This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/gcb.14928 This article is protected by copyright. All rights reserved DR. PIRKKO KORTELAINEN (Orcid ID : 0000-0002-1448-0688) DR. SARI JUUTINEN (Orcid ID : 0000-0002-7752-1950) Article type : Primary Research Articles Title Lakes as nitrous oxide sources in the boreal landscape Running title: N2O in boreal lakes Authors Pirkko Kortelainen1*, Tuula Larmola2, Miitta Rantakari1,3, Sari Juutinen4, Jukka Alm5 and Pertti J. Martikainen6 1*Finnish Environment Institute, Latokartanonkaari 11, 00790 Helsinki, Finland; tel: + 358 40 5356646 *e-mail pirkko.kortelainen@ymparisto.fi 2Natural Resources Institute Finland (Luke), Latokartanonkaari 9, 00790 Helsinki, Finland; e-mail tuula.larmola@luke.fi 3City of Helsinki, P.O. Box 58235, FI-00099 City of Helsinki; e-mail miitta.rantakari@hel.fi 4Faculty of Biological and Environmental Sciences, P.O. Box 65, FI-00014 University of Helsinki, Finland; e-mail sari.juutinen@helsinki.fi 5Natural Resources Institute Finland (Luke), P.O. Box 68, Joensuu 80101, Finland, e-mail A cc ep te d A rt ic le This article is protected by copyright. All rights reserved jukka.alm@luke.fi 6University of Eastern Finland, Department of Environmental and Biological Sciences, P.O. Box 1627, FI-70211 Kuopio, Finland; e-mail pertti.martikainen@uef.fi Correspondence *e-mail: pirkko.kortelainen@ymparisto.fi A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Abstract Estimates of regional and global freshwater N2O emissions have remained inaccurate due to scarce data and complexity of the multiple processes driving N2O fluxes the focus predominantly being on summer time measurements from emission hot spots, agricultural streams. Here we present four-season data of N2O concentrations in the water columns of randomly selected boreal lakes covering a large variation in latitude, lake type, area, depth, water chemistry and land use cover. Nitrate was the key driver for N2O dynamics, explaining as much as 78 % of the variation of the seasonal mean N2O concentrations across all lakes. Nitrate concentrations varied among seasons being highest in winter and lowest in summer. Of the surface water samples 71% were oversaturated with N2O relative to the atmosphere. Largest oversaturation was measured in winter and lowest in summer stressing the importance to include full year N2O measurements in annual emission estimates. Including winter data resulted in four-fold annual N2O emission estimates compared to summer only measurements. Nutrient rich calcareous and large humic lakes had the highest annual N2O emissions. Our emission estimates for Finnish and boreal lakes are 0.6 Gg and 29 Gg N2O-N y-1, respectively. The Global Warming Potential (GWP) of N2O cannot be neglected in the boreal landscape, being 35% of that of diffusive CH4 emission in Finnish lakes. Keywords climate change, trace gases, eutrophication, ecosystems, environmental change, lakes, landscape, nitrous oxide A cc ep te d A rt ic le This article is protected by copyright. All rights reserved 1 INTRODUCTION Lakes and streams acting as recipients of carbon, nitrogen and other nutrients transported from terrestrial ecosystems contribute to landscape greenhouse gas (GHG) balances emitting carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O). Freshwater N2O has received minor interest compared to carbon gases, CO2 and CH4. Consequently, both the variability of N2O concentrations and factors regulating N2O fluxes from lakes at regional and global scales have remained poorly constrained and estimates of freshwater N2O emissions are still highly uncertain due to sparse data (Soued, del Giorgio & Maranger, 2015; Deemer et al., 2016; DelSontro, Beaulieu & Downing, 2018). Majority of freshwater studies have focused on rivers and streams in N rich agricultural environments (Mulholland et al., 2008; Beaulieu et al., 2011; Hu, Chen & Dahlgren, 2016) excluding landscapes dominated by forests and peatlands, the most widely distributed ecosystems in the boreal zone. The main processes involved in N2O cycling are aerobic nitrification and anaerobic denitrification, which are regulated by several environmental factors like oxygen and organic matter content, pH, temperature, and the availability of ammonium and nitrate (Butterbach-Bahl et al., 2013). In contrast to most previous studies, which have predominantly focused either on a few lakes and/or summer time measurements we measured N2O concentrations for the four seasons in the water columns of 112 lakes in Finland covering different lake types, locating between the latitudes 60oN and 67oN (Fig. 1, Table 1). We examined how seasonal and spatial variation of N2O concentrations in lakes were associated with the characteristics of lakes (area, maximum depth, water chemistry, temperature, oxygen content) and catchments (area, elevation and land use cover) and compared spatiotemporal variation in N2O concentration with that of CO2 and CH4 measured simultaneously with the N2O (Kortelainen et al., 2006; Juutinen et al., 2009). 2 MATERIAL AND METHODS 2.1 Selection of the lakes and sampling A subpopulation of 87 boreal lakes in Finland (Fig. 1, Table 1) was randomly selected from the Nordic Lake Survey (NLS) data base (Henriksen et al., 1998; Mannio, Räike & Vuorenmaa, A cc ep te d A rt ic le This article is protected by copyright. All rights reserved 2000). Besides N2O lake chemistry and morphometry (lake area, max depth), catchment characteristics (catchment area, elevation, land use cover), land use cover (agricultural land %, peat %, forest %, urban %) and climate related variables (latitude, water temperature) were determined for each lake in order to identify key drivers contributing to seasonal variation of N2O concentrations in boreal lakes. The lakes were sampled once in winter, spring, summer and autumn during 1998-1999 from four depths at the deepest point of the lake for N2O concentrations and physical and chemical characteristics. For CO2 and CH4 analyses a subpopulation of 177 lakes was randomly selected from the NLS data (Kortelainen et al., 2006; Juutinen et al, 2009). Gas analyses were carried out in the laboratories of University of Eastern Finland, due to limited capacity for N2O analyses we had to exclude approximately 1/3 of the lakes, ending up with 87 randomly selected lakes for N2O measurements. Majority of boreal lakes are located in forested catchments with a relatively minor human disturbance. Consequently eutrophic lakes cannot be expected to be well represented in the randomly selected lake data base. In order to include also some eutrophic lakes in the study a subset of NLS lakes (n=25) with the highest total P concentrations was also sampled. Results from these lakes were included in the analyses to study the relationships between lake water chemistry and N2O concentrations and flux estimates from different lake types but excluded from the population estimates. Following similar approach as for CH4 we used modified Finnish lake typology required for the ecological lake status classification by the EU water framework to classify the lakes, e.g., according to nutrient concentrations, colour, size and depth (Juutinen et al., 2009). In order to avoid very small subgroups, we divided the lakes into four subgroups: Nutrient Rich and Calcareous (NRC, alkalinity > 0.4 mmol L-1, winter turbidity > 5 FTU), Humic Small (HSm, colour ≥30 Pt mg L-1, area ≤ 5 km2 or mean depth < 3 m), Humic Large (HL, colour ≥ 30 Pt mg L- 1, area > 5 km2), and Clear Water Lakes (CL, colour < 30 Pt mg L-1). 2.2 Calculation of gas fluxes N2O measurements were carried out by the headspace equilibration technique (McAuliffe et al., 1971). Ultra pure N2 gas (30 ml) was added to 60 ml syringes containing 30 ml water, vigorously the syringes were then shaken for 3 minutes. The headspace gas concentration was quantified with a gas chromatograph (Hewlett Packard Series II and Shimadzu GC-14-A) equipped with a FI-A cc ep te d A rt ic le This article is protected by copyright. All rights reserved detector. Lake–atmosphere gas fluxes were estimated from the surface-water gas concentrations according to the First Fick’s law of diffusion (Fick 1855, Wanninkhof & Knox, 1996): Fgas = kgas(Cgas – Ceq) (1) where Fgas is the lake–atmosphere flux of N2O, kN2O is the gas transfer velocity (m d–1), Cgas is the concentration of the gas in the surface water (μmol l–1), Ceq is the concentration of the gas (μmol l–1) in the surface water in equilibrium with the atmosphere. The in-situ equilibrium concentration of N2O in the lake surface water was calculated according to Henry’s Law, with appropriate corrections for the in-situ temperature, assuming atmospheric concentration of 324 ppb for N2O (319 ppb in 2005 in Hyytiälä SMEAR station, Finland) and a 0.8-ppb increase per year (Ciais et al., 2013). Since the data behind measured gas transfer coefficients (k values) are limited and the relationship between lake size and k values vary we used three approaches to estimate lake-atmosphere gas exchange (Vachon & Prairie, 2013; Heiskanen et al., 2014; Holgerson et al., 2017). We calculated the gas fluxes for those 94 lakes that had N2O concentrations measured in all four seasons. Some lakes are located in remote areas and were difficult to sample before and/or after ice melt. First, following the approach by Heiskanen et al. (2014) kN2O value was calculated, using the average values of kCO2 from a small Finnish lake during a four month period, which we transformed into the kN2O: kgas1/kgas2 = (ScGas1/ScGas2)–n (2) where Scgas is the Schmidt number for a given gas (Jähne, Heinz & Dietrich, 1987) and n is 1/2. The wind speed was assumed to be 3 ms-1, which is an average open water period wind speed at the height of 10 m for the inland measurement stations in Finland (Leinonen, 2000).A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Secondly, we applied Holgerson et al. (2017) approach for each of our 94 lakes, i.e., the size-class median estimates of k600 for 67 ponds and lakes across a size gradient measured with floating chambers or gas tracers. Thirdly, we calculated k600 values using an empirical model based on wind speed (U10) and lake area (LA) in 21 Canadian water bodies (Vachon and Prairie 2013): K600=2.51+1.48 U10 +0.39 U10 log10 LA (3) 2.3 Estimation of annual fluxes and upscaling First, following similar approach as for CO2 (Kortelainen et al., 2006), we estimated the annual N2O emission per surface area unit of each lake using the N2O concentrations from the depth of 1 m during the four sampling occasions as follows: the winter 1 m concentrations were used for half a month period assuming that the N2O accumulated under ice would be released to the atmosphere during the short spring circulation period. The spring N2O concentrations at the depth of 1 m were used for a 1.5 months period and the summer 1 m concentrations for a 3 months period representing the summer stratification between May and August. The samples taken during the autumn circulation were used for the period of 2 months representing the time after breakup of the summer stratification and before the permanent ice cover (altogether 7 months of ice free season). For comparison, we also estimated the evasion for the ice-free period by multiplying the median summer time N2O flux with the duration of ice-free period (7 months). Secondly, we estimated the annual median N2O evasion both for each lake size class separately (lake size specific evasion) and for all size classes combined (the mean, median and summer median of individual lakes) for Finnish and Boreal lakes. The randomly selected lakes, which could be sampled during all four seasons (n=71) were divided into the different size classes and the median annual evasion of each size class was multiplied with the area of the Finnish and boreal lakes belonging to the respective size class. Summing up the size class specific estimates results in an estimate for Finland and boreal region, respectively. The rationale was that when upscaling is based on the evasion estimated specific to each of the four lake size classes, the A cc ep te d A rt ic le This article is protected by copyright. All rights reserved impact of the numerous small lakes in our data on the regional estimate is not disproportionally large. Furthermore, we estimated the contribution of N2O emission from lakes to that of forests both in Finland and in the boreal zone. For Finland, we used the lake surface area of 32 663 km2 (Raatikainen & Kuusisto, 1990) and the forest area of 0.203 x 106 km2 (Vaahtera, 2018). For the boreal region, the lake surface area was estimated from MODIS data (1 422 448 km2), the forest area of 12.1 x 106 km2 was derived from Potter, Matson, Vitousek & Davidson (1996). 2.4 Water chemistry The lakes were sampled in each sampling occasion for dissolved oxygen, alkalinity, conductivity, pH, colour, total nitrogen (Ntot), nitrate (+ nitrite) nitrogen defined as (NO3-N, nitrite had negligible contribution to the total amount of nitrate and nitrite), ammonium nitrogen (NH4-N), total phosphorus (Ptot), phosphate phosphorus (PO4-P), TOC and total iron (Fetot). Water chemistry was analyzed from unfiltered samples in the accredited laboratories of the Regional Environment Centers. Ntot was determined by oxidation with K2S2O8, reduction of NO3-N to NO2-N in Hg-Cd (Cu-Cd) column and colorimetric determination of azo-color. The sum of NO3-N and NO2–N was measured by reduction of NO3-N to NO2-N in Hg-Cd (Cu-Cd) column, followed by colorimetric determination of azo-color. NH4-N was measured colorimetrically with hypochlorite and phenol. Ptot and PO4-P were measured colorimetrically. TOC was determined by oxidizing the sample by combustion and measuring inorganic C by IR-spectrophotometry (National Board of Waters, 1981). 2.5 Catchment characteristics The catchment areas of the NLS lakes were determined from the topographic maps, and the catchment boundaries were digitalized and combined with land use data based on satellite images using the Arc View georeferencing software. Lake area, catchment area, catchment to lake area - ratio, latitude, and the proportion of peatland, forest on mineral soil, agricultural land, water (consisting of the upstream water bodies and the lake itself) and built-up area in the catchments were determined.A cc ep te d A rt ic le This article is protected by copyright. All rights reserved 2.6 Statistical analysis The relationships between the N2O concentrations and lake chemistry, morphometry, latitude, and catchment characteristics were examined using Pearson's correlation coefficients using SAS 9.4 for Windows software. The variables were loge or square root transformed in order to normalize their distribution. Stepwise multiple linear regression models predicting N2O concentrations were carried out using lake chemical, physical and morphometric variables, climatic variables (temperature, latitude) and catchment properties as predictors. The cases with an absolute value of the Studentized residual exceeding 3 were excluded and only the independent variables with p-values less than 0.05 were included in the models. Linear mixed models were used to take into account that each lake was sampled four times and from four depths (interdependence of within-lake sampling). We run a linear mixed model with NO3-N, depth, season as fixed factors and lake as random factor. That is, the model analyzed the dependence of N2O concentration on NO3-N, and eliminated the effects of depth and season on the emerging N2O to NO3-N relationship. To examine the relationship between mean NO3-N and mean N2O concentrations in different lake types across the four seasons, a nonlinear regression model was fitted (Equation 4, see results). Seasonal mean values for each lake type (n=16) were obtained by averaging all measured depths across respective lakes. To examine how representative the patterns and drivers of N2O dynamics found in our lakes were, we used equation (4) to estimate N2O concentrations in the water samples from the independent nitrate-nitrite data set of 874 randomly selected Finnish NLS lakes sampled during the autumn overturn in 1998-1999. The modeled results were compared to the measured N2O-N concentrations in our study lakes (n = 1 542 seasonal water samples from 112 lakes). 3 RESULTS 3.1 Seasonal and spatial variation in N2OA cc ep te d A rt ic le This article is protected by copyright. All rights reserved During the open water season 71 % of the surface water samples were saturated with respect to the atmospheric equilibrium value of N2O, i.e., lakes were mostly sources of N2O. Nitrous oxide, similarly to CO2, peaked in winter. In contrast CH4 concentration was significantly higher in summer (Kruskal-Wallis 123.93, p<0.001, Fig. 2). Seasonal variation of N2O concentrations followed the variation of nitrate - highest in winter and lowest in summer despite differences in concentration levels across lake type classification (Fig. 3). Further, nitrate was the best predictor for N2O concentrations in the entire data (Table 2, Fig. 4) and across different seasons and depths despite the large variation in lake area (from 0.04 km2 to 44 km2), maximum depth (from 1 m to 42 m), latitude (from 60oN to 67oN) and land use cover. Nitrate concentrations were highest in nutrient rich, calcareous (NRC) lakes accompanied by highest N2O concentrations in winter and spring (Fig. 3). In summer and autumn N2O concentrations were more evenly distributed across the lake types. Depth profile distribution demonstrated similar N2O concentrations from surface to bottom (Fig. 5) in contrast to CO2 and CH4 which accumulated in bottom water (Kortelainen et al., 2006; Juutinen et al. 2009). There was no district correlation between N2O and water pH. Both the highest and the lowest N2O concentrations occurred around the median pH of 6.5 (Fig. 6). Even though there was a weak correlation between oxygen and N2O (Table 2), the variability in N2O concentrations across O2 gradient was large especially before ice melt among all lake types (Fig. 7). N2O and nitrate concentrations increased with decrease in elevation (Fig. 8a) accompanied by increasing coverage of agricultural land (Fig. 8d), which resulted in higher nitrate and N2O concentrations in large lakes (Fig. 8b) often surrounded by more intensive human impact compared to small headwater lakes, which in the boreal zone are predominantly surrounded by forests and peatlands. Lake chemistry predicted N2O better than catchment land use cover. In linear multiple regression models electron acceptors, nitrate and oxygen, and lake water temperature as independent variables predicted best (r2 = 0.55, stepwise procedure) the N2O in the entire data (n=1 396, all seasons and depths). Nitrous oxide in bottom water (all seasons) was best predicted by nitrate and oxygen concentration (r2 = 0.54). The surface water model (all seasons), had nitrate, temperature and the percentage of agricultural land in the catchment as the independent variables explaining A cc ep te d A rt ic le This article is protected by copyright. All rights reserved 58 % of the variation in N2O. The best model for winter (all depths) explained 58 % of the variability in N2O selecting nitrate, latitude and pH as the independent variables (Table 3). The linear mixed model results demonstrated that the significant relationship between nitrate and N2O remained (p<0.001) even after the influence of depth (p=0.096) and season (p<0.001) had been taken into account. 3.2 N2O evasion We estimated the median N2O evasion based on 71 lakes (< 100 km2) as 0.009 g N m-2 LA year-1. The seasonal median fluxes were 0.002 g N m-2 LA at the thaw (during 0.5 months), 0.002 g N m-2 LA in spring (1.5 months), 0.001 g N m-2 LA in summer (3 months), and 0.003 g N m-2 LA in autumn (2 months). For the largest lakes (> 100 km2) we used the evasion estimate from the 10- 100 km2 lake size class. Climate change scenarios predict increasing temperature and precipitation for northern Europe (Ciais et al., 2013) accompanied by shorter ice cover period in boreal lakes. The median evasion for the < 0.1 km2, 0.1-1 km2, 1-10 km2, 10-100 km2 lake size classes was estimated as 0.0047 g N m-2 LA, 0.007 g N m-2 LA, 0.018 g N m-2 LA and 0.02 g N m-2 LA, respectively (using k values from Holgerson et al. , 2017). Nitrous oxide data was not available for lakes larger than 100 km2; for this lake size class we used the median evasion estimate from the 10-100 km2 size class. N2O evasion per surface area unit was highest in the largest lakes reflecting the distribution of nitrate concentrations. In contrast, CO2 evasion estimates per surface area unit (Kortelainen et al., 2006) followed closely lake size distribution being largest in small lakes. Total annual N2O flux from Finnish lakes (total lake area 32 663 km2) was estimated as 0.6-0.8 Gg N2O-N y-1, based on the areas of the lake size distribution by Raatikainen and Kuusisto (1990) and the evasion estimates for the different lake size classes (Fig. 9). Both Holgerson et al. (2017) and Heiskanen et al. (2014) approaches resulted in an estimate of 0.6 Gg N2O-N y-1 for Finnish lakes, when the median N2O evasion estimates of different lake size classes were multiplied with the respective lake surface area distribution. The Vachon and Prairie (2013) approach resulted in a little bit larger estimate of 0.8 Gg N2O-N y-1 (Table 4, Supporting Information Table S1, S2). A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Freshwater N cycling integrates numerous simultaneous temperature dependent microbiological processes. Also, our data demonstrated large differences in N2O concentrations and estimated emissions among different lake types (Table 5, Supporting Information Table S2). Nevertheless, our data underline the key role of nitrate in regulating seasonal and spatial N2O concentrations across boreal lakes. Nitrate explained as much as 78 % of the variation in seasonal mean N2O concentrations across all lakes and depths (equation 4, Fig. 10). While O2 was the dominating driver for CO2 in our data, with similar explanation power, 78%, of the variation across all lakes and depths (Kortelainen et al., 2006). N2O-N (ng l-1) = 374 e0.003 NO3-N (µg l-1) (4) Using Equation (4), the modelled N2O concentrations for independent nitrate-nitrite data set of 874 randomly selected Finnish NLS lakes reproduced similar median N2O concentrations and an increasing trend by lake size further indicating that our lake data set is representative to Finnish conditions and can be expected to represent lakes over larger boreal landscape (Fig. 11). Large lakes turned out to be disproportionately important N2O sources among the lake population on the landscape level. Lakes larger than 10 km2 were estimated to contribute 77 % of the total N2O emission from Finnish lakes. In contrast, CO2 evasion estimates demonstrated that lakes smaller than 10 km2 dominated landscape CO2 evasion among the lake population while lakes larger than 10 km2 (representing 65 % of the total lake area distribution) represented only 45 % of the estimated CO2 evasion (Kortelainen et al., 2006). Our estimated N2O-N emissions from lakes represent 17 % of the N2O-N emissions from boreal forests, the dominating ecosystem in Finland where lakes cover 10 % of the total land area (Table 6). 4. DISCUSSION 4.1 Seasonal and spatial variation in N2O concentrations and fluxes N2O concentrations were high in lakes located in low elevation catchments accompanied by large coverage of agricultural land (Fig. 8a, b). Nutrient rich, calcareous lakes had high N2O concentrations, which was comparable with methane distribution (Juutinen et al., 2009) and A cc ep te d A rt ic le This article is protected by copyright. All rights reserved underlines close links between lake trophic state and GHG concentrations, further suggesting increasing GHG evasion as a result of eutrophication. Although calcareous lakes had high N2O concentrations, correlation between pH and N2O was negligible (Table 2, Fig. 6). The variability in N2O concentration was the largest around the median pH 6.5 (Fig. 6), which may reflect optimal pH of the accumulation of N2O, i.e., net production of N2O from nitrification and denitrification. Low pH inhibits the N2O reductase which increases the N2O to N2 ratio in denitrification (Richardsson, Felgate, Watmough, Thomson & Baggs, 2009). Nitrification is further inhibited at high C:N ratios (Her & Huang, 1995), typical for boreal Finnish lakes and often accompanied by low pH (Kortelainen et al., 2013). Supportingly, Humic large lakes, including only seven lakes in our data set, also had high N2O concentrations (Fig. 3, Fig. 10). Nitrous oxide peaked in winter similarly to CO2, while CH4 concentration was significantly higher in summer (Fig. 2). During the open water season 29 % of the surface water samples were under saturated with respect to the atmospheric equilibrium value of N2O showing that lakes can also act as N2O sinks, which supports results from Canadian freshwaters (Soued et al., 2015) and farm water bodies in the USA (Webb et al., 2019) underlining the uncertainty in the data presently used both in regional and global N2O budgets. On average, N2O concentrations were evenly distributed through the water column (Fig. 5), which contrasts the vertical distribution patterns of CO2 and CH4 that accumulate above the sediment (Kortelainen et al., 2006, Juutinen et al., 2009). Comparable distribution of GHGs has been reported from thirteen temperate lakes in Colorado Rocky Mountains (McCrackin & Elser, 2011). Generally N2O production is limited by low N turnover and low N mineralization in the high latitude N limited ecosystems (Potter et al., 1996). In N limited boreal terrain, N turnover is rapid and internal N cycling across forested ecosystems dominates the spatial nitrate distribution, which often reflects more closely catchment land use cover and topography than N deposition - in spite of N deposition being the major N source (Palviainen, Lehtoranta, Ekholm, Ruoho-Airola & Kortelainen, 2015). Nitrate is rapidly captured during growing season resulting in minor nitrate concentrations in downstream lakes in summer (Rekolainen, Mitikka, Vuorenmaa & Johansson, 2005). A cc ep te d A rt ic le This article is protected by copyright. All rights reserved The N2O/NO3-N ratio correlated negatively with nitrate (Fig. 4b) being lowest in large lakes (Fig. 8c) which had higher nitrate content than the small lakes. Increase in nitrate availability favours denitrification rate in lakes but the efficiency of denitrification declines with increasing N inputs (Mulholland et al., 2008; Rissanen, Tiirola, Hietanen & Ojala, 2013). Seasonal variation in N2O/NO3-N ratio was also strong - highest in summer and lowest in winter (Fig. 8f). Low temperature decreases the activity of N2O reductase more than the activity of other reductases in denitrification (Veraart, de Klein & Scheffer, 2011) which together with high nitrate availability favors N2O accumulation in the dormant season. The possible higher reduction of N2O to N2 as a result of low oxygen content (Richardson et al., 2009) in dormant season did not prevent N2O accumulation (Fig. 7). In summer low nitrate content and high N2O/NO3-N ratio resulted from nitrate being consumed in primary production, denitrification and other microbial processes. In freshwaters 0-4 % of N is generally released as N2O in denitrification (Mulholland et al., 2008; Seitzinger, 1988; Silvennoinen Liikanen, Torssonen, Stange & Martikainen, 2008). The explanation power of our statistical models for N2O concentrations in lakes (Table 3) is comparable to the power of the models developed for terrestrial N2O emissions (Leppelt et. al., 2014; Pärn et al., 2018). In the lake data set from boreal southern Norway and Sweden N2O concentrations correlated positively with nitrate in summer (Yang et al., 2015). In contrast, association between nitrate and N2O has been weak in temperate lakes. No significant correlation between nitrate and N2O could be found across temperate headwater Irish lakes in summer (Whitfield et al., 2016). Further, N2O fluxes could not be predicted by any measured environmental variables in aquatic network in temperate Quebec, Canada, where 40 % of the relatively pristine inland waters were under-saturated in N2O in summer (Soued et al., 2015). 4.2 Landscape scale patterns Key processes and feedbacks of landscape scale GHG fluxes have remained poorly quantified. Dynamics of N2O in our lakes at landscape scale did not follow those of CO2 (Kortelainen et al., 2006) and CH4 (Juutinen et al., 2009) (Fig. 1). Large lakes dominate the lake surface area distribution in Finland (Raatikainen & Kuusisto, 1990). Further, estimated N2O emissions per surface area unit were largest from large lakes (Fig. 9) reflecting the distribution of nitrate concentrations. In contrast, concentrations and estimated emissions of CO2 and CH4 decrease with A cc ep te d A rt ic le This article is protected by copyright. All rights reserved decrease in lake area (Kortelainen et al., 2006; Juutinen et al., 2009; Denfeld, Kortelainen, Rantakari, Sobek & Weyhenmeyer, 2016). Our results underscore simultaneous (Miettinen et al., 2015) long term CO2, CH4, and N2O measurements from freshwaters in order to better understand major controls of landscape GHG evasion. Freshwater GHG flux measurements have predominantly been carried out in summer, while our study underscores the importance of dormant season N2O measurements. The lower the temperature the higher N2O concentrations, which in our data underlines the link between seasonal variability of nitrate and N2O concentrations, i.e., lower concentrations in warm growing season compared to cold dormant period. Elevated N2O emissions during winter snow cover period have been measured also, e.g., in cropland (Groffman, et al., 2001). Higher N2O emission was shown to coincide with a greater number of freeze-thaw cycles that broke up soil macro aggregates and increased soil inorganic N pool (Ruan & Robertson, 2017). In a northern hardwood forest reduced N uptake by fine roots due to soil freezing was concluded to be the primary regulator of increasing nitrate export. Increasing nitrate and N2O concentrations during dormant period might thus ultimately reflect declining N uptake in terrestrial ecosystems resulting in increasing nitrate concentrations in downstream freshwaters (Campbell, Socci & Templer, 2014). Freeze-thaw related N2O fluxes were shown to be a major component of annual N2O emissions also in in boreal peatlands of Northeast China (Cui et al., 2016). Freshwater ecosystems have often been ignored in consideration of the landscape GHG fluxes. We used GWP (GWP100 = 265, IPCC 2014) (Ciais et al., 2013) to estimate the effect of N2O evasion from Finnish lakes on the climate as CO2 equivalents. Our estimate (0.6 Gg N2O-N y-1; Table 4) represents 35 % (the 25%-75% quartiles being 29-41%) of the GWP of Finnish lake diffusive CH4 emissions (Juutinen et al., 2009). It is estimated that ebullition may even double the methane emissions (Bastviken et al., 2004; Juutinen et al. 2009; Wik et al., 2014). When the uncertainty related to methane ebullition is taken into account (i.e., scenarios with and without ebullition included) our N2O estimate represents 15-41% of the GWP of Finnish lake CH4 emissions. Assuming similar atmospheric impact for boreal lakes in general gives an emission estimate of 29 Gg N2O-N y-1. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Annual mean and median N2O flux estimates from our lakes including seasonal data were surprisingly close to each other (Table 5) while estimates based only on summer measurements underestimated annual N2O emissions (Table 4). Further, the different annual N2O flux estimates resulting from the three approaches for gas transfer coefficients (k values) and their dependence on lake size (Vachon & Prairie, 2013; Heiskanen et al., 2014; Holgerson et al., 2017) demonstrate that there is uncertainty due to sparse/limited measured k values especially for large lakes. Recently, Webb et al. (2019) showed widespread unexpected undersaturation (67 %) of N2O in eutrophic farm waterbodies in the US, which according to our study might partly reflect the time of the sampling period. In our data 29 % of the samples were undersaturated, majority of these were sampled during summer. Our data demonstrated large seasonal variation of nitrate and N2O in boreal lakes and the important role of winter in annual emission estimates, the neglect of which results in underestimation of annual N2O flux estimates. During recent mild winters the ice cover period has been shorter and warm autumns have delayed the freezing day. Assuming one month shorter ice cover period (we extended autumn to three months, since especially autumn temperatures have been rising during recent years) resulted in 15 % larger evasion estimates. Climate change scenarios predict increasing temperature and precipitation for northern Europe (Ciais et al., 2013) accompanied by increasing frequency of freeze-thaw events which have been shown to result in enhanced nitrate and N2O fluxes (Cui et al., 2016). Rising temperature has further been shown to result in earlier spring snowmelt floods throughout northeastern Europe (Blöschl et al., 2017) which contributes to seasonal distribution of nitrate transport from land to lakes and further to the overall role of lakes as N2O sources in the boreal landscape. Acknowledgements Sincere thanks to our dear colleague Jari Huttunen, whose enthusiasm on N2O inspired us to finish this paper after his passing away, to two anonymous reviewers for the comments, which significantly improved the MS, to the personnel of the Water and Environment Districts for sampling and analyzing the water samples, to Antti Räike for preparing Figure 1 and to Tuija Mattson for her help with figures. Financial support was received from the Ministry of Environment and the Biowater project, funded by Nordforsk (Decision Number 82263) to PK and A cc ep te d A rt ic le This article is protected by copyright. All rights reserved from the Academy of Finland (Decision Number 57897 to PJM and JA, Decision Numbers 286731 and 319262 to TL and Decision Number 296423 to SJ). Competing financial interests The authors declare no competing financial interests. Data sharing and Data Accessibility The data that support the findings of this study are available from the corresponding author upon reasonable request. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved References Bastviken, D., Cole, J., Pace, M., and Tranvik, L. (2004). Methane emissions from lakes: Dependence of lake characteristics, two regional assessments, and a global estimate, Glob. Biogeochem. Cycles, 18, GB4009, doi:10.1029/2004GB002238. Beaulieu, J. J., et al. (2011). Nitrous oxide emission from denitrification in stream and river networks. Proc. Natl. Acad. Sci. USA 108, 214-219. Blöschl et al. (2017). Changing climate shifts timing of European floods. Science 357, 588-590. Butterbach-Bahl K, Baggs, E.M, Dannenmann M., Kiese R. & Zechmeister-Boltenstern S. Nitrous oxide emissions from soils: how well do we understand the processes and their controls? Phil Trans R Soc B 368, 20130122 (2013). Campbell, J.L., Socci, A.M. & Templer, P.H. (2014). Increased nitrogen leaching following soil freezing is due to decreased root uptake in a northern hardwood forest. Glob. Change Biol. 20, 2663-2673. Ciais, P., C., Bala, S. G., Bopp, L., Brovkin, J. Canadell, J., Chhabra, A. … Thornton, P., (2013): Carbon and Other Biogeochemical Cycles. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., S.K. Allen, J. Boschung, (eds.)]. Cambridge Univ. Press, Cambridge, United Kingdom and New York, NY, USA. Cui, Q., Song, C., Wang, X., Shi, F., Wang L. & Guo, Y. (2016). Rapid N2O fluxes at high level of nitrate nitrogen addition during freeze-thaw events in boreal peatlands of Northeast China. Atmospheric Environment 135 1-8. Deemer B. R. et al. (2016). Greenhouse Gas Emissions from Reservoir Water Surfaces: A New Global Synthesis. BioScience 66, 949-964. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved DelSontro, T., Beaulieu, J. J. & Downing, J. A. (2018). Greenhouse gas emissions from lakes and impoundements: Upscaling in the face of global change. Limnol. Oceanography Letters 3: 64-75. Denfeld B., Kortelainen P., Rantakari M., Sobek S. & Weyhenmeyer, G.A. (2016). Regional variability and drivers of below ice CO2 in boreal and subarctic lakes. Ecosystems 19, 461-476. Fick, A. (1855). On liquid diffusion. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science Series 4, 10, issue 63 V. Groffman, P.M., Driscoll, C.T. Fahey, T.J., Hardy, J.P., Fitzhugh, R.D. & Tierney, G.L. (2001). Colder soils in a warmer world: A snow manipulation study in a northern hardwood forest ecosystem. Biogeochemistry 56, 135–150. Heiskanen J.J. et al. (2014). Effects of cooling and internal wave motions on gas transfer coefficients in a boreal lake. Tellus B. 66, http://dx.doi.org/10.3402/tellusb.v66.22827. Henriksen, A. et al. (1998). Northern European Lake Survey, 1995: Finland, Norway, Sweden, Denmark, Russian Kola, Russian Karelia, Scotland and Wales. Ambio 27, 80-91. Her J. J. & Huang J. S. (1995). Influences of carbon source and C/N ratio on nitrate/nitrite denitrification and carbon breakthrough. Bioresource Technology 54, 45-51. Holgerson, M.A., Farr, E.R. & Raymond, P.A. (2017). Gas transfer velocities in small forested ponds. J. Geophys. Res. Biogeosciences 122, 1011-1021. Hu, M., Chen, D. & Dahlgren, R.A. (2016). Modeling nitrous oxide emission from rivers: A global assessment. Glob. Change Biol. 22, 3566-3582. Jähne, B., Heinz G. & Dietrich W. (1987). Measurements of the diffusion coefficients of sparingly soluble gases in water. J. Geophys. Res. 92, 10767-10776. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Juutinen, S. et al. (2009). Methane dynamics in different boreal lake types. Biogeosciences 6, 209- 223. Kortelainen, P., Rantakari, M., Huttunen, J. T., Mattsson, T., Alm, J., Juutinen, S., … Martikainen, P. J (2013). Carbon evasion/accumulation ratio in boreal lakes is linked to nitrogen. Glob. Biogeochem. Cycles 27, 363-374. Kortelainen, P. et al. (2006). Sediment respiration and lake trophic state are important predictors of large CO2 evasion from small boreal lakes. Glob. Change Biol. 12, 1554-1567. Leinonen L (ed.) (2000) Air Quality Measurements. SVT Official Statistics of Finland. Finnish Meteorological Institute, Helsinki. Leppelt, T., Dechow, R., Gebbert, S., Freibauer, A. Lohila, A., Augustin, J. … Strömgren, M. Nitrous oxide emission budgets and land-use driven hotspots for organic soils in Europe. Biogeosciences 11, 6595-6612 (2014). Mannio, J., Räike, A. & Vuorenmaa, J. (2000). Finnish lake survey 1995: Regional characteristics of lake chemistry. Verh. Int. Ver. Limnol. 27, 362-367. McAuliffe, C. C. GC (1971). Determination of solutes by multiple phase equilibration, Chem. Technol., 1, 46–51. McCrackin, M. L. & Elser, J. J. Greenhouse gas dynamics in lakes receiving atmospheric nitrogen deposition Glob. Biogeochem. Cycles 25, GB4005, doi:10.1029/2010GB003897 (2011). Miettinen, H., Pumpanen, J., Heiskanen, J., Aaltonen, H., Mammarella, I., Ojala, A., … , Rantakari, M. (2015). Towards a more comprehensive understanding of lacustrine greenhouse gas dynamics - two-year measurements of concentrations and fluxes of CO2, CH4 and N2O in a typical boreal lake surrounded by managed forests. Boreal Environment Research 20, 75-89. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Mulholland, P. J. et al. (2008). Stream denitrification across biomes and its response to anthropogenic nitrate loading. Nature 452, 202-205. National Board of Waters (1981). Methods of water analysis employed by the Water Administration (in Finnish with English summary), Helsinki, 136 pp. Palviainen, M., Lehtoranta, J., Ekholm, P., Ruoho-Airola, T. & Kortelainen, P. (2015). Land cover controls the export of terminal electron acceptors from boreal catchments. Ecosystems 18, 343- 358. Potter, C. S., Matson, P. A., Vitousek, P.M. & Davidson, E. A. (1996). Process modeling of controls on nitrogen trace gas emissions from soils worldwide. Journal of Geophysical Research, 101, 1361-1377. Pärn J., Verhoeven, T. A., Butterbach-Bahl, K., Dise, N. B., Ullah, S., Aasa, A. …Mander, Ü (2018). Nitrogen-rich organic soils under warm well-drained conditions are global nitrous oxide emission hotspots. Nature Communications 9, 1135 DOI: 10.1038/s41467-018-04197-6. Raatikainen M. & Kuusisto E. (1990). The number and surface area of the lakes in Finland. Terra, 102, 97-110. Rekolainen, S., Mitikka, S., Vuorenmaa, J. & Johansson, M. (2005). Rapid decline of dissolved nitrogen in Finnish lakes. Journal of Hydrology 304, 94-102. Richardson, D., Felgate, H., Watmough N., Thomson, A. & Baggs, E. (2009). Mitigating release of the potent greenhouse gas N2O from the nitrogen cycle – could enzymic regulation hold the key? Trends in Biotechnology 27, 388-397. Rissanen A.J., Tiirola M., Hietanen S. & Ojala A. (2013). Interlake variation and environmental controls of denitrification across different geographical scales. Aquatic Microbial Ecology 69: 1- 16.A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Ruan, L. & Robertson G.P. (2017). Reduced snow cover increases wintertime nitrous oxide (N2O) emissions from an agricultural soil in the Upper US Midwest. Ecosystems 20, 917-927. Seitzinger S.P. (1988). Denitrification in freshwater and coastal marine ecosystems: Ecological and geochemical significance. Limnology and Oceanography 33: 702-724. Silvennoinen H., Liikanen A., Torssonen J., Stange F. and Martikainen P.J. (2008). Denitrificaton and N2O effluxes in the Bothnian Bay (Northern Baltic Sea) river sediments as affected by temperature under different oxygen conditions. Biogeochemistry 88: 63-72. Soued, C., del Giorgio, P. A. & Maranger, R. (2015). Variability in N2O fluxes across boreal aquatic networks. Nature Geosci. 9, 116-120. Vaahtera E. et al. (2018). Finnish forest statistics Natural Resources Institute Finland, 189 pp. Vachon, D. & Prairie, Y.T. (2013). The ecosystem size and shape dependence of gas transfer velocity versus wind speed relationships in lakes. Can. J. Fish. Aquatic. Sci. 70, 1757-1764. Veraart, A.J., de Klein, J.J.M. & Scheffer, M. (2011). Warming can boost denitrification disproportionately due to altered oxygen dynamics. PLoS ONE 6, 1-6. Wanninkhof, R. & Knox M. (1996). Chemical enhancement of CO2 in natural waters. Limnol. Oceanogr. 41, 689–697. Webb, J.R., Hayes, N.M., Simpson G.L. Leavitt, P.R. Baulch, H.M. & Finlay, K. (2019). Widespread nitrous oxide undersaturation in farm waterbodies creates an unexpected greenhouse gas sink. Proc. Natl. Acad. Sci. USA 116 (20), 9814-9819. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Wik, M., Thornton, B. F., Bastviken, D., MacIntyre, S., Varner, R.K., Crill, P.M. (2014). Energy input is primary contoller of methane bubbling in aubarctic lakes. Geophys. Res. Letts. 41, 555- 560. 10.1002/2013GL058510. Whitfield, C. J., Aherne, J. & Baulch, H. M. (2011). Controls on greenhouse gas concentrations in polymictic headwater lakes in Ireland. Sci. Total Environ. 410, 217-225. Yang, H. et al. (2015). Greenhouse gas metabolism in Nordic boreal lakes. Biogeochemistry 126, 211–225. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Table 1. Average catchment characteristics and lake water chemistry of autumn 1 m samples for N2O lakes (sampled in 1998-1999) and for Nordic Lake Survey (NLS) lakes (sampled in 1995). N2O lakes, n = 87 NLS, n = 874 med mean min max med mean Lake area (km2) 0.31 3.3 0.04 44 0.22 20 CA/LA 14 49 2.8 610 16 72 Max depth (m) 6.8 9.1 1 42 NM NM Water (%) 9.2 11 0.3 35 8.8 10 Agric. land (%) 4.2 7.6 0.03 39 2.3 5.6 Peat (%) 8.3 14 0 84 14 18 Alkalinity (µmoll-1) 140 170 -70 990 130 150 pH 6.5 6.6 4.2 7.7 6.7 6.6 O2 (%) 82 79 47 95 NM NM pCO2 (µatm) 1100 1500 330 4500 860 990 Ntot (µg l-1) 500 560 170 2000 430 470 Inorganic N (%) 7.4 10 0.4 48 7.7 11 Ptot (µg l-1) 15 21 5.0 85 14 19 Inorganic P % 15 17 1.6 63 15 19 TIC (mg l-1) 2.0 2.5 0.6 11 1.6 2 TOC (mg l-1) 9.4 11 2.2 38 7.8 9 Fetot (µg l-1) 400 650 29 3600 360 560 A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Table 2. Correlation coefficient values between N2O and lake and catchment area, maximum depth, latitude, land use cover, elevation, water temperature and water chemistry in all lakes (randomly selected and eutrophic lakes), in surface water (all seasons), in bottom (all seasons) and in winter (all depths). NS not significant. All data Surface Bottom Winter Lake area 0.183 *** NS 0.248 *** 0.140 * Catchment area 0.294 *** 0.154 * 0.342 *** 0.267 *** Maximum depth 0.076 * NS NS NS Lake latitude -0.136 *** -0.194 *** -0.139 * -0.228 *** Forest % NS 0.100 * NS NS Peat % NS -0.123 * NS -0.155 ** Field % 0.211 *** 0.232 *** 0.190 *** 0.345 *** Built-up % NS NS NS 0.168 ** Water % -0.168 *** -0.209 *** -0.102 * -0.215 *** Lake elevation -0.195 *** -0.158 *** -0.193 *** -0.334 *** Water temperature -0.378 *** -0.558 *** -0.304 *** -0.239 *** Ntot 0.0692 * 0.332 *** NS NS NH4 NS 0.377 *** NS -0.178 ** NO3 0.593 *** 0.582 *** 0.612 *** 0.635 *** Ptot NS 0.194 *** NS 0.141 * PO4 NS 0.381 *** NS 0.115 * TOC 0.0737 * 0.110 * NS NS Water color NS 0.147 * NS NS O2 0.189 *** -0.105 * 0.206 *** 0.276 *** Fe NS 0.192 *** NS NS Conductivity 0.148 *** 0.260 *** NS 0.210 *** Alkalinity NS 0.143 * NS NS pH NS -0.120 * NS NS A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Table 3. Selected linear stepwise multiple regression equations for N2O (water chemistry, land use cover, climatic drivers and catchment topography were used as predictors; randomly selected and eutrophic lakes). Dataset n Dependent variable Intercept Parameter estimate Parameter Unit Model R2 All data 1396 ln N2O = 2.31 + 0.145 × ln NO3-N µg l-1 0.41 0.128 × ln O2 % 0.51 - 0.0303 × Temperature ºC 0.55 Surface 397 ln N2O = 2.74 + 0.0952 × ln NO3-N µg l-1 0.47 -0.0373 × Temperature ºC 0.56 0.0377 × √ Field % % 0.58 Bottom 408 ln N2O = 2.07 + 0.258 × ln NO3-N µg l-1 0.46 0.496 × ln O2 % 0.54 Winter 379 ln N2O = 123 + 0.394 × ln NO3-N µg l-1 0.52 -7.68 × ln Latitude ºN 0.57 -0.191 × pH 0.58 A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Table 4. Estimates of annual N2O flux (Gg N2O-N y-1) from Finnish and Boreal lakes. Annual estimates were calculated from our randomly selected set of lakes (n=71) using k values from (Holgerson et al. 2017) and different upscaling approaches: multiplying the area of the Finnish and boreal lakes with N2O flux estimates for each lake size class and for all size classes combined (the mean, median and summer median of individual lakes). For the median flux, the 25% and 75% quartiles and for the mean flux 95% confidence intervals, respectively, are given in parentheses. Two small humic lakes with fluxes of 863 and 22 085 mg N2O-N m-2 y-1 were excluded as outliers. Upscaling approach Finnish Lakes Boreal lakes median flux by lake size class 0.6 (0.5-0.7) 29 (26-32) median flux of individual lakes 0.3 (0.09-0.7) 12 (4-31) summer median flux of individual lakes 0.07 (-0.2-0.3) 3 (-10-13) mean flux of individual lakes 0.5 (0.3 -0.7) 23 (15-31) Area (km2) 32 6631 1 422 4482 1Lake area distribution (Raatikainen & Kuusisto (1990) 2Lake area distribution estimated based on MODIS data, excluding lakes <0.1 km2 A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Table 5. Annual N2O flux estimates (mg N2O-N m-2 y-1) by lake type based on all lakes that were sampled at all four seasons (n=94; the water quality data were missing from two lakes and the lake type could not be assigned): Annual fluxes for the randomly selected lakes (n=71), and for the subset of Eutrophic lakes with the highest total P concentrations (n=23). The annual fluxes (7- month ice-free season) consist of fluxes at the thaw (0.5 months), in spring (1.5 months), in summer (3 months), and in autumn (2 months) calculated using k values by Holgerson et al. (2017). Two small humic lakes with fluxes of 863 and 22 085 mg N2O-N m-2 y-1 were excluded as outliers. Lake type/group Mean Median SD CV% N Nutrient rich, calcareous 43 35 33.1 78 12 Clear water 9 11 11.7 126 14 Humic, large 60 41 44.2 74 7 Humic, small 18 8 29.7 164 59 All 23 11 32.2 142 94 Randomly selected 16 9 25.1 155 71 Eutrophic 43 35 42.6 99 23 A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Table 6. Estimates of annual N2O flux from Finnish and boreal lakes (calculated as median fluxes by lake size class, k values by Holgerson et al. (2017) compared to estimates of annual N2O flux from boreal forest in Finland and in boreal zone (using the estimate of 17.6 mg N2O-N m-2 y-1; Potter et al. 1996). Finland Boreal region Lake N2O-N flux (Gg y-1) 0.6 29 Forest N2O-N flux (Gg y-1) 41 2131 Lake area (km2) 32 6632 1 422 4484 Forest area (km2) 203 0003 21 200 0001 Lake N2O-N flux/Forest N2O-N flux (%) 17 14 1 Potter et al. (1996) 2 Raatikainen & Kuusisto (1990) 3 Vaahtera (2018) 4 Lake area distribution estimated based on MODIS data, excluding lakes <0.1 km2 A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Figure captions Figure 1. Location of the study lakes (n=112). Figure 2. Seasonal distribution (median, 1st and 3rd quartile) of the concentrations of N2O (n=87), CO2 (n=177) and CH4 (n=177), in randomly selected lakes, all depths. Minimum N2O concentrations were measured in summer in contrast to CO2 and CH4 distributions. Concentrations of CO2 and CH4 are based on the data from Kortelainen et al. (2006) and Juutinen et al. (2009). Note that y axis is on a log scale. Figure 3. Distribution of N2O (a) and NO3-N (b) across different lake types, randomly selected lakes, all seasons and depths. NRC Nutrient Rich and Calcareous, HSm Humic Small, HL Humic Large and CL Clear Water Lakes. Seasonal variation of N2O concentrations followed the variation of nitrate - highest in winter and lowest in summer despite differences in concentration levels across the lake type classification. Figure 4. The relationship between NO3-N and N2O (a), and NO3-N and N2O/NO3-N -ratio (b). All lakes, seasons and depths. Lake type identified, all data were log-transformed. Our randomly selected data across all lake types and seasons showed strong positive correlation between nitrate and N2O and strong negative correlation between nitrate and N2O/NO3-N –ratio. (Coloured online only). Figure 5. Distribution (median, 1st and 3rd quartile) of N2O across depth; randomly selected lakes, all seasons. Median N2O was evenly distributed across depth. Figure 6. N2O distribution across pH, different lake types identified as in Fig. 3; all lakes, depths and seasons. Both the highest and the lowest N2O concentrations occurred around the median pH of 6.5. (Coloured online only). Figure 7. N2O concentrations in different lake types before (a) and after ice melt (b) across oxygen saturation percentage; all lakes and depths. Oxygen was not a key driver for N2O in A cc ep te d A rt ic le This article is protected by copyright. All rights reserved Finnish boreal lakes. (Coloured online only). Figure 8. Distribution (median, 1st and 3rd quartile) of N2O across elevation (a) and the percentage of agricultural land in the catchment (d). Distribution (median, 1st and 3rd quartile) of NO3-N across lake area (b) and season (e). Distribution (median, 1st and 3rd quartile) of N2O/NO3-N ratio across lake area (c) and season (f), randomly selected lakes, all depths. Figure 9. Estimated annual N2O flux to the atmosphere across different lake size classes; randomly selected lakes (n=71) by the approaches of Holgerson et al. (2017) (a), Heiskanen et al. (2014) (b), and Vachon and Prairie (2013) (c). Nitrate and N2O concentrations were largest in large lakes resulting in largest emission estimates per surface area unit from largest lakes in contrast to CO2 and CH4. Figure 10. The relationship between mean NO3-N and mean N2O concentrations in different lake types across the four seasons (Equation 4). Seasonal mean values for each lake type (n=16) were obtained by averaging all measured depths across respective lakes. Our seasonal data from 112 boreal lakes in Finland underline the key role of nitrate in regulating seasonal N2O concentrations. Figure 11. Estimated N2O-N concentrations (by equation 4) in the Finnish NLS lakes (n = 874 autumn 1 m samples (grey symbols; see Table 1). The measured N2O-N concentrations in our study lakes (n = 1 542 seasonal water samples in 112 lakes) are shown for comparison (white symbols). The developed equation between nitrate and N2O well predicted N2O-N in the data of 874 randomly selected Finnish lakes sampled in autumn 1995 (NLS lakes). A cc ep te d A rt ic le gcb_14928_f1.eps This article is protected by copyright. All rights reserved A cc ep te d A rt ic le gcb_14928_f2.tif This article is protected by copyright. All rights reserved A cc ep te d A rt ic le gcb_14928_f3.tif This article is protected by copyright. All rights reserved A cc ep te d A rt ic le gcb_14928_f4.tif This article is protected by copyright. All rights reserved A cc ep te d A rt ic le gcb_14928_f5.tif This article is protected by copyright. All rights reserved A cc ep te d A rt ic le gcb_14928_f6.tif This article is protected by copyright. All rights reserved A cc ep te d A rt ic le gcb_14928_f7.tif This article is protected by copyright. All rights reserved A cc ep te d A rt ic le A cc ep te d A rt ic le A cc ep te d A rt ic le gcb_14928_f10.tif This article is protected by copyright. All rights reserved A cc ep te d A rt ic le gcb_14928_f11.tif This article is protected by copyright. All rights reserved A cc ep te d A rt ic le