Nesting Effects of Climate and Local Drivers on Long-Term Subalpine Fires in the Alps


 Background: The present article questions the relative importance of local- and large-scale processes on the long-term dynamics of fire in the subalpine belt in the western Alps. The study is based on soil charcoal dating and identification, several study sites in contrasting environmental conditions, and sampling of soil charcoal along the elevation gradient of each site. Based on local differences in biomass combustion, we hypothesize that local-scale processes have driven the fire history, while combustion homogeneity supports the hypothesis of the importance of large-scale processes, especially the climate. Results: The results show that biomass burning during the Holocene resulted from the nesting effects of climate, land use, and altitude, but was little influenced by topography (slope exposure: north versus south), soil (dryness, pH, depth), and vegetation. The mid-Holocene (6500–2700 cal BP) was an important period for climatic biomass burning in the subalpine ecosystems of the western Alps, while fires from about 2500 years ago appear much more episodic, prompting us to speculate that human society has played a vital role in their occurrence. Conclusion: Our working hypothesis assuming that the strength of mountain natural and local drivers should offset the effects of regional climate is not validated. The homogeneity of the fire regime between sites thus underscore that climate was the main driver during the Holocene of the western Alps. Long-term subalpine fires are controlled by climate at millennial scale. Local conditions count for little in determining variability at the century scale. The mid-Holocene was a chief period for climatic biomass burning in the subalpine zone, while fires during the late Holocene appear much more episodic, prompting the assumption that societal drivers has exercised key roles on their control.


Introduction
Events in recent years have dramatically shown that wild res are essentially under the control of extreme heatwaves. The overall regulation of wild res by climate is undoubted (e.g., Bond and Keeley 2005;Flannigan et al. 2009; Moritz et al. 2012) and this should be on the agenda of states and UNO stakeholders (IPCC 2021). However, other drivers also act on wild re regimes: soil granulometry, slope exposure, biomass quality and load, topography, altitude, landscape pattern, and, of course, past or present land uses all play their part (Loehman et al. 2013). The relative function of climatic and nonclimatic drivers can be tricky to unravel, while the local management of a re regime needs such ne distinctions in order to implement e cient management processes according to actual risks. Cold regions like boreal or temperate mountains present biomass with very similar elements in terms of climate (temperature, precipitation, seasonality, snow cover, etc.) and biota (functionally analogue woody species, structure of vegetation, etc.). However, they differ by their relief, which determines the fragmentation and the heterogeneity of landscape (slope degree and exposure, altitude, etc.). The altitude controls the moisture and the temperature (wetter and colder with increasing elevation) and the partial pressure of O 2 (pO 2 ) that is the biomass oxidizer during re (10% change of pO 2 per 1000 m in elevation). Further, the slope controls the water run-off and drainage, and the exposure regulates the solar radiation and the length of the growing season (Tian et al. 2001;Piedallu and Gégout 2007) with potentially signi cant effects on the re regime (Schumacher and Bugmann 2006). To sum up, as a rst approximation, a number of evident similitudes warrant some comparisons or assumptions between boreal and mountain systems, but the differences strongly encourage caution when considering management planning for wild re and forest resources (Loehman et al. 2014; Moritz et al. 2014).
Here we test whether the past climate controlled the re regime in a temperate mountain region over the last 11 millennia, i.e. during the Holocene -a warmer and more humid time in mountains compared to Glacial and Lateglacial periods that were colder and drier. The principles of the study are based on the hierarchy theory (O'Neill 1988; Allen 2009), which states that if climate is the main driver, different sites that are close to each other should reveal similar pattern of re regime (Yocom Kent et al. 2017). If they do not show such similarities, local drivers must be considered, notably slope exposure, altitude, soil type, and vegetation. Of course, historic land use must also be taken into account as an important re driver (Zumbrunen et al. 2009), but this has been di cult to prove in the past without resorting to speculation because accurate quantitative evidences of past land uses are often lacking, e.g., chronics of people density, their savoir faire, and technologies. Vegetation is under the direct in uence of mountain climate and soil types, through bedrocks that notably control the plant communities and their productivity. The vegetation is also largely shaped and controlled by traditional land uses that generally exclude woody plants and seek to sustain mountain grasslands that are not re-prone. When traditional land use in this context is abandoned, subalpine woody plants recover in decades (e.g., Tasser Lingua et al. 2008). For our purposes then, we here study three sites in close proximity, all located in the same valley of the western Alps. In each site, charcoals were extracted from several natural subalpine soils along the altitudinal gradient and further radiocarbon dated. The numerical analysis of 14 C datings allows reconstructing of biomass burning as a surrogate of re history (Payette and Gagnon 1985;Carcaillet 1998;Gavin et al. 2003), and enables us to determine homogeneities and differences by considering local drivers. Charcoal fragments were identi ed to reconstruct the past burned biomass, helping to test whether different woody communities were historically involved in these past res. Here, two sites -both situated on south-facing slopes but contrasted by their soil -were selected: acidic and well supplied with water versus alkaline with poor water storage. The third site is in between in terms of soil, but has north-facing slopes. This last condition (a north-facing site) should result in less biomass burning (more stochastic), while a south-facing site on poor water storage soil should in theory promote a more homogeneous chronology of biomass burning starting earlier during the Holocene, especially if vegetation was more re-prone.
We hypothesize that due to the strength of mountain contrasts natural local drivers should offset the effects of regional climate. Alternatively, the homogeneity of the re regime between sites would suggest that climate was the main historical driver. Con rmation of either hypotheses should result in different scenarios of land management, notably in planning the future of forest and woodlands facing the global warming that will almost certainly be accompanied by more frequent and severe drought in the Alps (Gao thicknesses were greater than 20 cm or when the soil horizon limits were not visible. About ten to 15 liters of dry ne material were sampled per layers of ca 20 cm. Each complete dried pro le weighed 30 to 100 kg depending on the soil texture, the organic content, and the horizon structures, and the depth of the pro les that determine the number of layers, from one to seven for 30 and 150 cm thick pro les, respectively. A otation procedure with an ascending water ow was used to extract charcoal from the soil matrix (Carcaillet and Thinon 1996). Flotation with wet sieving followed by manual sorting under a binocular microscope, allowed the nal separation of charcoal fragments from other soil particles. Only fragments larger than 0.4 mm in diameter were considered. Charcoal that is smaller than 0.4 mm may be transported over long distances, while the stand origin of charcoal larger than 0.4 mm is unequivocal Aussois (Carcaillet 1998). Because in this preliminary study of 1988 (i) no dates covered the recent period from 2200 calendar years before the present (hereafter, cal BP) above 2000 m asl at Aussois and above 1700 m asl at St Michel, and (ii) none was obtained before 6500 cal BP (Carcaillet 1998), the strategy was to increase signi cantly the number of datings to ll in these gaps in the chronologies, or to strengthen the already reported pattern. The point was to question whether these chronological gaps resulted in a problem of low numbers or to see if they correspond to an actual lack of burning in some sites at certain altitudes. The additional site (Lanslevillard) allows us to test whether these reported chronologies per altitude class were strictly local/artifactual or corresponded to a general pattern. In this region of the western Alps, no accurate sedimentary charcoal series are available below 2000 m asl (Leys and Carcaillet 2016), which prevents us formulating a response to this issue of re patterns according to altitude. We assume that re occurrences should be more frequent at lower altitudes and on south-facing slopes compared to higher altitudes and north-facing slopes due to greater dryness (precipitation increases with elevation and south-facing slopes receive more solar radiation) (Schumacher and Bugmann 2006), and posit their greater attractiveness to prehistoric humans who altered re patterns in response to the drivers prevailing thousands of years ago, e.g., climate, vegetation, topography, soil, etc. With a few exceptions (Table S1), each dating was measured on one charcoal fragment. Prior to AMS dating, charcoal fragments were cleaned using an ultrasonic wave generator under a microscope (×40) to remove small roots, fungi hyphae, and particles that could potentially alter the dating result. Further, each fragment was processed with a solution of Na 4 P 2 O 7 or NaP 2 O 4 to extract the organic compounds adsorbed in the charcoal porosity. The solution was changed daily up to the release end of organic compounds by the charcoal. Generally, ve to seven days were needed to chemically clean the charcoal fragments (Carcaillet and Thinon 1996).
The 14 C measurements were processed by several laboratories: the NSF Arizona Facility Laboratory (Arizona, USA), the Centre de Datation par le Radiocarbone (France), and the LMC14 (a national facility lab, also in France). Dating codes were respectively labeled AA-xxxx, Lyon(OxA)-xxxx, and SacA-xxxx (Table S1).
The radiocarbon measurements were calibrated against dendrochronological years using the CALIB program version 8.2, based on the IntCal20 dataset (Stuiver et al. 2021), and reported as intercept with 2 sigma and a probability range of 1.0. The probability distributions of 14 C datings were cumulated by site (e.g., Hajdas et al. 2007) and labeled local chronology. The probability distributions were obtained using the sum probability function of the CALIB program. Contrary to some assumptions (e.g., Payette et al. 2012;Couillard et al. 2021), such distribution cannot provide a chronology of re events, but rather a chronology of biomass burned. Important lags between the 14 C date (radiocarbon age of wood) and the age of re can exist, notably if the charcoal fragments used for dating came from the inner part of trunks for long-lived species such as those growing in subalpine forests (Carcaillet 1998). To bypass this problem -which Gavin (2001) called the 'inbuilt age' -and in the absence of a fuel load/consumption model to x it (Gavin 2001), charcoal fragments corresponding to twigs were here selected for dating, when possible to reduce the range of the so-called inbuilt age. However, a maximum of cumulated calibration does not correspond to a re event but to a stochastic period of burned biomass characterized by a concentration of 14 C dating and associated probability of calibration.
The statistical analyses of radiocarbon dating were based on their distribution revealed through violinplots (R package, ggplot2). The violin-plot is more informative than the box-plots sometimes used to establish dating distribution (e.g., Gavin et al. 2003), because a multimodal distribution is visual and illustrated by two bulges at least, whereas a unimodal distribution should present only one. The site distributions were statistically compared to each other based on ANOVA followed by Tukey HSD post-hoc test to determine whether differences occur between pairs of sites. Calendar dates cannot be directly used because a date corresponds to a complex range of values, sometimes to several ranges associated with different probabilities within the full calibration range of the date considered. Consequently, the mode of each calendar date was extracted from CALIB and used as representative of the dating for violinplots and ANOVA. The mode value divides the graph of the density function into two parts with equal probability, i.e. p = 0.5 (Michczyński 2007). This process is not perfect but better than to use the median (Michczyński 2007), which does not consider the probability variability within this calibration range (Telford et al. 2004). The mode (and the median) does allow for a unique value per date that is needed for the ANOVA.

Charcoal botanical identi cation and analysis
Extracted charcoal fragments were observed under an incident light microscope (×200, ×500, ×1000) to identify the woody taxa that burned. Identi cations were realized with the help of specialist reference wood anatomy books (e.g., Jacquiot 1955; Jacquiot et al. 1973;Schweingruber 1990), and some speci c studies on wood charcoal anatomy notably for Larix and Picea, for Betula, and for Empetraceae/Ericaceae. Because Larix and Picea can only be distinguished from each other in exceptional cases, they have been gathered under the taxon "type Larix-Picea" (Talon 1997). Betula trees can be distinguished from Betula shrubs (Hellberg and Carcaillet 2003). With regard to Ericaceae and Empetraceae, genera or species have been identi ed when possible based on the anatomy of the medulla: Arctostaphylos, Empetrum, Erica, Rhododendron, Vaccinium myrtillus, V. vitis-idaea, V. uliginosum (Talon 2004). Further, Pinus sylvestris (a mountain tree, growing up to 2000 m) and P. uncinata (a true subalpine tree growing from 1500 to 2500 m asl) cannot be distinguished based on their wood anatomy, and are thus labeled P. sylvestris/uncinata. The assemblages of charcoal are considered to correspond mainly to dead woody debris, especially that lying on the ground, and to living twigs of trees and shrubs which can easily burn due to their small diameters or their low height in the vegetation Charcoal assemblages were expressed as a percentage of the identi ed charcoal mass per taxon and per layer; unidenti ed charcoal fragments were not considered in this calculation. Unidenti ed charcoals were those fragments with too few anatomical features for de nitive identi cation, or glassy charcoals whose provenances still remain unknown despite experimentations (Théry-Parizot 2002). Taxa assemblages per layer were used for Correspondence Analysis (CA) to statistically distinguish the main organization of burned woody assemblages, and notably to test whether assemblages are distributed according to sites and to elevation. CA were computed using FactoMineR of the R package (Lê et al. 2008). For synthetic representation, charcoal concentrations per layers were averaged per taxon and per pro le to calculate the composition gradient along the elevation. This averaging was possible because each charcoal layer is not strati ed due to soil bioturbation (Carcaillet 2001 (Table S2).

Fire history
The three chronologies of burning present some differences of pattern at century scale ( Fig. 1), but all three sites show a common multimillennial motif. The main fact of note is that no dating covers the period before 7700 cal BP. At Lanslevillard, the rst occurred at about 7200 cal BP, while the rst at Aussois occurred at 6500 cal BP. Most datings of soil charcoal coincide with a vague mid-Holocene period, with important evidences of biomass burning from 6500 to 2700 cal BP, although at Lanslevillard this period is much less pronounced. It is mostly well gured and continuous at St Michel and at Aussois. The third period, from 2800 to the present, is characterized by repetition of rather brief episodes of burning intermingled with episodes without burning. These brief episodes concern two sites out of three together, notably at 2400-2100 cal BP (three sites) and 1700-1400 cal BP (three sites). From 1200 to the present the common pattern is less obvious, although biomass burned simultaneously at Lanslevillard and Aussois rstly between 1000 and 700 cal BP and secondly between 500 and 300 cal BP. Finally, the sites of St Michel and Aussois present a common burning phase during the last 300 years.
When observing the mid-Holocene, from 6500 to 2700 cal BP, it appears that soil charcoal dates occurred both in lower and higher altitudes at all sites (Fig. 1). The violin-plots of radiocarbon datings provide evidence that the distribution of biomass burning in St Michel is signi cantly different from those of Aussois and Lanslevillard ( Fig. 2; ANOVA and Tukey-test: F = 7.279; p < 0.01, Fig. 2). Further, the violin-plot of St Michel is rather regular, although a gentle bulge around 6000 cal BP illustrates a rst but minor phase of burning before the main bulge that is centred around 4200 cal BP. At Lanslevillard, the main bulge is around 2200 cal BP, but a minor one occurs around 5500 cal BP. Only Aussois shows a clear bimodal biomass burning chronology with a rst phase peaking around 4000 cal BP, and a second around 800 cal BP. These patterns demonstrate a certain independence in re histories between sites. Both sites situated on south-facing slopes present two different chronologies, and both sites situated in very different environments (north-versus south-facing slopes, soil quality) are not signi cantly different (ANOVA, p = 0.71).

Past burned biomass per site and per altitude
CA based on the percentage of taxa per layer shows that each site can be distinguished easily (Fig. 3a).
Axis-1 captures 25.4% of the variance of the CA, and 17.1% by axis-2, totalizing 42.5% on this rst factorial map. The barycenter of Lanslevillard assemblages is on the positive side of both axes, and at the opposite is situated the Aussois barycenter (Fig. 3a). Axis-1 distinguishes well all Lanslevillard assemblages from the others, with assemblages above 2000 m asl situated on the negative part of axis-2. Interestingly, all assemblages >2000 m asl of the three sites are concentrated on the negative part of axis-2, except one from Aussois. St Michel assemblages from altitudes below 2000 m are intermediate and much more dispersed than those of Aussois and Lanslevillard for the same altitude class, indicating a greater heterogeneity and diversity of plant assemblages in the lower subalpine belt. Axis-1 opposes Pinus cembra and Larix-Picea on the positive side and Pinus sylvestris-uncinata on the negative side. Axis-2 is characterized by Abies on the negative side (Fig. 3a). Alnus incana (axis-1), V. cf. vitis-idaea, and V. cf. uliginosum (axis-2) are indicated because there are strong correlations with these axes (Fig. 3a), but their actual contributions to the axes are lower than the average.  (Table S2). These three graphs clearly show that the burned biomass differs between sites in the lower subalpine belt, but the compositions converge and are homogeneous in the upper subalpine belt (Fig. 3c).
Charcoal concentrations per soil pro le (ppm) show a general decrease along the altitudinal gradient (Fig. 3b), but there are some differences between sites. If the decrease is signi cant for St Michel following negative exponential (r = 0.96, p <0.001), it is not signi cant for Aussois or Lanslevillard and much more stochastic. At Lanslevillard, charcoal concentrations are homogeneous along the gradient.
Considering all data, the altitudes of 2050-2100 m asl appear as a threshold in the charcoal concentrations, which drop below 120 ppm between 2100 and 2500 m (higher subalpine belt

Discussion
The present study shows that biomass burning, a surrogate re regime reconstruction based on soil charcoal dating, presents both similitudes and subtle differences between subalpine sites at millennial scale despite obvious contrasts in the local conditions of re spread. Indeed, a site characterized by conditions of climatic, edaphic, and topographic dryness, with burned biomass dominated by re-prone trees (Aussois), has a non-signi cant pattern than a wetter and colder site with poorly ammable species (Lanslevillard; Fig. 2). Surprisingly contrasting to one working hypothesis, these two sites did not reveal clear differences of burning periods ( Fig. 1 and 2), while the burned biomass was signi cantly different (Fig. 3). The reconstruction of burned biomass was a crucial condition of the present study to test whether fuel composition might have contributed to determine some similitudes or differences of biomass burning between sites. Finally, Aussois was actually dominated -as it is at present -by reprone vegetation with some species possessing re-resistance traits (notably the three pine species P. sylvestris, P. uncinata, and P. cembra), while burned biomass at Lanslevillard was characterized by poorly ammable species/community (Larix, Picea, Abies) with low re-resistance traits (Picea, Abies) (Fréjaville et al., 2013b, 2018b; Bär and Mayer, 2020). Finally, the most signi cant differences concern: (i) the two sites exposed to the south (Aussois and St Michel), both in terms of burning chronology ( Fig. 1 and 2) and burned biomass (Fig. 3); and (ii) the fact that on north-facing slopes the burning chronology seems episodic, whereas chronologies were more continuous on south-facing slopes whatever the dating number (Fig. 1). The following discussion develops over four paragraphs, namely: (i) the nesting of processes highlighted by the present study; (ii) the evidence of land use during the late Holocene; and, nally, (iii) the question of the absence of burning evidences during the early Holocene that cannot be the evidence of the absence of re.
Nesting broad and local patterns of subalpine biomass burning history If a multi-millennial pattern can be discerned at the valley scale, suggesting a control of biomass burning by the regional climate (three main similar periods at the three sites, no clear differences in burning between the two most contrasting sites), at times local drivers can offset this common pattern, notably by promoting re when other sites seem little or not all affected (i.e., around 7500-7000 at St Michel) or by suppressing res (i.e., since 2000 years ago in St Michel) while other sites still remain active. This suggests nesting processes in the control of re regimes in the subalpine forest ecosystem, involving large-scale and local drivers. Interestingly, the effect of north-versus south-facing slopes was highlighted at millennial scale based on sedimentary lacustrine charcoal from the same region of the western Alps (Leys and Carcaillet 2016), showing that re frequency was more homogeneous between sites on northfacing slopes than on south-facing slopes where it was more stochastic, perhaps due to natural drivers but chie y down to the effect of human activities. However, the main nding of the present study is that a drier site with more re-prone vegetation does not necessarily produce a longer or more continuous chronology. However, the colder, wetter site depicts a more episodic chronology, suggesting that it has been less frequently susceptible to re spread, which is also an output of one model-based study set out by Schumacher and Bugmann (2006).
Many have underscored the importance of human social processes as probable explanations of the alteration of re regimes in high mountain since 4000 or 4500 years ago (depending on the study, e.g., Leys  To avoid speaking about a 'period' with vague limits, it is preferable to evoke a humped-back pattern of biomass burning, peaking somewhere between 6000 and 2500 cal BP depending on local drivers, including the social. As classically evoked with paleo-data, it is not easy to determine what, between frequency, severity, or extent, explains the changes in biomass burning history. Several studies mentioned the decrease in available biomass to explain the changes, while frequency was little involved as driver but rather as a consequence of this change in forest biomass (Blarquez et al. 2012;Vannière et al. 2016). One study in the same valley as the present study evidenced a sharp increase in re frequency about 7500-7000, pre-dating and explaining the drop in biomass and the change in plant composition (Genries et al. 2009a). Similar results were obtained between 2500 and 1800 cal BP in other sites with nal total woodland suppression after the rise in re frequency, while from 7000 to 2500 cal BP the re frequency and the vegetation were very steady (Genries et al. 2009c , show long periods of burning starting around 7000 or 9000 cal BP, respectively, and lasting until ca 250 cal BP. Most of these charcoals came from north-, west-, or east-facing slopes. However, the best we can say is that the maximum of biomass burning occurred from 5000 to 500 cal BP (this will remain so, at least until a statistical analysis is conducted to accurately indicate the strength and the limits of these chronologies). An interesting fact of these studies is that most of the dated soil charcoal sampled in the higher subalpine belt (2200-2900 m asl) is concentrated during this 5000-500 cal BP period (Talon 2010;Saulnier et al. 2015), contrary to the present study where the upper subalpine belt (2100-2500 m) burned earlier, between 6500 and 3000 cal BP, mostly on south-facing slopes (Fig. 1). This highlights regional/local differences. However, based on sedimentary charcoal, the re frequency in the subalpine belt in the southern Alps attained a maximum (humped-back) about 4000-3700 cal BP, during a long period that started 7000 cal BP and ended at ca 3000 cal BP (Carcaillet and Blarquez 2017), matching with the present chronology (Fig. 1).
The obvious role of land use A general re motif of all these reported studies is that with time the biomass burning generally starts to decrease sometime between 4000 and 2000 cal BP until today in the western Alps, according to sites/valleys/massifs (e.g., Vannière et al. 2016). In the northwestern Alps (Fig. 1), the evidence of re in the upper subalpine became more and more rare after 3000 cal BP, probably pre guring the current subalpine grassland. Further, the brief episodes of biomass burning -common to two or three sites during the last 2500 years (Fig. 1) -could well be related to a phase of land use expansion using re to clear forest understoreys or to create grasslands, although brief climatic episodes favorable to natural re occurrence (i.e., warmer, drier periods) cannot be ruled out. The reporting of such motifs with the same dates in other areas of the western Alps would be needed to support a general climate interpretation.
Similarly, the absence of quantitative archaeological data in the study area hinders any possibility of reaching rm conclusions as to the social effect on the sequential motif of brief episodes in the last 2500 years.
In southern sites, this link between charcoal dates and altitude is less obvious because many datings have been obtained covering the last 3000 years (Talon 2010) or after 4500 cal BP (Saulnier et al. 2005), suggesting differences between sites/valley/massif. Associated to the altitudinal gradient is the charcoal concentration per soil pro le that shows a general decrease along the altitudinal gradient in response to increasing stress, although this pattern varies a lot between sites (Fig. 3b). These distributions were expected (Carcaillet and Talon 2001) and correspond to a common subalpine pattern (Carnelli et al. 2004;Talon 2010) well explained by the primary productivity that decrease with the elevation, and the assumed low resilience rate of woodlands to disturbances near the treeline.
The early Holocene: No re or bias due to residence time of soil charcoal?
On a rst approach, the early Holocene seems characterized by the absence of biomass burning, but this fact needs a careful and thorough interpretation, because the absence of evidence is not the evidence of absence. If we compare the chronology of St Michel with a charcoal sedimentary series available at the same place (Genries et al. 2009a), sampled from a pond at 150 m from the present three plots situated at 1950 m, the sediments demonstrate an abundance of charcoal from 9000 to 7000 cal BP, while soil reveals no charcoal. After 7000 cal BP, the sedimentary charcoal concentrations decrease, while soil charcoal starts to sequestrate (Fig. 1). Such similar differences between charcoal stocks in sediments and in soil have been previously reported (e.g., Ohlson

Conclusion
Our working hypothesis assuming that the strength of mountain natural and local drivers should offset the effects of regional climate is not validated. The homogeneity of the re regime between sites thus underscore that climate was the main historical driver during the Holocene. Long-term subalpine res are obviously controlled by climate at millennial scale. Local conditions count for little in determining variability the millennial to century scale. As already depicted in several studies, the mid-Holocene was a chief period for climatic biomass burning in the subalpine zone of the western Alps, while res since about 2500 years ago appear much more episodic, prompting the assumption that human society has exercised a key role on their control.

Declarations
Ethics approval and consent to participate Not applicable T Ann : mean annual temperature; T JJA : mean monthly temperature of June, July and August (summer); T DJF : mean monthly temperature of December, January and February (winter); P Ann : mean annual precipitation; P JJA : mean summer precipitation.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. CarcailletTabS2assemblages.xlsx