EFFECTS OF LAND USE ON NORTHERN BOREAL STREAMS

Degree thesis in biology, 15 ECTS

Bachelor’s programme in biology and geosciences, 180 ECTS

EFFECTS OF LAND USE ON NORTHERN BOREAL

STREAMS

A study of stream nutrient patterns in Röbäcksdalen, Umeå

Emma Stenlund

Effekter av markanvändning på boreala vattendrag

- en studie av näringsmönster i Röbäcksdalen, Umeå

Abstract

Nutrient loading to aquatic and marine ecosystems is a topic of interest, especially as the human population continues to grow and land use changes. Here, I examined the seasonal variability and relative amounts of different forms of nitrogen (N) and phosphorous (P) from four sites at Röbäcksdalen (Umeå, Västerbotten), an area influenced by both agriculture and partial urbanization. In addition, I studied how nutrient concentration varied with discharge during snowmelt. Overall, the results show that the seasonality of nutrient concentration did not differ drastically from what is expected in more pristine boreal catchments. However, the concentrations of dissolved inorganic N (DIN) and phosphate (PO4) were elevated. Land use activities in Röbäcksdalen also appear to be influencing inorganic N inputs to streams to a greater degree than inputs of inorganic P. Comparisons with more pristine boreal streams also reveal a fundamental change in the composition of the stream N pool, with greater dominance of DIN within the catchment. This pattern is likely due to inputs of nutrients in excess of biotic demand. Lastly, the results of the concentration-discharge analysis from Röbäcksdalen indicate that hydrological forcing rather than soil processes control the supply of nutrients to these streams during snowmelt. With concentrations being high, this also indicates that a considerable increase in the flux of nutrients from the area is expected with increased discharge.

Key words: nutrients, land use, agriculture, stream chemistry, discharge

Table of content

1 Introduction ... 1

1.1 Background ... 1

1.2 Purpose ...2

2 Method ... 3

2.1 Study site ... 3

2.2 Data sources ... 3

2.3 Data analysis ...4

3 Results ... 5

3.1 Nutrient concentrations throughout the year ... 5

3.2 Trends in composition of N ... 7

3.3 DIN:PO4 ratio ... 8

3.4 Concentration-discharge analysis ... 8

4 Discussion ... 10

4.1 Seasonal variability ... 10

4.2 Differences in seasonality ... 10

4.3 Implications for nutrient limitation ... 11

4.4 The relative role of DON ... 10

4.5 NH

and NO

relationship ... 10

4.6 Concentration-discharge relationships ... 12

5 Conclusion ... 13

6 Acknowledgements ... 14

7 References ... 14

1 Introduction

1.1 Background

The world is facing great challenges in terms of ecosystem degradation and a changing climate. A growing human population induces an increased demand for food resources and further increases the anthropogenic pressure on arable land (IPCC 2019). Between 1960 and 2015, agricultural production increased more than threefold (FAO 2017), made possible in part by an increased production of synthetic fertilizers (Galloway et al. 2008). Predictions of future population growth estimates that the human population could grow by more than 40%

between 2019 and 2100 (United Nations 2019), further increasing the demand for food.

Over the last decades, increasing concerns about nutrient pollution and eutrophication of freshwater and marine ecosystems have emerged with many studies conducted on the topic.

Extensive nutrient loading can disfavor certain species of organisms, while favoring algae that can cause toxic blooms and/or contribute to anoxic water conditions (Sukenik et al.

2015). At an even larger scale, fluxes of nutrients into the Baltic Sea are thought to be one of the reasons behind large-scale algal blooms causing oxygen depletion on the sea bottom (Conley et al. 2002). Yet, understanding what drives the flux of nutrients from terrestrial to aquatic systems remains an important research challenge. This flux depends on many factors including hydrology (Teutschbein et al. 2017), surficial geology and geomorphology

(Sandström et al. 2020; Thomas et al. 2016; Blackburn et al. 2017), nutrient use by plants and soil microbes (Peterjohn and Correll 1984), disturbance and succession (Vitousek and Reiners 1975), and the spatial distribution of nutrients within the soil (Abbott et al. 2018).

Given this range of factors, it can be difficult to resolve the main drivers, which are also likely to vary in space and time.

Nutrients are essential to all living organisms. In aquatic environments, nutrient limitation means that the supply of the limiting nutrient (that is biologically available) determines the rate and extent of growth by primary producers (Smith et al. 1999). Nutrient limitation can have effects on ecology in terms of favoring certain species and disfavoring others (Sukenik et al. 2015). Considering the ratio of different nutrients can be valuable in order to assess which of these is limiting a system (Sterner and Elser 2002). For example, the Redfield ratio

describes the molar carbon:nitrogen:phosphorous (C:N:P) ratio of nutrient demand for growth of oceanic planktonic biomass, which is considered to be approximately 106:16:1. The relationship was brought to light in an article from 1934 (Redfield 1934) and has been

supported by numerous studies since (e.g., Loladze & Elser 2011). However, studies also suggest that planktonic communities are too complex for a general ratio to be fully applicable (Ptacnik et al. 2010).

Nutrients can also be found in water in many different forms that differ in terms of sources and bioavailability (e.g., Soares et al. 2017). Nitrogen for example can exist in organic and inorganic forms, where the latter is more bioavailable. Boreal forests in Sweden are generally N-limited, and the retention of dissolved inorganic N (DIN) by plants is high (Högberg et al.

2017). Because of this terrestrial demand, the pool of N leeching into freshwater systems tends to be dominated by dissolved organic nitrogen (DON), which is less easily available to biota (Hedin et al. 1995; Sponseller et al. 2014). The main input sources of nitrogen to boreal forests in Sweden are atmospheric deposition and N² fixation from the air. These inputs are small in comparison to the cycling of nitrogen within the systems, which is strongly driven by the seasonality of plant growth (Högberg et al. 2017). Generally, lower concentrations of DIN are found in freshwaters during the growing season when the demand for nutrients is high, while higher concentrations are observed as temperature drops and hours of sunlight decrease during the winter, limiting plant growth. This pattern is true for boreal forests in Sweden and most of the northern hemisphere (Sponseller et al. 2014). In agricultural lands, these patterns may be interrupted by fertilizer additions and livestock waste, which have been shown to increase the inorganic N found in water (Hägg et al. 2011).

In addition to the relative contribution of organic N, DIN is also made up of different forms that can be diagnostic for specific catchment features and/or anthropogenic influences. For example, under oxic conditions with limited anthropogenic impact, the pool of DIN is mostly dominated by nitrate (NO3). Under these circumstances, ammonium (NH4) concentrations are low because it can be readily turned into NO3 by nitrifying bacteria. Small concentrations of nitrite (NO²) can also be found, but since this form of N exists as a transition state between NH⁴ and NO³, it is rarely found in high concentrations (Stanley and Maxted 2008). On the other hand, in low oxygen environments (e.g., wetland soils), NO3 is typically removed via denitrification, whereas NH4 can accumulate (Helton et al. 2015). In this way, the balance of these two inorganic forms can provide insight into catchment structure (Sponseller et al.

2018). However, catchment disturbance and/or application of organic and inorganic fertilizers may alter this balance in ways that are not predictable from thermodynamics.

To understand the transport of nutrients from terrestrial to aquatic systems, we need tools and metrics that capture key aspects of this flux: the concentration (C) of a given nutrient and the water flow, or discharge (Q). While organisms within the system are mainly affected by the instantaneous concentration, downstream lakes and coastal communities are affected by the total flux of nutrients from these systems, which is determined by the product

concentration and discharge (Moatar et al. 2017). In addition, evaluating the relationship between concentration of nutrients and the discharge can give insight to whether the input of nutrients is controlled by water flux or if processes in the soil are limiting this input. If the water flux is the main driver of nutrient supply, then the concentration of nutrient either increases with flow or does not change (Godsey et al. 2009). By contrast, if the flux to streams is limited by the amount of nutrients available in the soil, then concentrations will typically decrease (be diluted) with increasing discharge. These two different types of controls are represented as transport limitation and source limitation, respectively (Basu et al. 2010). How solute concentration varies with flow also depends on the chemical and physical properties of the solute (Moatar et al. 2017) and their sources (Godsey et al. 2009).

The concentration of geogenic solutes such as calcium (Ca), magnesium (Mg), natrium (Na) and silica (Si) have been found to remain somewhat constant with variations in discharge (Godsey et al. 2009), also known as chemostasis. However, for nutrients, concentrations may also be chemodynamic, either increasing or decreasing with discharge. For example, nitrate in heavily managed agricultural catchments has shown signs of transport limitation, where a greater flow is associated with increased or constant concentrations (Basu et al. 2010). Such responses are likely caused by an abundance of NO³ present in the soils that exceeds the biotic demand and can be flushed out when water levels increase (Moatar et al. 2017). In other systems, where NO3 is not abundant in the soil and biotic demand is high, dilution is expected as discharge increases. This pattern is often observed in forested catchments (Basu et al. 2010).

1.2 Purpose

The objective of this study is to characterize the composition and seasonal patterns of dissolved nutrients for agricultural streams at the Röbäcksdalen field station (Umeå,

Västerbotten), with the purpose of examining how land use influences the concentrations and make-up of N and P in boreal streams in northern Sweden. To provide insight into the

sources of nutrients to the catchment, I will study the proportions of different forms of N and P present (NO³, NH⁴, DON and phosphate (PO⁴)) and their seasonal variability. In addition, I will study the relative concentrations of N to P, which has implications for biological nutrient limitation. Comparisons between data from Röbäcksdalen and the forested Krycklan

catchment (Vindeln, Västerbotten) will be used to assess if land use affects these patterns.

Lastly, I will study the Q-C relationship for both N and P in Röbäcksdalen during snowmelt in order to examine whether the fluxes of nutrients from land to water are limited by

processes in the soil or rather by hydrological transportation. I am predicting to find seasonal differences in the make-up of the nitrogen pool that deviate from the ones observed in more pristine forested catchments (e.g. the Krycklan catchment). I also predict to find generally

higher concentrations of nutrients in the Röbäcksdalen streams. Based on prior findings for agricultural streams elsewhere (e.g. Basu et al. 2010), I further expect that the Röbäck data will exhibit a more chemostatic or enriching chemodynamic Q-C behavior than the Krycklan data, where more diluting processes would be expected.

2 Method

2.1 Study site

The data I used for my analyses has been obtained from 4 different sampling locations within the Röbäcksdalen field research station (Umeå, Västerbotten). The catchment is

characterized by agricultural activity and lies in close proximity to urban areas (SLU 2020a).

Site 4 is located in Degernäsbäcken, a stream that was constructed in the early 19th century for agricultural purposes. Site 6 is found in Röbäcken, which is of natural origin and drains into Degernäsbäcken. Upstream, Röbäcken passes through a small forested area. Site 1 and 5 are both located downstream where the two streams are merged. Within Röbäcksdalen, there are roughly 300 ha of agricultural land (SLU 2020a). The streams running through the area are therefore heavily influenced by agriculture. Both streams are also partially influenced by urbanization. The soil consists mainly of clay and silt, deposited when the area was

submerged in water during the latest ice age. The field station is managed by the Swedish University of Agricultural Sciences (SLU), and research has been conducted in the area for many years.

Figure 1. The Röbäcksdalen sampling sites (Umeå, Västerbotten).

2.2 Data sources

The nutrient data for Röbäcksdalen was provided by SLU and included concentrations of nitrate nitrogen (N-NO3), ammonium nitrogen (N-NH4), phosphate phosphorous (P-PO4) and total nitrogen (N-tot). N-tot was analyzed using a Shimadzu TOC-V^CPH analyzer

(Shimadzu), which applies a combustion catalytic oxidation method. N-NO3, N-NH4 and P- PO⁴ were analyzed using a SEAL Analytical AutoAnalyzer 3 (SEAL Analytical), which is a flow-segmented colorimeter analyzer (Burrows et al. 2016). From these data, DIN was determined as the sum of N-NO3 and N-NH4. The dataset did not include nitrite nitrogen (N- NO2) concentrations; hence this source of nitrogen is missing from the DIN pool. It can however be assumed that NO² concentrations are substantially smaller than the

concentrations of NO³ (Stanley and Maxted 2008). DON was determined as the difference between N-tot and DIN.

The discharge data were obtained from a thesis project from 2019 (Söderlund 2019), where three of the sites at Röbäcksdalen (Site 4, 5 and 6) were sampled at 10 occasions during snowmelt. Discharge was calculated from estimates of water velocity and measurement of cross-sectional area on each date. For eight of these dates, matching nutrient data could be obtained from the dataset provided by SLU, and these were used for the Q-C analyses.

Finally, data from the Krycklan catchment used in this report included the same parameters mentioned above, sampled throughout 2015 and 2016. These data were also provided by SLU using the same analytical methods and included samples from three separate sites. One of these is a forest stream (Site C2), the second drains a mire (C4) and the third is the outlet of a lake (C5) (Burrows et al. 2016).

2.3 Data analysis

To examine potential seasonal differences in nutrient concentrations in Röbäcksdalen, the samples were divided into four categories: winter (Dec-Feb), spring (Mar-May), summer (Jun-Aug) and autumn (Sep-Nov). Only samples with complete nutrient data were used for the analyses. Nutrient concentrations for each season in each of the four sites were displayed in box-whisker plots for comparison, with the number of samples (n) for each season ranging from 5 to 15. For each season, a box with samples from the three sites at the Krycklan

catchment combined was added alongside the Röbäcksdalen boxes for further comparison.

Beyond studying visual differences in the data, non-parametric ANOVAs (Kruskal-Wallis analyses) were carried out to test for significant differences between the seasons. Site 1 and Site 5 were excluded from these analyses since they are located downstream from Site 4 and Site 6, and hence are not independent. Pairwise comparisons were done using Dunn’s test to determine which individual seasons that differed from one another. Bonferroni correction was applied. These analyses were carried out using XLSTAT version 2020.2.2 (Addinsoft 2020).

To explore the composition of the nitrogen pool, %N-NO³, %N-NH⁴ and %DON (of total N) were plotted against the concentration of total N for each of the four Röbäcksdalen sites (following Stanley and Maxted 2008). One sample from 2018-05-02 was removed because the concentration of DIN was higher than the N-tot concentration, which indicates an error.

In addition, the concentration of N-NO³ was plotted against that of N-NH⁴ for all

Röbäcksdalen sites to explore whether the composition of the DIN pool corresponded to expectations based on thermodynamics (Helton et al. 2015). Finally, parallel calculations and plots were generated for all Krycklan sites combined, and these were plotted alongside the Röbäcksdalen data for comparative purposes.

To examine the N:P ratio, the molar concentration of DIN was divided by the molar concentration of P-PO4 for each sample. These parameters were chosen since no total phosphorous (P-tot) concentrations were available. Based on the Redfield ratio (Redfield 1934), a quotient >16 would indicate a system is limited by nitrogen while a quotient <16 would indicate phosphorous limitation.

For the discharge-concentration analysis, linear regression was used to assess the slope of the relationships. Only data from snowmelt was used for this (March 27-April 24 for

Figure 2. a) Concentration of N-NH4 in g/l for each site and season for Röbäcksdalen and Krycklan.

b) Concentration of N-NH4 in g/l for each season at Krycklan (all three sites combined). Mean values are marked with x and statistical outliers are shown as points.

Röbäcksdalen and April 7-May 31 for Krycklan), and all values were logged prior to this analysis. A slope within the boundaries of 0.2 was categorized as chemostatic, while slopes outside these boundaries were labeled as chemodynamic (Meybeck and Moatar 2012).

Significance was determined by the p-value obtained from the regression analysis ( = 0.1).

Due to low power in terms of few samples, the level of significance was set to 0.1. This allowed for observed patterns to be categorized to a greater extent. This slightly more

generous level of significance was taken into account when evaluating the results and caution was applied when drawing conclusions from this analysis. One sample from Site C5 was removed for P-PO4 since it had a value of only 0.01 g/l, which is far below the detection limit of the method for the chemical analysis (SLU 2020b).

3 Results

3.1 Nutrient concentrations throughout the year

For N-NH4, the highest average concentrations in Röbäcksdalen were found in winter (220- 383 µg/l) and spring (266-363 µg/l) while lower averages were found in summer (134-168 µg/l) and autumn (146-168 µg/l) (figure 2a). Large differences were also found between average N-NH⁴ in the Röbäcksdalen streams and the Krycklan sites, which average only 12-29 µg/l over the year (figure 2b).

Average N-NO³ concentrations were generally the highest in spring (440-693 µg/l) and autumn (307-439 µg/l) with some variance between the sites (figure 3a). Summer

concentrations were much lower (78-196 µg/l). The highest average concentration of N-NO3

was found in Site 4 for all seasons except winter. For all sites and seasons, the Röbäcksdalen concentrations were much higher than the concentrations found in Krycklan throughout the year (11-16 µg/l), with averages being more than tenfold (figure 3b).

N-NH4 concentration (µg/l)

0 20 40 60 80 100 120 140 160 180

Krycklan N-NH4 concentration (µg/l)

Winter Spring Summer Autumn

a b

Figure 3. a) Concentration of N-NO3

in g/l for each site and season for Röbäcksdalen and Krycklan.

b) Concentration of N-NO3 in g/l for each season at Krycklan (all three sites combined). Mean values are marked with x and statistical outliers are shown as points.

Figure 4. a) Concentration of DON in g/l for each site and season for Röbäcksdalen and Krycklan.

b) Concentration of DON in g/l for each season at Krycklan (all three sites combined). Mean values are marked with x and statistical outliers are shown as points.

Average DON concentrations were the highest in the spring for both Röbäcksdalen (358-425 µg/l) and Krycklan (452 µg/l) (figure 4a). Contrary to N-NO³ concentration, Site 4 exhibited the lowest average concentration of DON for all seasons. While there were no striking differences between the Röbäcksdalen and Krycklan concentrations, the Krycklan data exhibited higher average concentrations for all seasons – especially summer and autumn.

Seemingly, Röbäcksdalen displayed greater seasonal variation in DON concentration than Krycklan.

For P-PO4, the highest average concentrations were found in spring (21-27 µg/l) for the Röbäcksdalen sites (figure 5a). The lowest averages were found in summer (3-7 µg/l), followed by autumn (9-11 µg/l). For Krycklan, average concentrations ranged from 2-6 µg/l throughout the year, with spring concentrations being the lowest.

N-NO3 concentrations (µg/l)

0 10 20 30 40 50 60 70 80

Krycklan N-NO3 concentrations (µg/l)

Winter Spring Summer Autumn

DON concentration (µg/l)

0 200 400 600 800 1000 1200 1400

Krycklan DON concentration (µg/l)

Winter Spring Summer Autumn

a b

a b

Figure 5. a) Concentration of P-PO4 in

g/l for each site and season for Röbäcksdalen and Krycklan.

b) Concentration of N-NO3 in g/l for each season at Krycklan (all three sites combined). Mean values are marked with x and statistical outliers are shown as points.

For all parameters and sites at Röbäcksdalen, summer concentrations were among the lowest and least variable both between sites and between samples. Spring concentrations were generally the highest. These observations were strengthened by the statistical analysis. For all parameters, the Kruskal Wallis test showed a significant difference in concentration between the four seasons (p << 0.05). Looking at the pairwise comparisons, significance was less evident (table 1). Summer and spring concentrations were significantly different for all parameters. Significant differences were also found between autumn and spring for all parameters but N-NO³. Lastly, the difference between summer and autumn was significant for N-NO3.

Table 1. P-values from Dunn’s test determining significant differences in concentration between seasons. Only data from Site 4 and Site 6 was included in the analysis. The adjusted level of significance is  = 0.0084 due to Bonferroni correction. Significant p-values are highlighted in green, while p-values close to significance are highlighted in yellow.

3.2 Trends in composition of N

In Röbäcksdalen, an increase in N-tot concentration was accompanied by a decrease in the proportion of N-NH⁴ (figure 6a) and a greater dominance of N-NO³ (figure 6b). The

proportion of DON was also much smaller for the Röbäcksdalen sites (mostly ranging from 10-60%) than for Krycklan, where %DON mostly ranged from 80 to nearly 100% (figure 6c).

The concentrations of N-NO³ and N-NH⁴ were also generally much higher in Röbäcksdalen than in Krycklan. The relationship between N-NO³ and N-NH⁴ was positive for

Röbäcksdalen, meaning that higher concentrations of N-NH4 mostly coincided with higher concentrations of N-NO3 (figure 6d). No such relationship was evident for the Krycklan data, where high N-NH⁴ concentrations did not seem to coincide with high N-NO³ concentrations.

Pairwise comparison (Dunn's test) N-NH4 N-NO3 DON P-PO4

Winter - Spring 0.199 0.014 0.009 0.019

Winter - Summer 0.158 0.085 0.358 0.341

Winter - Autumn 0.200 0.680 0.766 0.924

Spring - Summer 0.001 <0.001 <0.001 <0.001

Summer - Autumn 0.900 0.007 0.123 0.282

Autumn - Spring 0.001 0.011 0.004 0.002

Bonferroni corrected significance level: 0.0083

P-PO4 concentration (µg/l)

0 5 10 15 20 25 30 35

Krycklan P-PO4 concentration (µg/l)

Winter Spring Summer Autumn

a b

Site 1 Site 4 Site 5 Site 6 Krycklan

For Site C4, which drains a wetland, average N-NH⁴ concentrations (34 g/l) were three times higher than N-NO³ concentrations (11 g/l).

Figure 6. a) Percentage N-NH4 against the concentration of N-tot b) Percentage N-NO3 against the concentration of N-tot c) Percentage DON against the concentration of N-tot d) Concentration of N-NO3 against the

concentration of N-NH4. Röbäcksdalen samples are divided by site, while the Krycklan sites are joined.

3.3 DIN:PO4 ratios

Great variance was observed for the ratio of DIN:P-PO4. The smallest quotient (DIN/P-PO4) observed was 23.12, the largest being 4968.28. The average quotient was 288.27 meaning that, on average, the molar concentrations of DIN found at the Röbäcksdalen sites were about 288 times higher than the molar concentrations of P-PO⁴. For Krycklan, the smallest value observed was 0.60 while the highest was 368.38. The average quotient for Krycklan was 36.26.

3.4 Concentration-discharge analysis

The results from the Q-C analysis show trends of chemostasis for all three Röbäcksdalen sites for DON (figure 7c) and N-NO³ (figure 7b), based on the slope being within the range of 0.2.

For N-NH⁴ (figure 7a) however, slopes of m=-0.331 and m=-0.954 were found in Site 5 and Site 6, respectively. This pattern would be classified as diluting chemodynamic, i.e. the concentration of NH4 decreases with increased discharge. Chemodynamic patterns were also observed for P-PO⁴ (figure 7d) in Site 4 , where m=0.619. The slope being positive indicates enrichment of PO⁴ with increased discharge. However, the only regression relationships from Röbäcksdalen that was significant (p<0.1) was the relationship between discharge and N- NH4 at Site 4 (p=0.057) and Site 6 (p=0.023) (table 2). Since regression tests the hypothesis of the slope of the linear relationship being significantly different from zero, a p-value greater

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 500 1000 1500 2000 2500 3000

%N-NO3

Concentration N-tot (µg/l)

%N-NO3

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 500 1000 1500 2000 2500 3000

%DON

Concentration N-tot (µg/l)

%DON

0 200 400 600 800 1000 1200 1400 1600

0 200 400 600 800 1000 1200 1400 1600

Concentration N-NO3 (µg/l)

Concentration N-NH4 (µg/l)

N-NO3 and N-NH4

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 500 1000 1500 2000 2500 3000

%N-NH4

Concentration N-tot (µg/l)

%N-NH4

a b

c d

than 0.1 would mean that the null-hypothesis of the slope being equal to zero cannot be rejected. Therefore, the lines with p-values >0.1 were considered chemostatic/possibly chemodynamic even though the slope was outside the 0.2 interval.

Figure 7. Concentration-discharge (Q-C) relationships for Röbäcksdalen (a-d) and Krycklan (e-h). A slope (m) within the boundaries of 0.2 is categorized as chemostatic, while slopes outside these boundaries are categorized as chemodynamic (enriching or diluting depending on sign of the slope).

m = 0,6192 R² = 0,0855 m = 0,0814 R² = 0,0076 m = 0,0816 R² = 0,0054

-0,5 0,0 0,5 1,0 1,5 2,0 2,5

-1,2 -1 -0,8 -0,6 -0,4 -0,2 0 0,2 0,4

log(P-PO4)

log(Discharge) Q-C log(P-PO4)

Site 4 Site 5 Site 6 m = -0,1507 R² = 0,0484

m = 0,0715 R² = 0,028 m = 0,0583 R² = 0,0629

0,00 0,50 1,00 1,50 2,00 2,50 3,00 3,50

-1,2 -1 -0,8 -0,6 -0,4 -0,2 0 0,2 0,4

log(N-NO3)

log(Discharge) Q-C log(N-NO3)

Site 4 Site 5 Site 6 m = -0,1481 R² = 0,4779

m = -0,3313 R² = 0,2327 m = -0,9545 R² = 0,6037

0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5

-1,2 -1 -0,8 -0,6 -0,4 -0,2 0 0,2 0,4

log(N-NH4)

log(Discharge) Q-C log(N-NH4)

Site 4 Site 5 Site 6

m = -0,1217 R² = 0,0505 m = -0,1074 R² = 0,1401 m = 0,1046 R² = 0,2039

0 0,5 1 1,5 2 2,5 3 3,5

-1,2 -1 -0,8 -0,6 -0,4 -0,2 0 0,2 0,4

log(DON)

log(Discharge) Q-C log(DON)

Site 4 Site 5 Site 6

m = 0,0191 R² = 0,0079 m = -0,2957 R² = 0,3357

m = -0,1656 R² = 0,059 _0,0

0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8 2,0

-4,0 -3,5 -3,0 -2,5 -2,0 -1,5 -1,0 -0,5 0,0

log(N-NO3)

log(Discharge) Q-C log(N-NO3)

Site C2 Site C4 Site C5

m = -0,1514 R² = 0,0385 m = -0,472 R² = 0,2907 m = -0,2648 R² = 0,1771

0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5

-4,0 -3,5 -3,0 -2,5 -2,0 -1,5 -1,0 -0,5 0,0

log(DON)

log(Discarge) Q-C log(DON)

Site C2 Site C4 Site C5 m = 0,1709 R² = 0,0414

m = -0,6166 R² = 0,6515

m = -0,32 R² = 0,1649 _0,0

0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8

-4,0 -3,5 -3,0 -2,5 -2,0 -1,5 -1,0 -0,5 0,0

log(N-NH4)

log(Discharge) Q-C log(N-NH4)

Site C2 Site C4 Site C5

m = -0,1686 R² = 0,1397 m = -0,8289 R² = 0,5355 m = -0,4563 R² = 0,2262

-1,2 -1,0 -0,8 -0,6 -0,4 -0,2 0,0 0,2 0,4 0,6 0,8 1,0

-4,0 -3,5 -3,0 -2,5 -2,0 -1,5 -1,0 -0,5 0,0

log(P-PO4)

log(Discharge) Q-C log(P-PO4)

Site C2 Site C4 Site C5

a e

b

c

f

g

d h

Röbäcksdalen Krycklan

For the Krycklan sites, the relationships were more variable for the different sites (table 2).

Site C2 (forest) behaved chemostatically for all parameters. Only P-PO⁴ showed significance, however. Site C4 (mire) was significantly diluting chemodynamic for all parameters. Lastly.

Site C5 (lake) was significantly diluting chemodynamic for DON, N-NH4 and P-PO4 (after outlier removal), while being chemostatic for N-NO3 (not significant).

Table 2. Summary of concentration-discharge analysis results (log-log linear regression) showing slopes (m), R2- values, p-values and Q-C pattern. The significance level was  = 0.1. Pattern abbreviations indicate whether the relationship is chemostatic (CST), enriching chemodynamic (CDN+) or diluting chemodynamic (CND-). Lack of significance means the relationship cannot be proven to not be chemostatic. Non-significant chemodynamic patterns are marked CST/CDN.

4 Discussion

The main goal of this study was to examine the effects of land use on nutrient seasonality, relative N and P supply, the make-up of the N-pool, and the concentration-discharge relationships in small boreal streams. Overall, the results show that the seasonality of DIN and PO4 does not differ drastically from what is expected in more pristine boreal catchments, contrary to my predictions. However, the concentrations of these forms of N and P were, on average, 17 respectively 4 times higher in Röbäcksdalen than in the Krycklan catchment. In addition, the land use activities in Röbäck appear to be influencing inorganic N inputs to streams to a greater degree than inputs of inorganic P. Comparisons with more pristine boreal streams also reveal a fundamental change in the composition of the stream N pool, with greater dominance of DIN within the catchment. Finally, the chemostatic behavior of the Q-C curves from Röbäcksdalen indicate that hydrological forcing rather than soil processes controls the supply of nutrients to these streams during snowmelt.

Site Parameter Slope R²-value p-value Pattern

4 DON -0.122 0.050 0.593 CST

N-NO3 -0.151 0.048 0.601 CST

N-NH4 -0.148 0.478 0.057 CST

P-PO4 0.619 0.086 0.482 CST/CDN+

5 DON -0.107 0.140 0.361 CST

N-NO3 0.072 0.028 0.692 CST

N-NH4 -0.331 0.231 0.226 CST/CDN-

P-PO4 0.081 0.008 0.838 CST

6 DON 0.105 0.204 0.261 CST

N-NO3 0.058 0.063 0.549 CST

N-NH4 -0.954 0.604 0.023 CDN-

P-PO4 0.012 0.005 0.863 CST

C2 DON -0.151 0.039 0.381 CST

N-NO3 0.019 0.008 0.695 CST

N-NH4 0.171 0.041 0.364 CST

P-PO4 -0.169 0.140 0.087 CST

C4 DON -0.472 0.291 0.014 CDN-

N-NO3 -0.296 0.336 0.007 CDN-

N-NH4 -0.617 0.651 <0.001 CDN-

P-PO4 -0.829 0.535 <0.001 CDN-

C5 DON -0.265 0.177 0.082 CDN-

N-NO3 -0.166 0.059 0.332 CST

N-NH4 -0.320 0.165 0.094 CDN-

Outlier removed P-PO4 -0.496 0.226 0.054 CDN-

Outlier included P-PO4 -0.281 0.031 0.482 CST/CDN-

4.1 Seasonal variability

Seasonal variability of nutrients in streams has been documented before (e.g. Abbott et al.

2018; Van Meter et al. 2019). For systems where no extensive inputs of nutrients or disruption of nutrient cycling is taking place, there is an expectation of finding greater concentrations of nutrients in bioavailable forms during winter (Sponseller et al. 2014). This is often considered to be due to microbial activity and biotic demand being low during the winter so that the nutrients cannot be retained within the soil and are instead flushed out into the waterways. In summer, plant demand (especially for the easily available DIN) retains nutrients more efficiently and hence, lower concentrations are found in water. Yet, this pattern could also reflect seasonal changes in hydrology that may dilute solutes and/or alter the connections with different soil strata in the catchment (Abbott et al. 2018; Van Meter et al. 2019). Such patterns were observed in the Krycklan catchment, where the highest average concentrations were found in winter and autumn for all parameters but for DON (figure 2b;

3b; 4b; 5b). As the DON pool is mostly composed of organic compounds that are not easily available to plants and microbes, the biological control of this form of N is not as strong. This is supported by findings by Hedin et al. (1995), where a vast majority of the N supplied to streams in forested catchments was organic, regardless of the strength of biological retention.

4.2 Differences in seasonality

One observable difference between the two catchments was the spring concentrations. In Röbäcksdalen, average spring concentrations were the highest for all parameters. In Krycklan however, spring concentrations were among the lowest for all parameters but for DON. This is likely the result of dilution from snowmelt in Krycklan, which is also supported by the Q-C analysis (figure 7e-h). Contrary to my predictions, the seasonality of NO3, NH4 and PO4 in Röbäcksdalen was not fundamentally different from what was observed in Krycklan apart from higher spring concentrations. The most evident differences between seasons in Röbäcksdalen, both visually and statistically, were found between spring and summer concentrations, where summer concentrations were much lower. This indicates that either ecological or hydrological processes exert some control on the nutrient flux from land that is similar to what happens in Krycklan. In a recent study on catchments with varying

anthropogenic impact within the Great Lakes Basin in North America, they found that the concentration of NO3 and PO4 in agricultural streams were often in-phase with seasonal hydrological variations, meaning that the annual variation in nutrient concentration co- varied with annual discharge patterns (Van Meter et al. 2019). However, streams draining catchments with higher degrees of urbanization and point source inputs of nutrients (e.g.

waste water treatment plants or nutrient-enriched groundwater input zones), more often showed out-of-phase tendencies, where high concentrations were associated with low flow and the lowest concentrations were found during high flow. Based solely on this study, the streams in Röbäcksdalen are seemingly more affected by agricultural activity than by the nearby urban areas.

Lastly, the seasonality in Röbäcksdalen is seemingly stronger than at Krycklan, at least for NO³. This is contrary to the findings in Van Meter et al. (2019), where agriculture was generally associated with more muted seasonality. The reasons behind this are unclear. It is possible that the soils are top loaded with DIN as a result of fertilizer application and that the connectivity between water and these soils varies throughout the year. Another explanation could be that plant demand retains a lot of the DIN during the growth season. It is also possible that the Krycklan seasonality is in fact greater than what can be seen from the box- plots and averages since these display the combined data from three sites with different properties.

4.3 Implications for nutrient limitation

My results suggest that the activities in Röbäcksdalen appear to be influencing the inputs of N to streams to a greater degree than the inputs of P. Yet, analyzing the DIN:P-PO4 ratio could be considered a blunt tool for assessing nutrient limitation, since findings show that

bioavailability is not restricted to inorganic species and that some fraction of the organic N and P can actually be taken up by bacterioplankton (Soares et al. 2017). In a recent study on four lakes in northern boreal Sweden, 27% of the DON and 36% of the dissolved organic phosphorous (DOP) was found to be bioavailable (Soares et al. 2017). Since DIN makes up such a great portion of the total dissolved nitrogen pool in Röbäcksdalen, the potential portion of bioavailable DON is likely to be small in comparison. Since total dissolved P concentrations are not known, it is more problematic to assume that organic phosphorous does not substantially contribute to the bioavailable P-pool. However, the average molar ratio for Röbäcksdalen shown by this analysis was approximately 288N:1P, which is well above the Redfield ratio of 16N:1P. The variance of ratios was also great, ranging from 23N:1P to

4968N:1P. This indicates that, at least for many of the sampling occasions, there was an excess of N in relation to P. This means that the system is likely not limited by N, but rather by P. This is contrary to the general belief that waters in northern boreal Sweden are limited by N (e.g., Bergström et al. 2013, Burrows et al. 2016). These findings highlight an

overlooked, but important, exception to the general assumption of N-limitation in northern boreal regions.

4.4 The relative role of DON

For undisturbed northern catchments, the prediction of DON being the dominant form of N found in water usually holds. As DON is not readily available to biota, it can leave soils without much retention (Stanley and Maxted 2008). This pattern was also observed at all Krycklan sites (figure 6c). In Röbäcksdalen, the proportion of DON for the majority of samples was below 50%. Only minor differences were found between the two catchments in terms of DON concentration, so the difference in the relative role of DON is likely due to heavily elevated levels of DIN found in Röbäcksdalen dominating the N-pool, especially in spring. These elevated levels of DIN are likely due to the addition of fertilizers containing inorganic N in amounts beyond plant demand.

Even though differences were slight, the Röbäcksdalen DON average concentrations were lower than Krycklan for every season. The reason for this is unknown, but one possible explanation is that the agricultural activity in Röbäcksdalen has reduced the amount of organic matter in the soils responsible for most of the DON supply (Sponseller et al. 2014). It could also be linked to differences in surficial geology. Röbäcksdalen sits on marine

sediments (SLU 2020) whereas most of the Krycklan catchment is above the highest coastline and is dominated by glacial till (Laudon et al. 2013).

4.5 NH

and NO

Prior studies have pointed out mechanisms that cause a trade-off between NO3 and NH4. In systems with oxic conditions and limited external inputs, NH4 concentrations are generally low since NH4 is turned into NO3 by nitrifying bacteria. However, under conditions where the concentration if dissolved oxygen is low in the water that feed streams, this process does not occur which can lead to accumulation of NH⁴ (Helton et al. 2015). Under such anoxic

conditions, NO3 in soil waters is used as a terminal electron acceptor which, in addition to biotic demand, strengthens the demand for NO3 (Helton et al. 2015), and reduces NO3 supply to streams. This pattern was observed in the Krycklan data, where site C4 exhibited

concentrations of N-NH⁴ three times higher than those of N-NO³. This pattern is likely explained by the fact that the Krycklan catchment is patchy and that Site C4 is highly

influenced by wetland (Sponseller et al. 2018). However, this type of trade-off is not found in Röbäcksdalen, where an increase in N-NH⁴ is mostly accompanied by an increase in N-NO³. One possible explanation for this could be that inputs of excess nutrients in the form of fertilizers are masking these mechanisms.

4.6 Concentration-discharge relationships

Analysis of the Q-C relationships suggests that the land use activities in Röbäcksdalen have altered some of the basic relationships expected for catchments within this region.

Specifically, compared to Krycklan streams, Röbäcksdalen sites showed mostly chemostatic tendencies during snowmelt (figure 7a-d). In Krycklan however, a tendency towards dilution was more common (figure 7e-h), especially for Site C4 and C5 (table 2). These relationships have both been found elsewhere (Basu et al. 2010). This indicates that streams affected by agricultural land use in northern boreal regions are likely to display similar tendencies as streams in other regions.

The chemostatic relationships in Röbäcksdalen mean that the measured concentrations remained rather constant with changing flow. What this also means is that the flux of

nutrients from the system is increasing with increasing discharge. For streams like the one at the Krycklan C2 site, which also showed signs of chemostasis, this is not a problem since concentrations are low. For streams where concentrations are elevated for different reasons however, large nutrient fluxes during floods could have effects on downstream lakes or coastal areas.

Chemostasis has been observed for geogenic solutes such as Ca, Mg, Na and Si that are evenly distributed throughout the soil and are released due to weathering (Godsey et al. 2009).

Finding the same pattern for almost all parameters and sites at Röbäcksdalen is somewhat surprising, considering that fertilizer inputs are not continuous. However, studies have shown that both phosphorous and nitrogen can remain as legacies in soils and groundwaters for many years (McCrackin et al. 2018; Van Meter et al. 2018). This is thought to be one possible explanation for the transport limited supply of nutrients to many agricultural streams (Basu et al. 2010). This may also serve as a time lag when it comes to reducing the flux of nutrients to freshwater and marine systems and must be considered when setting goals for mitigation efforts. To truly investigate whether the reason behind the observed chemostatic behavior in Röbäcksdalen is due to legacy nutrients within the soil, it would be interesting to study how, or if, the Q-C relationships change throughout the year.

Snowmelt contributes a lot of water to these streams in spring and could dilute the stream water, turning Q-C relationships chemostatic that would otherwise be enriching

chemodynamic. It could also do the opposite by increasing the connection between water and soils. Since Degernäsbäcken is an artificial stream, purposed to drain both the agricultural fields and nearby rural area, it is likely constructed in such a way that it does not flood regularly. During major hydrological events (such as snowmelt) however, increased contact with surficial soil strata may increase the transport of nutrients to steams. This could be an explanation behind the high spring concentrations.

To find out if any of these (or a combination of both) processes are responsible for the observed chemostasis, studying other (smaller) hydrological events would be interesting. It would, for example, be interesting to see if the low summer concentrations are proportional to summer discharge or if biotic demand is simply retaining DIN more efficiently in summer.

A greater sample size would also be desirable for such an analysis. It is plausible that some of the Q-C slopes that were non-significant chemodynamic could be chemodynamic with greater statistical power. If the tendency towards transport limitation is observed all year around, legacy nutrients in soil or groundwater could be an explanation.

5 Conclusion

Most research on agricultural waters in Sweden has been focused on southern systems. This is to no surprise since, as of today, most agricultural activity is focused in the southern parts of Sweden (SCB 2018). My results highlight the importance of also considering the effects that land use can have on northern boreal streams. Since waters in northern boreal Sweden are usually assumed to be N-limited, the findings from Röbäcksdalen highlight an important exception where high concentrations of bioavailable N are transported from land to water.

How great the nutrient loading on downstream waterbodies is seems to be determined by the magnitude of the flow. Since northern boreal lakes are generally N-limited, great fluxes of dissolved N from streams could cause increased production in these otherwise unproductive lakes. Looking exclusively at N, the effect from agriculture may actually have more drastic effects in northern systems than in the south where P-limitation is more common (Bergström et al. 2005).

The observed domination by DIN in Röbäcksdalen, along with a weakened trade-off between NO3 and NH4, suggests that the supply of these forms of N are likely not mainly controlled by ecological processes but rather by external inputs of inorganic fertilizers and hydrology. The chemostatic, or transport limited, patterns found in Röbäcksdalen during snowmelt also indicate that hydrology rather than soil processes control inputs of dissolved nutrients to the streams. This pattern could be due to a legacy of nutrients stored within the catchment.

However, the transport limitation found during snowmelt could also be associated with increased connectivity with nutrient top loaded soils. To further investigate the reasons behind the observed seasonal patterns, one could study how, or if, the Q-C relationship changes throughout the year and with different magnitudes of flood events.

Since climate change is expected to alter precipitation regimes, changes in discharge and timing of large flows can be expected (Moatar et al. 2017). Over the last decades, there has been observations of increased annual discharge in most of Scandinavia (Saaltink et al. 2014;

Wilson et al. 2010). Future changes in temperature, precipitation regimes and land use have the capability of changing the nutrient loading from northern boreal regions to aquatic and marine systems. Hence, it is of interest to understand the present dynamics to set standards and plan mitigation efforts, not least if legacy nutrients are generating time lags.

6 Acknowledgements

First off, I would like to thank my supervisor Ryan Sponseller who has been giving me a lot of valuable input throughout the writing of this thesis. Your support and never-ending desire to help has been worth a lot. I also want to thank Erik Söderlund and Johanna Wallsten for providing me with data and support. Lastly, my appreciation goes out to Jesper Hassellöv for supporting me on a personal level and for discussing ideas.

7 References

Abbott, B. W., Moatar, F., Gauthier, O., Fovet, O., Antoine, V., and Ragueneau, O. 2018.

Trends and seasonality of river nutrients in agricultural catchments: 18 years of weekly citizen science in France. Science of The Total Environment 624: 845–858. doi:

10.1016/j.scitotenv.2017.12.176

Addinsoft. 2020. XLSTAT statistical and data analysis solution. New York, USA.

https://www.xlstat.com

Basu, N. B., Destouni, G., Jawitz, J. W., Thompson, S. E., Loukinova, N. V., Darracq, A., Zanardo, S., Yaeger, M., Sivapalan, M., Rinaldo, A. and Rao, P. S. C. 2010. Nutrient loads exported from managed catchments reveal emergent biogeochemical

stationarity. Geophysical Research Letters 37(23). doi: 10.1029/2010gl045168

Bergström, A.-K., Faithfull, C., Karlsson, D., and Karlsson, J. 2013. Nitrogen deposition and warming - effects on phytoplankton nutrient limitation in subarctic lakes. Global Change Biology 19(8): 2557–2568. doi: 10.1111/gcb.12234

Blackburn, M., Ledesma, J. L. J., Näsholm, T., Laudon, H., and Sponseller, R. A. 2017.

Evaluating hillslope and riparian contributions to dissolved nitrogen (N) export from a

boreal forest catchment. Journal of Geophysical Research: Biogeosciences 122(2): 324–

339. doi: 10.1002/2016jg003535

Burrows, R. M., Laudon, H., Mckie, B. G., and Sponseller, R. A. 2016. Seasonal resource limitation of heterotrophic biofilms in boreal streams. Limnology and Oceanography 62(1): 164–176. doi: 10.1002/lno.10383

Conley, D. J., Humborg, C., Rahm, L., Savchuk, O. P., and Wulff, F. 2002. Hypoxia in the Baltic Sea and Basin-Scale Changes in Phosphorus Biogeochemistry. Environmental Science & Technology 36(24): 5315–5320. doi: 10.1021/es025763w

FAO. 2017. The future of food and agriculture – Trends and challenges. Rome.

Galloway, J. N., Townsend, A. R., Erisman, J. W., Bekunda, M., Cai, Z., Freney, J. R., Martinelli, L. A., Seitzinger, S. P. and Sutton, M. A. 2008. Transformation of the

Nitrogen Cycle: Recent Trends, Questions, and Potential Solutions. Science 320(5878):

889–892. doi: 10.1126/science.1136674

Godsey, S. E., Kirchner, J. W., and Clow, D. W. 2009. Concentration-discharge relationships reflect chemostatic characteristics of US catchments. Hydrological Processes 23(13):

1844–1864. doi: 10.1002/hyp.7315

Hedin, L. O., Armesto, J. J., and Johnson, A. H. 1995. Patterns of Nutrient Loss from Unpolluted, Old-Growth Temperate Forests: Evaluation of Biogeochemical Theory. Ecology 76(2): 493–509. doi: 10.2307/1941208

Helton, A. M., Ardón, M., and Bernhardt, E. S. 2015. Thermodynamic constraints on the utility of ecological stoichiometry for explaining global biogeochemical patterns. Ecology Letters 18(10): 1049–1056. doi: 10.1111/ele.12487

Hägg, H. E., Humborg, C., Swaney, D. P., and Mörth, C.-M. 2011. Riverine nitrogen export in Swedish catchments dominated by atmospheric inputs. Biogeochemistry 111(1-3): 203–

217. doi: 10.1007/s10533-011-9634-7

Högberg, P., Näsholm, T., Franklin, O., and Högberg, M. N. 2017. Tamm Review: On the nature of the nitrogen limitation to plant growth in Fennoscandian boreal forests. Forest Ecology and Management 403: 161–185. doi: 10.1016/j.foreco.2017.04.045

IPCC. 2019. Summary for Policymakers. In: Climate Change and Land: an IPCC special report on climate change, desertification, land degradation, sustainable land

management, food security, and greenhouse gas fluxes in terrestrial ecosystems [P.R.

Shukla, J. Skea, E. Calvo Buendia, V. Masson-Delmotte, H.- O. Pörtner, D. C. Roberts, P.

Zhai, R. Slade, S. Connors, R. van Diemen, M. Ferrat, E. Haughey, S. Luz, S. Neogi, M.

Pathak, J. Petzold, J. Portugal Pereira, P. Vyas, E. Huntley, K. Kissick, M. Belkacemi, J.

Malley, (eds.)]. In press.

Laudon, H., Taberman, I., Ågren, A., Futter, M., Ottosson-Löfvenius, M., and Bishop, K.

2013. The Krycklan Catchment Study-A flagship infrastructure for hydrology, biogeochemistry, and climate research in the boreal landscape. Water Resources Research 49(10): 7154–7158. doi: 10.1002/wrcr.20520

Loladze, I., and Elser, J. J. 2011. The origins of the Redfield nitrogen-to-phosphorus ratio are in a homoeostatic protein-to-rRNA ratio. Ecology Letters 14(3): 244–250. doi:

10.1111/j.1461-0248.2010.01577.x

McCrackin, M. L., Muller-Karulis, B., Gustafsson, B. G., Howarth, R. W., Humborg, C., Svanbäck, A., and Swaney, D. P. 2018. A century of legacy phosphorus dynamics in a large drainage basin. Global Biogeochemical Cycles 32.

https://doi.org/10.1029/2018GB005914

Meybeck, M. and Moatar, F. 2012. Daily variability of river concentrations and fluxes:

indicators based on the segmentation of the rating curve. Hydrological Processes 26(8):

1188-1207.

Moatar, F., Abbott, B. W., Minaudo, C., Curie, F., and Pinay, G. 2017. Elemental properties, hydrology, and biology interact to shape concentration-discharge curves for carbon, nutrients, sediment, and major ions. Water Resources Research 53(2): 1270–1287. doi:

10.1002/2016wr019635

Peterjohn, W. T., and D. L. Correll. 1984. Nutrient Dynamics in an Agricultural Watershed:

Observations on the Role of A Riparian Forest. Ecology 65(5): 1466–1475.

doi:10.2307/1939127.

Ptacnik, R., Andersen, T., and Tamminen, T. 2010. Performance of the Redfield Ratio and a Family of Nutrient Limitation Indicators as Thresholds for Phytoplankton N vs. P Limitation. Ecosystems 13(8): 1201–1214. doi: 10.1007/s10021-010-9380-z

Redfield, A. C. 1934. On the proportions of organic, derivatives in sea water and their relation to the composition of plankton. James Johnstone Memorial Volume: 176–192.

Saaltink, R., Velde, Y. V. D., Dekker, S. C., Lyon, S. W., and Dahlke, H. E. 2014. Societal, land cover and climatic controls on river nutrient flows into the Baltic Sea. Journal of

Hydrology: Regional Studies 1: 44–56. doi: 10.1016/j.ejrh.2014.06.001

Sandström, S., Futter, M. N., Kyllmar, K., Bishop, K., Oconnell, D. W., and Djodjic, F. 2020.

Particulate phosphorus and suspended solids losses from small agricultural catchments:

Links to stream and catchment characteristics. Science of The Total Environment 711.

doi: 10.1016/j.scitotenv.2019.134616

SCB. 2018. Agricultural statistics 2018 including food statistics - tables.

https://www.scb.se/contentassets/0a1612046e24433ebbdcea80aceb8ae0/jo1901_2017 a01_br_jo02br1801.pdf (Accessed 2020-05-02)

SLU. 2020a. SITES Röbäcksdalen. https://www.slu.se/institutioner/norrlandsk-

jordbruksvetenskap/infrastruktur/sites-robacksdalen/mer-om-sites/ (Accessed 2020- 05-28)

SLU. 2020b. Fosfatfosfor, PO4-P. https://www.slu.se/institutioner/vatten-

miljo/laboratorier/vattenkemiska-laboratoriet/detaljerade-metodbeskrivningar/fosfat/

(Accessed 2020-05-29)

Smith, V., Tilman, G., and Nekola, J. 1999. Eutrophication: impacts of excess nutrient inputs on freshwater, marine, and terrestrial ecosystems. Environmental Pollution 100(1-3):

179–196. doi: 10.1016/s0269-7491(99)00091-3

Soares, A. R. A., Bergström, A.-K., Sponseller, R. A., Moberg, J. M., Giesler, R., Kritzberg, E.

S., Jansson, M. And Berggren, M. 2017. New insights on resource stoichiometry:

assessing availability of carbon, nitrogen, and phosphorus to

bacterioplankton. Biogeosciences 14(6): 1527–1539. doi: 10.5194/bg-14-1527-2017 Sponseller, R. A., Temnerud, J., Bishop, K., and Laudon, H. 2014. Patterns and drivers of

riverine nitrogen (N) across alpine, subarctic, and boreal

Sweden. Biogeochemistry 120(1-3): 105–120. doi: 10.1007/s10533-014-9984-z

Stanley, E. H., and Maxted, J. T. 2008. Changes In The Dissolved Nitrogen Pool Across Land Cover Gradients In Wisconsin Streams. Ecological Applications 18(7): 1579–1590. doi:

10.1890/07-1379.1

Sterner, R. W., and J. J. Elser. 2002. Ecological Stoichiometry: the Biology of Elements from Molecules to the Biosphere. Princeton: Princeton University Press.

Sukenik, A., Quesada, A., and Salmaso, N. 2015. Global expansion of toxic and non-toxic cyanobacteria: effect on ecosystem functioning. Biodiversity and Conservation 24(4):

889–908. doi: 10.1007/s10531-015-0905-9

Söderlund, E. 2019. Snowmelt flushing of dissolved organic carbon (DOC) from urban boreal streams: A study of stream chemistry in Degernäsbäcken and

Röbäcken. Bachelor’s thesis. Department of Ecology and Environmental Science, Umeå

University, Sweden. http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-163172 (Accessed 2020-03-28)

Teutschbein, C., Sponseller, R. A., Grabs, T., Blackburn, M., Boyer, E. W., Hytteborn, J. K.

and Bishop, K. 2017. Future Riverine Inorganic Nitrogen Load to the Baltic Sea From Sweden: An Ensemble Approach to Assessing Climate Change Effects. Global

Biogeochemical Cycles 31(11): 1674–1701. doi: 10.1002/2016gb005598

Thomas, Z., Abbott, B. W., Troccaz, O., Baudry, J., and Pinay, G. 2016. Proximate and ultimate controls on carbon and nutrient dynamics of small agricultural

catchments. Biogeosciences 13(6): 1863–1875. doi: 10.5194/bg-13-1863-2016 United Nations, Department of Economic and Social Affairs, Population Division. 2019.

World Population Prospects 2019: Highlights (ST/ESA/SER.A/423).

Van Meter, K. J., Cappellen, P. V., and Basu, N. B. 2018. Legacy nitrogen may prevent achievement of water quality goals in the Gulf of Mexico. Science 360(6387): 427–430.

doi: 10.1126/science.aar4462

Van Meter, K. J., Chowdhury, S., Byrnes, D. K., and Basu, N. B. 2019. Biogeochemical asynchrony: Ecosystem drivers of seasonal concentration regimes across the Great Lakes Basin. Limnology and Oceanography 65(4): 848–862. doi: 10.1002/lno.11353

Vitousek, P. M., and Reiners, W. A. 1975. Ecosystem Succession and Nutrient Retention: A Hypothesis. BioScience 25(6): 376–381. doi: 10.2307/1297148

Wilson, D., Hisdal, H., and Lawrence, D. 2010. Has streamflow changed in the Nordic countries? – Recent trends and comparisons to hydrological projections. Journal of Hydrology 394(3-4): 334–346. doi: 10.1016/j.jhydrol.2010.09.010