the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Jun 2026
The physical and cultural traces of a 19th century volcano-collapse tsunami: New constraints on the magnitude and impacts of the 1888 Ritter Island tsunami, Papua New Guinea
Abstract. The lateral collapse of Ritter Island volcano, Papua New Guinea, on 13th March 1888, generated a devastating tsunami with observed effects at distances of up to 500 km. As the largest recorded volcano lateral collapse, the event is significant in improving our general understanding of volcano instability and associated hazards. Proximal records of the tsunami are limited to post-event reports of extensive coastal damage of nearby islands. Here, we present new field observations of tsunami deposits on these coastlines, supplemented by and interpreted alongside oral accounts from coastal communities (from 2004) and a reanalysis of contemporary reports. Our data reveal the widespread presence of massive and chaotic conglomerates deposited by the tsunami at sites up to 30 km from Ritter, with thicknesses of up to 2.0 m and extending up to 400 m from the shoreline. Such deposits may be indicative of large-magnitude landslide-generated tsunamis, and are generally structureless, coarsest and thickest near the coast, and contain mixed terrigenous and marine clasts, including corals, shells and benthic foraminifera. Isolated large coral blocks are also widespread tsunami inundation markers, reaching >200 m from the shore. The deposits extend to approximately half the maximum inundation distance indicated by oral accounts; the latter preserve precise detail consistent with the timing, spatial characteristics and impacts of the Ritter tsunami interpreted from prior geological surveys, distal accounts and simulations. Collectively, this study indicates that the impacts of the 1888 tsunami on proximal shorelines were larger than has been previously inferred from post-event reports. The maximum run-up height was likely many tens of metres on Sakar and Umboi and extensively exceeded 20 m on western New Britain. From the collated descriptions, we also estimate that ~2000–3000 deaths were caused directly by the tsunami. These new observations add to previous submarine surveys of the event, allowing the coupled process of landslide motion and tsunami generation at Ritter Island to be investigated more comprehensively.
- Preprint
(1256 KB) - Metadata XML
-
Supplement
(485 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-2101', Anonymous Referee #1, 18 Jul 2026
-
AC1: 'Reply on RC1', Sebastian Watt, 09 Sep 2026
RESPONSE:
We thank the reviewer for their constructive comments and positive notes on the content and presentation of the manuscript. On their specific major comments:
Reliability of oral accounts: It is correct that any oral accounts are removed from the original event by 2-3 generations (although we note that most accounts are derived from elderly individuals who generally stated they were two generations removed, and in one instance, only one generation), and that this may lead to information gaps, potential changes in detail due to transmission within and between communities, or potential complexities in distinguishing details that may relate to other events (which may have been older or younger). This cannot be fully accounted for, but we consider that the derivation of compatible event details from accounts spanning multiple communities, across multiple islands, gives confidence that the information preserved within these oral histories is a robust reflection of the events observed at the time. This robustness is implied by various observations: for example, accounts on New Britain (to the east) consistently refer to a sea withdrawal, but this is not the case in accounts from Umboi and elsewhere. These are differences which are fully consistent with what is expected for a landslide-sourced tsunami entering the ocean to the west, implying that specific information can be accurately preserved within the region’s oral histories. Details of relative timing is also generally consistent across all accounts; in the few instances where these were not, we note that the accounts may represent a younger tsunami, and did not include those accounts within our broader analysis. We also note that specific local geographic details recorded within individual accounts (e.g. the water reaching a specific local landmark) are plausible and consistent with modelled wave impacts (e.g. Karstens et al., 2019), and that these specific details themselves imply the continuity of transmission within local communities. It is thus the collective set of information from multiple independent accounts that lends confidence to the details preserved within them.
More generally, we sought to highlight the value of the oral histories as a significant additional source of information, which doesn’t replace traditional field methods but adds to and enhances insights that can be obtained from geological observations. Most oral accounts did not provide specific quantitative detail, but where they did, we considered it useful to present and compare these against other observation types, which was the intention of Figure 6. The insights obtained from geological datasets and oral histories are not, however, directly equivalent but are complementary. For example, the maximum inundation is expected to substantially exceed the maximum extent of chaotic deposits (as has been observed in historical tsunamis). The oral accounts described here corroborate this, with all estimates from oral accounts exceeding local deposit measurements. This observation lends further strength to our contention that meaningful specific information is preserved within oral histories. This information is not precise, but that does not imply that it is inaccurate. There is not an objective way to estimate what this precision is, but the accuracy can be evaluated by interpreting this information alongside geological field data, as described above. To explore these challenges and the value of information within the oral histories, we propose to add some brief additional discussion around the precision and consistency of quantitative information in the oral accounts. We would also be happy to add a table that summarises quantitative detail from these accounts and sets it alongside the field data (essentially, the data presented in Figure 6), if this enables a clearer comparison of different information types.
Other landslide-related tsunamis: We would be happy to expand the discussion on comparable historical events, including the examples mentioned. Nevertheless, we note that the scale of the Ritter collapse distinguishes it from other historical events, but agree that additional discussion would add useful context. Examples that can be drawn on include tsunamis at Stromboli and larger events at Anak Krakatau, Oshima-Oshima and Harimkotan (but for most historical events, specific details on tsunami impacts and on associated deposits are often very limited, which is again a distinctive feature of the constraints we are now able to place on the Ritter event).
Small sketches: We agree that the presentation of Fig 3 in the pre-print version of the article seen by reviewers is not of the quality we had intended, and apologise for not recognising this ahead of submission. The figures themselves were produced at higher resolution, but the compressed versions in the file are not easily legible. We suggest that this issue will be resolved by us providing the higher resolution original figures. Fig 3 and 4 were designed to show multiple logs on a single plot and enable comparison, and this is our preferred format. However, we would be happy to split these into more figures if this would improve legibility further (although with the higher resolution images this may not be an issue). One benefit of this would also be that it would enable us to show photos next to the plot (see the response below and Reviewer 2 comments). The depicted logs are already a subset of all the studied field sites, but if needed, we can provide some of these logs these as supplementary data (or alternatively, keep the summary figures, but provide larger versions of each individual log as supplements).
Campaign photos: We can see the value of providing higher resolution and larger photos of specific field sites next to the relevant log, and we would be able to reorganise figures 3 and 4 to show the information in this way (see above comment). The comment on the grainsize distributions is reasonable and it was also noted by Reviewer 2 that we didn’t draw on this information. We didn’t provide the data in a quantitative format, and can certainly provide this as a supplementary table. We will also add brief discussion of what insights can be drawn from this data, with additional comparison with chaotic tsunami deposits described in the literature.
Source island imagery: We agree that providing additional imagery of Ritter island would help the reader understand the broader context of the site. We suggest producing an introductory figure that combines pre-event sketches of the island (from Jacobs; cf. Ward and Day, 2003) and subsequent aerial images of the island, to highlight the morphological changes and scale of the collapse for readers who are not familiar with the event.
Discursive results section: We acknowledge that the results sections contain some discursive elements. To some extent, we would argue this is necessary to retain a clear narrative within the paper, particularly because it draws on combinations of qualitative (oral history) and quantitative detail, seeking to bring these together in a way that can be easily followed by the reader. We would be happy to revisit the text in these sections to both shorten them and to minimise discursive elements, moving interpretation as far as possible to the subsequent section.
Citation: https://doi.org/10.5194/egusphere-2026-2101-AC1
-
AC1: 'Reply on RC1', Sebastian Watt, 09 Sep 2026
-
RC2: 'Comment on egusphere-2026-2101', Gianfilippo De Astis, 22 Jul 2026
This paper deals with relevant scientific and/or technical questions within the scope of NHESS.
May be the review I made is in an intermediate position between minor and major revisions.
Anyway, since I believe that is closer to minor, I selected minor. The "poor presentation quality" is especially related to some Figures and the photos contained in these Figures. I think that it should be easily recovered.
All my comments, notes and suggesions - both general and specific on the text or figures of this manuscript - are reported in the attached .pdf file containing some general comments in the beginning and then a critical review page by page.
The pdf file is: Egusphere-2026-2101_Full Gianfi review.I hope I have been helpful and stimulating in the growth and final publication of this important multi-disciplinary article.
Thanks and my best wishes to the Authors and the journal,G. De Astis
-
AC2: 'Reply on RC2', Sebastian Watt, 09 Sep 2026
We thank the reviewer for their positive comments and constructive suggestions.
The main overarching comment regarded poor presentation quality, particularly of photos. We acknowledge that this is an issue within the pre-print. We hadn’t identified the poor resolution of Fig. 3 in the original submission (see comment above) and can certainly easily resolve this. With the photos, we can provide higher resolution and larger images of the field photographs than those which were initially submitted, although there are unfortunately limits to this because the photos were taken on a film camera in 2004 and all we have available are scans from the original images. The resolution of these is good enough to illustrate the detail that we seek to show and to highlight the macroscopic features of the tsunami deposits, which is why we decided to draw on them within the paper despite their limited resolution. The photos in Fig. 5 can certainly be shown at a larger size, and this will improve clarity, and we can also provide additional photos as supplementary files. For the field photos (Fig 2 and 5) we are satisfied that this will show details of the deposits clearly. Unfortunately, the foraminifera photos in Fig 2 cannot be significantly improved; they are all that remains available from the original work examining these after the 2004 fieldwork. We felt they were important to include, because they unequivocally demonstrate the presence of marine microfossils within the deposit – a key indicator of a marine derived depositional process – but acknowledge that they are relatively low-quality microscope images. Given their importance in helping confirm the interpretation of these units as tsunami deposits, we would still like to include them even though they are relatively low resolution.
Specific comments in the separate PDF file are also addressed here:
The reviewer noted:
This paper is notable for the integration of new field surveys on the onshore tsunami deposits and for the systematic analysis of oral histories of local communities collected in 2004. Important developments have been made regarding: - the actual height of the run-up and the trimline (vegetation stripping limit); - - - the death toll (victims): a historically very uncertain number of deaths is now better fixed; the importance of comparing and integrating Oral Histories vs. Geological data; Identification of "Chaotic Conglomerates" (Proximal Chaotic Conglomerates) and Layered Conglomerates due to Backwash Flow. And more things about 1888 event are also improved… Conversely, I don’t see any convincing steps forward regarding the "conflict" on eruptive activity before or during the sector collapse of the Ritter volcano (i.e. the presence of submarine vs. onshore pumice clasts). Although colonial accounts of the time reported that the coasts of New Britain and Kelana were covered with pumice after the disaster, this study indicates that no traces of it were found onshore. The lack of pumices in the deposits observed on land and related to the 1888 tsunami is one of the most surprising findings presented in this new paper (2026), which therefore challenges some previous interpretations. Yet, there is no trace of this discovery in the Conclusions and this is surprising.
We thank the reviewer for these insightful comments and are pleased that the new constraints on the Ritter lateral collapse and tsunami have been recognised. The comment on evidence of accompanying eruptive activity identifies an important ambiguity that remains about this event. This does indeed have broader significance regarding the association between volcano lateral collapse and magmatic activity, and in determining the relationship between volcano failure and eruptive processes. We discuss this in Section 4.5 (line 493 onwards), but would be happy to expand this discussion, and to also address these observations within the abstract and conclusions. The absence of pumice clasts within the deposits is intriguing, and implies that any explosive activity preceding the event was insufficient to have caused extensive fallout or floating pumice rafts. This is not unexpected, given that Ritter’s documented activity has produced mafic magmas and would not be expected to produce large-magnitude pumice-forming eruptions. However, the submarine samples described in Watt et al. (2019) highlight that a vesicular, silicic component is present in the uppermost (youngest) part of the submarine deposits, providing an intriguing hint that a pumice-forming explosive event may have followed the collapse, and thus have been triggered by it (i.e. unloading of the edifice enabling eruption of a shallow silicic magma). Such an observation is potentially compatible with the contemporary post-event observations or shorelines impacted by the tsunami (i.e. a post-collapse pumice-forming explosive eruption would be likely to produce floating pumice rafts, which may have accumulated on nearby shorelines which had already been devastated by the prior tsunami, and may be explain observations of pumice on shorelines referred to by post-event surveyors. Such deposits, derived from the stranding of a pumice raft formed after the tsunami, would be coastal. They would not have been preserved within or carried on top of the chaotic tsunami deposits documented here. What is clear is that there is no evidence of a coarse-grained pumiceous fallout deposit on top of the tsunami deposits. This is also compatible with any accompanying pumice-forming eruption occurring after the collapse, because by that stage the vent was several hundred metres below sea level and would not have formed an atmospheric eruption column or airfall deposit. We can be relatively confident that any airfall tephra deposit would have been identified, since tsunami deposit exposures have been examined in multiple directions, from Tolokiwa to New Britain, with no evidence being found for such an airfall event. We further note that there is no evidence of such deposits in marine sampling from 2016 (Watt et al., 2019), and that the contemporary post-event surveys documented by Steinhauser describe undamaged inland forests and villages, sharply contrasting with the coastal destruction, with no reference to the type of destruction that would be expected from an extensive tephra fall deposit, suggesting that an airfall origin for any pumice is unlikely.
Based on the above considerations, observations of pre-collapse volcanic activity (fire, explosions) can be best explained as volcanic activity typical of Ritter (i.e. basaltic magmas, Strombolian styles; cf. Johnson, 2013), perhaps in line with the intense but not atypical activity observed at Anak Krakatau ahead of its collapse in 2018 (Cutler et al., 2022). The descriptions imply that the volcano may have been erupting at the time of its failure. Such activity (i.e. Strombolian eruptions of mafic magma) is not likely to have produced widely dispersed or thick tephra fall deposits, particularly at the field-site distances which are tens of kilometres from the vent. We also note again that the post-event surveys, which document the devastated coastline, make no reference to forest damage beyond this coastal region, which would be expected in the case of a widespread fall deposit (e.g. defoliation of trees due to ashfall).
The only contemporary observation that is difficult to reconcile with the above interpretations is that of a “fine, barely perceptible ash”, reported by Steinhauser in Finschhafen. While it could be argued that vigorous explosive eruptions of basaltic magma in the build up to the collapse could have led to very fine ash reaching such distances (e.g. fine ash was deposited in Java, 60 km away, from activity at Anak Krakatau in 2018), this seems unlikely given both the distance (120 km) and the column height that would be required to transport ash to such distances, and it would also imply much thicker deposition (and accompanying deposits and vegetation damage) closer to source. Again, we note this was not observed in the post-event survey by Steinhauser, and that no impacts of this type are referred to in the oral accounts. What is also clear is that the reference to a fine, barely perceptible ash couldn’t be from an eruption accompanying the collapse, because the travel times would be very different (i.e. the report is of ash accompanying the tsunami, but ash transport through the atmosphere would have been substantially slower than the tsunami propagation). Balancing these various observations, and noting that there are no other observations of tephra deposition (from proximal oral histories, distal contemporary observations or post-event field surveys, or geological deposits), we conclude that there is no evidence for a large-magnitude pre-collapse eruption. The account from Finschhafen thus remains difficult to interpret. One possibility is that Steinhauser’s reference to ashfall was influenced by phenomena that he expected to see, if his expectations were influenced by knowledge of the Krakatau eruption and tsunami, occurring just five years previously. Potential sources of bias within all the accounts need to be considered in any interpretation, but we again note that the self-consistent details identified across multiple oral histories from different locations lends weight to inferences drawn from these.
We would be happy to expand Section 4.5 with details from the above discussion.
The reviewer also noted:
Finally (and including the paragraph 4.5), it’s not clear if the Authors of the article have at least considered phreatomagmatism (and not a subplinian/Plinian eruption) as a possible contributory cause in the dynamics of Ritter’s collapse. Since a specific study was performed on the sediments composing the sandy matrix of the chaotic coastal conglomerates from the 1888 event, and the authors do not report the presence of fresh juvenile volcanic ash (co-eruptive from 1888) within the analysed matrix, it would be fair to conclude that syn-collapse (eruptive) activity is highly unlikely. However, doubts remain regarding the ash layers (tephra) on which the tsunami deposit rests in several locations (Barang and Kampalap, at Umboi, Fig.2) or the pumice fallout in New Guinea, which are reported in some articles (e.g. Karstens et al. 2019). What would then be the origin of these layers of ash? These doubts are further supported by several accounts in the supplementary material (to name just a few): - smoke and fire at Kabi (Umboi), where at dawn (between 4-6 a.m.), a woman went to the beach and saw "smoke and fire coming from Ritter." This observation prompted the entire village to evacuate inland before the tsunami, saving almost the entire population; - historical data collected in New Britain (particularly Sackley's 1974 account mentioned in the text) indicate that the collapse was preceded by 2-3 days of continuous eruptive activity; - in Alairo (Sakar): a very clear chronology has been handed down where the inhabitants first saw the Ritter explode and, subsequently, saw the western half of the island collapse and end up in the sea; - the phenomenon of bizarre fire-like (brick-red) and "Roman soldier helmet" shaped flares on wave crests is consistently reported in several distant villages, such as Kilenge and Tauali on the west coast of New Britain, and Alairo on the island of Sakar: how is this explained other than by an optical reflection of the explosive eruption on the approaching wave crests? How else? Therefore, the Ritter collapse is not clearly documented as a purely gravitational "cold" event (such as classic Bandai, from the same year...) and there are probabilties to have been triggered and/or associated with phreatomagmatic activity. Do the Authors think that was possible or not? Is the Ritter’s event a “hybrid” and more complex case than the classic and purely "cold" model of Bandai? In my opinion, a clear position on this temporal and causal relationship between the cone collapse and the eventual explosive activity must emerge. Given that, the study and in-depth knowledge of the data is the prerogative of the Authors and I can only suggest slightly different and more critical perspectives, with the sole aim of hopefully enriching this scientific article.
We thank the reviewer for these additional comments and agree with them that this collective body of descriptions in the oral accounts lend strong support to the occurrence of some form of explosive volcanism in the build-up to and, possible, during the event. Our observations here builds on our comments above. It is clear from much earlier navigational reports that Ritter was a highly active volcano (Johnson, 2013), dominated by mafic magmas (as has been confirmed by petrological studies; e.g. Watt et al. 2019) and typified by broadly Strombolian eruption styles. It would not be surprising if the collapse had occurred during an eruption (again, some analogies could be drawn with Anak Krakatau; Cutler et al., 2022), and we agree that the multiple references in oral accounts of sounds, fire, smoke etc. lend strong support to this. Indeed, it is difficult to argue for an absence of prior/accompanying volcanic activity, given these multiple reports. Whether these events were magmatic or phreatomagmatic cannot easily be distinguished. The oral observations were made at distances of ~12-30 km; even relatively powerful explosive activity of mafic magma is unlikely to have produced significant fall deposits at these distances, and that would regardless be strongly dependent on wind direction. We note that none of the proximal oral accounts refer to any sort of airfall deposit or darkness. Given this, the absence of a prominent volcaniclastic component (e.g. vesicular ash) in the matrix of the tsunami deposit is not strong evidence for either the presence or absence of accompanying volcanic activity. However, we return to the point above about reports of pumice: the absence of pumice lapilli in any of the tsunami deposits, or of tephra fallout on top of these deposits, does argue against the occurrence of a large-magnitude explosive eruption producing a convective column (e.g. sub-Plinian or larger) preceding or accompanying the event. Since field observations were made in multiple directions, we would expect to have identified at least some evidence of such an event, in the form of pumice lapilli within or on top of the tsunami deposits, had it occurred. We therefore conclude that Ritter was likely erupting in the build up to the event, but we have no evidence to support that the eruption style was different to prior activity or that it had started to erupt more silicic magmas. As discussed above, we do have evidence that a more silicic eruption followed the collapse, and this is potentially compatible with the observations of pumice on shorelines made by post-event surveyors.
Certain details in the oral accounts remain difficult to explain. The reference to a Roman helmet on wave crests are unusually specific and a perhaps unexpected analogy, but occur in more than one account. The reference may potentially allude to Christian imagery which may have been familiar to communities. It seems unlikely that this descriptive detail would emerge independently at more than one location, and thus suggests possible transmission of accounts between communities (however, as discussed in response to reviewer 1, there are nevertheless other specific details, such as those referencing local geographic sites, that imply the fidelity of transmission at specific locations). Whether the description can be inferred to relate to eruptive activity (e.g., fire-like glows or reflections), as suggested by the reviewer, is uncertain; it is not a detail that we feel able to currently interpret with confidence, but we note that the appearance of glows on wave crests does not necessarily imply reflections linked to eruptive activity at Ritter. Since the eruption occurred close to dawn, for communities on New Britain shorelines (facing west), the forward slope of waves approaching the shoreline may simply have been reflecting or scattering light from the rising sun, lending an unusual appearance to the disturbed sea surface ahead of wave arrival. While this is speculative, it offers an alternative potential origin for the appearance of unusual light patterns or glows on the incoming sea surface (the account from Alairo, Sakar, also refers to glows, but wave diffraction around the slope of Sakar and slowing over its fringing reef would also lead to wave arrivals broadly from the west).
We agree with the reviewer that it would be helpful to draw a more explicit conclusion on the nature of the collapse, in terms of whether this was purely “cold” (e.g. Bandai-type) or eruption-associated. The body of evidence (marine samples with syn/post-event silicic material; multiple oral accounts referencing details consistent with subaerial eruption) suggests that Ritter was erupting both in the immediate build-up and after the event. We would be happy to extend the discussion on this point and to highlight this important conclusion more clearly. Whether the event was truly eruption-triggered cannot, however, be determined; the interplay between magma ascent and edifice failure are complex, as has been shown for the Anak Krakatau 2018 collapse and elsewhere, and we do not have sufficient observations to reach a firm conclusion on the trigger of the collapse and the role of eruptive activity.
Finally, the reviewer refers to both the fine ash observed at Finschhafen (already discussed above), and to the tephra fall deposits recorded beneath tsunami deposits at Barang and Kampalap. For the latter, we conclude that these are older deposits, unrelated to the Ritter 1888 event. Rip-up clasts derived from the underlying ash layer were observed on Umboi (we can make this clearer within the descriptions), suggesting the layer was indurated, and not freshly deposited. This also suggests an erosional base to the chaotic tsunami deposit, which may have removed surficial soil, leading to a misleadingly close stratigraphic association between the tephra and the overlying tsunami deposit. Nevertheless, we note that the logs on Umboi identify an intervening silty soil; the primary tephra deposit and the tsunami deposit are therefore not in direct contact. We also note the prior point, that although several oral accounts refer to fire, sounds etc., none of them refer to ashfall or darkness, which would be expected if a decimetre-scale ash deposit had been formed in the build up to the collapse.
Further comments were noted on:
PROBLEMS with Figures and Pictures quality Along the article review (attached pdf) there are several comments and suggestions related to the (low) quality and often illegibility of some figures. This is a problem that cannot be ignored and together with that I invite the Authors to significantly improve the description of the components reported in the Legends of Figure 3 and 4, both in terms of pattern/drawing and textual description that identifies them. Grain Size Distribution Curves (data) appear to be completely disconnected and not functional for Discussion and Conclusions. Given that the curves are merely described without any associated meaning, it's difficult to understand why they were performed and presented. Perhaps the Authors didn't adequately explain their intentions, or I just wasn't able to grasp them.
We fully accept this comment, and have noted above that Figure 3 in particular is reproduced at poor resolution. We would be happy to redesign these figures both to show additional detail, and to provide accompanying photographs at higher resolution. We also acknowledge that the grain-size data, which was collected to characterise the nature these chaotic tsunami conglomerates (which, in general, are very poorly described and documented in scientific literature), could be more extensively described and evaluated. We would be happy to add this to the manuscript.
And a final comment on:
Observational Gaps Unfortunately, due to the destruction and access difficulties (if I’ve fully understood), there remain serious data gaps in the proximal areas most affected by the tsunami: on the coasts closest to Ritter (eastern Umboi) - where the hydrodynamic models of Karstens et al. (2020) indicate catastrophic wave heights of 60–100 meters - there are no oral accounts from surviving eyewitnesses and subsequent generations who have inherited stories... And also, the geological records along the coast between Kabi and Aupwel (eastern to north-eastern sector of Umboi) seem to be lacking or limited, even if the slopes of the coast are very gentle in that stretch. Probably, the collection of one or two logs more documenting tsunami deposits at a certain distance from the coastline (assuming that for this stretch of the hinterland only erosion was active and involved, being so close to Ritter), could have been very useful. These gaps leave several margins of uncertainty and knowledge.
We agree with the reviewer that there do unfortunately remain several observational gaps, particularly from Sakar and the northern coast of Umboi. These are due both to the likely extreme impacts of the tsunami in these areas (resulting in no surviving oral accounts derived from direct eyewitnesses) and to the coastal environment on parts of Umboi, where mangroves inhibited both access to and potentially the formation of clastic tsunami deposits (observations from other historical events show that the nature of deposits is highly dependent on coastal environment and sediment sources). We consider it unlikely that these gaps can be filled, and thus some uncertainties will remain about the event. Nevertheless, our new observations are compatible with models (such as Karstens et al., 2019) and, as the reviewer notes, add significant new constraints on the event sequence, impacts and magnitude, highlighting the scale and impacts of volcanic-landside tsunamis of this type. Alongside existing marine observations and distal accounts, this is now a remarkably rich dataset for a significant historical tsunami, despite the remaining gaps. We would be happy to add brief comment to the manuscript around remaining data gaps.
Citation: https://doi.org/10.5194/egusphere-2026-2101-AC2
-
AC2: 'Reply on RC2', Sebastian Watt, 09 Sep 2026
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 62 | 78 | 9 | 149 | 21 | 9 | 7 |
- HTML: 62
- PDF: 78
- XML: 9
- Total: 149
- Supplement: 21
- BibTeX: 9
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The study focuses on the 1888 lateral collapse of the Ritter Island volcano, which generated a devastating landslide-related tsunami. The authors present new field observations of (relatively) proximal tsunami deposits found up to ca. 30 km from Ritter Island. These geological findings illustrate a greater and more detailed impact than that based on previously known reconstructions. These new data are supplemented by oral histories collected from nearby communities in 2004, based on oral tradition passed down for two-three generations. The integration of geological and anthropological data provides an improved understanding of this catastrophic event. The English is clear and the text reads well.
Major Comments
Reliability of oral accounts: The manuscript utilizes oral accounts shared by coastal communities in 2004. These interviews in my opinion are affected by a significant uncertainty because they were orally transmitted for two-three generations (i.e., ca. 115 years). Figure 6 summarizes all information, but a barplot or a table would be useful to put all sources side-by-side, quantitatively. Have you tried to add an uncertainty buffer to the presented data?
Other landslide-related tsunamis: The authors should consider referencing and discussing other well-documented works on past landslide-related tsunamis. Specifically, drawing comparisons with geological, historical and recent events at Stromboli (Italy) and Kolumbo 1650 (Greece) would provide valuable context.
Small sketches: Studied stratigraphic sections are currently represented using sketches and fence diagrams. The authors should enlarge these sketches significantly to improve legibility, possibly splitting them in more figures and moving the less important ones to Supporting Material.
Campaign photos: Some campaign photos are present but compressed in Figure 2. I suggest showing the photos side-by-side with the related sketches. Moreover, grain-size distributions are graphically reported, but I could not find tabulated percentile values, or at least the median and sorting.
Source island imagery: Figure 1 rightly shows the regional context, including the locations where field surveys and oral accounts were gathered. However, the manuscript notably lacks a dedicated figure or photograph of the source, Ritter island itself, and, possibly, an idea of its shape before and after the catastrophic event.
Discursive Results Section: The Results section is currently written in a highly narrative format. Because of this, it reads more like a "pre-discussion". The authors should make an additional effort to streamline this section a bit more.