Over recent decades, palaeolimnological records from remote sites have provided convincing evidence for the onset and development of several facets of global environmental change. Remote lakes, defined here as those occurring in high latitude or high altitude regions, have the advantage of not being overprinted by local anthropogenic processes. As such, many of these sites record broad-scale environmental changes, frequently driven by regime shifts in the Earth system. Here, we review a selection of studies from North America and Europe and discuss their broader implications. The history of investigation has evolved synchronously with the scope and awareness of environmental problems. An initial focus on acid deposition switched to metal and other types of pollutants, then climate change and eventually to atmospheric deposition-fertilising effects. However, none of these topics is independent of the other, and all of them affect ecosystem function and biodiversity in profound ways. Currently, remote lake palaeolimnology is developing unique datasets for each region investigated that benchmark current trends with A celebration of Prof. Rick Battarbee's contributions to palaeolimnology, edited by Holmes et al.This paper has been written as a contribution to celebrating Rick Battarbee's influence on palaeolimnology. Some of us have benefitted from his leadership (and friendship) in transnational European projects during the last decade (e.g., ALPE, ALPE2, MOLAR, CHILL-10000, EMERGE, EUROLIMPACS), which together with some other initiatives spawned pan-European remote lake research. Others have respected Rick as a teacher, colleague and a friend. To some extent, this review follows the chronological order of topics addressed in these projects, which also respond to the growing social awareness about each issue. Rick also facilitated bridges between North American and European schools, and beyond. We expect his attitude towards collaboration will pervade and persist through the palaeolimnological community for years to come, and global change will certainly provide stimulating and challenging questions with which to do so.

Summary 1. Species assemblages of diatoms, rotifers, chydorids, planktonic crustaceans and chironomids were studied in 235 alpine lakes in the Alps, Pyrenees, Tatras (Western Carpathians), Retezat (Southern Carpathians) and Rila Mountains (Balkans). 2. For all taxonomic groups we found a hierarchical structure in the community assemblage using distinct scales of lake clustering (number of k‐means groups) based on species composition similarity (Hellinger distance). We determined the optimal partition in assemblage types (i.e. number of lake clusters) for each taxonomic group by maximising the sum of the taxon indicative value (IndVal) and performed discriminant analyses, using environmental variables not conditioned by geographical patterns. Relevant environmental variables differed among and within taxonomic groups. Therefore the assemblages respond to a complex environmental mosaic, with the exception of diatom assemblages, which followed an acid–base gradient. 3. The significant environmental variables could be grouped into four general factors: lake size, tropho‐dynamic status, acid–base balance and ice‐cover duration (i.e., altitudinal gradient). Lake size was significant for the highest number of assemblage types; however, the most significant factor differed among taxonomic groups: acid–base balance for diatoms, lake size for rotifers, ice‐cover duration for chydorids and planktonic crustaceans and tropho‐dynamic status for chironomids. No single environmental typology accounted for the assemblage structure of all taxonomic groups. 4. However, defining ecological thresholds as values within environmental gradients at which the rate of change in assemblages is accelerated relative to points distant from that threshold, we were able to find specific threshold values for each of the four main general environmental factors identified, which were relevant across several taxonomic groups: 3 ha for lake area; 0.6 mg L−1 for dissolved organic carbon; 190 days for ice‐cover duration and 200 μeq L−1 for acid neutralising capacity. Above and below these values ecosystem organisation change substantially. They have direct applications in establishing lake typologies for environmental quality and biodiversity conservation programmes, and in improving predictions about global change impacts.

1. We carried out a coordinated survey of mountain lakes covering the main ranges across Europe (including Greenland), sampling 379 lakes above the local tree line in 2000. The objectives were to identify the main sources of chemical variability in mountain lakes, define a chemical classification of lakes, and develop tools to extrapolate our results to regional lake populations through an empirical regionalisation or upscaling of chemical properties. 2. We investigated the main causes of chemical variability using factor analysis (FA) and empirical relationships between chemistry and several environmental variables. Weathering, sea salt inputs, atmospheric deposition of N and S, and biological activity in soils of the catchment were identified as the major drivers of lake chemistry. 3. We tested discriminant analysis (DA) to predict the lake chemistry. It was possible to use the lithology of the catchments to predict the range of Ca 2+ and SO 4 2) into which a lake of unknown chemistry will decrease. Lakes with lower SO 4 2) concentrations have little geologically derived S, and better reflect the variations in atmospheric S loading. The influence of marine aerosols on lakewater chemistry could also be predicted from the minimum distance to the sea and altitude of the lakes. 4. The most remarkable result of FA was to reveal a factor correlated to DOC (positively) and NO 3 ) (negatively). This inverse relationship might be the result either of independent processes active in the catchment soils and acting in an opposite sense, or a direct interaction, e.g. limitation of denitrification by DOC availability. Such a relationship has been reported in the recent literature in many sites and at all scales, appearing to be a global pattern that could reflect the link between the C and N cycles. 5. The concentration of NO 3 ) is determined by both atmospheric N deposition and the processing capacity of the catchments (i.e. N uptake by plants and soil microbes). The fraction of the variability in NO 3 ) because of atmospheric deposition is captured by an independent factor in the FA. This is the only factor showing a clear pattern when mapped over Europe, indicating lower N deposition in the northernmost areas. 6. A classification has been derived which takes into account all the major chemical features of the mountain lakes in Europe. FA provided the criteria to establish the most important factors influencing lake water chemistry, define classes within them, and classify the surveyed lakes into each class. DA can be used as a tool to scale up the classification to unsurveyed lakes, regarding sensitivity to acidification, marine influence and sources of S.

1.A survey of c. 350 remote high altitude and high latitude lakes from 11 different mountain regions was undertaken to explore species distribution across Europe at a scale not previously attempted. 2. Lakes were sampled for planktonic crustaceans, rotifers, littoral invertebrates and subfossil chironomids, diatoms and cladocerans. Each lake was characterised in terms of water chemistry, morphology, catchment attributes and geographical location. 3. Separate TWI NSPAN TWI NSPAN analyses were undertaken on diatom, chironomid, planktonic crustacean, littoral invertebrate and cladoceran (chydorids only) data to classify sites according to taxonomic composition. For most datasets there was a spatial component to the classification with distinct geographical groups emerging -Norway and Scotland, Finland and Central ⁄ Eastern Europe. 4. Constrained ordination methods were employed to examine how species responded to a range of environmental factors, which were aggregated into a series of component groups -proximal environment (the chemical, trophic and physical attributes of the lake), catchment characteristics and geographical location. Several key environmental gradients were identified, which explained significant levels of the variance across several of the biological groups including dissolved organic carbon (chironomids, planktonic crustaceans), temperature (chironomids and littoral invertebrates), chloride ⁄ sea-salt (littoral invertebrates, diatoms and rotifers), lake morphology (all groups), calcium ⁄ pH (diatoms), nitrate (chydorids, littoral invertebrates, rotifers and planktonic crustaceans) and fish (littoral invertebrates). In some cases these statistical relationships are likely to represent direct ecological constraints and, in others, it is probable that the environmental variable is acting as a surrogate for some other attribute or process. 5. Variance partitioning was undertaken to quantify how much of the variation in each biological group could be uniquely attributed to variables representing the proximal environment, catchment characteristics and geographical location. For most groups the Correspondence: Martin Kernan, Environmental

1. We surveyed the distribution of several trace elements in contemporary and preindustrial sediment s in 275 lakes in alpine and arctic lake districts across Europe including the Pyrenees, Alps, the Rila Mountains, Retezat, Julian Alps, Tatras, Scottish mountains, Central Norway and Greenland. 2. Sediment cores were collected at the deepest part of each lake and analysed at two depths (surface sediment and at 15-17 cm depth) for Ti, Pb, Cd, Zn, Cu, As, Hg and Se. 3. The concentrations of trace elements found in the lakes included in the survey are comparable to those reported in aquatic sediments receiving higher contamination loads. With the exception of Greenland, a large percentage of lakes showed enrichment factors for most elements well above 1.5, indicating atmospheric contamination. The influence of contamination has increased the co-distribution of trace elements in sediments, with the exception of As. 4. Pb is the element that shows the highest contamination level at the European scale, followed by Hg and As. Zn, Cd, Cu and Se contamination is detectable to a lower degree. 5. The Tatra Mountains and Scotland seem to be most affected. Natural mechanisms leading to the formation of highly organic, metal-binding sediments may be the cause of the high levels in Scotland, whereas those in the Tatras appear to be due to elevated deposition. 6. The Retezat and Central Norway appear to be least polluted. 7. In the Alps, enrichments in Pb, Hg and Zn are higher in southern than in central areas suggesting a flux of these pollutants from the south. In the Pyrenees, the high natural levels of As are remarkable. Metal enrichments in the Rila Mountains are comparable to those in the Tatras, but concentrations are much lower. 8. In general terms, the increase in trace elements in modern with respect to pre-industrial sediments reflects the history of a long range contamination affecting the remotest locations in Europe.

Major fluxes of sulphur and dissolved inorganic nitrogen were estimated in Central European mountain ecosystems of the Bohemian Forest (forest lakes) and Tatra Mountains (alpine lakes) over the industrial period. Sulphur outputs from these ecosystems were comparable to inputs during a period of relatively stable atmospheric deposition (10-35 mmol m -2 yr -1 ) around the 1930s. Atmospheric inputs of sulphur increased by three-to four-fold between the 1950s and 1980s to ~140 and ~60 mmol m -2 yr -1 in the Bohemian Forest and Tatra Mountains, respectively. Sulphur outputs were lower than inputs due to accumulation in soils, which was higher in forest soils than in the sparser alpine soils and represented 0.8-1.6 and 0.2-0.3 mol m -2 , respectively, for the whole 1930-2000 period. In the 1990s, atmospheric inputs of sulphur decreased 80% and 50% in the Bohemian Forest and Tatra Mountains, respectively, and sulphur outputs exceeded inputs. Catchment soils became pronounced sources of sulphur with output fluxes averaging between 15 and 31 mmol m -2 yr -1 . Higher sulphur accumulation in the forest soils has delayed (by several decades) recovery of forest lakes from acidification compared to alpine lakes. Estimated deposition of dissolved inorganic nitrogen was 53-75 mmol m -2 yr -1 in the Bohemian Forest and 35-45 mmol m -2 yr -1 in the Tatra Mountains in the 1880-1950 period, i.e. below the empirically derived threshold of ~70 mmol m -2 yr -1 , above which nitrogen leaching often occurs. Dissolved inorganic nitrogen was efficiently retained in the ecosystems and nitrate export was negligible (0-7 mmol m -2 yr -1 ). By the 1980s, nitrogen deposition increased to ~160 and ~80 mmol m -2 yr -1 in the Bohemian Forest and Tatra Mountains, respectively, and nitrogen output increased to 120 and 60 mmol m -2 yr -1 . Moreover, assimilation of nitrogen in soils declined from ~40 to 10-20 mmol m -2 yr -1 in the alpine soils and even more in the Bohemian Forest, where one of the catchments has even become a net source of nitrogen. In the 1990s, nitrogen deposition decreased by ~30% and DIN output decreased to <70 and 35 mmol m -2 yr -1 in the Bohemian Forest and Tatra Mountains, respectively. New steady-state conditions, with negligible nitrogen export, could be reached in future but at lower nitrogen depositions than in the 1930s.

No abstract

Ninety-one lakes distributed along the Tatra Mountains (most of lakes > 1 ha and 65% of lakes > 0.01 ha) were sampled and analysed for ionic and nutrient composition in September 2004 (15 years after reduction in acid deposition). Eighty-one lakes were in alpine zone and ten lakes in Norway spruce forest. The results were compared to similar lake surveys from 1994 (the beginning of water recovery from acidification) and 1984 (maximum acidification). Atmospheric deposition of SO + and Al concentrations occurred in the most acid lakes. On a regional basis, no significant change was observed for total phosphorus, total organic nitrogen, and dissolved organic carbon (DOC) in the 1994-2004 period. However, these parameters increased in forest lakes, which exhibited an increasing trend in DOC concentrations, inversely related (P < 0.001) to their decreasing ionic strength (30% on average in [1994][1995][1996][1997][1998][1999][2000][2001][2002][2003][2004].