/10.5194/egusphere-2025-3419-RC1 - AC1: 'Reply on RC1', Leila Rahimikhameneh, 24 Dec 2025
-
RC2: 'Comment on egusphere-2025-3419', Anonymous Referee #2, 29 Oct 2025
Peer Review Comments
This study introduces the concept of incorporating normalized soil moisture (NSM) into rainfall-triggered landslide threshold models to enhance early warning performance. The integration of long-term inclinometer data, NOAA precipitation records, and NASA SMAP soil moisture observations provides a valuable multidisciplinary framework. However, from the standpoint of scientific rigor and generalizability, the manuscript still presents notable limitations in terms of data representativeness, statistical validation, spatial-scale consistency, and geotechnical mechanism interpretation. Major revision is required before the manuscript can be considered for publication. The authors are encouraged to strengthen the temporal resolution, address spatial resolution discrepancies, and provide quantitative uncertainty analyses to improve the reliability and applicability of the conclusions.
-The manuscript frequently uses the first-person pronoun, which should be removed for formal scientific writing. For instance,
“We aim to improve existing rainfall thresholds for landslides along highways by incorporating antecedent soil moisture conditions.”
should be revised to
“This study aims to improve existing rainfall thresholds for landslides along highways by incorporating antecedent soil moisture conditions.”
-The use of quarterly inclinometer readings results in low temporal resolution. This makes it impossible to accurately link specific rainfall events with landslide occurrences, severely weakening the temporal correspondence between rainfall–duration (I–D) thresholds and soil moisture response, and consequently reducing statistical correlation and causal interpretability.
-The manuscript does not clarify the training background or consistency evaluation of the operators involved in inclinometer data acquisition, which could introduce subjective bias.
-The spatial resolution of the SMAP data (9 km × 9 km) is too coarse to represent local-scale topographic and soil moisture variability, limiting the applicability of the model in mountainous areas with small-scale landslides.
-The sample size and regional representativeness are limited. The dataset is small and highly localized, resulting in weak generalizability of the conclusions and making it difficult to substantiate the claim of a “transferable threshold.”
-Figure 3 fails to clearly demonstrate the relationship between rainfall, soil moisture, and landslide occurrence. The graphical evidence does not convincingly support the authors’ interpretation and should be clarified with enhanced visualization or statistical quantification.
-Geological and soil parameters were not quantitatively controlled. Although the manuscript classifies strata into three lithologic types, key physical parameters (e.g., permeability, cohesion, plasticity index) are not incorporated, which weakens the geotechnical basis of the proposed thresholds.
-The innovation is more phenomenological than mechanistic. The study focuses on statistical correlation without sufficient exploration of the underlying hydro-geotechnical processes that govern the observed trends, reducing the theoretical depth of innovation.
-The applicability boundaries of the proposed approach are not explicitly discussed. This omission reduces the methodological rigor and limits understanding of the model’s valid domain.
-The rainfall event classification is overly simplified, and the threshold selection may be too lenient, as it does not account for the independence of consecutive dry periods or short-duration, high-intensity rainfall events.
-The discussion section focuses primarily on whether prediction accuracy improved, but lacks an in-depth analysis of data sources, model structure, and scale compatibility. A more comprehensive discussion of these factors is needed; substantial revision of this section is recommended.
-Performance metrics are reported without confidence intervals or statistical significance testing. Consequently, the claimed improvements cannot be validated statistically. The authors should incorporate cross-validation or independent testing.
-The normalization procedure may obscure extreme moisture conditions. Averaging across long periods can reduce contrast between very wet and very dry states, thereby weakening the detection of extreme antecedent conditions that critically influence landslide initiation.
Citation: https://doi.org/10.5194/egusphere-2025-3419-RC2 - AC2: 'Reply on RC2', Leila Rahimikhameneh, 24 Dec 2025
-
RC3: 'Comment on egusphere-2025-3419', Yichuan Zhu, 06 Nov 2025
The manuscript presents research on improving existing rainfall thresholds for landslide prediction along highways by incorporating antecedent soil moisture conditions. The authors established an inventory containing landslide and non-landslide events, precipitation data from NOAA, and soil moisture data from NASA's SMAP. Rainfall thresholds from the literature for landslide forecasting were examined using the inventory data. Furthermore, the research proposed incorporating normalized soil moisture into the development of rainfall thresholds, which shows potential for reducing false positives in prediction. The work is highly practical and will be of interest to practitioners in landslide assessment and management. However, the overall quality of the manuscript needs improvement before it can be accepted for publication. Below are the comments on the manuscript:
Major Comments:
- In the Discussion section, the authors mention spatial resolution issues, which remain a significant concern for the current study. NASA's SMAP operates at a 9-km by 9-km resolution, while CONUS data has a spatial resolution of 28-km by 28-km. How do these resolutions align with the site-specific study presented in the manuscript?
- The soil moisture data from NASA's SMAP satellite may require calibration before incorporation into the working pipeline. Based on the reviewer's experience, systematic bias between SMAP and in-situ soil moisture monitoring can exhibit seasonal patterns. It would strengthen the manuscript if the authors could provide additional justification regarding this potential issue.
- The current work adopts a previous threshold of 5 mm to distinguish landslide from non-landslide events. While a reference is provided, it would be beneficial to include rationale for this threshold in the current manuscript. From the reviewer's perspective, whether internal movement of 5 mm should be classified as "landslide" is worth discussion. Such movement could represent only localized slope displacement rather than strain bifurcation or connection into a plastic zone. Please justify why the 5 mm threshold is effective for classifying sites as landslide locations.
- Regarding performance metrics, is it possible within the current research framework to plot a Receiver Operating Characteristic (ROC) curve and compute the Area Under the Curve (AUC)?
- The current work uses normalized soil moisture-dependent thresholds. How does this approach affect the uncertainty or sensitivity of predictions across space? Future work could include variogram or Bayesian analysis to investigate spatial uncertainty.
Minor Comments:
- Please add references for Pandas, NumPy, OS, and Matplotlib as a way to support the open-source community.
- For Figures 2 and 3, consider plotting moving averages to better illustrate seasonal or annual changes in soil moisture.
- In Figures 6 and 7, the legend shows the thresholds as shaded blocks, while the plot presents them as lines. Please make these representations consistent.
- In Figure 9, the legend notation "1.25 to<2.15" reads awkwardly. Consider using a simpler format such as "1.25–2.15."
Citation: https://doi.org/10.5194/egusphere-2025-3419-RC3 - AC3: 'Reply on RC3', Leila Rahimikhameneh, 24 Dec 2025
Peer review completion
Journal article(s) based on this preprint
Data sets
Inventory of Landslides Along Alabama Highways L. Rahimikhameneh et al. https://www.designsafe-ci.org/data/browser/public/designsafe.storage.published/PRJ-5979
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 9,130 | 1,213 | 239 | 10,582 | 210 | 248 |
- HTML: 9,130
- PDF: 1,213
- XML: 239
- Total: 10,582
- BibTeX: 210
- EndNote: 248
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Abraham Alvarez Reyna
Jack Montgomery
Frances O'Donnell
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2460 KB) - Metadata XML
General comment
The manuscript aims to optimize rainfall thresholds by using soil moisture data. The topic is of high interest as it deals with the interaction of those phenomena with road infrastructures. The idea of improving those thresholds is of high interest, too. However, the manuscript needs further improvements before its acceptance. Its structure of the manuscript is not well defined and needs careful rewriting as several sections contain sentences belonging to other parts. The following lists show the other, major and minor, issues referenced with line(s) numbers.
Major comments
Minor comments: