Skeletonema marinoi ecotypes show specific habitat-related responses to fluctuating light supporting high potential for growth under photobioreactor light regime.

Diatoms are a diverse group of phytoplankton usually dominating areas characterized by rapidly shifting light conditions. Because of their high growth rates and interesting biochemical profile, their biomass is considered for various commercial applications. This study aimed at identifying strains with superior growth in a photobioreactor (PBR) by screening the natural intraspecific diversity of ecotypes isolated from different habitats. We investigated the effect of PBR light fluctuating on a millisecond scale (FL, simulating the light in a PBR) on 19 ecotypes of the diatom Skeletonema marinoi isolated from the North Sea-Baltic Sea area. We compare growth, pigment ratios, phylogeny, photo-physiological variables and photoacclimation strategies between all strains and perform qPCR and absorption spectra analysis on a subset of strains. Our results show that the ecotypes responded differently to FL, and have contrasting photo-physiological and photoprotective strategies. The strains from Kattegat performed better in FL, and shared common photoacclimation and photoprotection strategies that are the results of adaptation to the specific light climate of the Kattegat area. The strains that performed better with FL conditions had a high light (HL)-acclimated phenotype coupled with unique nonphotochemical quenching features. Based on their characteristics, three strains were identified as good candidates for growth in PBRs.

Land use change and coastal water darkening drive synchronous dynamics in phytoplankton and fish phenology on centennial timescales.

At high latitudes, the suitable window for timing reproductive events is particularly narrow, promoting tight synchrony between trophic levels. Climate change may disrupt this synchrony due to diverging responses to temperature between, for example, the early life stages of higher trophic levels and their food resources. Evidence for this is equivocal, and the role of compensatory mechanisms is poorly understood. Here, we show how a combination of ocean warming and coastal water darkening drive long-term changes in phytoplankton spring bloom timing in Lofoten Norway, and how spawning time of Northeast Arctic cod responds in synchrony. Spring bloom timing was derived from hydrographical observations dating back to 1936, while cod spawning time was estimated from weekly fisheries catch and roe landing data since 1877. Our results suggest that land use change and freshwater run-off causing coastal water darkening has gradually delayed the spring bloom up to the late 1980s after which ocean warming has caused it to advance. The cod appear to track phytoplankton dynamics by timing gonadal development and spawning to maximize overlap between offspring hatch date and predicted resource availability. This finding emphasises the importance of land-ocean coupling for coastal ecosystem functioning, and the potential for fish to adapt through phenotypic plasticity.

Drivers of copper sensitivity in copepods: A meta-analysis of LC50s.

Copper is both an essential trace element and a potent pesticide. The use of copper as an antifoulant has increased in the last decades in line with the expanding aquaculture and shipping industries. In aquatic environments, it also affects non-target taxa. One of which are copepods, which constitute the central link in the marine food web. Despite their ecological importance, there are no systematic reviews of the lethal concentration range and drivers of copper toxicity in this taxon. Here, we combined literature data from 31 peer-reviewed articles recording the Lethal Concentration 50 (LC50) for copper in copepods and the experiments' respective environmental, developmental, and taxonomic parameters. The LC50 is a traditional endpoint for toxicity testing used in standardized toxicity testing and many ecological studies. In total, we were able to extract 166 LC50 entries. The variability in the metadata allowed for a general analysis of the drivers of copper sensitivity in copepods. Using a generalized additive modeling approach, we find that temperature increases copper toxicity when above approximately 25℃. Counter to our expectations; salinity does not influence copper sensitivity across copepod species. Unsurprisingly, nauplii are more susceptible to copper exposure than adult copepods, and benthos-associated harpacticoids are less sensitive to copper than pelagic calanoids. Our final model can predict sensible specific-specific copper concentrations for future experiments, thus giving an informed analytical approach to range testing in future dose-response experiments. Our model can also potentially improve ecological risk assessment by accounting for environmental differences. The approach can be applied to other toxicants and taxa, which may reveal underlying patterns otherwise obscured by taxonomic and experimental variability.

Effects of Maternal Reproductive Investment on Sex-Specific Pollutant Accumulation in Seabirds: A Meta-Analysis.

Avian egg production demands resources such as lipids and proteins. Relative egg size and mass varies across species, reflecting differences in maternal investment. This variability may affect the maternal transfer of anthropogenic pollutants including lipophilic polychlorinated biphenyls (PCBs) and protein-associated per- and polyfluoroalkyl substances (PFASs) and mercury (Hg). We conducted a meta-analysis on seabirds and investigated whether interspecies variation in maternal investment contributes toward skewed pollutant concentration ratios between males and females, as Cmale/Cfemale (80 studies). Overall concentrations of PCBs and perfluorooctanesulfonic acid (PFOS) were 1.6 and 1.3 times higher, respectively, in males than females, whereas mercury was similar between sexes. Few studies compared females and eggs ( n = 6), highlighting a knowledge gap. We found that an increasing maternal investment as a clutch-to-female mass ratio resulted in lower PCB concentrations in females than in males during the incubation period, but no sex-specific differences were observed for mercury and PFOS. Egg production is both a lipid dominated and protein-limited process. Females transfer lipophilic pollutants more easily to eggs, and to a higher degree with increasing maternal investment, but feeding ecology may be more important. Interspecies variation in maternal pollutant transfer may lead to negative effects scaling from an offspring to population level.

Potential Effect of Migration Strategy on Pollutant Occurrence in Eggs of Arctic Breeding Barnacle Geese ( Branta leucopsis).

Arctic-breeding geese acquire resources for egg production from overwintering grounds, spring stopover sites and breeding grounds, where pollutant exposure may differ. We investigated the effect of migration strategy on pollutant occurrence of lipophilic polychlorinated biphenyls (PCBs) and protein-associated poly- and perfluoroalkyl substances (PFASs) and mercury (Hg) in eggs of herbivorous barnacle geese ( Branta leucopsis) from an island colony on Svalbard. Stable isotopes (δ13C and δ15N) in eggs and vegetation collected along the migration route were similar. Pollutant concentrations in eggs were low, reflecting their terrestrial diet (∑PCB = 1.23 ± 0.80 ng/g ww; ∑PFAS = 1.21 ± 2.97 ng/g ww; Hg = 20.17 ± 7.52 ng/g dw). PCB concentrations in eggs increased with later hatch date, independent of lipid content which also increased over time. Some females may remobilize and transfer more PCBs to their eggs, by delaying migration several weeks, relying on more polluted and stored resources, or being in poor body condition when arriving at the breeding grounds. PFAS and Hg occurrence in eggs did not change throughout the breeding season, suggesting migration has a greater effect on lipophilic pollutants. Pollutant exposure during offspring production in arctic-breeding migrants may result in different profiles, with effects becoming more apparent with increasing trophic levels.

Predation Risk Potentiates Toxicity of a Common Metal Contaminant in a Coastal Copepod.

To examine whether natural stressors like predation risk affect responses to anthropogenic contaminants, we exposed nauplii of the copepod Tigriopus brevicornis to chemical cues from fish (kairomones) and copper (Cu). We tested effects of these treatments, singly and combined, on copepod age and size at maturity, and development stage sensitivity, while controlling for effects of genetic heterogeneity (clutch identity). Predation risk, Cu and clutch identity interacted in their effect on development time. Predation risk alone had minor effects, but potentiated Cu toxicity in the combined treatment by doubling the delay in age at maturity, as compared to Cu exposure alone. This potentiating effect on developmental delay appeared already at the first copepodite stage. The specific strength of response varied among nauplii from different females' clutches. There were no differences in copepod size at maturity among treatments. We did, however, find an interaction between the effect of Cu and clutch identity on copepod growth. Our results demonstrate the importance of ecological interactions for potentiating the toxicity of environmental contaminants. We also demonstrate the need to consider genetic heterogeneity in ecotoxicology. Natural variation in stressor responses has implications for the interpretation of results from toxicological studies using single-clone or inbred culture populations.

Increased hypertrophic response with increased mechanical load in skeletal muscles receiving identical activity patterns.

It is often assumed that mechanical factors are important for effects of exercise on muscle, but during voluntary training and most experimental conditions the effects could solely be attributed to differences in electrical activity, and direct evidence for a mechanosensory pathway has been scarce. We here show that, in rat muscles stimulated in vivo under deep anesthesia with identical electrical activity patterns, isometric contractions induced twofold more hypertrophy than contractions with 50-60% of the isometric force. The number of myonuclei and the RNA levels of myogenin and myogenic regulatory factor 4 were increased with high load, suggesting that activation of satellite cells is mechano dependent. On the other hand, training induced a major shift in fiber type distribution from type 2b to 2x that was load independent, indicating that the electrical signaling rather than mechanosignaling controls fiber type. RAC-α serine/threonine-protein kinase (Akt) and ribosomal protein S6 kinase ß-1 (S6K1) were not significantly differentially activated by load, suggesting that the differences in mechanical factors were not important for activating the Akt/mammalian target of rapamycin/S6K1 pathway. The transmembrane molecule syndecan-4 implied in overload hypertrophy in cardiac muscle was not load dependent, suggesting that mechanosignaling in skeletal muscle is different.

The impact of irradiance on optimal and cellular nitrogen to phosphorus ratios in phytoplankton.

Phytoplankton acclimates to irradiance by regulating the cellular content of light-harvesting complexes, which are nitrogen (N) rich and phosphorus (P) poor. Irradiance is thus hypothesised to influence the cellular N : P ratio and the N : P defining the threshold between N and P limitation (the 'optimal' N : P). We tested this hypothesis by first addressing the response of the optimal N : P to irradiance in a controlled experiment with Chlamydomonas reinhardtii. Then, we did a meta-analysis of experimental data on optimal and cellular N : P ratios across light gradients to test the generality of an N : P to light response within species. In both the experiment and the meta-analysis, N : P ratios decreased with irradiance, indicating that factors affecting underwater irradiance, like depth and the composition of the water, may influence the relative N : P requirement. The effect of irradiance did not differ between optimal and cellular N : P ratios, but observations of optimal N : P were on average 2.8 times higher than observations of cellular N : P.

Selection in the host structures the microbiota associated with developing cod larvae (Gadus morhua).

Marine fish larvae are immature upon hatching, and share their environment with high numbers of bacteria. The microbial communities associated with developing fish larvae might be structured by other factors than those important in developing terrestrial animals. Here, we analysed the beta (ß)-diversity of the microbiota associated with developing cod larvae and compared it with the bacterial communities in water and live feed by applying pyrosequencing of bar coded v4 16S rDNA amplicons. A total of 15 phyla were observed in the cod larval microbiota. Proteobacteria was the most abundant, followed by Firmicutes, Bacteroidetes and Actinobacteria. The composition and diversity of the cod larval microbiota changed considerably with age. The temporal and spatial patterns of ß-diversity could not be explained by stochastic processes, and did not coincide with changes in the rearing conditions. Furthermore, the larval microbiota was highly distinct from the water and the live feed microbiota, particularly at early developmental stages. However, the similarity between larval and water microbiota increased with age. This study suggests that strong selection in the host structures the cod larval microbiota. The changes in community structure observed with increasing age can be explained by altered selection pressure due to development of the intestinal system.

Seasonal diversity and dynamics of haptophytes in the Skagerrak, Norway, explored by high-throughput sequencing.

Microalgae in the division Haptophyta play key roles in the marine ecosystem and in global biogeochemical processes. Despite their ecological importance, knowledge on seasonal dynamics, community composition and abundance at the species level is limited due to their small cell size and few morphological features visible under the light microscope. Here, we present unique data on haptophyte seasonal diversity and dynamics from two annual cycles, with the taxonomic resolution and sampling depth obtained with high-throughput sequencing. From outer Oslofjorden, S Norway, nano- and picoplanktonic samples were collected monthly for 2 years, and the haptophytes targeted by amplification of RNA/cDNA with Haptophyta-specific 18S rDNA V4 primers. We obtained 156 operational taxonomic units (OTUs), from c. 400.000 454 pyrosequencing reads, after rigorous bioinformatic filtering and clustering at 99.5%. Most OTUs represented uncultured and/or not yet 18S rDNA-sequenced species. Haptophyte OTU richness and community composition exhibited high temporal variation and significant yearly periodicity. Richness was highest in September-October (autumn) and lowest in April-May (spring). Some taxa were detected all year, such as Chrysochromulina simplex, Emiliania huxleyi and Phaeocystis cordata, whereas most calcifying coccolithophores only appeared from summer to early winter. We also revealed the seasonal dynamics of OTUs representing putative novel classes (clades HAP-3-5) or orders (clades D, E, F). Season, light and temperature accounted for 29% of the variation in OTU composition. Residual variation may be related to biotic factors, such as competition and viral infection. This study provides new, in-depth knowledge on seasonal diversity and dynamics of haptophytes in North Atlantic coastal waters.

Temperature and developmental responses of body and cell size in Drosophila; effects of polyploidy and genome configuration.

Increased adult body size in Drosophila raised at lower temperatures could be attributed both to an increase in the cell volume and cell number. It is not clear, however, whether increased cell size is related to (or even caused by) increased nuclear volume and genome size (or configuration). Experiments with Drosophila melanogaster stocks (Oregon-R and w1118) raised at 16, 22, 24, and 28°C resulted in larger adult body and wing size with lower temperature, while eye size was less affected. The increase in wing size reflected an increase in cell size in both males and females of both stocks. The nucleus size, genome size, and DNA condensation of adult flies, embryos, and Schneider 2 cells (S2 cells, of larval origin) were estimated by flow cytometry. In both adult flies and S2 cells, both nucleus size and DNA condensation varied with temperature, while DNA content appears to be constant. From 12% to 18% of the somatic cells were tetraploid (4C) and 2-5% were octoploid (8C), and for the Oregon strain we observed an increase in the fraction of polyploid cells with decreasing temperature. The observed increase in body size (and wing size) at low temperatures could partly be linked with the cell size and DNA condensation, while corresponding changes in the haploid genome size were not observed.

Multigenerational genomic responses to dietary phosphorus and temperature in Daphnia.

Temperature and nutrient availability are both hypothesized to affect organisms at the cellular and genomic levels. In this multigenerational study, Daphnia magna (D. magna) and Daphnia pulex (D. pulex) were maintained at high (20 °C) and low (10 °C) temperatures and nourished with phosphorus (P)-sufficient (50 µmol/L) and P-deficient (2 µmol/L) algae for up to 35 generations to assess the multigenerational impacts on genome size and nucleus size. Analysis by flow cytometry revealed significant increases in nucleus size for both species as well as genome size for D. magna in response to a low temperature. The degree of endoreplication, measured as cycle value, was species specific and responded to temperature and dietary composition. Under dietary P deficiency, D. magna, but not D. pulex, showed an apparent reduction in haploid genome size (C-value). These genomic responses are unlikely to reflect differences in nucleotide numbers, but rather structural changes affecting fluorochrome binding. While the ultimate and proximate causes of these responses are unknown, they suggest an intriguing potential for genomic responses that merits further research.

Do DOM quality and origin affect the uptake and accumulation of a lipid-soluble contaminant in a filter feeding ascidian species (Ciona) that can target small particle size classes?

The widely reported increase of terrestrial dissolved organic matter (terrDOM) in northern latitude coastal areas ("coastal darkening") can impact contaminant dynamics in affected systems. One potential impact is based on differences in chemical adsorption processes of the molecularly larger terrDOM compared to marine DOM (marDOM) that leads to increased emulsification of lipophilic contaminants with terrDOM. Filter feeders filter large amounts of water and DOM daily and thus are directly exposed to associated contaminants through both respiration and feeding activity. Thus, increased exposure to terrDOM could potentially lead to an increase in bioaccumulation of lipid soluble contaminants in filter feeders. To assess the effect of DOM on bioaccumulation in filter feeders, we exposed the mucous based filter feeding ascidian Ciona intestinalis (formerly known as Ciona intestinalis Type B), to the lipophilic veterinary drug teflubenzuron (log KOW: 5.39) in combination with four DOM treatments: TerrDOM, marDOM, a mix of the two called mixDOM, and seawater without DOM addition. The exposure lasted for 15 days, after which the individuals in all DOM treatments showed a trend towards higher bioaccumulation of Teflubenzuron than those in the seawater control. However, there was considerable overlap in posterior distributions. Against our expectations, marDOM resulted in the highest bioaccumulation factor (BAF), followed by mixDOM, with terrDOM resulting in the lowest BAF except for seawater (kinetic BAF L/kg median, 2.5 %-97.5 % percentile marDOM 94, 74-118; mixDOM 82, 63-104; terrDOM 79; 61-99; seawater 61, 44-79). All BAFs were below the level of concern according to the EU REACH regulation (BAF < 2000 L / kg) and, therefore, likely not environmentally problematic in the examined context. However, the results show that DOM can act as a dietary vector; thus, different combinations of contaminants, DOM, and filter feeding organisms should be tested further.

Spatial predictors and temporal forecast of total organic carbon levels in boreal lakes.

Browning of Fennoscandian boreal lakes is raising concerns for negative ecosystem impacts as well as reduced drinking water quality. Declined sulfur deposition and warmer climate, along with afforestation, other climate impacts and less outfield grazing, have resulted in increased fluxes of Total Organic Carbon (TOC) from catchments to freshwater, and subsequently to coastal waters. This study assesses the major governing factors for increased TOC levels among several catchment characteristics in almost 5000 Fennoscandian lakes and catchments. Normalized Difference Vegetation Index (NDVI), a proxy for plant biomass, and the proportions of peatland in the catchment, along with surface runoff intensity and nitrogen deposition loading, were identified as the main spatial predictors for lake TOC concentrations. A multiple linear model, based on these explanatory variables, was used to simulate future TOC concentration in surface runoff from coastal drainage basins in 2050 and 2100, using the forecasts of climatic variables in two of the Shared Socio-economic Pathways (SSP): 1-2.6 (+2 °C) and 3-7.0 (+4,5 °C). These scenarios yield contrasting effects. SSP 1-2.6 predicts an overall decrease of TOC export to coastal waters, while SSP 3-7.0 in contrast leads to an increase in TOC export.

Do DOM quality and origin affect the uptake and accumulation of lipid-soluble contaminants in coastal filter feeders? An experimental simulation of teflubenzuron exposure to blue mussels.

The increased export of terrestrial dissolved organic matter (terrDOM) to coastal marine ecosystems may affect local filter feeders and the local food web via the altered uptake of organic material and associated contaminants. To compare terrDOM to marine DOM (marDOM) as contaminant vectors to coastal biota, we exposed blue mussels (Mytilus sp.) to the different DOM types in combination with teflubenzuron, a widely applied lipophilic aquaculture medicine targeting salmon lice (Lepeophtheirus salmonis). A 16-day exposure of the blue mussels to DOM and teflubenzuron was followed by a depuration phase of 20 days without teflubenzuron. We calculated teflubenzuron adsorption rates and bioaccumulation factors (BAF) using a Bayesian model, expecting teflubenzuron uptake to be greater with terrDOM than marDOM due to the higher prevalence of large amphipathic humic acids in terrDOM. Humic acids have strong absorption properties and are able to envelope lipophilic molecules. Thus, humic acids can function as an efficient contaminant vector when taken up by filter feeders. Although there were varying degrees of overlap, the mussels tended to accumulate higher amounts of teflubenzuron in the DOM treatments than in the seawater control (bioaccumulation factor [BAF] in seawater: median 106 L/kg; 2.5 %-97.5 % percentile: 69-160 L/kg). Contrary to expectations, mussels exposed to marDOM showed a trend toward more bioaccumulation of teflubenzuron than those exposed to terrDOM (BAF marine 144 L/kg; 102-221 L/kg versus BAF terrestrial: 121 L/kg; 82-186 L/kg). The highest teflubenzuron accumulation was observed with the 50:50 mixture of marDOM and terrDOM (BAF mix: 165 L/kg; 117-244 L/kg). The slight difference in DOM-type accumulation rates observed in this experiment-especially the accumulation rate of terrDOM compared to that of the seawater-only treatment type-was not considered environmentally relevant. Further studies are necessary to see if the observed trends transfer to complex environmental systems.

"Trophic overyielding": phytoplankton diversity promotes zooplankton productivity.

Diversity-productivity relationships at the primary producer level have been extensively studied, especially for terrestrial systems. Here, we explore whether the diversity of aquatic primary producers (phytoplankton) has effects on higher trophic levels (zooplankton). We investigated the effect of phytoplankton diversity on an artificial zooplankton community in a laboratory experiment where phytoplankton biomass and elemental composition (carbon-to-phosphorus ratio) were kept constant. Phytoplankton diversity increased the means of both zooplankton growth rate and abundance while suppressing their variability, and sustained higher zooplankton diversity. Likely explanations include resource complementarity effects among phytoplankton species as food entities, as well as niche complementarity effects among Daphnia species as competitors. By affecting the productivity as well as the variability of the next trophic level, biodiversity of primary producers may have far-reaching consequences in aquatic food webs.

Prey-predator dynamics in rotifers: density-dependent consequences of spatial heterogeneity due to surface attachment.

Classical models of prey-predator interactions assume that per capita prey consumption is dependent on prey density alone and that prey consumption (functional response) and consumer proliferation (numerical response) operate on the same timescales and without time lags. Several modifications have been proposed for resolving this timescale discrepancy, including variants where the functional response depends on both prey and predator densities. A microcosm system with the rotifer Brachionus 'Nevada' feeding on the prasinophyte Tetraselmis sp. showed significant (P < 0.0005) increases in steady-state biomasses of both prey and predators with increasing carrying capacity (represented by total phosphorus of the growth medium), which is inconsistent with predictions based on the traditional prey-only-dependent functional response. We provide data indicating that surfaces where the predator can attach provide a high-quality habitat for rotifers, which can result in a predator-dependent functional response. We also show that partitioning between the attached and free-swimming habitats was fast compared to the timescale of the numerical response. When attached to surfaces, rotifers maximized net energy gain by avoiding the high cost of swimming and by increased food capture due to reduced viscous drag. A mathematical model with prey-dependent functional response and wall-attached and free-swimming fractions of the population describes our data adequately. We discuss the implications of this finding for extrapolating microcosm experiments to systems with other surface-to-volume ratios, and to what extent our findings may apply to other popular model organisms for prey-predator interaction.

Drivers and variability of CO2:O2 saturation along a gradient from boreal to Arctic lakes.

Lakes are significant players for the global climate since they sequester terrestrially derived dissolved organic carbon (DOC), and emit greenhouse gases like CO2 to the atmosphere. However, the differences in environmental drivers of CO2 concentrations are not well constrained along latitudinal and thus climate gradients. Our aim here is to provide a better understanding of net heterotrophy and gas balance at the catchment scale in a set of boreal, sub-Arctic and high-Arctic lakes. We assessed water chemistry and concentrations of dissolved O2 and CO2, as well as the CO2:O2 ratio in three groups of lakes separated by steps of approximately 10 degrees latitude in South-Eastern Norway (near 60° N), sub-Arctic lakes in the northernmost part of the Norwegian mainland (near 70° N) and high-Arctic lakes on Svalbard (near 80° N). Across all regions, CO2 saturation levels varied more (6-1374%) than O2 saturation levels (85-148%) and hence CO2 saturation governed the CO2:O2 ratio. The boreal lakes were generally undersaturated with O2, while the sub-Arctic and high-Arctic lakes ranged from O2 saturated to oversaturated. Regardless of location, the majority of the lakes were CO2 supersaturated. In the boreal lakes the CO2:O2 ratio was mainly related to DOC concentration, in contrast to the sub-Arctic and high-Arctic localities, where conductivity was the major statistical determinant. While the southern part is dominated by granitic and metamorphic bedrock, the sub-Arctic sites are scattered across a range of granitic to sedimentary bed rocks, and the majority of the high-Arctic lakes are situated on limestone, resulting in contrasting lake alkalinities between the regions. DOC dependency of the CO2:O2 ratio in the boreal region together with low alkalinity suggests that in-lake heterotrophic respiration was a major source of lake CO2. Contrastingly, the conductivity dependency indicates that CO2 saturation in the sub-Arctic and high-Arctic lakes was to a large part explained by DIC input from catchment respiration and carbonate weathering.

Disentangling effects of natural and anthropogenic drivers on forest net ecosystem production.

Net Ecosystem Production (NEP) of forests is the net carbon dioxide (CO2) fluxes between land and the atmosphere due to forests' biogeochemical processes. NEP varies with natural drivers such as precipitation, air temperature, solar radiation, plant functional type (PFT), and soil texture, which affect the gross primary production and ecosystem respiration, and thus the net C sequestration. It is also known that deposition of sulphur and nitrogen influences NEP in forest ecosystems. These drivers' respective, unique effects on NEP, however, are often difficult to be individually identified by conventional bivariate analysis. Here we show that by analyzing 22 forest sites with 231 site-year data acquired from FLUXNET database across Europe for the years 2000-2014, the individual, unique effects of these drivers on annual forest CO2 fluxes can be disentangled using Generalized Additive Models (GAM) for nonlinear regression analysis. We show that S and N deposition have substantial impacts on NEP, where S deposition above 5 kg S ha-1 yr-1 can significantly reduce NEP, and N deposition around 22 kg N ha-1 yr-1 has the highest positive effect on NEP. Our results suggest that air quality management of S and N is crucial for maintaining healthy biogeochemical functions of forests to mitigate climate change. Furthermore, the empirical models we developed for estimating NEP of forests can serve as a forest management tool in the context of climate change mitigation. Potential applications include the assessment of forest carbon fluxes in the REDD+ framework of the UNFCCC.