Over the last millennium, anthropogenic landscape opening has greatly accelerated global soil erosion dynamics. However, the long-term link between land openness and erosion remains unclear. Existing soil erosion models assume that contemporary calibrations and current observations of the link between land openness and erosion apply to the past. However, one may wonder whether large-scale anthropogenic land clearance during the Holocene intensified or not caused soil erosion beyond a linear relationship, partly due, for example, to increased incision processes. In this study, we investigate the link between soil erosion and land openness, and refine erosion estimates during the Holocene in the northwestern Alps via a vegetation abundance model (REVEALS), pollen assemblage data from lake sediment archives, and terrigenous accumulation rates in lakes as a proxy of catchment soil erosion. Using empirical cumulative distribution functions (ECDF) and correlation tests, we show that erosion indeed follows a nonlinear response to increasing land openness in the northwestern Alps, potentially challenging current model representations and current regional and global erosion estimates. We attribute this nonlinear relationship to increased incision and remobilization of deeper soils above critical thresholds of increased land openness, as rapid land clearance heightened surface soil vulnerability to hydrological events and erosion processes. These findings suggest that revisiting erosion models could be crucial to better understand long-term landscape evolution impacts, especially on soil erosion and its potentially significant downstream effects on carbon exports.

Rapid glacier retreat is transforming mountain landscapes, exposing terrain where postglacial ecosystems emerge while new opportunities for infrastructure development also arise. These newly ice-free areas are therefore becoming socio-ecological frontiers in which ecosystem development, land use and conservation frameworks increasingly overlap. Yet spatially explicit evidence of where these overlaps occur remains scarce, limiting the ability of governance systems to anticipate emerging conflicts. Here, using the European Alps as a data-rich case study, we pair satellite-derived land cover mapping with inventories of ski infrastructure, hydropower reservoirs and protected areas within the end-of-Little Ice Age glacier footprint. We show that newly exposed terrain already contains emerging vegetated and freshwater habitats, whose regional distribution is mainly modulated by temperature gradients. Concomitantly, ski and hydropower infrastructures already occupy parts of postglacial terrain, even within strictly protected areas, revealing mismatches between formal conservation status and land-use decisions. Emerging postglacial ecosystems fall into a de facto legal vacuum, as current tools encourage reactive, piecemeal decisions that fail to safeguard ecosystem functions and services. We argue for anticipatory stewardship that recognizes glacier and postglacial ecosystems in law and planning as a single transforming system. These regional findings foreshadow a global governance challenge that will intensify as deglaciation expands in regions where deglaciation exposes land to competing ecological, economic and political interests.

Climatic variability in the Southern Hemisphere is largely controlled by the latitudinal position of the Southern Hemisphere Westerly Winds (SHW), whose migration influences precipitation, temperature, and Antarctic upwelling. This study presents the results of analyses of two lacustrine sediment cores from Lake Armor, located on the subantarctic Kerguelen Islands (49 degrees 15 ' S, 69 degrees 10 ' E), within the SHW belt. Lipid biomarkers (Glycerol Dialkyl Glycerol Tetraethers, n-alkanes, and their hydrogen isotopes) were used to reconstruct mean annual air temperature above freezing (MAF) and humidity conditions. These records are compared with a high-resolution diatom-based summer sea surface temperature (SST) reconstruction from marine core MD11-3353, situated 150 km southwest of Lake Armor. In the late glacial and Early Holocene, our results reveal a period of warm air temperature, comparable to current values and very warm sea surface temperature, 5 degrees C above the current values. Around 9000 cal a BP, an abrupt transition occurred, marked by a cooling of 5 degrees C in SST and 1.5 degrees C in MAF, interpreted as a northward migration of the SHW and associated oceanic fronts. The Mid-to-Late Holocene period is characterized by pronounced MAF variability, including a notably warm interval between 3000 and 2000 cal a BP, when n-alkane dD suggests the prevalence of wetter conditions. Since similar to 250 cal a BP, a southward migration of the SHW has produced a 2.5 degrees C rise in MAF. Our findings are overall consistent with previous studies from the Indian Ocean, but permit us to go a step further as by comparing SSTs and air temperatures. This suggests that SST is not a reliable predictor of air temperature on the Kerguelen Islands, particularly during the Early Holocene. We hence argue that Kerguelen air temperature is predominantly controlled by the position of westerly winds, as an indicator of reorganisations in air mass trajectories.

Over the past millennia, human activities and land clearing have significantly altered land cover, accelerating both diffuse and concentrated erosion of soils, ultimately threatening soil ecosystem services. While numerous studies have emphasized on diffuse erosion, the understanding of concentrated soil erosion through incisions remains limited and has rarely been quantified over long-term periods. This study quantifies both concentrated and diffuse erosion to differentiate their respective contributions to the total erosion within the Lake La Thuile catchment in the northwestern Alps, linking spatial and temporal data. The RUSLE model was used to estimate past diffuse erosion. The RUSLE model was temporally constrained with Land Use Land Cover (LULC) reconstruction from the pollen-driven LOVE model to assess the effect of land opening on soil erosion during the Holocene. Concentrated erosion was evaluated by quantifying contemporary land incision using high-resolution LIDAR land surface elevation data. Furthermore, basin sedimentation was assessed with both paleoerosion signal and geophysical investigations. Our results highlight a significant contribution of concentrated erosion, corresponding to up to 46% of the total Holocene erosion exports. This underscores the importance of accurately quantifying the contributions of intensive erosion processes in long-term assessments of soil erosion and/or nutrient exports to avoid misestimating terrestrial fluxes in the critical zone.

Snow acts as an insulating layer on soils, preserving microbial function and promoting soil organic matter (SOM) mineralization over winter. Climate change is expected to increase the frequency of winter drought in temperate mountain ecosystems leading to snow-free winter, exposing soils to freezing and drying conditions that can disrupt microbial activity and key biogeochemical processes. However, the consequences of extreme snow drought event on microbial communities and associated C and N dynamics remain poorly understood, particularly from a functional and compositional perspective. This study aimed to investigate the ecological consequences of an extreme snow drought in subalpine grasslands by experimentally excluding all winter snowfall. By isolating the effects of a snow-free winter, without the confounding influences of warming or vegetation change, we were able to trace its impacts on ecosystem functioning from winter through the subsequent spring and summer. We observed a sharp spike in N2O emissions (+700 %) and a significant drop in CO2 fluxes (-70 %) during the snow-free winter, measured through discrete greenhouse gas flux sampling throughout the year, including winter. These changes coincided with immediate soil freezing and were linked to shifts in microbial community composition and function, assessed at three key periods-winter, spring, and peak growing season-using a combination of DNA-based community profiling, biomass quantification, and enzymatic assays. Functional markers showed widespread declines in microbial activity, including respiration, decomposition, and ammonification, along with a compositional shift toward anaerobic taxa and increased denitrification. These functional disruptions were further reflected in SOM mineralization dynamics, characterized via infrared spectroscopy and labile carbon fractions, and in reduced nitrogen cycling, measured through NH4+, NO3- content, and resin bag analyses. Although an extended growing season and compensatory microbial responses partially offset winter impacts, full functional recovery was not achieved by the end of the growing season. These findings highlight how snow-free winters, though extreme, can profoundly disrupt soil functioning, leaving lasting carryover effects that last into subsequent seasons.

Background Soil health degradation is a major threat to European food security, biodiversity, and climate stability. While scientists have debated how to define soil health during recent decades, a quantifiable framework for monitoring, management, and policy remains lacking.Aim We introduce SHERPA (Soil Health Evaluation, Rating Protocol, and Assessment) as a framework for discussion and present a first quantitative soil health assessment across Europe.Methods All major soil degradation processes (with the exception of organic contamination) were scored, averaged, and subtracted from the intrinsic soil health resulting in quantitative final scores.Results As reported before, cropland soils throughout Europe are highly degraded. Surprisingly, soil health of grasslands is also very negatively impacted. Soil erosion, nutrient surplus, and pesticide risk are largely driving poor soil health aligning with reported high biodiversity loss in agricultural land. Forest soils are also surprisingly low in health, mainly because of nitrogen surplus, reflecting documented widespread forest decline from nutrient imbalances. Interactive maps highlight specific threats to soil health across Europe, offering valuable insights for targeted action.Conclusions SHERPA is able to quantify soil health across Europe. However, at the current state of data availability, soil health is likely to be overestimated. Monitoring data of soil structure, compaction, pesticide spread and, in forest ecosystems, disturbance of humus layer are urgently needed for final assessment of soil health.

ABSTRACT Polycyclic aromatic hydrocarbons (PAHs) are ubiquitous persistent organic pollutants that accumulate in soils because of their high affinity for soil organic matter (SOM). As these pollutants are toxic to humans and the environment, a better understanding of their fate in the environment is required. This study aimed to assess the PAH distribution within soils according to different soil fractions: the free particulate organic matter (fPOM), the occluded particulate organic matter (oPOM) and the mineral‐associated organic matter (MaOM). PAH contents were measured in bulk soils and SOM fractions of alpine soils along an elevation gradient in the French Alps (Lautaret) from 1920 m to 2840 m a.s.l. A specific PAH distribution was identified, with highest PAH contents in the oPOM, followed by the fPOM, then the MaOM. Organic matter (OM) contents of each fraction can partly explain this distribution, but results of nuclear magnetic resonance (NMR) spectroscopy on fPOM and oPOM also highlighted a correlation between the PAH contents and the degree of decomposition of SOM. This indicates that the PAH distribution may be linked to the formation and transformation of fractions: (i) PAHs in the fPOM correspond to relatively recent deposits and mainly reflect the background contamination, (ii) in the oPOM are the PAHs that resist biodegradation during the transformation of fPOM into oPOM and accumulate in the oPOM; this accumulation may be further enhanced by the formation of aggregates. Finally, (iii) in the MaOM, the lower PAH contents can be explained by the different formation pathway of this fraction and its high degree of decomposition. As the PAH distribution may have an impact on their dynamics in soils, it should be taken into consideration in future research.

Estimating SOC stocks and stability, as well as modeling their response to rising temperatures, is crucial for predicting climate change impacts. This is particularly true in mountainous regions, where low temperatures slow down SOC decomposition, resulting in higher SOC stocks compared to soils at lower elevations. However, these stocks are also more vulnerable to warming, increasing the risk of SOC depletion. Such conditions create the potential for a positive feedback loop in which warming accelerates SOC losses, further amplifying climate change impacts on these sensitive ecosystems. To better understand the factors controlling SOC stocks and stability in mountain soils, we sampled 170 soil profiles along 29 elevation gradients in the western Alps from 280 to 3160 m a.s.l. We assessed SOC stocks and chemical composition using mid-infrared spectroscopy method and SOC stability with Rock-Eval (R) thermal analysis. Our findings, based on an unprecedented dataset, reveal a clear elevational pattern in SOC properties. SOC stocks increase with elevation up to the montane belt (1200-1500 m a.s.l.), remain relatively stable through the subalpine zone, and then decline beyond the subalpine/alpine boundary (2200-2400 m a.s.l.). Notably, this transition is also marked by a significant drop in SOC stability, suggesting a shift in the dominant stabilization processes at higher elevations. Our results also indicate that SOC stocks and stability are influenced by a complex interplay of factors. At higher elevations, climate emerges to be the dominant factor, whereas lithology and weathering play a more significant role at lower elevations. These results suggest that at high-elevations, harsh climatic conditions favor stabilization of SOC, while less developed soils limit organo-mineral interactions. In contrast, at warmer, lower elevations with higher carbon fluxes, more developed soils facilitate organo-mineral interactions, thereby enhancing SOC stability in the long term. Consequently, alpine grasslands, which contain substantial stocks of labile carbon stabilized by climatic conditions, appear to be particularly vulnerable to the effects of climate warming.

The evolution of soil organic carbon (SOC) stocks is critical for both food production and climate change mitigation. This study uses advanced electron paramagnetic resonance (EPR) spectroscopy to investigate the spatial localisation and characterisation of organic carbon in speleothems, with a particular focus on methodological advances in recent decades. A speleothem sample from the Choranche cave in France was analysed using UV laser-induced fluorescence (LIF) and continuous wave EPR spectroscopy. The LIF analysis identified three main types of organic compounds- aromatic amino acids, aliphatic aromatics and larger aromatic compounds- distributed throughout the sample. EPR spectroscopy revealed the presence of Fe3+ and Mn2+ ions, along the entire sample for Fe3+ and more localised for Mn2+. When radical organic matter (ROM) is detected, first and second harmonic EPR imaging shows its collocation with Fe3+ and Mn2+, suggesting specific embedding conditions or source events. The study highlights a significant discrepancy between fluorescent organic matter (FOM) and ROM, challenging previous assumptions about their co-transfer from soil to speleothems. The results suggest that ROM is likely to be associated with specific soil redox conditions or high-energy events, whereas FOM represents a continuous background transfer. This distinction is crucial for accurate interpretations of soil organic carbon loss and its environmental implications. Future research should integrate detailed spectroscopic and isotopic analyses to better quantify organic carbon dynamics and their environmental proxies. Our results highlight the importance of distinguishing between different types of organic matter in speleothems to improve our understanding of soil organic carbon fluxes in relation to climate and human land use.

Alpine and subalpine grasslands experience strong seasonal climatic variations, with snow cover for over six months maintaining steady soil temperature and moisture. This seasonal structure limits plant growth and strongly influences microbial activity, which together control key ecosystem functions like soil organic matter (SOM) inputs, mineralization, and greenhouse gas fluxes, ultimately influencing the composition and quantity of SOM. In our study, we monitored soil pedoclimate (temperature and moisture) and net ecosystem exchange (CO2 flux) at both plot and local footprint scales, using discrete measurements and flux tower, to characterize the seasonal context of alpine and subalpine grasslands. Meanwhile, we investigated the seasonal properties of topsoil organic matter (SOM) at six key times throughout the hydrological year: before snow cover, before snowmelt, after snowmelt, during the growing season, at vegetation peak, and during senescence. SOM properties were analyzed through the combination of methods including DRIFT spectroscopy, RockEval® thermal analysis, water-extractable organic carbon and permanganate-oxidizable carbon (POxC). Finally, soil incubations were conducted to assess microbial respiration sensitivity to temperature and moisture across these six periods, enhancing our understanding of seasonality’s impact on microbial features. Our study integrates in situ and in vitro measurements across multiple scales (soil sample, plot, and landscape), traditionally analyzed separately. This approach bridges microbial mechanisms with SOM quality and links them to ecosystem-scale carbon exchanges. Our findings highlighted a clear seasonality in SOM properties, offering valuable insights into the functioning of these grasslands. We identified a labile seasonal pool of SOM that persists through the winter due to low temperatures and low-carbon outputs, maintaining its availability for mineralization at the onset of the growing season—when primary producers have the highest nutrient demand. This labile pool decreases over the growing season, as microbial activity peaks and organic matter inputs decline. Additionally, seasonal shifts in microbial responses to temperature and humidity indicate functional acclimations: enhanced cold tolerance in winter, waterlogged tolerance during snowmelt, and increased capacity to degrade complex organic molecules during the growing season.

Over the past four decades, seasonal snow cover has declined rapidly in temperate alpine regions. However, the fine-scale dynamics of snowmelt preceding the ongoing warming period remain largely unknown, limiting our understanding of the long-term influence of past snow cover on alpine ecosystems. Here we rely upon the spatial similarities in melt-out patterns and a temperature-based model of fractional snow cover area, to reconstruct fine-scale snow cover changes over the past 250 years in instrumented catchments of the southwestern Alps. We provide evidence that, until the 1980s, prolonged snow cover in many late-lying snowfields delayed ecosystem development and explain why current vegetation cover, soil organic matter content, and mineral weathering are significantly lower in these areas than in surrounding ecosystems. These findings highlight the long-term legacy of snow cover on alpine landscapes and underscore the need to re-evaluate its effects on ecosystem structure, functioning, and responsiveness to ongoing changes.

Alpine ecosystems, shaped by cold temperatures and prolonged snow cover, are warming twice as fast as lowlands, making them particularly vulnerable to climate change. This rapid warming alters ecosystem functioning by increasing soil temperatures and shifting snow regimes, leading to shorter snow cover periods and longer growing seasons. Such changes impact soil organic matter (SOM), which regulates carbon storage and soil fertility through microbial mineralization. We investigated long-term SOM changes following experimental warming in alpine grasslands. By transplanting soil plots downslope (2470 m a.s.l. to 1920 m a.s.l.), we increased mean annual temperature by 3 degrees C and extended the growing season by 77 days. Seven years later, we analyzed soil nutrients content, SOM characteristics (chemistry, labile pools, and thermal stability) and conducted discrete flux measurements to determine net ecosystem exchange (NEE). Soil incubations assessed microbial traits and their acclimation to warming. Our results indicate that, even after seven years, alpine soil microbial activities showed limited acclimation to warming, contributing to SOM destabilization and soil nutrients enrichment by boosting mineralization. Indeed, warmed plots acted as carbon sources, with an 18 % decrease in SOC stocks and increased NEE. Carbon losses exceeded C gains from plant productivity, primarily depleting labile pools. This may create a positive feedback loop between carbon cycling and climate warming. These findings highlight the long-term consequences of temperature increases and snow regime shifts on alpine ecosystem functioning and suggest that soil carbon losses in warming mountain environments may continue over time.

How ecosystems will provide ecosystem services in the future given uncertain changes in climate and land use is an open question that challenges decision-making on adaptation to climate change. Prospective assessments of ecosystem services should carefully include and communicate the sources of uncertainties that affect the predictions. We used the ecosystem service of soil protection against erosion in the Maurienne Valley (French Alps) as a case study to illustrate how several sources of uncertainties can be integrated into an assessment of future ecosystem service supply. We modeled future erosion rates in the Maurienne Valley for years 2020 and 2085 using the Revised Universal Soil Loss Equation (RUSLE) and six climatic and socioeconomic scenarios. We quantified how the ecosystem service supply will be likely affected by climate and land-use change, separately and jointly. We assessed the effects of different sources of uncertainty on projected erosion rates: scenarios, climate models choice, and methods to parametrize the ecosystem service model. Land-use change increased erosion (+ 3.3 ton.ha-1.yr-1 on average, with significant increases in 81 % of the study site), while climate change contributed to a slight reduction (-0.21 ton.ha-1.yr-1 on average with significant decrease 20 % of the study site). The uncertainty of the ecosystem service model parameterization explained 93 % of the variance in erosion values. Furthermore, uncertainty linked to climate models and future scenarios contributed almost equally to the variability in the direction (positive or negative) of erosion change (41 % and 38 % respectively). The uncertainties surrounding the direction of future changes in ecosystem services come mainly from uncertainties in climate models and future scenarios rather than from uncertainties in the ecosystem service model parameters. Assessing the likelihood of future changes in ecosystem services helps prioritize locations where adaptation solutions are likely to be needed.

ABSTRACTMountains are particularly vulnerable to climate change, as they are warming at a rate that exceeds the global average, significantly impacting cold‐adapted ecosystems. In these environments, soil organic matter (SOM) stocks are often considerably larger than at lower elevations. These stocks are therefore highly susceptible to global warming and the associated risk of greenhouse gas (GHG) (CO₂, CH₄, N₂O) emissions driven by temperature‐induced increases in SOM mineralisation. In order to quantify these emissions and the change of mineralisation rates under warming, it is necessary to gain an understanding of the annual mineralisation balance. We investigated how warming impacts the duration and intensity of mineralisation in different seasons. The main aim of this study is to quantify alpine SOM mineralisation rates and GHG production under a range of seasonal conditions, including those associated with warming. An in vitro approach was employed to expose alpine topsoils (0–10 cm) to the conditions of key seasonal periods: snow cover, growing season and rainfall/snowmelt. This was achieved by experimentally varying temperature and inflow of precipitation water. Additionally, the soil samples were subjected to a temperature increase of 4°C. The short‐term responses of carbon (C), nitrogen (N) and phosphorus (P) mineralisation and GHG production were monitored. The results demonstrated that alpine soil respiration rates exhibited a twofold increase with a 4°C warming, while the relative proportion of labile SOM demonstrated a decline with rising temperatures. Water saturation from simulated rain and snowmelt played a crucial role in organic matter mineralisation and increased the mineralisation of carbon (+12% to +53%), nitrogen (+20% to +80% of net ammonification) and phosphorus (+50% of net phosphate production). This suggests that nutrients present in the snowpack or the rain were added to the soil. In contrast, soil–water saturation decreased net nitrate production by between 10% and 90%. The results of this study highlight the potential for alpine soil warming to release labile SOM and demonstrate the influence of the snow regime on nutrient and carbon fluxes.

Agropastoral activities have impacted the habitable part of our planet-the "Critical Zone"-for thousands of years, triggering a major increase in soil erosion in mountain environments. Understanding and quantifying the impact of these activities on soil is central to the well-being of our societies. Here, we investigate the isotope ratios of the trace element lithium in detrital sediments of Lake Bourget, European Alps, and provide a reconstruction of the impact of human activities on the evolution of alpine soil during the Holocene. We demonstrate that during the Early Holocene, soil formation was altered by the development of pastoralism followed by tillage. This led to three major erosive surges (3.8 to 3.0, 2.8 to 1.6, and 1.6 ky cal BP to modern times), thinning soils down to a state close to that of their early development 10,000 y ago. The detailed study of the lithium detrital signal reveals the appearance of an altitudinal decoupling in the response of the Critical Zone in the Alps following the development of the agropastoral activities during the Iron Age. The onset of agropastoral activities disrupted the balance between soil formation and erosion, leading to erosion rates 3 to 10 times faster than soil production since the end of the Ice Age.

Increasing climate warming and summer droughts are known to affect mountain plant communities, their functional traits and life strategies. However, little is known about how strongly and efficiently communities respond to climate change, and how tightly plant responses are linked to responses of ecosystem functions. To test this, we transplanted alpine plant communities to subalpine conditions, exposing them to warming and drying. We compared these transplanted communities to alpine and subalpine control communities to assess their responses. Five years after transplantation, we found slower growth (e.g. lower leaf nitrogen) and more outsourcing strategies (e.g. lower specific root length) in the warmer and drier subalpine control communities compared to the alpine controls, probably due to drought. Traits of warmed alpine communities shifted toward subalpine controls. However, neither below- nor aboveground traits nor productivity of plants fully acclimated to subalpine conditions. Nevertheless, standard litter decomposition rates, arbuscular colonization and bacterial biomass showed no acclimation lag to the subalpine controls. Significant but insufficient acclimation of plant functional traits and strategies is prone to result in maladapted plant productivity, impairing competitiveness with better adapted subalpine species and leading to the temporally delayed loss of ecosystem features specific to alpine environments.

In harsh environments such as Alpine screes, the substrate represents one of the main selective constraints for plants. The substrate’s influence on species distributions has been well characterised using major lithological classes. However, our understanding of the local heterogeneity of substrates and the consequent impact on plants, particularly in mountainous environments, remains limited. By analysing multiple facets of the local substrates, we tried to identify if ecological barriers separate scree specialist species. We sampled substrates in contact with the rhizosphere of plants from a species complex within the genus Noccaea, in 53 sites distributed across the Alps. We analysed the composition and properties of the substrates, in addition with climatic and topographic variables. We compared multiple aspects of substrate between pairs of species and discuss the potential impact of edaphic diversity in terms of plant physiology. The analyses demonstrated that the edaphic diversity forms a continuum rather than distinct classes, and shelter a very large diversity of conditions, particularly within carbonated rocks. We found that Alpine endemic Noccaea species grow in substrates that significantly differ in chemical properties and concentrations in toxic elements, while the topography and climate occupied by different species remain comparatively similar. This study demonstrates that the local diversity of edaphic conditions in alpine screes has been underestimated. The fine-scale differences in the concentration of nutrient or toxic elements appear to be a key factor in the distribution and adaptation of plants, and possibly a reason for the high degree of endemism observed in scree habitats within the Alps.

Polycyclic aromatic hydrocarbons (PAHs) are toxic and persistent organic pollutants that are ubiquitous in the environment. In particular, their presence and persistence in soils represent public health and environmental problems. However, no regulation of the levels of these pollutants in soils has been introduced in France. Despite a few nationwide monitoring studies of PAHs, little is known about the contamination by these pollutants in ecosystems such as mountainous areas. The aim of this study was therefore to assess PAH soil contamination in the northern French Alps and to improve our understanding of (i) their spatial distribution and (ii) their dynamic on sites submitted to very local past and present contamination. To this end, PAH levels were measured in the topsoil (0–10 cm) along eight elevational gradients (220–2700 m), the latter allowing us to study a wide range of climatic and ecological conditions and different levels of PAH exposure over a limited spatial area. The results revealed that the distance from sources played a major role in determining the PAH distribution. In addition, sites subject to major past and present PAH-emitting industrial activities were studied. At the site subject to current high emissions, PAH levels were extremely high, with a specific signature. In contrast, at site that was no longer subject to any particular source, the PAH levels were close to the regional background. The complementary study of these different types of sites allowed us to improve our knowledge of the spatial and temporal dynamics of PAHs in mountainous areas.

Soils host a quarter of all terrestrial species, and the pivotal role played by earthworms in soil ecosystem functioning is thought to be paramount. Here, we aimed to quantify the causal influence of changing environments, especially climate, soil conditions, and vegetation type and structure, on earthworm diversity and the subsequent impacts on ecosystem functions. French Alps. 2016–2021. Earthworms and plants. Along 17 elevational gradients, we sampled climate, soil conditions, plant relevés, ecosystem functions, and earthworm diversity from soil environmental DNA. Through a causal inference framework and structural equation modelling, we quantified how climate conditions structure soil conditions and vegetation structure along the gradients, how these three compartments shape earthworm diversity, and how, in turn, earthworm diversity modulates ecosystem functions in addition to direct environmental impacts. Vegetation was the most important driver of earthworm diversity, with acquisitive plant strategies contributing to an increase in earthworm diversity, followed by soil organic matter. While climate was important, its impact on earthworm diversity was only indirect, through cascading effects mediated by vegetation and soil. In addition to the abiotic drivers, earthworm diversity had an important effect on three out of the four studied ecosystem functions (i.e., carbon stock, plant primary productivity and carbon and nitrogen limitation of microbes). In closed habitats, earthworm diversity was positively linked to proxies of carbon stock and carbon availability, while in grasslands, it was positively linked to proxies of nitrogen availability. Interestingly, aboveground productivity was found to be independent of earthworm diversity. Our study underscores the central role of earthworm diversity in linking environmental drivers to ecosystem functioning. We emphasise the importance of incorporating earthworm diversity in models aiming to elucidate the cascading effects of climate change across various ecosystem compartments, ultimately shaping ecosystem functioning.