Diatom-oxygen isotopic record from high-altitude Petit Lake ( 2200 m a . s . l ) in the Mediterranean Alps : shedding light on a climatic pulse at 4200

The 4.2 kyrs event, used as a marker of holocene stratigraphy, has been described as a rapid climate change in the northern hemisphere triggering droughts in the Mediterranean region. However, the severity and geographical extent of this event are still the subject of investigation considering the small number of palaeoclimatic records for this time period, and the presence of contrasted climatic expressions between areas. At Petit Lake (France, Mediterranean Alps, 2200 m a.s.l) a 15 multiproxy study of Holocene lake sediments has revealed major changes in erosion processes and phytoplanktonic assemblages in the lake ecosystem around 4200 cal. BP. According to pollen analysis, deforestation is unlikely to be the main explanation of environmental changes as the watershed was covered by open vegetation for the duration of the study period. To test the implication of climate, our study presents an analysis of oxygen isotopes (δO) in diatoms describing hydrological modalities during the 4.2 kyrs event in the Mediterranean Alps. The highest values of δOdiatom occur from 20 4400 to 3900 cal. BP and are interpreted as an increase in water evaporation and/or a decrease in freshwater inputs to the lake system. Changes in water balance might have been associated with a change in precipitation sources towards a greater influence of precipitation coming from the Mediterranean area. These results are concomitant to an increase in erosion in the watershed and high representation of very low-dispersal pollen in the sediments suggesting the presence of intense runoff. This new isotopic record together with previously-published proxy-data, allows us to describe the 4.2 kyrs event at Petit 25 Lake as an increase in Mediterranean climate influences in the region, amounting to a general dry period punctuated by episodes of intense runoff occurring on the catchment slopes.

recorded in ice cores and in numerous worldwide palaeoenvironmental records. These are: the Younger Dryas (13.500-11.500 cal. BP) at the end of the Late Glacial, and the 8.2 kyrs event in the beginning of the Holocene (Alley et al., 1997;Brauer et al., 1999;Tinner and Lotter, 2001). Other climatic events were described during the Holocene but were interpreted as less intense or regionally limited.
However, although their geographical extent is still under discussions, these climatic events triggered some substantial 5 impacts on the environment. One of them, the "4.2 kyrs event" has been recognised in various studies as an abrupt climate change (Bond et al., 2001;Booth et al., 2005;Huang et al. 2011;Thompson et al., 2002;Staubwasser et al. 2003) which is now commonly used as a marker of Holocene stratigraphy (Walker et al., 2012). In the Mediterranean area, the 4.2 kyrs event is recorded as a complex period lasting a maximum of several hundred years, with contrasted palaeohydrological expression between regions (Bruneton et al., 2002;Digerfeldt et al., 1997;Drysdale et al., 2006;Kharbouch, 2000;Magny et 10 al., 2009;Miramont et al., 2008;Zanchetta et al., 2011). In the Eastern Mediterranean, this climatic event is recognised to be responsible for severe droughts and was likely involved in the fall of the Akkadian civilisation (Weiss, 1993;Dean et al., 2015;Cullen et al., 2000). In Central Mediterranean, while speleothems from southern Italy (Renella, Corchia Cave) recorded dry conditions from ca. 4300 cal. BP to 3800 cal. BP, dry conditions were less expressed in records from northern Italy. In the Alps, an opposite trend has been described, and the same time period is characterised by cool and wet conditions 15 (Zanchetta et al., 2011(Zanchetta et al., , 2016. Sedimentary records of past lake levels also mirrored changes during this period showing different climatic expression along a latitudinal gradient. At Ledro Lake and Accesa Lake in Italy (respectively 45° N and 42° N) the transition from mid to late Holocene (ca. 4500 cal. BP) is recorded as a transition period towards higher lake levels. However, the opposite trend has been found for the same period at Preola Lake in Sicily (37° N) (Magny et al., 2012).
Finally, a high-resolution record (Accesa Lake, Italy) allowed to interpret the 4.2 kyrs climatic event as a tripartite climatic 20 oscillation characterised by a phase of drier conditions from 4100 to 3950 cal. BP bracketed by two phases of wetter conditions from 4300 to 4100 cal. BP and from 3950 to 3850 cal. BP (Magny et al., 2009). Overall, new palaeoclimatic records from different longitudes and altitudes in the Mediterranean area are needed to better constrain the regional expression of the 4.2 kyrs event.
In the Southern Alps, the high-altitude Petit Lake (Massif du Mercantour, France, 2200 m a.s.l) offers pollen and diatom-rich 25 sediments covering the last 5000 years. A multiproxy analysis, including sedimentological and geochemical measurements (XRF, ICP-AES) as well as pollen and diatom morphological analysis clearly revealed two phases separated by a major shift around 4200 cal. BP. This major shift was characterised by a detrital pulse (Brisset et al., 2012(Brisset et al., , 2013 followed by a longlasting change in tychoplanktonic diatom assemblages (Cartier et al., 2015). Over the last 5000 years, the vegetation around the lake has been reconstructed as open, rejecting the hypothesis of a massive deforestation in the catchment as the 30 explanation for the detrital pulse. Therefore, the involvement of a rapid climate change either in precipitation regime or temperature leading to increasing soil erosion and runoff around 4200 cal. BP was proposed (Brisset et al., 2012(Brisset et al., , 2013Cartier et al., 2015). Clim. Past Discuss., https://doi.org/10.5194/cp-2018-103 Manuscript under review for journal Clim. Past Discussion started: 3 September 2018 c Author(s) 2018. CC BY 4.0 License.
In order to test the latter hypothesis, we measured the oxygen isotopic composition (δ 18 O) in diatoms (δ 18 O diatom ) taken from the Petit Lake 5000-present day sedimentary core previously used for multiproxy reconstruction. Because δ 18 O values are a function of lake water isotopic composition (δ 18 O lake ) and temperature, δ 18 O diatom records are commonly used for climatic reconstructions (e.g. Barker et al., 2001;Leng et al., 2006;Quesada et al., 2015). Previous δ 18 O records from Mediterranean lakes were discussed in terms of changes in precipitation and lake water balance, which depend on lake location and 5 watershed properties (Roberts et al., 2010). Here, from the δ 18 O diatom record we aim to assess the last 5000 years of hydrological changes at Petit Lake and build assumptions on climatic changes that may have occurred around 4200 cal. BP in the Southern Alps.

Site settings
In the Mercantour range, alpine and mediterranean influences produce a climate marked by mild winters and dry summers. 10 Mean annual temperature at 1800 m a.s.l. is 5 °C, varying from 0.3 °C in winter to 9.9 °C in summer (Durand et al., 2009), with rainfall occurring mainly in spring and autumn. Mean annual precipitation is 1340 mm at 1800 m a.s.l. Snow depths in winter are relatively important (150 to 250 cm at 2400 m a.s.l.) due to moisture from the nearby Mediterranean Sea. Snow cover duration is about 185 days at 2100 m a.s.l. mainly from November to April (Durand et al., 2009).
Petit Lake (2200 m a.s.l; N 44°06.789; E 7°11.342) is a small circular body of water 150 m in diameter located in the 15 Southern French Alps about 60 km from the Mediterranean Sea. Petit Lake is at the lowest elevation of a chain of five lakes that were partly formed by glacier retreat (Fig. 1). The lakes are connected in the spring by ice meltwater but remain unconnected for the rest of the year. The lake catchment (area: 6 km 2 ) culminating at 2600 m a.s.l. is composed of crystalline bedrock (gneiss and migmatites) and is largely covered by alpine meadows; the upper tree line (Larix sp.) being located at about 2100 m a.s.l. The lake surface is usually frozen from October to April. The depth of Petit Lake is up to 7 m in the wake 20 of the snow-melt in late spring and decreases to 6.5-6 m at the end of summer.
Because it is located at the extreme south-west of the Alps, Petit Lake is strongly influenced by precipitation originating from the Mediterranean region during the summer, and by precipitation from the Atlantic in the winter (Bolle, 2003). In Southern France, precipitation is mostly generated by the clash between the warm, humid air of Mediterranean or mixed Atlantic-Mediterranean origin and cool air masses coming from the North (Celle-Jeanton, 2000). Precipitation of 25 Mediterranean origin has a weighted annual mean (δ 18 Op) of −4.33 ‰ (standard deviation s=1.72 ‰), and precipitation from the Atlantic has a δ 18 Op of −8.48 ‰ (standard deviation s=3.51 ‰) (from April 1997to March 1999Celle-Jeanton et al., 2004). This is reflected in the seasonal weighted mean δ 18 Op in the Alps (Fig. 1

Material and methods
Sediment core PET09P2 (144 cm-long) was sampled in 2009 in the deepest part of the lake using a UWITEC gravity corer.
Core PET09P2 is organic-rich (total organic carbon represents 9 % of the dry weight on average) and has a high abundance 10 in biogenic silica (averaging 65 % of the dry weight) (Brisset et al., 2013). Diatoms (D) represent the major contribution of biogenic silica in the sedimentary record. Only a few cysts of Chrysophyceae (C) were identified (C/D ratio = 0.01). The age-depth model covering the last 4800 years is based on short-lived 210 Pb and 137 Cs radionuclides and seven 14 C ages obtained from terrestrial macro-remains (see Brisset et al. [2013] for further details).
Twenty diatom samples (1 cm 3 ) were sub-sampled from core PET09P2. Each diatom sample of 1cm 3 includes on average 36 15 years (min: 11 years; max: 55 years) of sedimentation according to the age-depth model. Diatom samples were weighed after drying at 50 °C. To remove carbonates and organic matter, the samples were first treated using standard procedures (bathed in a 1:1 mixture of H 2 O 2 : water, a 1:1 mixture of HCl: water, and repeatedly rinsed in distilled water). Following these steps, the identification and counting of diatom species for palaeoenvironmental reconstruction were performed; and the data were reported in Cartier et al. (2015). Then, diatom silica was cleaned from remaining detrital particles by following the protocol 20 developed by Crespin et al. (2008), which includes 7 steps based on chemical treatments and physical separation. This protocol has already been used successfully in previous studies (Alexandre et al., 2012;Crespin et al., 2010;Quesada et al., 2015). The purity of each sample was then checked using optical and scanning electron microscopy (SEM) together with micro-X-ray fluorescence (XRF) measurements (5 measurements per sample). The hardware used for these analyses consisted of a HORIBA XGT-5000177 microscope equipped with an X-ray guide tube capable of producing a focused, high-25 intensity beam having a 100 µm spot size (detection limit: 2 ppm). The following compounds were detected via XRF: SiO 2 , h held constant at 1020 °C, 2h cooling down to 400 °C) under a continuous dry N 2 flow. Oxygen extractions were then performed using the IR Laser-Heating Fluorination Technique (Alexandre et al., 2006;Crespin et al., 2008). No ejection occurred during the analysis. The oxygen gas samples were sent directly to and analysed by a dual-inlet mass spectrometer (ThermoQuest Finnigan Delta Plus). Measured δ 18 O values were corrected on a daily basis using a quartz lab standard (δ 18 O Boulangé 50-100 µm ) calibrated on NBS28 (9.6 ± 0.3 ‰; n=11) (Alexandre et al., 2006;Crespin et al., 2008). The values were 5 expressed in the standard δ-notation relative to V-SMOW. The long-term precision of the quartz lab standard is ± 0.2 ‰ (1s; n=50). The final δ 18 O values for each sample is the average of two replicates, yielding a reproducibility of better than ± 0.2 ‰ (1s). The age-depth model of PET09P2 (Brisset et al., 2013) was constructed using the R package Clam (Blauuw et al., 2011). For the purpose of this study, we have recalculated the age-depth model using the Bacon R package (Blaauw and Christen, 2011) and implemented the function "proxy.ghost" (square resolution:200) in order to take into account the 10 chronological uncertainties associated with the proxy representation. The result is a range of possible ages for each sample depth, each of which is represented on the graph. The darkest grey is assigned to the most likely value within the entire core (normalised to 1). Lower age probabilities are coloured in lighter grey.

Results
Oxygen isotopes values (‰ vs V-SMOW) measured on the 20 sedimentary diatom samples (table 1)  A zoom on the period 4800-3000 cal. BP period, taking into account the uncertainties of the age-depth model is presented in 25

Climatic interpretation of the δ 18 O diatom record
Effects of human occupation on ecosystems of the Mediterranean Alps, since the middle Holocene, have been widely documented (De Beaulieu, 1977;Mocci et al., 2008;Walsh et al., 2007;Walsh and Mocci, 2016). Human impacts can contribute to changes in erosion and vegetation and along with climate change, must be taken into account when interpreting 5 erosion and vegetation proxies (Giguet-covex et al., 2011;Jalut, 2009;Roberts et al., 2010;. By contrast, lake temperature and hydrological balance as reflected by δ 18 O lake water are the only factors governing changes in δ 18 O diatom . δ 18 O lake water itself depends on the δ 18 O composition of the moisture source, the precipitation amount and regime, and in a lesser extent, the transport pathway followed by moving air masses (Dansgaard, 1964;Gat, 1996). These factors, that are inherent to the climate system, can be evaluated from an accurate interpretation of the high δ 18 O diatom values from 4400 to 3900 cal. 10 BP at Petit Lake. In addition, the lake hydrological balance has to be assessed. At a minimum, potential isotope fractionation related to the dissolution of the diatom spicules during sedimentation may occur (Dodd et al., 2017). However, SEM observations of our diatom samples did not show significant dissolution features (Fig. 3C).
Water inflows to Petit Lake consist of direct precipitation (rain and snow) and intermittent streams that form during the spring ice melt. The outlet of Petit Lake (today blocked by a dam) is an intermittent surface outlet and is non-active when the 15 lake level is decreasing of 1 meter (Figure 1). Therefore, the hydrological regime alternates between two states: an open system when the outlet is active during snow melt and a closed system during summer months when most water losses are due to evaporation. In these systems, changes in water balance are commonly recognised to be the main factor triggering large excursions in the isotopic signal (Leng and Barker, 2006;Roberts et al., 2008). Summer months correspond to the season during which diatoms grow after the first spring blooms. During this time of year and given the shallowness of the 20 lake, waters are well mixed under the effect of wind (Cartier, 2016). From that perspective, it is reasonable to assume that an increase in lake water evaporation, due to drier climate conditions during the summer months, may have triggered an increase in δ 18 O diatom . A rapid change in water conditions around 4200 cal. BP is supported by the dominance of the diatom species Staurosirella pinnata (Fig. 4), a species with high tolerance to rapid changes in alkalinity and conductivity in unstable aquatic environments (Cartier et al., 2015). 25 An increasing contribution of precipitation coming from the Mediterranean area, with high δ 18 Op, can also contribute to an increase in δ 18 O lake water and consequently in δ 18 O diatom . Today, Mediterranean precipitation favours runoff and erosion in steep areas (Kosmas et al., 2002). Geochemical data showing high terrigenous inputs to Petit Lake between 4400 and 4000 cal. BP (Fig. 4), interpreted as an increase of runoff in the watershed (Brisset et al., 2013), are thus consistent with a greater seasonal variability of the Mediterranean climate characterised by intense precipitation occurring in fall and spring and significantly 30 drier periods in the summer months (Durand et al., 2009). according to the thermo-dependent fractionation coefficient between temperature and diatoms of -0.2 ‰/°C (Brandriss et al., 1998;Crespin et al., 2010;Moschen et al., 2005). This inferred-temperature is not in agreement with air temperature estimates based on chironomids and pollen assemblages from the Swiss Alps and Europe which suggest that temperature variations did not exceed 2 °C during the Holocene (Davis et al., 2003;Heiri et al., 2003). Thus, the δ 18 O diatom shift cannot be explained by a drop in summer month temperatures and temperature doesn't appear to be the main factor of δ 18 O variability 5 at Petit Lake. In summary, the rapid increase in δ 18 O diatom from 4400 to 3900 cal. BP is most likely the result of an increase in water evaporation possibly associated with a shift in precipitation origin and distribution over the year. This state lasted for ca. 500 years.
After 3900 cal. BP, δ 18 O diatom values decreased and remain relatively constant for 3300 years suggesting less water evaporation/humid conditions during the Neoglacial period. However, the low resolution of the record in this part might 10 limits the identification of short-term events. A last major excursion impacting the δ 18 O diatom record, but not the other proxies, is recorded around 310 cal. BP. It consists of a rapid drop in the δ 18 O diatom values (Fig. 3A). Conversely to what may have happened during the time interval 4400-3900 cal. BP, lower evaporation during the summer months, increasing precipitation from Atlantic sources, and/or an increase in summer air temperature, appear plausible. This time span falls within the Little Ice Age (450-50 cal. BP), which is known to be a cold and humid period in the Southern Alps according to tree-ring records 15 (Corona et al., 2010), fluvial activity reconstruction (Miramont et al., 1998;Sivan et al., 2006) and glacial tongue advances (Holzhauser et al., 2005;Ivy-ochs et al., 2009). Therefore, the decrease in δ 18 O diatom at Petit Lake is most likely the response of increased humidity and Atlantic precipitation influences. These results are concomitant with a strong decrease in δ 18 O measured on ostracods from Allos Lake sediments (Cartier et al., in prep) indicating the expression of a regional climate change. Other cold periods such as the Late Antiquity recorded ca. 1700 cal. BP in Southern France were not identified in 20 the record, suggesting that the climatic effects of the LIA were of a greater magnitude in the studied area. However, better resolution is needed to confirm this observation.

Expression of the 4.2 kyrs climatic event and regional comparison
At Petit Lake the highest values of δ 18 O diatom , from 4400 to 3900 cal. BP were broadly interpreted as an increase in water evaporation/decrease in water inputs to the lake. A strong influence of Mediterranean precipitation characterised by higher 25 δ 18 Op might also have been an additional factor favouring increased δ 18 O lake at this time. By considering the other palaeoenvironmental proxies, these results appear concomitant to high terrigenous inputs to the lake and chemical weathering of soils (Fig. 4). These minerogenic inputs together with a high representation of very low-dispersal alpine meadow pollen were interpreted as the result of intense runoff on the catchment slopes (Brisset et al., 2013, Fig. 4). In the lake ecosystem, the dominant diatom species, Staurosirella pinnata, was then replaced by other species like 30 Pseudostaurosira robusta (Fig. 4). Results provided by measuring the oxygen isotopic composition of fossil diatoms strongly support the hypothesis of a rapid climate change which triggered the shift in the environmental history of the watershed. Local responses to this climatic event probably happened in several stages: (1) a presence of more intense runoff Clim. Past Discuss., https://doi.org/10.5194/cp-2018-103 Manuscript under review for journal Clim. Past Discussion started: 3 September 2018 c Author(s) 2018. CC BY 4.0 License. in a general drier context; (2) increased erosion in the watershed and terrigenous inputs to the lake; (3) a change in lake ecosystem properties (e.g. water transparency, conductivity) and aquatic assemblages. An indirect effect of climate on the lake ecosystem is supported by the presence of an offset between changes in δ 18 O diatom and diatom assemblages. Indeed, the shift in diatom species occurred after the first increase in δ 18 O diatom and before the decrease of δ 18 O diatom showing the end of the 4.2 kyrs event period. Indirect effects have been more likely modulated by changes in watershed properties like changes 5 in erosion processes as shown in previous studies (Jeppesen et al., 1997;McQueen et al., 1989). To sum up, the 4.2 kyrs climatic event is expressed at Petit Lake as a period of general dry climatic conditions during which intense rainfall occurred.
In a Mediterranean context, strong erosion on dry soils is increased by high seasonal variability and the presence of extreme episodes (droughts, flashflood events) particularly conducive to denudation processes (Brisset et al., 2017;Nearing et al., 2004;Yaalon, 1997). At Petit Lake, the 4.2 kyrs event by changes precipitation regime and water balance, led to a long-10 lasting change in the lake trajectory.
Close to Petit Lake, a palaeoenvironmental record at Grenouilles Lake also recorded high percentages of Poaceae, Chenopodiaceae and Caryophyllaceae around 4200 cal. BP even if the 4.2 kyrs climatic event was not discussed in the interpretation (Kharbouch, 2000). At Allos Lake, located in the Mercantour Massif and at a similar altitude to Petit Lake, there is no evidence of a major detrital supply at 4200 cal. BP. However, this period is generally characterised by high 15 lacustrine production (represented by an increase in organic sedimentation rate) and rapid shifts in the percentages of the benthic diatom Ellerbeckia arenaria suggesting a high variability in lake levels with potential periods of droughts during the presence of these diatom-rich laminae (Cartier et al., 2018). At Saint Léger Lake (Alpes-de-Haute-Provence), a reconstruction of past lake level changes argues, this time, for a moderate rise of the lake level from 4500 to 3000 cal. BP (Digerfeldt et al., 1997). Other detrital or flood events have been recorded between 4500 and 3000 cal. BP in the Alps for 20 example at Bourget Lake (Arnaud et al., 2005;2012) and a review of dated-landslides revealed a cluster around 4200 cal. BP supporting the presence of heavy precipitation (Zerathe et al., 2014;Fig. 5). In the Massif du Mont Blanc (46° N) a moraine contemporary with the 4.2 kyrs climatic event has been dated suggesting a glacier advance during this period (Le Roy et al., 2017). At a broader scale, stable isotope and trace element data from a calcite flowstone located in northern Italy (Buca della Renella; 44° N) have shown that the 4.2 kyrs event was expressed in mid-latitude Europe by dry climatic conditions 25 (Drysdale et al., 2006) (Fig. 5). A similar trend has been found at Preola Lake (37° N) in Sicily (Magny et al., 2012).
However, this period of rapid climatic changes is also documented in some lacustrine records by shifts from low stands to higher lake levels around 4000 cal. BP, for example at Cerin Lake, Ledro Lake (45° N) and Accesa Lake (42° N) (Magny et al., 2013) (Fig. 5).
Overall, few isotopic records with sufficient resolution for studying the 4.2 kyrs climatic event exist in the Western 30 Mediterranean. Most of the records on stalagmites or lake sediments are located in central Italy or in the Eastern part of the Mediterranean region. Concerning the use of other palaeoenvironmental proxies, the difficulty lies in separating local changes in land use from climatic effects. The potential of Petit Lake to record past climatic events is probably enhanced by its location in the head of the watershed and the sparse vegetation which exposes soils to erosion. In addition, the semi-Clim. Past Discuss., https://doi.org /10.5194/cp-2018-103 Manuscript under review for journal Clim. Past Discussion started: 3 September 2018 c Author(s) 2018. CC BY 4.0 License.
closed lacustrine system might have increased the responsiveness of Petit Lake to changes in water regime. Local watershed properties, a lack of isotopic records and age-depth model accuracy might explain the difficulties in identifying similar trends during the 4.2 kyrs event, which still requires further attention at a regional scale.

Conclusion
Measurements of oxygen isotopes in diatoms (δ 18 O diatom ) from Petit Lake were performed on the last 5000 years of the 5 sedimentary record in order to investigate the influence of changes in climate on this alpine watershed. A major and rapid shift in the environmental history was recorded at 4200 cal. BP in both terrestrial and lacustrine proxy-data. The system turned from a steady-state without soil erosion to a new state dominated by a degradation trend for both slopes and vegetation cover. Additionally, a long-lasting change in diatom assemblages highlighted major variations in lacustrine living conditions. The new δ 18 O diatom record for Petit Lake was used to reconstruct past hydrological changes and decipher climatic  Clim. Past Discuss., https://doi.org/10.5194/cp-2018-103 Manuscript under review for journal Clim. Past Discussion started: 3 September 2018 c Author(s) 2018. CC BY 4.0 License.

Competing interests
The authors declare that they have no conflict of interest.

Acknowledgements
This work was supported by the ECCOREV research federation (HOMERE program led by F. Guiter and C. Paillès). The PhD thesis work of R. Cartier (Aix-Marseille University) was funded by the French Ministry of Education. 5 We thank C. Vallet-Coulomb (CEREGE, France) for the oxygen isotope analysis of modern Petit Lake waters and P.