Organic matter (OM) is a key component of lake sediments, affecting carbon, nutrient, and trace metal cycling at local and global scales. Yet little is known about long‐term (millennial) changes in OM composition due to the inherent chemical complexity arising from multiple OM sources and from secondary transformations. In this study we explore how the molecular composition of OM changes throughout the Holocene in two adjacent boreal lakes in central Sweden and compare molecular‐level information with conventional OM variables, including total carbon, total nitrogen, C:N ratios, δ13C, and δ15N. To characterize the molecular OM composition, we employed a new method based on pyrolysis‐gas chromatography/mass spectrometry (Py‐GC/MS), which yields semiquantitative data on >100 organic compounds of different origin and degradation status. We identify large changes in OM composition after deglaciation (circa 8500 ± 500 B.C.), associated with early landscape development, and during the most recent 40–50 years, driven by degradation processes. With molecular‐level information we can also distinguish between natural landscape development and human catchment disturbance during the last 1700 years. Our study demonstrates that characterization of the molecular OM composition by the high‐throughput Py‐GC/MS method is an efficient complement to conventional OM variables for identification and understanding of past OM dynamics in lake‐sediment records. Holocene changes observed for pyrolytic compounds and compound classes known for having different reactivity indicate the need for further paleo‐reconstruction of the molecular OM composition to better understand both past and future OM dynamics and associated environmental changes.

Monitoring of surface waters in the boreal region over the last decades shows that waters are becoming browner. This timeframe may not, however, be sufficient to capture underlying trajectories and driving mechanisms of lake-water quality, important for prediction of future trajectories. Here we synthesize data from seven lakes in the Swedish boreal landscape, with contemporary lake-water total organic carbon (TOC) concentrations of 1.4-14.4 mg L −1 , to conceptualize how natural and particularly human-driven processes at the landscape scale have regulated lake-water TOC levels over the Holocene. Sediment-inferred trends in TOC are supported by several proxies, including diatom-inferred pH. Before $ 700 CE, all lakes were naturally acidic (pH 4.7-5.4) and the concentrations of inferred lake-water TOC were high (10-23 mg L −1 ). The introduction of traditional human land use from $ 700 CE led to a decrease in lake-water TOC in all lakes (to 5-14 mg L −1 ), and in four poorly buffered lakes, also to an increase in pH by > 1 unit. During the 20 th century, industrial acid deposition was superimposed on centuries of land use, which resulted in unprecedentedly low lake-water TOC in all lakes (3-11 mg L −1 ) and severely reduced pH in the four poorly buffered lakes. The other lakes resisted pH changes, likely due to close connections to peatlands. Our results indicate that an important part of the recent browning of boreal lakes is a recovery from human impacts. Furthermore, on a conceptual level we stress that contemporary environmental changes occur within the context of past, long-term disturbances.

On the annual and interannual scales, lake microbial communities are known to be heavily influenced by environmental conditions both in the lake and in its terrestrial surroundings. However, the influence of landscape setting and environmental change on shaping these communities over a longer (millennial) timescale is rarely studied. Here, we applied an 18S metabarcoding approach to DNA preserved in Holocene sediment records from two pairs of co-located Swedish mountain lakes. Our data revealed that the microbial eukaryotic communities were strongly influenced by catchment characteristics rather than location. More precisely, the microbial communities from the two bedrock lakes were largely dominated by unclassified Alveolata, while the peatland lakes showed a more diverse microbial community, with Ciliophora, Chlorophyta and Chytrids among the more predominant groups. Furthermore, for the two bedrock-dominated lakes—where the oldest DNA samples are dated to only a few hundred years after the lake formation—certain Alveolata, Chlorophytes, Stramenopiles and Rhizaria taxa were found prevalent throughout all the sediment profiles. Our work highlights the importance of species sorting due to landscape setting and the persistence of microbial eukaryotic diversity over millennial timescales in shaping modern lake microbial communities.

Sedimentary environmental DNA (sed-eDNA) coupled with metabarcoding is increasingly exploited for ecological studies, but application of the method to resolve fish dynamics in lakes still needs better validation. This study (1) evaluated the sed-eDNA yields from the commonly used DNeasy PowerSoil DNA Kit from mineral-rich and organic-rich sediments and (2) examined the viability of fish sed-eDNA recovery and detection in surface sediment samples from 13 Swedish mountain lakes, with organic contents of 18–52%, by using conventional PCR and droplet digital PCR. Based on concurrent fish-population surveys these lakes contain arctic char and brown trout. We show that, compared to other specifically designed lysis buffers, the DNeasy PowerSoil DNA Kit is less effective to recover DNA from organic-rich sediments and almost 50% of the extracted DNA was lost during purification steps. The amplification of fish sed-eDNA using conventional PCR with teleo primers failed to detect positive signals; whereas ddPCR assays enabled quantification of amplifiable DNA in all the extracts. However, further molecular cloning of the positive ddPCR droplets from one sediment sample revealed amplified sequences of unidentified origin that cannot be aligned well to fish. Thus the performance of the teleo primers for quantification of fish sed-eDNA detection requires further examination. For detection of fish sed-eDNA for ecological studies, we suggest that DNA extraction methods and primers should be carefully selected and the performance of ddPCR to detect DNA at low quantities needs to be further scrutinized to circumvent the pitfalls of false positives.