Seasonal and interannual variability of precipitation and rainfall erosivity in Northeastern Japan: insights from 14 years of observations after Fukushima Dai-ichi nuclear accident (2011&ndash;2024)

Chalaux-Clergue, Thomas; Chaboche, Pierre-Alexis; Wakiyama, Yoshifumi; Evrard, Olivier

doi:10.5194/egusphere-2025-6030

Preprints

https://doi.org/10.5194/egusphere-2025-6030

Preprints

20 Jan 2026

| 20 Jan 2026

Status: this preprint is open for discussion and under review for Earth Surface Dynamics (ESurf).

Seasonal and interannual variability of precipitation and rainfall erosivity in Northeastern Japan: insights from 14 years of observations after Fukushima Dai-ichi nuclear accident (2011–2024)

Thomas Chalaux-Clergue, Pierre-Alexis Chaboche, Yoshifumi Wakiyama, and Olivier Evrard

Abstract. The Fukushima Dai-ichi Nuclear Power Plant (FDNPP) accident in March 2011, led to the deposition of about 2.0 PBq of ¹³⁷Cs across Fukushima Prefecture, resulting in widespread long-term environmental contamination. The accumulation of ¹³⁷Cs in the uppermost soil layer, led rainfall-driven soil erosion to become a major mechanism for ¹³⁷Cs redistribution across the landscape. Previous research (Laceby, et al., 2016b), which analysed data over 1995–2015, identified the June–October period as critical for rainfall erosivity. However, this analysis covered only the immediate five years period after the accident, leaving spatial and temporal patterns, event-scale dynamics, and long-term trends unquantified in the longer post-accidental context. To address this gap, the current research analysed 10-minute precipitation records from 58 weather stations located within a 110-km radius of the FDNPP, to include most of ¹³⁷Cs deposition onland, during the 14 years that followed the accident (2011–2024), calculating event-level precipitation amount and rainfall erosivity (EI₃₀), and aggregating metrics across annual, monthly, seasonal, and extreme-event timescales. Between 2011 and 2024, median annual precipitation was 1318 mm, with erosive events contributing 917 mm and generating 3,290 MJ mm ha^-1 h^-1 of erosivity. The June–October period accounted for 61 % of erosive events, representing 46 % of annual precipitation (602 mm) but 86 % of annual erosivity (2,698 MJ mm ha^-1 h^-1), with a clear precipitation-erosivity relation (r² = 0.62). Erosivity was highly concentrated during a few extreme events: the three most erosive events of each year contributed 55 % of total annual erosivity (1,753 MJ mm ha^-1 h^-1; 224 mm), whilst the single most erosive event alone accounted for 28 % (912 MJ mm ha^-1 h^-1; 94 mm). In 2011, 2015, and 2019, single events dominated annual erosivity, generated 38–44 % of annual erosivity. These findings quantify the erosive drivers of ¹³⁷Cs remobilisation since the accident with at large spatial and a high temporal resolution, supporting the retrospective reconstruction and prediction of long-term contamination dynamics. The predominance of extreme rainfall events emphasises their role in controlling sediment and contaminant transport in the post-accidental Fukushima context. All datasets, reproducible analytical workflows, and scripts (R package RainErosivity) are openly available via Zenodo and GitHub, supporting research on rainfall-driven ¹³⁷Cs redistribution, sediment transfer, erosion risk assessment and long-term geomorphological evolution of catchments.

Received: 03 Dec 2025 – Discussion started: 20 Jan 2026

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, 16750 KB)

Supplement (53 KB)

Download & links

Thomas Chalaux-Clergue, Pierre-Alexis Chaboche, Yoshifumi Wakiyama, and Olivier Evrard

Status: open (until 27 Apr 2026)

Post a comment Subscribe to comment alert

CC1:
'Comment on egusphere-2025-6030', Jr-Chuan Huang, 30 Mar 2026 reply
This manuscript presents a 14-year (2011-2024) characterization of precipitation and rainfall erosivity (EI₃₀) following the Fukushima Dai-ichi Nuclear Power Plant (FDNPP) accident across 58 weather stations within a 110 km radius of the Fukushima Dai-ichi Nuclear Power Plant (FDNPP). The work is technically fit, methodologically transparent, and the open-source RainErosivity package contribution is appreciated. I appreciate the long-term monitoring work. This work provides very fine spatiotemporal resolution of precipitation and ¹³⁷Cs dataset, which is very valuable to improve our understanding in erosivity, erodibility, and erosion. This work aligns with and extend the previous work (Laceby et al. (2016a)) and so the merit should address what we can learn from such high resolution dataset, but not just follow the previous one. For example, the difference between the erosivity derived from hour- or 10min-based calculations should be discussed. Moreover, what we can learn from the fine spatial resolution? Where is something special that only can be found in your dataset? What’s the interaction between erosivity and erodibility? So far, the authors put much effort to present their dataset without good digestion. Besides, the figures contain too much information which can’t show the most important merit within it. Please try to make the figures more meaningful.
Significance and Originality

The paper's primary originality lies in extending the Laceby et al. (2016a) dataset from 1995–2015 (five years after the accident) to a long-term period of 2011–2024, with finer event-scale resolution and a broader station network (58 stations). However, the manuscript largely confirms rather than advances the existing findings. The Discussion sections 4.1–4.3 repeatedly reported that results are "consistent with" prior studies. The limited expression of the new findings will affect the originality of this work. The abstract [P-1, L-10–22] does not clearly differentiate the novel contributions from existing knowledge. A concise, explicit statement of what this study discovers that was previously unknown would substantially strengthen the manuscript's readability.
Contribution to the Field

The connection between the erosivity metrics and their actual implications for ¹³⁷Cs dynamics is the motivation for the entire study. Section 4.4 [P-22–23, L-488–529] reads largely as a review of existing literature on ¹³⁷Cs transfers rather than as an original synthesis linking the new erosivity dataset to updated contamination dynamics. No new quantitative analysis of sediment or radiocesium fluxes is presented using the 2011–2024 erosivity data. Readers would benefit from at least a demonstration of how the ¹³⁷Cs export in the updated dataset during key erosive years, such as 2015 and 2019.
iii. Clarity and Organization
The manuscript is logically structured. However, several issues impair clarity. Figures 3, 4, 6, 8, and 9 are printed sideways (rotated 90°), making them extremely difficult to read. These multi-panel spatial maps are central to the paper's spatial analysis and would require correction before review is possible. The Results section is very data-dense, with extensive repetition of numerical ranges across subsections. For example, the values for the June–October seasonal contribution appear in near-identical form in the abstract, Results section 3.2.1, and Discussion section 4.2. Restructuring these repetitions would substantially improve readability.
Language and Readability

I am not a native speaker and so can’t judge it very precisely.

Technical Comments
[P-6, L-119–120]: The formula description states "et was calculated using the RUSLE2 corrected formula of Brown and Foster (1987) Yin et al. (2017) (0.082 vs. 0.05)." The parenthetical comparison of coefficients is unclear without explanation. Correcting this one would improve readability.
[P-7, L-150–151]: The significance threshold is described as: "pairs with a p value greater than α = 0.5 were considered to be non-statistically different." The threshold α = 0.5 appears to be a typographic error — it should almost certainly be α = 0.05. This is a critical methodological error if left uncorrected, as it drastically alters the interpretation of which year-pairs are considered statistically distinguishable. The authors must clarify the intended threshold and, if α = 0.5 is intentional, provide a strong justification for this highly unconventional choice.
[P-7, L-155–156]: The software version RStudio 2025.09.1+401 is cited, which post-dates the study period and suggests the manuscript was revised after the main analysis. If this version was used for final revisions, the version used for primary analysis should be clearly stated.
[P-16, L-351]: "in 2011, 2015, and 2015 the annual highest erosive event maxima…" — the year 2015 is repeated twice. This should be corrected.
[P-38, L-699]: The Nishii reference is incomplete: "Nishii: The 2022 Heavy Rainfall Disaster Caused by Stagnated Linear Rainbands in the Kaetsu Region, Niigata." No journal, volume, pages, DOI, or year to be provided. This reference must be completed.
Suggestions for Improvement
Abstract [P-1, L-1–22]: The abstract would benefit from one explicit sentence stating what is new relative to Laceby et al. (2016a), beyond merely covering a longer period. The current framing reads as an incremental update; refining the novelty statement would improve its acceptance.
Introduction [P-2–3, L-25–75]: The contextual framing of the ¹³⁷Cs problem is thorough but lengthy relative to the actual focus of the paper (precipitation/erosivity). Several paragraphs on radiocesium soil binding and plant transfer [L-32–38] are only tangentially relevant to the rainfall erosivity analysis and could be condensed to one paragraph.
Methods [P-5–7, Section 2.3–2.4]: The event identification criteria [P-5, L-108–111] are well described. Describing briefly would strengthen methodological credibility.
Section 2.4 [P-6, L-141–145]: The sign-test grouping methodology is novel, but the practical interpretation of groups (e.g., what it means to be in "group b, f, i") is not properly explained. A brief worked schematic in the supplementary material would help readers unfamiliar with this approach.
Data availability [P-24, L-565–572]: The data availability statement is exemplary, with Zenodo DOIs and GitHub links clearly provided, which is commended.
In summary:
The study offers a valuable update on the regional erosivity in the Fukushima area following the 2011 FDNPP accident. However, moderate revisions are recommended to:
(1) more clearly articulate novelty beyond incremental temporal extension
(2) strengthen the relationship between the new erosivity dataset and ¹³⁷Cs dynamics with at least some new quantitative synthesis

Reply
Citation: https://doi.org/10.5194/egusphere-2025-6030-CC1
- AC1: 'Reply on CC1', Thomas Chalaux-Clergue, 09 Apr 2026 reply
  
  Dear Jr-Chuan Huang,
  We sincerely thank you for your thorough and insightful review of our manuscript. Your comments will be valuable to improve the clarity, scientific rigour, and impact of our work. Below, we address the points that you raised in detail and outline the revisions that we propose to implement, should we be granted the opportunity to revise our manuscript.
  We fully acknowledge the challenges of presenting 14 years of high-resolution data across 58 stations while maintaining readability. To address the issues that you raised regarding figure clarity, we will reformat all figures into portrait orientation and optimise the arrangement of subpanels to maximise space and legibility. Given the quantity of data compiled, we are going to try to find a better subpanel arrangement in portrait format increasing each individual subpanel size. Additionally, we will provide high-resolution or scalable versions of all figures and individual subpanels in the supplementary materials, allowing readers to explore the data in greater detail.
  One of the key points that you raised was the need to emphasise the novelty of our work more clearly. Our primary objective was to identify the most relevant temporal scales (among annual, seasonal, event-based) for characterising precipitation and rainfall erosivity (EI₃₀) in the context of the FDNPP accident contamination redistribution. However, we recognise that we did not sufficiently highlight this in the abstract and throughout the manuscript, which we are going to improve. The investigation of different temporal scales allowed us to identify the most relevant scale to identify the occurrence of correlation between precipitation and rainfall erosivity (i.e. from seasonal scale), which was not dealt with in Laceby et al. (2016a).
  You also noted that the Results section is data-dense, with repetitive numerical ranges across subsections/section. We agree that this may reduce readability, and we plan to improve it by reporting only key statistics once in the Results section. Additionally, we recognize that reporting both medians and means it is not very clear for the readers. Nevertheless, we do believe that reporting both values provides valuable statistics to avoid low-frequency extreme events to disproportionately influence general trends (e.g., 2015 or 2019). While prior studies focus on means, our dual approach ensures comparability while highlighting the skewed nature of erosivity distributions in this region.
  Regarding the connection between erosivity and ¹³⁷Cs dynamics, we recognise that our initial focus was the investigation of the most appropriate scale to report about rainfall and erosivity metrics as interchangeable proxies. While we deliberately excluded ¹³⁷Cs monitoring data to avoid scale mismatches, we agree with the need to compare our dataset with published ¹³⁷Cs dynamics. In the revised manuscript, we are going to add such a comparison. However, as it was shown that individual events tend to show a rather high spatial variability, this regional dataset may then show some limitation regarding comparison with measured sediment and ¹³⁷Cs dynamics.
  We will also incorporate all the technical corrections that you highlighted (illustration of sign tests and event identification rules, correction of alpha = 0.05 not 0.5. RStudio version…).
  Again, we are grateful for your constructive feedback, which will help us strengthen the scientific contribution of our work, if the opportunity to revise our manuscript is granted by the editor. Thank you again for your time and expertise.
  Best regards,
  
  Thomas Chalaux-Clergue
  
  on behalf of the authors
  
  Reply
  
  Citation: https://doi.org/10.5194/egusphere-2025-6030-AC1
RC1:
'Comment on egusphere-2025-6030', Anonymous Referee #1, 10 Apr 2026 reply

The manuscript examines precipitation and rainfall erosivity over the period 2011–2024 within a 110 km radius of the Fukushima nuclear accident site. Overall, the paper is very well presented: the analyses are thoughtfully designed, and both the results and discussion are clearly articulated. With only a few minor suggestions, the manuscript appears close to being ready for publication.
That said, I have some reservations regarding whether the manuscript fully meets the scope and expectations of an original research article for EsfD. In its current form, the paper reads somewhat like a highly detailed meteorological report. It provides a comprehensive synthesis of station-based and regional trends, patterns, and anomalies in annual, seasonal, and event-scale precipitation and rainfall erosivity, contextualized within the Fukushima accident and the role of precipitation-driven soil erosion in mobilizing radiocesium.
However, the manuscript itself acknowledges that the remobilization of radiocesium through soil erosion and sediment transport has already been extensively studied. In this sense, the present work primarily contributes valuable data that may support future research, a point also noted in the discussion. At the same time, the study does not fully develop into a more in-depth climatological analysis. For instance, the finding that a small number of events account for the majority of annual erosivity, while relevant, is generally well established within the meteorological community. Although the discussion briefly highlights implications for erosion and radiocesium mobilization (and notes, for example, that much radiocesium is stored in forested areas that may be susceptible to landslides), these aspects are not explored in sufficient depth (the authors do mention that it is out of scope).
There are also some attempts to compare patterns across stations, but these analyses remain limited (an yield rather homogeneous results). A more detailed exploration, for example, of spatial variability or potential drivers such as orographic effects would strengthen the manuscript, particularly if it were to meet the standards of a more specialized meteorological or climatological study. Note that the research questions are very much oriented towards an application on geosciences (so lacking understandably lacking valid research question common to climatology), but at the same time the is no apparent geoscience application. The discussion mentions ways forwards for such geoscience applications. In addition, the paper would benefit from a stronger engagement with the broader international climatological literature, as the current framing is largely centered on studies conducted in Japan.
In summary, I am not fully convinced that the stated objective, namely “to assess major interannual trends and similarities, and to compare the relevance of annual versus seasonal scales in deriving representative erosivity metrics” sufficiently explores a scientific gap. The manuscript builds on previous work with a similar aim, particularly the paper of Laceby et al. (2016), which analyzed rainfall erosivity over the period 1995–2015. While the present study extends the temporal coverage to 2011–2024, it does so without fully integrating the earlier dataset. Given the overlap in authorship between the two studies, this appears to be a missed opportunity. In climatology, longer time series (e.g., ~30 years) are typically preferred to derive robust averages and trends.
To the editor. You may observe that Laceby et al 2026 has been able to publish their results in an EGU journal. But clearly that was soon after the Fukushima accident and had the advantage of urgency.
Laceby, J. P., Chartin, C., Evrard, O., Onda, Y., Garcia-Sanchez, L., and Cerdan, O.: Rainfall erosivity in catchments contaminated with fallout from the Fukushima Daiichi nuclear power plant accident, Hydrology and Earth System Sciences, 20, 2467–2482, https://doi.org/10.5194/hess-20-2467-2016, 2016a.
I leave it t the editor for a final advice. If the scope satisfies the journal scope, it is definitely ready for publication.

Reply

Citation: https://doi.org/10.5194/egusphere-2025-6030-RC1
- AC2: 'Reply on RC1', Thomas Chalaux-Clergue, 14 Apr 2026 reply
  
  Dear Referees and Editor,
  We would like to thank you sincerely for your thoughtful and constructive feedback on our manuscript. We are grateful for the time and expertise you have dedicated to evaluating our work, and we acknowledge the validity of your concern regarding its suitability for Earth Surface Dynamics and its suitability as a research paper.
  As you rightly note, our study is highly data-driven and descriptive, primarily focusing on updating and refining regional precipitation and erosivity metrics in the context of the FDNPP accident. Although we initially framed this work as a contribution to the geosciences in the context of the FDNPP accident, we acknowledge that the manuscript does not develop the climatological issues to the extent required.
  Nonetheless, we believe that our dataset and analysis are valuable in providing a more comprehensive, high-resolution (10-minute) characterisation of precipitation and rainfall erosivity in the Fukushima region to date. Our work extends the record by nearly a decade and captures recent and unprecedented extreme events (e.g. in 2015 and 2019), which have significantly influenced erosivity trends. It also establishes a baseline for future research, particularly for studies investigating sediment and radiocesium dynamics driven by erosivity at finer spatial and temporal scales. While we agree that the manuscript would be better suited to a report format, we believe that the underlying data will support ongoing and future work in the region.
  Taking the reviewers’ feedback into account, we have decided to withdraw the manuscript and revise it for submission to a more appropriate journal. We will incorporate suggestions to improve clarity, reduce redundancy and better contextualise the dataset's utility for geoscience applications.
  We are grateful to the reviewers for their insightful critiques, which have helped us to refocus the manuscript and identify a more suitable journal for publication. Their feedback has also highlighted ways in which we can strengthen the broader impact of our work, and we will ensure that future iterations of this research address these points.
  Thank you again for your time and consideration.
  Best regards,
  
  Thomas Chalaux-Clergue
  
  on behalf of the authors
  
  Reply
  
  Citation: https://doi.org/10.5194/egusphere-2025-6030-AC2
RC2:
'Comment on egusphere-2025-6030', Anonymous Referee #2, 13 Apr 2026 reply

This paper reports on the precipitation and rainfall erosivity from 2011 to 2024 within a 110 km radius of the Fukushima Dai-ichi Nuclear Power Plant. The results show that rainfall erosivity was highly concentrated during a few extreme events. Previous studies have used rainfall erosivity to assess the export of 137Cs, suggesting that extreme events play a dominant role in the transport of sediments and 137Cs. Therefore, the authors argue that longer-term monitoring of rainfall data can support future research on sediment and 137Cs transfers in the Fukushima region.
I agree with the opinion of another anonymous reviewer, who noted that this paper is more like a report on rainfall data, rather than scientific paper. The abstract and the body of the paper used considerable space to discuss the transport of 137Cs after the Fukushima Dai-ichi Nuclear Power Plant accident. However, the paper does not provide any data on the export of 137Cs after the accident, nor does it offer new insights into the flux or residence time of 137Cs export. The discussions are based on existing observations and do not provide substantial new understanding. The paper mainly outlines the temporal and spatial variability of precipitation and rainfall erosivity, with extreme events dominating erosivity. However, the connection between erosivity, sediment, and 137Cs is not established. Therefore, any claims regarding the influence of erosivity on the transport forms and flux of 137Cs are speculative. Based on this, I believe that this study is not suitable for publication in Earth Surface Dynamics.

Reply

Citation: https://doi.org/10.5194/egusphere-2025-6030-RC2
- AC3: 'Reply on RC2', Thomas Chalaux-Clergue, 14 Apr 2026 reply
  
  Dear Referees and Editor,
  We would like to thank you sincerely for your thoughtful and constructive feedback on our manuscript. We are grateful for the time and expertise you have dedicated to evaluating our work, and we acknowledge the validity of your concern regarding its suitability for Earth Surface Dynamics and its suitability as a research paper.
  As you rightly note, our study is highly data-driven and descriptive, primarily focusing on updating and refining regional precipitation and erosivity metrics in the context of the FDNPP accident. Although we initially framed this work as a contribution to the geosciences in the context of the FDNPP accident, we acknowledge that the manuscript does not develop the climatological issues to the extent required.
  Nonetheless, we believe that our dataset and analysis are valuable in providing a more comprehensive, high-resolution (10-minute) characterisation of precipitation and rainfall erosivity in the Fukushima region to date. Our work extends the record by nearly a decade and captures recent and unprecedented extreme events (e.g. in 2015 and 2019), which have significantly influenced erosivity trends. It also establishes a baseline for future research, particularly for studies investigating sediment and radiocesium dynamics driven by erosivity at finer spatial and temporal scales. While we agree that the manuscript would be better suited to a report format, we believe that the underlying data will support ongoing and future work in the region.
  Taking the reviewers’ feedback into account, we have decided to withdraw the manuscript and revise it for submission to a more appropriate journal. We will incorporate suggestions to improve clarity, reduce redundancy and better contextualise the dataset's utility for geoscience applications.
  We are grateful to the reviewers for their insightful critiques, which have helped us to refocus the manuscript and identify a more suitable journal for publication. Their feedback has also highlighted ways in which we can strengthen the broader impact of our work, and we will ensure that future iterations of this research address these points.
  Thank you again for your time and consideration.
  Best regards,
  
  Thomas Chalaux-Clergue
  
  on behalf of the authors
  
  Reply
  
  Citation: https://doi.org/10.5194/egusphere-2025-6030-AC3

Thomas Chalaux-Clergue, Pierre-Alexis Chaboche, Yoshifumi Wakiyama, and Olivier Evrard

Supplement

https://doi.org/10.5194/egusphere-2025-6030-supplement

Data sets

Precipitation and rainfall erosivity datasets for Northeastern Japan over 2011-2024: 10-minute precipitation records, rainfall event characteristics, and processed summary tables (year, month, typhoon season, single most and three most erosive events) from 58 JMA weather stations located within a 110-km radius around Fukushima Dai-ichi Nuclear Power Plant. Thomas Chalaux-Clergue et al. https://doi.org/10.5281/zenodo.17573969

Model code and software

RainErosivity: Customisable Tools to Calculate Precipitation Event Rainfall Erosivity Index Thomas Chalaux-Clergue https://doi.org/10.5281/zenodo.14745960

Thomas Chalaux-Clergue, Pierre-Alexis Chaboche, Yoshifumi Wakiyama, and Olivier Evrard

Viewed

Total article views: 309 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
187	104	18	309	26	21	30

HTML: 187
PDF: 104
XML: 18
Total: 309
Supplement: 26
BibTeX: 21
EndNote: 30

Views and downloads (calculated since 20 Jan 2026)

Month	HTML	PDF	XML	Total
Jan 2026	107	34	12	153
Feb 2026	28	8	2	38
Mar 2026	42	52	3	97
Apr 2026	10	10	1	21

Cumulative views and downloads (calculated since 20 Jan 2026)

Month	HTML	PDF	XML	Total
Jan 2026	107	34	12	153
Feb 2026	28	8	2	38
Mar 2026	42	52	3	97
Apr 2026	10	10	1	21

Viewed (geographical distribution)

Total article views: 288 (including HTML, PDF, and XML) Thereof 288 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 14 Apr 2026

Short summary

Long term analysis of precipitation and rainfall erosivity provide insights into erosive rainfall regime could affect ¹³⁷Cs redistribution in Fukushima-impacted catchments. Over 2011–2024, June–October produced 46 % of precipitation and 86 % of erosivity, driven by few intense events. The three most erosive events contributed 55 % of annual erosivity, and the most exceeding 40 %. This study provides long-term basis for regional assessments of rainfall-driven erosion and contaminant transfer.


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