The prevailing climate at Nam Co is characterized by strong seasonality, with long, cold winters and short but moist summers. During winter, the westerlies control the general circulation and lead to cold and dry weather, with daily temperature minima below -20 ∘C. In springtime, the TP heats up and allows the meltwater to percolate to deeper soil layers. The drought situation increases gradually until the monsoon rains arrive, typically between May and June. During autumn, weather shifts again to clear, cold and dry conditions (Yao et al., 2013). The mean annual temperature measured at the NAMORS research station (Fig. 1) between 2006 and 2017 was -0.6 ∘C, and the annual precipitation was between 291–568 mm (mean = 406 mm), with the majority occurring during the monsoon season from May to October (Table 1). The onset and strength of monsoonal precipitation varies substantially between individual years and can be delayed by up to 6 weeks, depending on the altitude and latitude on the TP (Miehe et al., 2019). Precipitation rates are subject to spatial variations due to the > 7000 m high Nyainqêntanglha range, which represents the southern border of the lake catchment. This leads to considerably larger glacial areas in the southwestern area (∼700 km2) than in the northeastern area of the mountain range (∼100 km2 ) (Bolch et al., 2010).

The rise of satellites such as Envisat, CryoSat, and ICESat and the increasingly widespread availability of their data has enabled the accurate study of lake and glacier parameters as far back as the early 1970s (Wu and Zhu, 2008; Zhu et al., 2010b; Liao et al., 2013). The size of Nam Co and the extent and distribution of glaciers in the Nyainqêntanglha range have been the subject of many publications over the recent years (Yao et al., 2007; Frauenfelder and Kääb, 2009; Bolch et al., 2010; Wang et al., 2013; Fig. 2; Table 1). Due to different data sources with varying resolutions as well as different mapping procedures, the estimated glacier area varies between different studies (Fig. 2; Table 1), as the delineation of debris- and snow-covered glaciers is rather subjective (Wu et al., 2016). This is especially true for the first glacier inventory (Li et al., 2003 in Bolch et al., 2010), which has been discussed in various studies due to inaccuracies and the quality of its base data (Frauenfelder and Kääb, 2009; Bolch et al., 2010). Nevertheless, recent studies show glacier shrinkage in the Nyainqêntanglha range at a rate of 0.3 %–0.5 % yr-1, as measured since 1970 when the first satellite images were acquired (Fig. 2; Table 1). As a result of this glacier melting, the lake surface area has expanded from ca. 1930 km2 to ca. 2018 km2 at a rate of 2.1 km2 yr-1 (Fig. 3a), and the lake level rose at a rate of 0.3 m yr-1 until approximately 2009 and at lower rates since then (Fig. 3b). The initial rising trends of both lake level and surface area are mirrored by most lakes in the southern part of the TP, but the slowdown of this trend observed at Nam Co around 2009 seems unique (Jiang et al., 2017). This suggests that the lakes on the TP react to changing environmental parameters in a variety of different ways, and that geographical proximity among lakes does not necessarily produce similar reactions to change. The effects on freshwater input to the lake are discussed in Sect. 2.3. Although changes in monsoonal precipitation and wind direction may influence glacial retreat rates (Wang et al., 2013), rising temperatures remain their primary cause (Ji et al., 2018). The total contribution of glacial meltwater as surface runoff to this lake level increase has been estimated as ranging from 10 % to 53 % (Zhu et al., 2010b; Lei et al., 2013; Wu et al., 2014; Li and Lin, 2017), with recent studies being at the lower end of this spectrum. Increased precipitation is estimated to be responsible for 50 %–70 % of lake growth (Zhu et al., 2010b; Lei et al., 2013). Whether there is a change in evaporation remains unclear, as studies for approximately the same time period have suggested both a slightly increasing and a slightly decreasing evaporation rate since the late 1970s (Lazhu et al., 2016; Ma et al., 2016).

The rises in temperature and precipitation are also affecting permafrost soils that extend over an area of ca. 1.4 million km2 (Yang et al., 2004) on the TP. The permafrost layers can be described as relatively warm and thin, with temperatures mostly > -1.5 ∘C and < 100 m thickness (Wu et al., 2010). The mean annual soil temperature of permafrost in particular areas of the TP has increased by 0.1–0.3 ∘C between 1970 and 1990 (Cheng and Wu, 2007). Simulation studies have shown that due to climate warming the permafrost extent may decrease by 9 %–19 % by 2049 and by 13 %–58 % by 2099 (Li and Cheng, 1999; Nan, 2005). Although there is no clear estimate of permafrost extent in the Nam Co basin, Tian et al. (2009) reports a lower limit of permafrost at an elevation around 5300 m a.s.l. along the northern slopes of Mt. Nyainqêntanglha (7162 m). A frost lens was also encountered 9 m below the surface (4738 m a.s.l.) while sampling an outcrop along the right bank of the Gangyasang Qu close to the northwestern lake shore in 2005 (Schütt et al., 2010). Thus, due to increasing temperatures, permafrost degradation may serve as an additional recharge factor to groundwater, resulting in increased subsurface inflow into the lakes.

Focusing on Nam Co, the hydraulic interaction between lake and groundwater is still uncertain, as previous studies either neglected or ignored the influence of groundwater due to a lack of reliable data (Zhang et al., 2011). However, recent studies revealed a water imbalance, which was explained by lake water seepage with an estimated outflow of 1.9×109 and 1.5×109 m3 during 1980–1984 and 1995–2009, respectively (Zhou et al., 2013; Du et al., 2018).

The maximum recorded depth of Nam Co is 122 m (Li et al., 2008a), with brackish water characterized by an alkaline pH of 7.8–9.5 and a conductivity of 1920 µS mm-1 (Keil et al., 2010). The chemical composition of a lake is essentially a function of its climate (which affects its hydrology) and the basin geology. Increased freshwater input from precipitation, melting glaciers and thawing permafrost alters the chemical composition of the lake water and enhances surface runoff, infiltration rates and subsurface flow. Together with the input of freshwater, streams transport dissolved organic matter (DOM), which is composed of a wide range of dissolved components and particles (≤0.45 µm), thus affecting the water chemistry in the lake (Spencer et al., 2014). Excessive landscape disturbance through removing vegetative cover causes higher rates of DOM leaching, more erosion and increasing water runoff velocity, resulting in additional input of minerals and nutrients into the lake. Since the process of DOM leaching and translocation itself is largely dependent on water and sediment cycles (Kaiser and Kalbitz, 2012), it represents both the seasonal and interannual variation in an ecosystem as well as its long-term trend. As the glaciers on the TP retreat, highly bioavailable DOM may provide additional nutrients to downstream environments and amplify the trend of eutrophication of lotic and lacustrine ecosystems. Furthermore, the rivers on the TP have been shown to transport dissolved organic carbon from thawing permafrost areas (Qu et al., 2017), which is likely rapidly degraded via microbial activity, resulting in CO2 emissions, thus potentially producing a positive feedback on global warming. However, the research into DOM as an important allochthonous source of nutrients and as a capture of biodiversity and geodiversity of its respective catchment area is largely lacking for High Asia. The concentration and ratios of different ions in the water have a regulatory impact on the structure of biotic communities (microbes, invertebrates and fish) that can best tolerate abiotic conditions (Wrozyna et al., 2012). In Nam Co, water conductivity has been regarded as the most important environmental factor for shaping communities such as archaea, bacteria, phytoplankton, and micro-invertebrates (Hu et al., 2010; Wang et al., 2011). Studies demonstrated that ammonia-oxidizing archaea (autotrophic microorganisms) are key contributors to ammonia oxidation in deep and oligotrophic lakes (Callieri et al., 2016). This has implications for CO2 fixation in the hypolimnion or the benthic zone, where there is insufficient irradiance to support photosynthesis, implying that archaea would perform the final step in the decay of organic matter via methanogenesis, resulting in carbon dioxide accumulation (e.g., when they decrease during winter). Although nitrification does not directly change the inventory of inorganic nitrogen in freshwater ecosystems, it constitutes the only known biological source of nitrate and as such represents a critical link between mineralization of organic N and its eventual loss as N2 by denitrification or anaerobic ammonia oxidation to the atmosphere (Herber et al., 2020). Ultimately, the changes in the communities of primary producers could alter the lake's trophic structure, which also affects the top predators of the ecosystem. The primary productivity, as an indicator of nutrient supply and a longer growing period associated with a shorter ice cover duration, has increased markedly at Nam Co within the last 100 years (Lami et al., 2010). Wang et al. (2011) reported the increasing abundance of the diatom species Stephanodiscus minutulus during the last few decades (ca. 1970–2001). This species is generally viewed as an indicator of water phosphorus enrichment, suggesting increasing inputs from the lake's catchment and stronger mixing in spring season. To predict future consequences of ongoing climate change, it is essential to understand the responses of biotic communities to hydrological variations. Thus, long-term monitoring is needed to adequately address the feedbacks of recent environmental changes, while climatic conditions of the past can be reconstructed through the study of organisms such as diatoms and ostracods that are sensitive to hydrologic and chemical variations (see Sect. 3.2).

Nam Co is located in the transitional zone between the central Tibetan Kobresia pygmaea pastures and the northwestern alpine steppe ecosystem (Miehe et al., 2019) (Fig. 4). Situated on the northern slope of the Nyainqêntanglha range, the vegetation pattern changes according to elevation, moisture availability and temperature. Grazing intensity and abundance of small rodents, such as the plateau pika (Ochotona curzoniae), may contribute to the shaping of the vegetation cover (Dorji et al., 2014; Miehe et al., 2014). The area close to the lake (< 4800 m) is covered mainly with alpine steppe vegetation consisting of Artemisia, Stipa, Poa, Festuca and Carex (Li, 2018; Nölling, 2006). Soils developed in the drier steppe areas consequentially tend to show lower organic carbon contents, naturally lowering their total C sink or source potential, as indicated by a study from Ohtsuka et al. (2008). Only one evaluable soil investigation exists from the area of Nam Co. Wang et al. (2009b) investigated two lake terrace sites, situated in the alpine steppe biome. According to their findings, the soils reflect the cold semiarid climate of the area by showing low biologic activity, while the influence of physical weathering is dominant. The soils showed several decimeter-thick layers of loess in which mainly the A horizons were developed. Although only very sparse to moderate vegetation cover occurs, an almost 30 cm thick organic-rich topsoil with granular structure was developed there (Wang et al., 2009b). Further organic-rich buried horizons were found and dated in both profiles, showing phases of climatic conditions enabling the buildup of organic material related to warm wet periods in the past (before 2.4 and 1.6 ka cal BP) and interchanging with phases of erosion, leading to, e.g., sheet erosion and the formation of gullies and alluvial fans supposedly during colder periods. These results fit well to climate reconstructions presented in Sect. 3.3 of our review. In accordance with the World Reference Base for Soil Resources (WRB) classification, we propose that the soils described by Wang et al. (2009b) can be classified as Calcisols, as there is evidence of carbonate translocation.

Higher up the slope (4800–5200 m), the alpine steppe is replaced by Kobresia pygmea pasture. Wang et al. (2007) and Kaiser et al. (2008) investigated the relationship between plant communities and development of soil types on the High Asian Plateau and for pasture soils in the wider area. Vegetation strongly controls the input of organic material into the soil but beyond that also stabilizes fine materials (< 0.1 mm) and governs the degree of chemical weathering. The authors found soils with stronger signs of biologic activity and chemical weathering (e.g., Cambisols) associated with alpine pasture sites. Kobresia root mats are usually developed in up to 40 cm thick loess layers and form a distinctive felty horizon that protects against erosion. The genesis of this felty root mat is attributed to Kobresia pygmaea, since this shallow-rooted, small plant allocates most of its biomass belowground and is able to reproduce vegetatively, making it well adapted to the high grazing pressure (Miehe et al., 2008). The curious dominance of K. pygmaea is often linked to grazing: (i) K. pygmaea replaces taller plants at sites where grazing pressure is increased experimentally. (ii) Several enclosures show that other grasses and shrubs increase in dominance after grazing competition ceased (Miehe et al., 2008). Hence, the felty root mat can be seen as an effect of an anthropozoogenic plagioclimax. At higher elevation (5200–5900 m), only sparse alpine vegetation associated with initial soil processes occurs (Ohtsuka et al., 2008).

Where water availability is abundant, alpine swamps with Carex sagensis and Kobresia schoenoides are formed, especially at source areas, along riverbanks and in waterlogged depressions, some of which can cover large areas (Li et al., 2011). Concerning soil development in alpine wetlands, the database is sparse compared to the alpine pasture and steppe biomes. It was pointed out for alpine pastures that a strong relationship exists between plant communities and (top)soil genesis. This relationship probably also holds true for alpine wetlands, with the exception that the influences of waterlogging and seasonal fluctuations and frost–melt cycles in the water table are likely to have an effect on soils. This can be expressed in terms of formation of gleyic features, frost turbations, heaves or other azonal features related to the soil-forming effects of water (Chesworth et al., 2008). It still needs to be clarified how these waterlogged areas affect the cycling and processing of organic matter and nutrients. There is no evidence of tree species, only the evergreen shrubs of Juniperus pingii var. wilsonii, which are mainly found on the south-facing slopes of the northern Nam Co catchment, and shrubs of Salix spp. in the Niyaqu Valley in the eastern lake catchment (Li, 2018). Alpine steppe is comprised of more plant species compared to pasture and marsh ecosystems, which are predominantly covered with Carex spp. and Kobresia spp. (Miehe et al., 2011b). Alpine pastures are often described as “golf-course-like” (Miehe et al., 2014) with the intention of illustrating their unique plane surface. However, small-scale structures such as thufa or hummocks are also present. The origin of these structures around Nam Co remains unclear; however, frost heave and permafrost degradation processes are seen to play a major role (Adamczyk, 2010). The landscape, generally dominated by endemic Kobresia pygmea sedges, harbors only a few other species (Miehe et al., 2019), but the additional microhabitats provided by thufa and hummocks enable rarer and less competitive species to settle in niches in these heterogeneous structures (Vivian-Smith, 1997). Compared to the surroundings, the microtopography of thufa possesses different degrees of wetness, exposition and insulation; depth of soil material; and type of topsoil. Local studies of the Nam Co area state that slightly degraded bare soil patches and gullies are often areas where plants have the chance to evade the suppression of the closed Kobresia pygmea root mat (Schlütz et al., 2007; Dorji et al., 2014). Thus, the genesis of thufa and mild, limited degradation processes are likely to increase species richness and diversity by cracking open the closed root mat of alpine pastures. These structures can also be formed by grass species that grown in tussocks (i.e., clumps, bunches or tufts), such as the endemic species Stipa purpurea (Liu et al., 2009) or Kobresia tibetica (Yu et al., 2010) and Kobresia schoenoides (Nölling, 2006).

The often-cited degradation of alpine pastures is likely initiated by natural polygonal cracking (Miehe et al., 2019), which can occur through drying (Velde, 1999) and then tends to be amplified by livestock trampling and plateau pikas using the cracks as highways (Liu et al., 2017b; Hopping et al., 2016). Overgrazing in alpine pastures is one of the most frequently mentioned causes of pasture degradation (Unteregelsbacher et al., 2012; Harris, 2010; Miehe et al., 2008), as excessive trampling by livestock might aggravate the initial conditions of polygonal cracking (Miehe et al., 2019). This effect, however, seems to be limited to the direct vicinity of herder's settlements and camps (piosphere centers), and many factors that are usually attributed to degradation are instead proof of environmental control, especially in drier areas (Wang et al., 2018b). Some researchers argue that climate change is the dominant or even sole driver of degradation (Wang et al., 2007), although the effects of rising temperatures and increasing precipitation appear to be an intensifier rather than the cause of degradation (Zhou et al., 2005; Harris, 2010). In turn, both Wang et al. (2018b) and Cao et al. (2019) point out that a multitude of effects might be in play, with a locally differing magnitude or even reversion, while usually moderate grazing was not found to cause degradation. Certainly there are more factors than just grazing pressure, and there might be site-specific effects leading to nonequilibrium behavior of the study object, be it pasture or steppe (Wang and Wesche, 2016). Plot-level experiments from the Nam Co area found warming to have significant effects on the shallow-rooted Kobresia pygmaea by reducing the number of flowers and delaying its reproductive phenology. These changes were provoked by simulating increasing precipitation by means of snow addition (Dorji et al., 2013) and also by maintaining a moderate level of grazing combined with snow addition (Dorji et al., 2018). This underlines the importance of climate forcing on the terrestrial systems in the Nam Co catchment. Grazing should not be seen as a disturbance but as an integral part of a non-steady-state but plagioclimax environment. Currently there are no estimates of the extent of degraded land at Nam Co, but the degradation of wide areas of alpine pastures is not without consequences for pastoralist communities. The severe degradation and sloughing off of the whole topsoil removes the basis for business and might lead to unknown consequences for the lake ecosystem by means of enhancing or terminating nutrient exchanges. The economic rationale of herders might be to increase the numbers of livestock, as this represents a form of social security (Simpson et al., 1994). The bottom line is that conflicts arise as less land is available for grazing (Hopping et al., 2016).

The Chinese government has favored policies such as sedentariness and fostered the construction of stationary settlements, which have, in turn, created hotspots of overgrazing (Miehe et al., 2008). In these hotspots, large portions of the topsoil are lost by erosion and denudation, leaving only an area of humic material or subsoil, thus being called “black beach” (Miehe et al., 2008) or “black-soil patch” (Liu et al., 2017a). The remaining landscapes are usually dry, poor in plant cover and prone to further degradation. Increasing areas of bare soil patches enhance evapotranspiration, causing earlier cloud cover formation, especially before noon. This may, in turn, lead to reduced radiation and temperature at the surface, thus hampering photosynthesis and consequently overall carbon sequestration (Babel et al., 2014). However, the evolution of grasslands on the TP has been accompanied by herbivore communities; thus, the plants have developed coping mechanisms to persist under continuous grazing pressure (Miehe et al., 2011a). According to the intermediate disturbance hypothesis, species diversity is higher under moderate disturbances, which suggests the positive effect of intermediate level of grazing pressure. Indeed, a plant clipping experiment to simulate grazing demonstrated that under the effect of climate warming, the grazing activities mitigated the negative effects of rising temperature by maintaining a higher number of plants (Klein et al., 2008). Many studies hold the traditional nomadic practice to be a sustainable one (Miehe et al., 2008; Babel et al., 2014; Hafner et al., 2012), but the current policy of removing pastoralist lifestyles from certain regions could potentially reduce overall species richness.

Changes in temperature and moisture have a significant effect on the biotic community structure with feedbacks on ecosystem productivity. Alpine meadows respond with increased plant productivity to warming, while productivity may be hampered in alpine steppe ecosystems (Ganjurjav et al., 2016). As soil moisture governs the community response to warming, negative effects of warming on plant productivity likely occur due to limited water availability (Ganjurjav et al., 2016). Warming was also reported to have a negative effect on plant species diversity in both alpine meadow and steppe ecosystems (Klein et al., 2008; Ganjurjav et al., 2016). Possible explanations for a decline in plant species diversity include changes in small mammal activity, storage of belowground nutrient resources and water stress and microclimate in general (soil temperature and moisture) (Ganjurjav et al., 2016; Klein et al., 2008, 2004). Thus, climate change may reduce the habitat quality for the local populations of grazers and reduce the well-being of the pastoralists by diminishing the abundance of palatable and medicinal plant groups. The changes in the plant productivity levels, as well as community changes, affect the local carbon cycle. Alpine grassland root mats on the TP are estimated to store up to 10 kg of carbon (C) per square meter (Li et al., 2008b), summing up to roughly 2.5 % of the global terrestrial carbon stocks (Wang et al., 2002). At Nam Co, the topsoils contain an almost 30 cm thick organic-rich layer (Wang et al., 2009b), thus representing considerable soil organic carbon (SOC) stocks. Due to higher plant productivity, alpine meadows in general represent a CO2 sink; however, the interannual and seasonal uptake is highly variable (Kato et al., 2004, 2006; Gu et al., 2003). Like plant productivity, the CO2 uptake depends on water availability and temperature, which exhibit diurnal, seasonal and annual fluctuations. The overall great importance of water availability and temperature on ecosystem–atmosphere CO2 exchange in the central Tibetan alpine Kobresia meadows was demonstrated in several studies through eddy covariance measurements (Zhang et al., 2018), chamber measurements (Zhang et al., 2018; Zhao et al., 2017), decomposition of cellulose cotton strips (Ohtsuka et al., 2008) and altitudinal transplantation experiments (Zhao et al., 2018). Similarly, carbon fluxes in alpine steppe biomes are driven by precipitation and temperature on a daily to seasonal and annual timescale. The interannual flux variability follows the varying monsoonal precipitation, showing a stronger tendency towards functioning as a C sink in wetter years and as a C source in drier years (Wang et al., 2018a; Zhu et al., 2015b). Soils that develop in the drier steppe areas tend to show lower organic carbon contents, therefore lowering the total C sink and source potential (Ohtsuka et al., 2008). Although the production of plant biomass may be hampered in steppes, the ecosystem may still act as a carbon sink through microbial CO2 fixing activities, as shown by a recent study on the TP that reported relatively high CO2 fixation capacity (29 mg kg-1 soil d-1, Zhao et al., 2018). Interestingly, this study also found that alpine steppe soils demonstrated significantly higher microbial CO2 fixation capacity compared to meadow soils (29 vs. 18 mg kg-1 soil d-1, respectively).

As a result of increasing precipitation and glacier runoff, wetlands in the Nam Co area are expanding, thus increasing emissions of CH4, which is 28 times more climate active than CO2 (IPCC, 2013). A study conducted in the alpine wetlands around Nam Co reported that CH4 emissions have increased exponentially with increasing precipitation, especially when soil moisture exceeded 80 % (Wei et al., 2015). However, there was a large difference between swamp meadows and swamps (67 and 1444 µg CH4 m-2 h-1, respectively). Swamps are permanently inundated, while swamp meadows are usually seasonally inundated. Furthermore, SOC stocks are higher in swamps compared to swamp meadows (Wei et al., 2015). Large amounts of SOC in combination with anoxic conditions are the main precursors for methanogens activity, which results in increasing CH4 emissions to the atmosphere (Kato et al., 2013). Thus, the saturated soils with high SOC content produce higher CH4 emissions (Deng et al., 2013). Observations from 2008 to 2013 at Nam Co have shown that alpine steppe and alpine meadows show annual uptake rates of 72 and 59 µg CH4 m-2 h-1, respectively (Wei et al., 2015); however, the corresponding emission rates are much higher. Generally, it is expected that the alpine wetland acts as a CH4 source, while the aerated soils of alpine steppe and alpine meadow act mainly as a CH4 sink.

As the grasslands on the TP are widely used for yak and sheep grazing, carbon cycling is influenced particularly through human activities and the degree of degradation. The intensity of grazing has a significant effect on CO2 fluxes, with moderate grazing enhancing belowground allocation of carbon (Hafner et al., 2012), while adversely affecting the C-sink potential through reduction of aboveground and belowground biomass at higher grazing intensities (Babel et al., 2014). Overgrazing, along with the increase in burrowing pikas in the Tibetan grasslands, may increase nitrous oxide (N2O) emissions (Zhou et al., 2018), an important greenhouse gas with 297 times larger warming potential compared to CO2 (IPCC, 2013). Despite several studies focusing on greenhouse gas emissions on the TP, the magnitude of the N2O emissions in different ecosystems has not yet been estimated. Experimental studies on the eastern TP demonstrated that the rate of N2O emission may increase with increasing soil temperature and soil moisture under a future climate change scenario (Yan et al., 2018; Yingfang et al., 2018). Expanding wetland areas provide anoxic conditions for the release of methane and, due to the greater temperature sensitivity of permafrost areas, subsurface SOC is at high risk of loss, which may decrease the carbon sequestration potential in the region (Li et al., 2018). Besides carbon cycling through decomposition processes, responses to changing temperature and precipitation depend on the composition of decomposer communities (Glassman et al., 2018). Thus, the conclusive effects and feedback mechanisms (i.e., positive vs. negative loop) on warming are complex and not always clear.

Topography, geological context, climate and their complex interplay are key determinants for the distribution of organisms. In general, the ecoregion can serve as a proxy for community- and species-level biodiversity, which best describe communities of mammals, birds and plants (Smith et al., 2018). The TP forms a distinctive zoographical region, an “ecological island” (Deng et al., 2019), characterized by fauna that are adapted to high altitudes, drought, low temperatures and low oxygen levels (He et al., 2016). The TP forms unique high-altitude biogeographical biota by also harboring many unique lineages of other organisms, with a higher endemism of low dispersal species (Yang et al., 2009; Clewing et al., 2016). As mountain building has been directly associated with the development of biodiversity (Hoorn et al., 2013; Antonelli et al., 2018), the biodiversity hotspots are located in the south and southeast of the TP especially. There is also a pattern of increasing biodiversity from west to east, which correlates positively with increasing precipitation. In contrast, the harsh central areas of the TP show much lower richness but nevertheless harbor various endemics (Päckert et al., 2015). Throughout the geological formation of the TP, the mountainous southeastern parts have been hypothesized to serve as center of species diversification (Mosbrugger et al., 2018), although the core TP region is also suggested to represent a center of origin (Deng et al., 2011). The TP has been a source area for several mammalian lineages (“Out of Tibet hypothesis”; Deng et al., 2011), including the snow leopard and the arctic fox (Wang et al., 2015), as well as birds, such as redstarts (Voelker et al., 2015), and plants, such as Gentiana (Favre et al., 2015). These mountainous areas may also have acted as refugia, which preserved unique lineages over long periods (López-Pujol et al., 2011; Lei et al., 2014). Whether some endemic taxa represent relics of a formerly more diverse clade or have never extensively diversified remains unclear (Päckert et al., 2015). Besides being a center of origin, the TP may represent a center of accumulation as proposed by the examples of Saxifraga (Ebersbach et al., 2017), warblers (Johansson et al., 2007) and hynobiid salamander (Zhang et al., 2006). Overall, the regional biota of the TP is comprised mainly of Palearctic and Oriental species, Nearctic species from the Bering land bridge, as well as species from speciation in situ and postglacial recolonization from adjacent areas. The evolution of biodiversity on the TP has been affected by the combination of geological and climatic changes over the time of the uplift phases (Mosbrugger et al., 2018). Although many studies have associated recent in situ radiations to different uplift phases of the TP, Renner (2016) pointed out that the evidence for recent rapid uplift (9–8 or 3.6–2.6 Ma) remains doubtful and controversial. As proposed by the “mountain geobiodiversity” hypothesis, the evolution of biodiversity on the TP is a result of an increasing local geodiversity in combination with rapid climatic oscillations and steep ecological gradients (Mosbrugger et al., 2018).

The combination of geological, climatic and ecological changes has left its footprint in the history of Nam Co. There are at least seven different levels of continuous terraces around Nam Co, with the highest being over 30 meters above the current lake level, corresponding well with the elevation of the natural spillway in the northeast of Nam Co. Several authors claim the existence of a much larger fluvial lake system called Old Qiangtang Lake, which covered an area of around 30 000–50 000 km2 or more (Li et al., 1981; Zhu et al., 2002). The connections provided by a large lake allowed gene flow between drainages, which is reflected, for example, by the closely related clades of schizothoracine fish (Cyprinidae, Osteichthyes) from Nam Co and the surrounding lakes, compared with more distant parts of the TP (He et al., 2016). In contrast, due to a vector-mediated passive dispersal across large areas, other aquatic taxa, such as freshwater snails, seem to have been less influenced by drainage histories (Oheimb et al., 2011). Higher lake terraces are older, suggesting a long-term reduction in lake level (Zhu et al., 2002). This may be associated with an evolution from wet to dry phase, which Li et al. (1981) connects to the gradual uplift of the plateau from early Pleistocene to the Holocene. However, there is an alternative suggestion to this interpretation of a rather modern uplift proposed by Renner (2016) who states that large parts of the TP had already reached average heights of 4000 m and more during the mid-Eocene (∼40 Ma ago). Recent findings of palm leave fossils on the central part of the TP, dated to ca. 25.5±0.5 million years, do not suggest a presence of such a high plateau before the Neogene (Su et al., 2019). Thus, although it is suggested that the final large lake phase took place ca. 40–25 ka cal BP (Lehmkuhl et al., 2002; Zhu et al., 2002), the complex relationship between evolution of the TP and the development and the temporal existence of Old Qiangtang Lake are not completely resolved.

Lake sediments contain important indicators or proxies that can be used to reconstruct limnological and (hydro-) climatic conditions over long time periods (Zhu et al., 2010a; Wrozyna et al., 2010). Widely used environmental indicators include communities of diatoms (Bacillariophyceae) and ostracods (Crustacea) as they are abundant and usually preserve well in sediments (Kasper et al., 2013). For example, the investigations of Quaternary ostracods, modern assemblages, and stable isotopes from Nam Co and nearby water bodies represent the most detailed application of ostracod analysis in the south–central region of the TP (Mischke, 2012). Different approaches (stratigraphy, paleoecology, etc.) detected several climatic fluctuations between warmer–cooler and wetter–drier periods (Fig. 5). In general, higher lake levels based on aquatic fauna suggest a more humid environment during the early and middle Holocene, which displayed a shift pattern compared to the northern TP (Wünnemann et al., 2018). Together with the indicator species approach and the application of transfer functions for Nam Co sediments, different stages can be recognized. In Stage I (8.4–6.8 ka cal BP), the climate changed from warm humid to cold arid with water depth being much lower than today (Zhu et al., 2010a). In Stage II (6.8–2.9 ka cal BP), environmental conditions returned to warm and humid (Zhu et al., 2010a). During 4–2 ka cal BP, lake water depth initially remained much shallower than today but then gradually increased due to high rates of precipitation (Frenzel et al., 2010). The presence of the diatom taxa Stephanodiscus in this stage indicates stronger monsoon activity and higher availability of nutrients (Kasper et al., 2013). Finally, in stage III (2.9 ka cal BP to present), the climate again became warm humid, with a cold dry event between 1.7 and 1.5 ka cal BP (Zhu et al., 2010a). Between 2 and 1.2 ka cal BP, benthic diatoms inferred a lower water level and drier climate (Kasper et al., 2013). Subsequently, wetter conditions and an increase in lake level was detected (1.2 ka cal BP until 250 cal BP), possibly corresponding to the Medieval Warm Period (MWP), with high planktonic diatom species and high ostracod diversity (Kasper et al., 2013). During the late Holocene, the minimum water level occurred throughout the Little Ice Age (LIA) (∼1490 and 1760 CE) (Frenzel et al., 2010). However, the lake level increased towards the present, which is plausibly linked to the melting of the glacier due to the current warming.

Although a large number of studies describe profound hydrological changes and general climate fluctuations, there are several uncertainties regarding taxonomy, resolution and proxy sensitivities. For example, modern ostracod data detects several morphological variations, characterized by different nodding or shell sizes, which could lead to an erroneous ecological interpretation and later, vague paleoenvironmental conclusions in relation to salinity changes (Fürstenberg et al., 2015). In paleo-studies, different sedimentation rates and uncertainties in the core chronologies also cause a lack of correspondence between signals detected by different proxies (Wang et al., 2012). For this reason, it is surrogate to understand the precise causal relationships between a complex environmental gradient (e.g., water depth, water chemistry, temperature, etc.) and the response of bioindicators. Although ecological information is still poorly known for many species, ostracod and diatom assemblages represent reliable proxies to trace the climatic history of Nam Co.

Further emphasis should be placed on combining morphology and DNA analysis to corroborate the classification of the species already described. Furthermore, experiments with living individuals should be performed under controlled environmental variables to allow the setup of a transfer function that could be used to evaluate quantitative data for paleo-reconstructions.

The comparison of modern pollen assemblages with those from sediment cores allows the reconstruction of floristic diversity and distribution across various timescales. Vegetation patterns contribute to the reconstruction of past climate and the assessment of the degree of local human influence. Modern vegetation belts around Nam Co reveal that alpine steppes contain mostly species of Artemisia (Asteraceae) and Poaceae, while alpine meadows and swamps are dominated by Cyperaceae (Li et al., 2011). The sedimentary pollen ratio of Artemisia to Cyperaceae (A/Cy) can, within certain limitations, be used to reconstruct past climates (Li et al., 2011; Li, 2018; Zhu et al., 2015a) provided that vegetation belts move with altitude during climate change. For example, when the climate is warmer and drier, alpine steppe reaches higher up the mountain, displacing alpine meadow into areas further away from the lake, leading to a higher input of Artemisia pollen into the nearby lake and consequently a higher A/Cy pollen ratio in the sediments. However, the A/Cy pollen ratio and abundance of tree pollen originating from a short distance can be altered by human-driven change of plant composition, hence the beginning of pastoral economy might limit the explanatory power of pollen records (Adamczyk, 2010; Miehe et al., 2014). Pollen composition inferred from sediment cores reveals a downward shift of the altitudinal vegetation belts since 8.4 ka BP (Li et al., 2011). A major extension of alpine pasture and alpine sparse vegetation closer to the lake shore during the late Holocene is corroborated by a pollen-based climate reconstruction from a peat core near Nam Co (Herrmann et al., 2010) and two other pollen records from the eastern lake shore (Adamczyk, 2010). They found a trend of increasing temperatures from the late glacial until the early Holocene, accompanied by an extension of alpine steppe, tree and shrub vegetation. Already in this period, synanthrope taxa pollen are increasing in the data used by Adamczyk (2010) with the only small occurrence of, for example, Plantago lanceolata in the whole profile. This very early signal shows that a lot of room still exists for studies of pollen archives around Nam Co, with much doubt still persisting at present. Climate fluctuated between dry and humid from 8.5 to 4.8 ka BP, with an intense cold regression between 8.1 to 7.8 ka BP. The onset of human activity at Nam Co is dated to 5.6 ka BP according to synanthrope taxa proxies (Li et al., 2011; Herrmann et al., 2010). Between 4.8 and 0.7 ka BP, a relatively stable climate with predominantly humid conditions developed (Fig. 5), and the vegetation pattern already showed trends of a human-made steppe biome, potentially a plagioclimax (Adamczyk, 2010). Since 0.7 ka BP, drier conditions prevailed.

Whether and to what extent the central Tibetan Plateau was forested and what caused the forest decline are the subject of ongoing discussion (Miehe et al., 2006, 2019). This matter is closely related to the prior discussed onset of more intense human activity in the area, since parts of the discussion involve a human-made forest clearing in combination with a natural forest decline. As stated, there are only occurrences of shrubs (Juniperus pingii var. wilsonii and Salix, Nölling, 2006) in the Nam Co area. No remains and no reliable evidence of a once tree-rich vegetation have yet been found in the Nam Co catchment. According to locals, there exist several caves with potentially (pre)historic tree depictions of unknown age. Unfortunately, there is no verification of their existence nor any dating approach. Since the area of Damxung still does feature larger occurrences of Juniperus pingii var. wilsonii and, around 4250 m a.s.l., tree stands of Juniperus tibetica in enclosed areas, there is the potential to discuss that these species have been more numerous in this area in the past (i.e., last tooth theory). Miehe et al. (2019) show locations of forest relicts and give a drought line of 200–250 mm precipitation and elevations between 3600 and 4000 m a.s.l. as the upper tree line. Questions arise as to whether there has been an expansion of J. tibetica into the Nam Co catchment in earlier times, which would be feasible within certain limitations according to the presented thresholds. Charred micro remains as a potential sign of fire-driven forest decline are missing in one of the profiles of Adamczyk (2010) but can be found throughout the Holocene until 1 ka cal BP (Herrmann et al., 2010). The authors attribute the size and shape of the charcoal remains to local, small-scale burning of wood and leaves, not showing signs of larger forest clearings. In addition to the burning of Juniperus trees for religious reasons (Miehe et al., 2006), trees and shrubs may have been burned for heating or clearing of pastures by nomads. Following the presumptuous argumentation of some authors, the trees were previously able to spread again due to sufficient precipitation provided by the summer monsoon. Furthermore, the occurrence of synanthropic taxa has been observed in the nearby Damxung valley since 8.5 ka cal BP, corroborating the strong anthropogenic influence on the formation and restructuring of the vegetation patterns in the area (Schlütz et al., 2007). The decrease in summer precipitation and temperature, in conjunction with ongoing human activity, ultimately led to the total disappearance of trees and the formation of the alpine grasslands and steppe as we know them today (see Sect. 2.4). Furthermore, the occurrence of synanthropic taxa has been observed in the nearby Damxung valley since 8.5 ka cal BP (Schlütz et al., 2007). This corroborates the strong anthropogenic influence on the formation and restructuring of vegetation patterns in the area but leaves a time gap of almost 3 ka between the evidence from Damxung valley and Nam Co. Hence, further research is needed to address the question of onset of human activity and degree of landscape modification.