The Réal torrent, located in the south of France, has been instrumented since September 2010 . Three monitoring stations are distributed along the channel. Figure  shows the station locations. The first one, station S1, is located on a 20 m wide check dam as seen in Fig. a and is the most upstream. Stations S2 and S3 are located in the middle reach and at the outlet of the torrent and are both on natural cross-sections. In Table , a summary of the main physical features of the stations is shown (drawn from ). The purpose of the installation is to monitor the flow height, rainfall and seismic activity during sediment activity from the bed load to debris flow. A thorough study of the station can be found in and in . The methodology presented above has been applied to these three stations, and the results are presented further in this paper.

In essence, each station is equipped with (i) a tipping-bucket rain gauge with 0.201 mm resolution (Campbell), (ii) an ultrasonic or radar flow stage sensor (Paratronic), and (iii) a set of three vertical geophones (Geospace GS20DX0) each spaced out ≈100 m apart from each other upstream, midstream and downstream of the flow height sensors.

Images of the channel and flow proved to be useful to facilitate the interpretation of the signals . Two cameras have been added to stations S1 and S2 (Campbell CC640; replaced in 2018 by a Reconyx PC900 and Canon EOS1200D, respectively). Data are recorded using an environmental data logger (Campbell CR1000) powered by a solar panel and are stored in a compact flash module (Campbell CFM100).

In Fig. a, a complete set of measurements for one debris-flow event at station S1 exemplifies the data analysis for one event. Out of these raw measurements, the best suited signals are chosen by the user, as seen in Fig. b:

For flow height along the event, if multiple flow height signals are available, the most reliable one is chosen, i.e. the flow height sensor that does not present any artefact (e.g. unphysical values, very noisy signal). Consistently choosing the same sensor across all events when it does not have any malfunctions is preferable. Here, only one is available.

This leads to Fig. b with only the datasets used for the determination of the hydraulic values of interest. For each of these measurements, surges are identified and their features are computed. The user cross-controls the measurements and eventually goes for the visual method if the cross-correlation does not provide satisfying results (irrelevant value of velocity, low correlation coefficient or inconsistent velocity when compared to a first quick manual computation). The visual method consists in manually inputting the date of the onset of the surge on each geophone and considering the difference as the lag (see Fig. S3 in the Supplement). This visual method was used marginally, i.e. for one surge in our case, and was confirmed using image processing.

These sensors and post-processing allow for having the following for each event: (i) seismic activity at three different points around the station with a frequency of 5 or 10 Hz, (ii) rainfall data every 5 min (not used directly in this work), (iii) flow height with a frequency of 5 or 10 Hz, and (iv) imagery of the event (when possible) with a 0.2 or 1 Hz frequency.

Further in the paper, the value of the effective flow height is taken as the following:

in the case of a controlled section, the mean value between the two assumptions for the section shape, as described in ;

For the construction of the database, only significant events were considered to ensure the analysis of mature debris flows: a threshold of flow height above 1 m was selected for this catchment. This threshold is arbitrarily chosen from our experience on this particular catchments. Overall, 34 events were considered for the Réal station for the period 2011–2020. Table  shows when those events occurred, the number of surges passing at each station and the availability of the describing parameters. Over the 34 surges, most, i.e. 26, are recorded in upstream station S1, while only four surges reached S2 and only two reached S3, the most downstream station. The lack of events in the period 2014–2018 is partially due to the natural variability in event sizes but also due to faulty sensors during that time period.

One of the main interests in having an integrative dataset is to allow access to field ranges of hydraulic values of interest, such as Froude numbers and volumes of surges. In Fig. , different cumulative distribution functions (CDFs) of the datasets are presented. Froude numbers range from 0.25 to 1.6, showing the range of regimes found in debris flows in our site. Whether this is a site-specific feature or it can be shown at more sites that Froude numbers are typically critical would be a strong take-home message for the community.

Surge volumes range from 200 to 4500 m3 (Fig. c; quantiles at 25 %, 50 % and 75 % of 390, 640 and 1460 m3, respectively). Surges are relatively small, typically from 1000 to 2000 m3 km-2 (recall that this is a surge scale and an event may comprise several of them, e.g. 1–4 in our observations of Table , as well as some diluted runoff). Maximal flow height is most of the time lower than 2 m (Fig. a: quantiles at 25 %, 50 % and 75 % of 1.1, 1.25 and 1.6 m). The peak discharge ranges between 6.2 and 91.8 m3 s-1 (Fig. b; quantiles at 25 %, 50 % and 75 % of 10.8, 17.5 and 27.9 m3 s-1). The unit peak discharge is thus typically 0.775 to 7.65 m3 s-1. Froude numbers range from 0.25 to 1.6 (Fig. d; quantiles at 25 %, 50 % and 75 % of 0.48, 0.65 and 0.95); i.e. they are typically near critical. The complete dataset is available in Table S1 in the Supplement.

Finally, relationships between these hydraulic values may be explored with a wider dataset and a more thorough description of each event. Figure  shows for instance the relationship between a few key variables (Froude numbers, volume of each surge normalized by the catchment area, front height and velocity). The surge volume was normalized by the catchment area not only to cross-compare measurements performed at different stations but also to help transfer these results to other catchments.

A slight trend can be seen in Fig. a with an increasing Froude number for an increasing specific surge volume. While no clear conclusion can be drawn, there are no surges with large specific volumes (>1000 m3 km-2) which have clearly subcritical Froude numbers (all Froude numbers are above 0.8). Most of these surges have near-critical Froude numbers. It seems that debris-flow surges of a large volume require a strong inertial input to flow, as there are no subcritical Froude numbers for volumes of the selected range. Their heavy granular content, increasing their macroscopic viscosity, would cause subcritical, slower flows with high volumes to stop or deconstruct. On the other hand, smaller surges can flow more easily and do not need strong inertial inputs to maintain steady flow. The fact that most of these surges are near critical might in part be due to the sampling at the stations and not the possibility for them to exist: very fast surges with high volume and high inertia are very rare in this catchment. Indeed, the hydrology of the catchment allows for sediment transfers to occur rather often (see ), and the moraine material and steep slopes lead to a low yield criterion of the accumulated sediments. This means that the surges with high volume that are passing at the stations meet the “minimum requirements” for flow. One surge with a supercritical Froude number and high volume is still detected.

If surges would all be of the same hydrograph shape and mixture composition, surge volume would be highly correlated with flow height. However, maximum flow height is quite variable with surge volume (Fig. b). This supports the argument that debris-flow hydrographs vary widely.

Similarly, no clear correlation seems to appear between front velocity and flow height (Fig. c). Literature data have been displayed, drawing from , , , , , , and .

Our dataset range has similar Froude numbers as the literature, with most points between Fr=0.5 and Fr=1.5. A point from would plot out of the figure (maximal values of velocity: 4 m s-1, flow depth: 4.5 m, rendering a subcritical Froude number of Fr=0.6). Two points from the dataset have similar features, notably Froude numbers close to 0.6, and would also plot out of the figure. Most datasets show values similar to the Réal torrent with the notable exception of the dataset provided by that has generally higher Froude numbers. This is attributed to specificities of this catchment which do not have the slow laminar features that can be found on the reach like the Réal torrent. Overall, all Froude numbers displayed stay under Fr=3.

We interpret this lack of a clear trend or correlation as evidence of varying surge mixture composition between events. The sample size remains however relatively small and site-specific, calling for careful interpretation of these data. We believe it will be of high interest if several other sites could be added to a similar analysis. Fitting a relationship between Froude numbers and surge volume could be a very interesting asset for numerical and experimental modelling.

Figures  and  show the ranges of the different features in the database for the Réal torrent. Specific volumes range from 101 to 2237 m3 km-2. In comparison to specific volumes given by , which range from 171 to 7690 m3 km-2 (catchment size: 11.69 km2), these are much smaller. One of the key reasons why there is such a difference – apart from differences in geological and rheological makeup – is the method employed: classically, available volumes can contain multiple surges and diluted tails, and thus, volumes are not as restrictive as in the method employed in this paper. Specific volumes of the Réal catchment being much smaller is consistent with the difference in the hypothesis of each method. In , the Gadria catchment monitoring is described and the method employed is much more comparable. In that case, specific volumes of surges range from 35 to 952 m3 km-2, when taking the catchment size as 6.3 km2, which is a range similar to our dataset.

For smaller specific volumes (<1000 m3 km-2), Froude numbers range from 0.2 to 1.2 with most surges being clearly subcritical with a value of <0.8. Flow conditions for smaller volumes require less inertial input. For the same specific volume, a wide range of subcritical Froude numbers are found, showing that volume is not the main driver to flowing conditions and that surge mixture composition varies widely in surges of a low volume, i.e. <1000 m3 km-2. This composition of the mixture changes the mobility of surges.

The initial expectation for Fig. b would be that surges of a higher volume render higher maximal flow height. This would be the case if they hydrograph shape was consistent for all events. Debris flows have very variable flow hydrographs (, among others) due to a wide range of flow mixture. This leads to similar volumes of debris-flow surges caused by different types of flow hydrographs: a shallow surge which lasts for a long duration or very intense high but short surges.

Figure c shows no definitive relationship between proxies for inertial and potential inputs in the flow. This is yet another argument to point out that surge granular content and mixture composition might differ widely from one event to another in the same catchment. The idea that composition of the debris-flow surges changes between events is supported by . A study of the surge content in boulders and coarse grain and of their interstitial fluid rheology would be complementary to support this idea but is at the moment not possible with the available data.

In Fig. b and d, the valley bottom landforms bear the footprint of high morphological activity due to debris flows, more specifically in the reach between S1 and S2 where landforms such as abandoned channels, levees and lobes can be seen (Fig. b–c). Figure  exemplifies these changes in the channel morphology directly downstream of station S1 at five different dates. An erosion–deposition cycle of the channel incising and refilling is highlighted over 6 years of field pictures. Such processes explain why many debris flows are measured at station S1, while many fewer are observed further downstream.

In Fig. , volumes of all events are shown along time. If the geomorphic cycle exemplified in Fig.  was detectable by this method, pseudo-cycles of cumulated volume surges at station S1 would be less frequently exported as surges of a higher volume at station S2 (or as many small-volume surges at S2 in the following years); i.e. if it were possible to see this geomorphic cycle, the cumulated volumes of the surges passing at S1 would be found to be equal to the cumulated volume at S2 over the years. Any amount of the deposits at S1 or between S1 and S2 would then be exported downstream. It can be seen that the two surges reaching station S3 are indeed of a relatively high volume, but the data lacking between 2015 and 2019 prevent us to draw further observations. With the current data, we can simply conclude that higher volumes of debris-flow pass station S1 than further downstream. The system is thus storing sediment in the valley through aggradation and/or also exporting sediment volume through a process other than mature debris flows. This is in agreement with the analysis in which concludes that the sediment activity can be transfer, erosion or deposition in these positions in the reach and in this range of slope (0.11–0.18 m m-1; see Table ). The applicability of this approach to study the sediment cascade is limited by multiple aspects: the first is that the data of interest are kept at the surge scale and focus on mature debris flows (threshold height of >1 m). Due to the way the data have been processed, studies on global sediment balance are not possible with this analysis, as the events of the bed load and wash load are not taken into account. Indeed, despite its high debris-flow activity, the Réal torrent experiences other processes causing long-term morphological changes such as bed load transport and debris flood that have a meaningful impact on morphological changes and sediment fluxes in various parts of the catchment .

A key interest of having three different monitoring sub-stations in the same torrent is the possibility of studying cascading sediment transfers. Figure  shows the analysis of volumes, flow rates, Froude numbers and flow height of each event that could be found at more than one of the stations. One could expect to see consistent relationships between upstream and downstream characteristics, but results are more complicated.

Volumes passing stations S1, S2 and S3 are generally very different at the same date (Fig. ). In some cases, the debris-flow surges were growing, recruiting sediment from the bed (V2>V1 and/or V3>V2), showing the profound morphological changes debris-flow passage can lead to. In other cases, some deposition occurred (V2<V1), but erosion might still appear downstream. For the subset of events happening on the same date at the three stations, no particular relationship between the four parameters studied in Fig.  was identified.

In Fig. a and b, volumes and peak discharge should consistently grow if the surges were consistently eroding from upstream to downstream of the reach. Events like the 30 April 2012 surges show increasing volumes, with a potential agglomeration of the surges between S1 and S2 (accumulated volumes at S1 are smaller than the volume at S2). This shows deep erosion is possible between the two stations, which is consistent with the morphological changes shown in Fig. b. Nonetheless, for this event, peak discharge is not increasing between the two stations. This specificity points out how the pure measurement data and analysis benefit from more specific event data and description.

Similarly, maximum surge depth can also be either lower upstream (30 March 2013, Fig. c) or higher at the first station (events of summers 2011 and 2014, Fig. c). The Froude number also varies from upstream to downstream with some events having a lower downstream Froude number and others not (Fig. d). Froude numbers could be expected to be consistent from upstream to downstream: the ability of the surge to flow would be driven by the interplay between kinetic and potential inputs. Erosion and deposition processes of the surge along the reach will influence the Froude number by changing both the volume and the composition of the surge. This is in agreement with the fact that the slopes in this section are in a sediment transfer regime, as stated by .

The observations of volumes, discharges and surge heights, as well as the much stronger frequency of mature debris flow passing S1 against those passing S2 or S3 (26, 4 and 2, respectively), highlight that strong processes of erosion and deposition occur in the catchment.

While analysing data from three different stations located on such a small and active catchment is interesting, events detected on multiple stations are scarce: most surges detected upstream tend to deposit or to attenuate while travelling such that they are not detected as a mature surge downstream. On the opposite end of this spectrum, a surge that was under the detection threshold on the upstream station might have become fully formed in the downstream stations (see the events of 10 June 2014 and 28 October 2018 that were detected at S1, not detected at S2 and again detected at S3, Table ).

On the other hand, surges that are detected at multiple stations are also difficult to link to one another, and although volume comparison could be interesting, actual quantitative comparison relies on the hypothesis that the exact same surge between upstream and downstream stations is comparable, i.e. that along the journey, only marginal changes in the process occurred, which is known to be a crude hypothesis of this first work. In essence, the data shown in this paper are interesting because they are actual field observations with quantitative measurements, but the analysis of the catchment sediment transfers is not possible. However, the dataset does demonstrate how strong and intense the processes of erosion and deposition in debris-flow-prone catchments are. An analysis seeking to determine rainfall-triggering conditions of debris flows would for instance draw different conclusions depending on which station is used (but see , partially addressing this issue). We believe that further effort should be put on better understanding not only debris-flow-triggering factors but also propagation through headwaters and intermediate reaches.

Additional multitemporal high-resolution images would help in drawing conclusions on this temporal investigation, and such field campaigns would help answer some of the remaining questions such as the remobilization of the deposited material and evidence of global pseudo-cycles (e.g. ).

Comparing the present data to the literature shows the ranges of volumes and flow rates found in the Réal torrent to be consistent with empirical fits proposed in previous works , even though the measurements of volumes were done with debris-flow levees in these previous works rather than direct measurements as in our contribution. More precisely, these fits are using the full-scale debris-flow event rather than a single debris-flow surge. In Fig. , three values are always plotted for the Réal database: they compare the maximizing, the minimizing and the value of the wetted area chosen to be saved in the database. The effect of the choice of the assumption stays relatively marginal for the upstream station but does have a significant effect on the natural cross-sections, as expected. This highlights the importance of these assumptions in the processing of raw data.

According to Fig. , the peak discharge of the Réal catchment for various volumes of debris-flow surges seems closer to the empirical fit related to granular debris flows of or the fit proposed by . Peak discharges associated with muddy debris flows are lower than those measured at the Réal catchment for equivalent volumes. These results are consistent with the work of , who already showed this concordance using an analysis considering the full debris-flow event with a former version of this methodology.