Mountain glaciers are well known as sensitive indicators for climate variability. Glaciers in the Alps are currently retreating due to the observed temperature increase. At some glacier termini in the Alps, this retreat have been uncovering peat fragments as well as remnants of long-lived trees, e.g. counting up to some 700 tree rings. Radiocarbon dating as well as tree-ring analyses prove that these organic samples are usually date back to the early and middle Holocene indicating long glacier retreat periods. I.e., the tree-ring series established prove multi-centennial to millennial long periods with glacier extents smaller than ca. 2000/2010 AD.

Do these finds and results show that the current state of climate is still “normal” and within the Holocene climate variability range? The ongoing strong glacier retreat that begun in the 1980s after a short re-advance phase indicate that current glacier extents are not in equilibrium with climate conditions. Model runs with the glacier flow model OGGM using the climate data of the last ca. 30 years as forcing shows that Alpine glaciers will continue their retreat over decades to reach equilibrium conditions with significantly reduced extents. I.e., no further increase in temperature is necessary to explain the possibility of tree growth in glacier forefields that are now becoming free of ice but had such conditions in former periods of the Holocene.

Lake Telestskoye is a unique water body due to its location at the center of large mountain region in Eurasia. It is the deepest (maximum depth is 323.3 m) lake in Altai Mountains of tectonic origin with low trophicity. The surface area is 223 km2, the catchment with 70 permanent and 150 ephemeral tributaries – 227 km2. The lake of 41.1 km3 volume lies in a narrow mountain valley and has thick bottom sediments which reflect the climate and environmental changes in the region. The average sedimentation rate for the lake bottom sediments is 1.45 mm yr-1, the underwater Sofia Lepneva Ridge – 0.3 mm yr-1 (Selegei et al., 2001; Kalugin et al., 2009). The composition and abundance of diatom algae hosted by the sediments are good indicators of climate, therefore time series of temperature, pH, etc. can be reconstructed. More than 200 taxa of diatom frustules in different layers of 1940-mm core (2240 BC-2006 AD) from the underwater ridge were revealed. The number of diatoms varied from 0.86 to 75.6 with an average of 22.3±0.58 mln frustules/g. The peaks of the diatom abundance, namely the representatives of cold-water assemblages, including the dominant Aulacoseira subarctica (O. Müller) Haworth, were observed in the periods with low water temperature. The correlation coefficient of diatom number with temperature and precipitation for the period of 1965-2006 was 0.89 and -0.83, respectively. The pH value reconstructed from the proportion of pH-indicating diatom species number varied within 7.48-7.98 (average 7.620 ± 0.004), changed insignificantly and corresponded to a weakly alkaline environment (7.0–8.5). At the deepest layers of the core, the pH values and proportion of planktonic and bottom species changed more significantly, that points to slightly different climate conditions 4000 years ago and is in good agreement with data from other Altai paleoarchives.

The Altai Mountains are located on the boundary between contrasting climatic and vegetation zones, i.e. the boreal forest zone in the north and the vast steppe belt of Kazakhstan to the west and south-west, and Mongolia to the south and south-east. A variety of environmental conditions determines the presence of various objects that can be used as paleoindicators for reconstruction and prognosis of climate changes. Bottom sediments in deep lakes are informative paleoarchives due to their thickness, long accumulation period and presence of a large number of stored biological proxies. Every layer can usually characterize some years of lake’s history. The sediments of the 1940-mm core from the underwater Sophia Lepneva Ridge in deep mountain Lake Teletskoye reflect the period of 2240 BC-2006 AD, every layer – ∼11 years. The spatial and temporal variations in temperature, pH, vegetation or aridity that occurred during the Holocene in the Altai Mountains and the surrounding areas were reconstructed based on composition and number of biological proxies as from deep Lake Teletskoye (Kalugin et al., 2009; Mitrofanova et al., 2016; Rudaya et al., 2016) or 30 peats and small lakes from the south-east Altai (Zhang, Feng, 2018) and shallow Lake Manzherokskoye (Blyakharchuk et al., 2017). The tree rings, ice core from glaciers, speleotherms in caves are indicative archives for a year or even a season reconstruction. Analysis of the physical and chemical data from some Altai glaciers revealed that they were formed during the Younger Dryas and survived the Holocene Climate Optimum and the Medieval Warm Period (Eichler et al., 2009; Aizen et al., 2016), while the one of biological proxies in the ice core from Belukha allowed identifying the main sources of them deposited onto the incoming to the glacier surface (Papina et al., 2013).

River flooding is among the most destructive of natural hazards. Societies are currently under increasing threat from such floods, predominantly from increasing exposure of people and assets in flood-prone areas, but also as a result of changes in flood magnitude, frequency, and timing. Accurate flood hazard and risk assessment are therefore crucial for the sustainable development of societies worldwide. With a paucity of hydrological measurements, evidence from the field (i.e. natural archives such as speleothems, tree-rings, lake and fluvial sediments) offers the only insight into truly extreme events and their variability in space and time. In this context, I strongly contributed to data production of Alpine millennia-long (paleo)flood series and methodological developments in the field of paleoflood hydrology. Based on this acknowledged contribution to the field, I participated to initiate in 2015 the PAGES Floods Working Group, gathering today more than 100 international researchers working on historical and natural flood archives as well as statisticians, hydrologists and modelers in order to identify and tackle interdisciplinary challenges aiming to improve flood hazard and risk assessments. After two review papers on the different paleoflood archives and recent methodological advances (Wilhelm et al., 2018, 2019), we enter a new synthesis project on the global to regional flood variability during past warm and cold periods. This work is expected to better understand climate-flood linkages at different time and space scales and to contribute to the next IPCC AR6. Hence, my contribution to the workshop will provide knowledge on flood-hazard changes in the context of past climate changes.

- Wilhelm B. and 22 authors (2019) Interpreting historical, botanical, and geological evidence to aid preparations for future floods, WIRES Water, e1318

- Wilhelm B. and 7 authors (2018) Recent advances in paleoflood hydrology: from new archives to data compilation and analysis, Water Security 3, 1-8

Alpine flood variability and past climate changes: a view from lake sediment records

A team of French, German, Swiss and Italian colleagues is currently working together to synthesize the knowledge gained from millennia-long paleoflood chronicles reconstructed from ca. 25 alpine lake sediment sequences. This dataset is of particular interest to document the flood variability to past climate changes because i) it covers the response of a broad panel of hydrosystems (from torrential, highland catchments to fluvial, lowland catchments) to past climate changes and ii) all series are continuous in time. First results clearly highlight a dominant role of large-scale atmospheric processes (i.e. rather than temperature) on flood-frequency changes at centennial timescale. Regarding changes in flood magnitude, the few lake sediment records documenting that feature reveal more complex trends, seemingly depending on the atmospheric-related flood processes.

In order to investigate the variation of glacier extend during the Little Ice Age (LIA) in the Eastern Alps, we studied a multi-proxy sediment record from a small high alpine lake spanning over 10 kyr. The lake is located at 2790 m a.s.l. in Southern Tyrol, Italy, in a terrain with active glaciers. Proxies include organic carbon and nitrogen, dry-weight and wet-density; sediment geochemistry at 1 cm resolution measured on ground dry samples with ED-XRFA, and at 1 mm resolution on core halves with ITRAX-XRF core-scanner; P-wave velocity, gamma density, and magnetic susceptibility measured at 1 mm resolution with a GEOTEK Multi-Sensor Core Logger; biological microfossils. With the Bayesian age-depth modelling software Bacon, we established an age-depth model based on 15 radiocarbon dates using plant macrofossils, and 210Pb and 137Cs dating of the top sediment layers. Due to turbidites and visible changes in sediment accumulation rate, establishing a robust chronology remained challenging.

While the lake has currently lost its fluvial connection to the local glaciers, the periods of glacier meltwater input are evident in the sediment record. Before 10 000 cal. yr BP the lake first lost its connection to the glaciers; around 9500 ± 500 cal. yrs BP the lake shifted from a turbid to a clear water lake, and sediment layers became very homogenous and dominated by organic sedimentation for a the major part of the Holocene. After 1100 ± 220 cal. AD, we observed another clear system shift; distinct changes in sediment organic content, density, turbidites, and geochemical composition separate periods with glacier meltwater input of variable intensity and frequency and thus mark the onset of the LIA in this region, and the subsequent fluctuations in glacier extent. Around 1880 ± 20 cal. AD the glaciers started to retreat, and by 1950 ± 7 cal. AD the lake again became a clear water lake without direct glacier meltwater inflow.

VULPES (www.vulpesproject.com) is a multidisciplinary and multi-scale project that aims at exploring the capacity of mountain tree species to persist in microrefugia. In this presentation I will illustrate the aim of this project with the case of the Atlas cedar (Cedrus atlantica) which is currently restricted to a relatively small range in the Moroccan mountains between 1400 and 2200m elevation.

Fossil pollen records indicate that Atlas cedars occupied in the past a wider range at lower elevations when winter temperatures were lower than today and the water availability was higher. The modern population genetic data reveal very low overall divergence but high allelic richness in the northern populations. Model simulations show that the range of the Atlas cedar decreased tremendously over the last 50 years.

All together, these past climatic and ecological reconstructions, genetic data, and modeling simulations indicate that the Atlas cedar forests formed very large populations. Their range became fragmented in the Moroccan mountains over the past millennia by extinctions at lower elevations and their modern decreasing trend is threatening those populations that have the highest genetic diversity which may lead to a potential extinction of the species at more or less short term if the modern microrefugia are protected.

The approach developed within the VULPES project highlights the importance of combining information from different but complementary disciplines within a multi-scale approach for identifying areas where modern threatened species may have persisted and for building a scientific basis that may contribute to their future conservation.

Accurately dated information on Alpine glacier extent during the last couple of millennia are mostly limited to point observations during glacial advances. Little to none information is available on the glacier extents during phases of retreat, for example during the Roman warm period or the Medieval Climate Anomaly. We will use the Open Global Glacier Model to simulate selected Alpine glaciers for the last ca. 2500 years. Tree-ring based temperature reconstructions will be used to drive the model, and model results will be compared to established glacier extent reconstructions. A preceding study will show the model's capabilities to simulate past glacier changes for the last 170 years. Here we use historical instrument records as model driver and can compare results with frequent glacier observations. We conclude by identifying and describing the challenges and limitations of our approach, and discuss possible ways forward.