Variations and controls of nitrogen stable isotopes in particulate ...

Oecologia (2009) 160:421–431 DOI 10.1007/s00442-009-1323-z

C O N C E P T S , R E V I E WS A N D S Y N T H E S E S

Variations and controls of nitrogen stable isotopes in particulate organic matter of lakes B. Gu

Received: 11 July 2008 / Accepted: 27 February 2009 / Published online: 8 April 2009 © Springer-Verlag 2009

Abstract Nitrogen stable isotope (
15N) data of particulate organic matter (POM) from the literature were analyzed to provide an understanding of the variations and controls of
15NPOM in lakes at the global scale. The
15NPOM variability characterized by seasonal mean, minimum, maximum, and amplitude (deWned as
15NPOM maximum ¡
15NPOM minimum) from 36 lakes with seasonal data did not change systematically with latitude, but was signiWcantly lower in small lakes than in large lakes. The seasonal mean
15NPOM increased from oligotrophic lakes to eutrophic lakes despite large variations that are attributed to the occurrences of nitrogen Wxation across the trophic gradient and the diVerences in
15N of dissolved inorganic nitrogen (DIN) in individual lakes. Seasonal mean
15NPOM was signiWcantly correlated with DIN concentration and
15NDIN in two subsets of lakes. Seasonal minimum
15NPOM in individual lakes is inXuenced by nitrogen Wxation and
15NDIN while seasonal maximum
15NPOM is inXuenced by lake trophic state and
15NDIN. As a result of the dominance of non-living POM in the unproductive surface waters, seasonal
15NPOM amplitude was small

Communicated by Ulrich Sommer. Electronic supplementary material The online version of this article (doi:10.1007/s00442-009-1323-z) contains supplementary material, which is available to authorized users. B. Gu Fisheries College, Guangdong Ocean University, 40 East Jiefang Road, 524088 Zhanjiang, Guangdong, China Present Address: B. Gu (&) Everglades Division, South Florida Water Management District, 3301 Gun Club Road, West Palm Beach, FL 33406, USA e-mail: [email protected]

(mean = 4.2‰) in oligotrophic lakes of all latitudes. On the other hand, seasonal
15NPOM amplitude in eutrophic lakes was large (mean = 10.3‰), and increased from low to high latitudes, suggesting that the seasonal variability of
15N in the phytoplankton-dominated POM pool was elevated by the greater spans of solar radiation and thermal regimes at high latitudes. The
15NPOM from 42 lakes with no seasonal data revealed no consistent patterns along latitude, lake area, and trophic gradients, and a greater than 2‰ depletion compared to the lakes with seasonal data. Along with the large seasonal variability of
15NPOM within lakes, these results provide insightful information on sampling design for the studies of food web baseline in lakes. Keywords
15N · Nitrogen Wxation · Seasonal variability · Trophic state

Introduction Aquatic ecologists have used nitrogen stable isotopes (
15N) of particulate organic matter (POM) to track the sources and sinks of various nitrogen compounds, major biogeochemical processes (Hodell and Schelske 1998; Ostrom et al. 1998; Gu et al. 2006), and to establish trophic baselines for pelagic food webs in lakes (Kling et al. 1992; Gu et al. 1994; Cabana and Rasmussen 1996). Although many studies have reported that
15NPOM varies greatly over production cycles (Gu et al. 1994; McCusker et al. 1999; Syväranta et al. 2008), fewer studies have compared the variations in
15NPOM among lakes of diVerent latitudes, morphometries and trophic states. Gu et al. (1996) found a non-linear relationship between total phosphorus (TP) concentration and planktonic
15N in ten Florida lakes with diVering trophic states. Patoine et al. (2006) reported a

123

422

close relationship between
15NPOM and the abundance of N2-Wxing cyanobacteria in six lakes of the northern Great Plains. Data from these studies conducted on a limited number of lakes are not suYcient to reveal general patterns and controls of
15NPOM to be representative on a global scale. The
15NPOM in lakes is inXuenced by the source of organic matter in the surface water, phytoplankton species composition and growth rate, concentration, form, and isotopic composition of dissolved inorganic nitrogen (DIN). Phytoplankton growth rate is aVected by seasonal changes in solar radiation, water temperature, and nutrients. Major nitrogen cycling processes such as nitrogen Wxation, nitriWcation, and denitriWcation have severe impacts on the isotopic fractionation during microbial nitrogen transformation and phytoplankton assimilation (Owens 1987; Hadas et al. 2009). Some of these processes frequently act synergistically or against each other to inXuence
15NPOM, making data interpretation especially diYcult. Particulate organic matter and dissolved inorganic nutrients in the surface water of lakes originate from autochthonous and allochthonous sources, each of which often possesses distinct
15N. Lakes surrounded by watersheds with strong human inXuences may receive more contributions from agriculture and domestic discharges than lakes in the remote areas (Carpenter et al. 1998; Anderson and Cabana 2005). Therefore, lakes of diVerent geographical locations and morphometries may be diVerent in the sources and isotopic composition of particulate and dissolved matter that can profoundly inXuence the
15NPOM. It has been recognized that the nitrogen cycling processes play pivotal roles in controlling
15NPOM. NitrogenWxing cyanobacteria do not have signiWcant fractionation against 15N during the uptake of atmospheric N2 (Hoering and Ford 1960; Delwiche and Steyn 1970) that is the international reference for
15N determinations. Therefore, the
15N of POM dominated by N2-Wxing cyanobacteria is close to 0‰. DenitriWcation often causes signiWcant isotopic fractionation (Owens 1987), leading to 15N enrichment in the residual nitrate pool. Phytoplankton utilizing this pool will be enriched in 15N (Gu 1993; Hodell and Schelske 1998; Hadas et al. 2009). In well-oxygenated waters, nitriWcation converts isotopically light ammonium to nitrite and nitrate. Depending on which form of nitrogen is used, phytoplankton may either be enriched or depleted in 15N (Lehmann et al. 2004; Syväranta et al. 2008). Lake productivity is a fundamental factor controlling
15NPOM in surface waters. It has been known that fastgrowing phytoplankton has less isotopic fractionation during nutrient assimilation (Wada and Hattori 1978; Owens 1987) and may deplete the nutrient pool, leading to further 15 N enrichment (Peterson and Fry 1987). However, a linear relationship between lake trophic state and
15NPOM is not

123

Oecologia (2009) 160:421–431

always observable because N2-Wxing cyanobacteria that often dominate the eutrophic lakes is depleted in15N (Estep and Vigg 1985; Gu et al. 1996, 2006). The
15NPOM data have been used to establish trophic baselines to assess the feeding relationship between consumers and preys, and energy Xow pathways that support the lake pelagic food webs (Kling et al. 1992; Gu et al. 1994; Bootsma et al. 1996). Many previous studies relied on
15NPOM from a few samples, which may potentially misinterpret the true trophic links and main energy sources fueling the aquatic ecosystems. A better understanding of the pattern and magnitude of seasonal variability of
15NPOM is a must before stable isotope technique can be used in food web analysis with conWdence. In this review, I analyzed cross-system variations of
15NPOM in 36 lakes with seasonal data along a wide range of latitudinal, morphometric, and trophic gradients. Prior to this analysis, I hypothesize that increasing trophic state would increase
15NPOM with the exception of lakes dominated by N2-Wxng cyanobacteria. Increasing trophic state would also increase seasonal variability of
15NPOM and the highest variability would be found in high latitude eutrophic lakes where extreme environmental forcing often has the greatest seasonal impacts on lake primary productivity. In addition, 42 lakes without seasonal
15NPOM data, i.e., with data from one to two sampling events were compared with the 36 lakes with data from multiple sampling events to evaluate the impacts of sampling frequency on
15NPOM distribution.

Materials and methods Data on
15NPOM were assembled from published studies on lakes from around the globe. In most cases, only isotope data labeled as being from POM were selected, but occasionally isotope data labeled as phytoplankton, plankton, or bulk plankton were also used when POM data were not speciWed. Examinations of the sampling methods from these studies revealed that these samples were actually suspended particles or POM. Along with
15NPOM, data on
15N of DIN, latitude, surface area, mean depth, pH, TP, total nitrogen (TN), DIN (NH4+ + NO3¡ + NO2¡), and chlorophyll a (Chl a) were recorded. Since not all the data were always reported from a single study, additional publications or database were searched to Wnd the missing data. Despite this eVort, missing data are unavoidable in this dataset. I found 78 lakes with
15NPOM, one to two morphometric and one to Wve biogeochemical variables from open literature. Seasonal
15NPOM data from multiple sampling events, typically bimonthly or monthly covering the entire or a major part of the open period were available in 36 lakes.

Oecologia (2009) 160:421–431 15

n=7

n = 21

10

n=8 Seasonal Non-seasonal

AB

A

B

AB AB

5

0 n = 12

n = 18

n = 12

-5 Low

Mid

High

Latitude 20

b

n = 16

n = 11

n=9

15

15

10

B

B A

A

A

A

5 0 n = 23

-5

n = 12

1-100

Surface area (km2) 16

Results

a

AB

δ NPOM (‰)

These data were used to calculate seasonal mean, maximum, minimum, and amplitude for
15NPOM. Seasonal
15NPOM amplitude is deWned as the diVerence between maximum
15NPOM and minimum
15NPOM. These parameters were used, along with latitude, morphometry, and trophic state indices, to assess the patterns and controls of
15NPOM in these lakes. Because missing data for latitude, morphometry, nutrients, and Chl a concentrations were common in this dataset, geographic, morphometric, and trophic state classiWcations (Wetzel 2001), instead of the respective numerical values, are used to assess the patterns of
15NPOM among lakes. In addition to the 36 lakes with seasonal data, I also found 42 lakes with
15NPOM data from one to two sampling events per lake. These data reXect major biogeochemical processes in the surface waters at a very short time-scale and were analyzed separately in association with the environmental variables. The
15NPOM from the 36 and the 42 lakes were compared to assess the potential impacts of sampling design on
15NPOM. Statistics were used to detect diVerences among sample means, and to perform correlation analysis using SPSS software (Version 10). Non-parametric analyses (MannWhitney test for two sample comparison and Spearman’s  correlation for correlation) were used when sample size was small. Statistics were considered signiWcant at P < 0.05.

423

c

n = 13

n=5

12

n = 18

AB B

The lakes in both datasets are situated along a wide latitudinal gradient from tropical to arctic, but with a majority in the temperate region (Fig. 1a). Stable isotope values and selected environmental variables varied widely among lakes (Tables 1, 2). The mean and median for lake area, mean depth, TP, and Chl a concentrations diVered greatly due to the inXuences of a small number of large lakes and extremely high trophic state values in some lakes. The seasonal mean
15NPOM from the 36 lakes varied from –0.5 to 13.3‰, while the
15NPOM from the 42 lakes varied from –2.5 to 11.6‰. These large variations are consistent with the large diVerences in morphometry and trophic state found among lakes at the global scale.

8

A

AC

4

A C

0 -4

n = 13

n=8

Oligotrophic Mesotrophic

n = 21 Eutrophic

Trophic state

Seasonal variability of
15NPOM

Fig. 1 Box plots of a latitude (Low < 10°S · 30°N, Mid > 30°N < 50°N, and High ¸ 50°N), b surface area, c trophic state and
15NPOM for study lakes with seasonal mean (open boxes, n = 36) and non-seasonal (shaded boxes, n = 42) data. Box boundaries indicate the 25th and 75th percentiles; whiskers the 10th and 90th percentiles; the inner horizontal line is the median; circles indicate the 5th and 95th percentiles. ox bars with diVerent letters indicate statistical diVerence in
15NPOM (Mann–Whitney test, P < 0.05)

The mean
15NPOM for the 36 and 42 study lakes averaged 6.0 and 3.8‰, respectively. Both datasets displayed slight increases from low to high latitude, but these increases are not signiWcant (Fig. 1a). The
15NPOM in small lakes (·1 km2) were 3.1 and 3.7‰ more depleted than in the mid-size and large lakes in the 36 lake dataset, respectively, while no diVerence along the area gradient for the 42 lake

dataset is found (Fig. 1b). The seasonal mean
15NPOM increased from 4.6‰ in the oligotrophic lakes to 7.3‰ in eutrophic lakes (Fig. 1c). This diVerence was statistically signiWcant despite large variability within each trophic class. The highest seasonal mean
15NPOM (13‰) was reported in Lake Jyväsjärvi which is a mesotrophic–eutrophic

123

424

Oecologia (2009) 160:421–431

Table 1 Descriptive statistics of selected morphometric and biogeochemical variables, and stable nitrogen isotopes of particulate organic matter (POM) and dissolved inorganic nitrogen (DIN) in 36 study lakes with seasonal data Variable


15NPOM (‰) APT

Area (km2)

Mean

Max

Mean

6.0

10.0

2.3

7.7

5.6

5,122

Median

6.8

9.2

2.0

6.0

4.8

15

SD

Min


15NDIN (‰)

3.4

5.8

2.9

5.6

3.9

15,697

Minimum

¡0.5

2.0

¡2.9

1.4

1.1

0.04

Maximum

13.3

27.0

10.0

25.0

12.1

68,800

Count

36

36

36

36

12

36

Mean depth (m)

pH (SU)

TP (
g L¡1)

TN (mg L¡1)

DIN (
g L¡1)

TN:TP (Molar ratio)

Chl a (
g L¡1)

40

7.9

71.6

1.23

775

62

28

8

8.3

33.8

1.08

398

36

13

101

1

72.4

1.11

973

74

33

1

1

5

0.08

24.9

12

1

299

128

23

32

570

9.1

280

4.70

3,690

33

32

26

24

23

APT: seasonal amplitude for
NPOM 15

Table 2 Descriptive statistics of selected morphometric and biogeochemical variables, and stable nitrogen isotopes of particulate organic matter (POM) in 42 study lakes with data from 1 to 2 sampling events
15NPOM (‰)

Area (km2)

Mean

3.8

5,218

Median

3.1

11

SD

2.9

16,786

Variable

Minimum

¡2.5

0.01

Maximum

11.6

82,367

Count

42

29

Mean depth (m)

pH (SU)

TP (
g L¡1)

TN (mg L¡1)

49.5

7.6

56.5

1.25

738

51

2.2

7.9

23.0

0.99

398

58

8

148.8

1.0

85.1

1.09

957

31

70

TN:TP (Molar ratio)

Chl a (
g L¡1)

0.9

35

5.5

3.1

0.01

25

1

1

740

9.6

440

3.61

3,690

95

270

28

37

36

18

23

15

25

lake (Syväranta et al. 2008), while the lowest seasonal mean
15NPOM (¡0.5‰) was reported in the oligotrophic African Great Lake, Lake Tanganyika (Sarvala et al. 2003). Depleted
15NPOM ( 30°N < 50°N, and High ¸ 50°N), b surface area, c trophic state and seasonal minimum
15NPOM (n = 36). Box boundaries indicate the 25th and 75th percentiles; whiskers the 10th and 90th percentiles; the inner horizontal line is the median; circles indicate the 5th and 95th percentiles. ox bars with diVerent letters indicate statistical diVerence in
15NPOM (Mann–Whitney test, P < 0.05)

Concentration and
15N of DIN, and
15NPOM I found 23 lakes with DIN concentrations from the 36 lakes with seasonal data. A plot of DIN concentration and
15NPOM revealed a positive relationship despite of large

0 Oligotrophic Mesotrophic

Eutrophic

Trophic state Fig. 3 Box plots of a latitude (Low < 10°S · 30°N, Mid > 30°N < 50°N, and High ¸ 50°N), b surface area, c trophic state and seasonal maximum
15NPOM (n = 36). Box boundaries indicate the 25th and 75th percentiles; whiskers the 10th and 90th percentiles; the inner horizontal line is the median; circles indicate the 5th and 95th percentiles. ox bars with diVerent letters indicate statistical diVerence in
15NPOM (Mann–Whitney test, P < 0.05)

isotopic variability along the nutrient gradient (Fig. 5). I also found 12 lakes with
15N data of ammonium, nitrate or DIN from the seasonal dataset. Because these data are not suYcient for correlation analysis when used separately, isotope data for ammonium and nitrate from the same lake were averaged and reported as
15NDIN here. This might be

123

426

Oecologia (2009) 160:421–431 25

a

14

n=7

n = 22

n=7 12

20 A

A

10

A

δ15NPOM (‰)

10

15

8 6 4

5

2

0 Low

Mid

High

0

Latitude

b

n=9

n = 16

n = 11

25 20 15

A A

10

1000

10000

Fig. 5 Plot of dissolved inorganic nitrogen (DIN) concentration and seasonal mean
15NPOM for study lakes with seasonal data. Data for both DIN concentrations and
15NPOM are available for 23 lakes from literature. The regression equation is
15NPOM = 2.60DIN–0.39 (R2 = 0.27). Note logarithmical scale is used in x-axis

A 14

5

12

0 >1-100

Surface area (km2)

c

n = 13

n=5

n = 18

20

10

δ15NPOM (‰)

30°N < 50°N, and High ¸ 50°N), b surface area, c trophic state, and seasonal
15NPOM amplitude (n = 36). Box boundaries indicate the 25th and 75th percentiles; whiskers the 10th and 90th percentiles; the inner horizontal line is the median; circles indicate the 5th and 95th percentiles. ox bars with diVerent letters indicate statistical diVerence in
15NPOM (Mann–Whitney test, P < 0.05)

a bias because ammonium and nitrate concentrations in the surface waters are often diVerent. The
15NDIN varied from 1.1 to 12.1‰, with a mean and a range similar to those of
15NPOM (Table 1). A plot of
15NDIN versus
15NPOM revealed a positive relationship with an intercept of 3‰ (Fig. 6).

123

Discussion Controls of seasonal mean
15NPOM This literature analysis revealed higher seasonal mean
15NPOM in productive lakes than in the unproductive lakes, consistent with theoretical consideration (Peterson and Fry 1987; Goericke et al. 1994) and laboratory culture experiments (Wada and Hattori 1978; Montoya and McCarthy 1995; Waser et al. 1998) which demonstrate decreases in 15 N fractionation with increases in the uptake rate of nitrogenous nutrients. This suggests that
15NPOM is indicative of lake trophic state. However, large variations in
15NPOM are

Oecologia (2009) 160:421–431

shown in each trophic class, indicating that the
15NPOM variation along the trophic gradient is also inXuenced by other processes. In fact, a linear relationship between trophic state and
15NPOM has not been reported in previous studies of individual lakes (Lehmann et al. 2004; Gu et al. 2006). Gu et al. (1996) found that planktonic
15N increased as Chl a concentration increased from oligotrophic to eutrophic lakes, but decreased in hypereutrophic lakes, possibly a result of the increasing dominance of N2-Wxing cyanobacteria. A previous study indicated that the
15N of Anabaena cylindrica did not change with the changes of algal growth rate in a pure culture (Minagawa and Wada 1986). This implies that seasonal increases in primary productivity in lakes dominated by N2-Wxing cyanobacteria do not lead to increases in
15NPOM. The occurrence of N2-Wxing cyanobacterial blooms have been reported in many eutrophic lakes from this dataset. For example, Gu and Alexander (1993) reported 15N depletion in A. Xos-aquae in Smith Lake, Alaska, during the spring bloom period. Gu et al. (2006) attributed the low
15NPOM to an annual cyanobacterial bloom (Cylindrospermopsis sp.) in Lake Wauberg, Florida. Vuorio et al. (2006) isolated individual algae from three eutrophic lakes in Finland and found low
15N values (¡0.5 to 2.0‰) in the N2-Wxing cyanobacteria Anabaena spp. and Aphanizomenon spp. The occurrences of N2-Wxing cyanobacteria were also reported in several other eutrophic lakes in this dataset (Chapman and Schelske 1997; Havens et al. 1998). Nitrogen Wxation also contributes to the 15N depletion in some oligotrophic lakes such as Lakes Michigan (MacGregor et al. 2001), Pyhäjärvi (Vuorio et al. 2006), Mekkojärvi (Jäntti 2007), and Malawi (Gondwe et al. 2008) where
15NPOM was low (¡0.2 to 2.0‰). Thus, the 15N depletion in these oligotrophic lakes is likely the combined result of low productivity and high N2 Wxation. Similarly, the 15N measurements of suspended particles suggested a major contribution by N2 Wxation to the nitrogen budget of the oligotrophic North Atlantic Ocean (Montoya et al. 2002). The large variation in
15NPOM along the trophic gradient may reXect in part the diVerences in isotopic composition and the form of DIN utilized across lakes. The close relationship between
15NDIN and
15NPOM in the small dataset from this analysis suggests that the variation in
15NPOM was controlled partly by the isotope baseline of DIN in each lake. The
15NDIN is aVected by its source and major biogeochemical processes such as nitriWcation and denitriWcation in lakes. For example, Ostrom et al. (1998) revealed that the isotopically depleted nitrate (¡4‰) that was derived from atmospheric deposit led to low
15NPOM in Lake Superior. Wang et al. (2008) found low and high
15NPOM in a Chinese lake from the spring and winter period, respectively, which were inXuenced by isotopically light and heavy nitrate from in situ nitriWcation in spring

427

and industrial wastewater discharge in winter. Syväranta et al. (2008) suggested that the enriched
15NPOM in the Finnish Lake Jyväsjärvi was due to the uptake of ammonium that was isotopically elevated by the preferential removal of 14N during nitriWcation. Gu (1993) attributed the high
15NPOM during the spring turnover period to the uptake of nitrate with enriched 15N due to denitriWcation in Smith Lake, Alaska. It is evident that the large variation in
15NPOM along the trophic gradient is inXuenced by multiple factors including primary productivity, nitrogen Wxation, sources, and isotope composition of DIN (Hadas et al. 2009). Great attention must be paid when
15NPOM is used as an indicator for lake trophic state. InXuence of lake area Lake size is an important factor determining the concentrations of nutrients and organic matter that in turn aVect the isotope composition of POM. Post (2002) found a positive relationship between lake area and
13C of a benthic primary consumer that Wlters POM from the water column. He attributed this relationship to more allochthonous contribution in small lakes than large lakes. Bade et al. (2004) found a decrease in
13C of dissolved inorganic carbon with a decrease in lake area. This is due to the relative increase in terrestrial inputs of organic matter as the surface area of lakes decreases. In this analysis, large lakes (>1 km2) are signiWcantly more enriched in
15NPOM than small lakes. Because TP concentration decreased with the increase in lake area (Spearman’s  correlation, P < 0.05, n = 26), a reduction of nutrient loading may lead to the decreases in isotope fractionation in large lakes. On the other hand, high TP concentration in small lakes may cause nitrogen limitation and increasing dominance of N2-Wxing cyanobacteria (Gu et al. 1994, 2006; Jäntti 2007). Controls of seasonal minimum and maximum
15NPOM The seasonal minimum
15NPOM reXects the greatest isotope depletion in individual lakes over a production cycle. The similar seasonal minimum (»2‰) for lakes of diVerent trophic states may be attributed to the occurrence of nitrogen Wxation in lakes across the trophic gradient. In many lakes, nitrogen Wxation takes place during a certain time period of the production cycle, causing temporary 15N depletion in POM. For example, low
15NPOM associated with A. Xos-aquae blooms was observed in Smith Lake, Alaska, during the brief spring period (Gu et al. 1994). Lehmann et al. (2004) also attributed the low
15N in June to nitrogen Wxation by Aphanizomenon Xos-aquae in Lake Laguna. Patoine et al. (2006) studied the contribution of nitrogen Wxation in six lakes of the northern Great Plains and found that
15NPOM decreased with increasing

123

428

cyanobacterial abundance as season progressed. Another cause for 15N depletion is the uptake of isotopically light DIN which may be derived from external loading (Ostrom et al. 1998; Wang et al. 2008) or in situ nitriWcation (Syväranta et al. 2008; Wang et al. 2008; Hadas et al. 2009). The
15NPOM in approximately 50% of the lakes with no seasonal data is less than 3‰ which is likely due to nitrogen Wxation. In fact, N2-Wxing cyanobacterial blooms have been reported from several lakes in this dataset (Estep and Vigg 1985; Chapman and Schelske 1997; Havens et al. 1998; Gondwe et al. 2008). The seasonal
15NPOM maximum represents the highest isotope enrichment in individual lakes that is inXuenced by at least two major biogeochemical processes, i.e., primary productivity and
15NDIN. Productivity-induced 15N enrichment is evidenced by the stepwise increments in 15N along the trophic classes (Fig. 3c). However, the highest seasonal
15NPOM maximum (27‰) is not found in the most productive lakes, but in a meso-eutrophic lake (Syväranta et al. 2008). Recently, Hadas et al. (2009) reported 15–30‰ jump in
15NPOM in the mesotrophic Lake Kinneret. Several oligotrophic lakes (e.g., Lake Ontario and Lake Loch Ness) were also highly enriched with 15N. The use of 15 N-enriched DIN is likely the cause for isotope enrichments in these lakes. The seasonal
15NPOM maximum was 19 and 20‰ for Lake Baldeggersee and Lake Lugano, respectively. The two lakes are eutrophic, but do not possess the highest productivity based on their TP concentrations. It is likely that the high
15NPOM in these lakes were inXuenced by both high productivity and the 15N-enriched DIN.

Oecologia (2009) 160:421–431

seasonal variability of temperature and solar radiation increases with the increase in latitude, which aVects the growth rate and isotope fractionation in phytoplankton. This hypothesis is conWrmed when data for seasonal
15NPOM amplitude in eutrophic lakes are grouped based on their latitudinal locations. The seasonal variability is small in oligotrophic lakes regardless their geophysical origin (Fig. 7). The highest variability of
15NPOM occurs in the high latitude eutrophic lakes as a combined result of high phytoplankton biomass in POM that is sensitive to environmental changes and the greater seasonal Xuctuations in the external physical drivers. InXuence of nitrogen concentration he variations in
15N of suspended and sedimented organic matter have been related to surface water nitrate concentration and utilization in lakes (Owen et al. 1999; Teranes and Bernasconi 2000; Lehmann et al. 2004). In fact, several previous studies that are included in this analysis display an inverse relationship between DIN concentration and
15NPOM in individual lakes (e.g., Gu 1993; Owen et al. 1999; Teranes and Bernasconi 2000). Other studies failed to Wnd a consistent relationship (Lehmann et al. 2004; Gu et al. 2006). By contrast, my cross-lake comparison revealed increasing trend of
15NPOM with increasing DIN concentrations (Fig. 5). Most successful studies using
15N as a proxy of surface water nitrogen concentration and utilization have been conducted in the marine environment where nitrogen is a limiting nutrient. In many freshwater lakes, phosphorus, not

InXuences of trophic state and latitude on seasonal
15NPOM amplitude

123

n = 12

n=5

n=4

20

n=7 B

AB 16 AB

12

15

Seasonal δ NPOM amplitude (‰)

The seasonal
15NPOM amplitude is a measure of biogeochemical variability in lakes and changes with lake’s trophic state and geographical settings. Greater seasonal
15NPOM amplitude is found in eutrophic lakes than in oligotrophic lakes from this dataset. This is because POM in eutrophic lakes is dominated by phytoplankton that is sensitive to environmental changes. Conversely, low variability is found in oligotrophic lakes due to high proportion of non-living organic matter in POM. Unless the source of non-living organic matter changes seasonally, its isotope ratio should remain fairly consistent over time. This is supported by the small seasonal
15NPOM amplitude (1.5– 8.5‰; mean = 4.2‰) in oligotrophic lakes as compared to the large variability (2.6–25.0‰; mean = 10.3‰) in eutrophic lakes (Fig. 4c). The seasonal and non-seasonal
15NPOM for oligotrophic lakes also share very similar averages (Fig. 1c). Based on these observations, one may expect that seasonal
15NPOM amplitude in eutrophic lakes will increase along the latitude gradient. This is because

24

8 A 4

0 Oligotrophic lakes

Low

Mid

High

Eutrophic lakes

Fig. 7 Box plot of seasonal
15NPOM amplitude for oligotrophic lakes of diVerent latitudes and eutrophic lakes grouped by latitudinal regions (Low < 10°S · 30°N, Mid > 30°N < 50°N, and High ¸ 50°N). Box boundaries indicate the 25th and 75th percentiles; whiskers the 10th and 90th percentiles; the inner horizontal line is the median; circles indicate the 5th and 95th percentiles. ox bars with diVerent letters indicate statistical diVerence in
15NPOM (Mann–Whitney test, P < 0.05)

Oecologia (2009) 160:421–431

nitrogen limits primary productivity (Goldman et al. 1990). Phosphorus limitation in the study lakes is indicated by the high average molar TN/TP ratio (Tables 1, 2). As a result, the increase in
15NPOM along the trophic gradient can be attributed to increasing primary productivity driven by phosphorus. The concomitant increase in DIN concentration with
15NPOM may be considered as increasing nutrient supplies needed to sustain the primary productivity. A recent review suggests that freshwaters could also be limited by both phosphorus and nitrogen (Elser et al. 2007). Under the colimitation scenario, increases in nitrogen supply may stimulate phytoplankton growth and hence reduce 15 N fractionation. However, in lakes with cyanobacterial presence, nitrogen limitation often triggers nitrogen Wxation, leading to 15N depletion in POM. Therefore, the use of
15N of suspended or sedimented POM to predict present or past surface-water nutrient concentrations in lakes must be conducted with other pertinent data such as algal species composition and N:P ratio. InXuence of sampling frequency on foodweb baseline estimate Many earlier investigations of the consumer-diet relationship for the pelagic waters were based on a single to a few POM samples. The consistent
15NPOM depletion in lakes with no seasonal sampling as compared to lakes with seasonal sampling pointed to the necessity of a sampling design that must capture the seasonal variability of
15NPOM. The mean seasonal
15NPOM amplitude in the 36 study lakes is 7.7‰ with a maximum as high as 25‰ (Table 1), indicating that any sample taken in a given time only represents a snapshot of the large seasonal variability in
15NPOM (Gu et al. 1994; Yoshioka et al. 1994; Syväranta et al. 2008). Obtaining samples that cover the seasonal
15NPOM cycle during the growth season is critical considering the large seasonal variability and the small isotope enrichment (average = 3.4‰) between consumers and their prey (Post 2002). An average of greater than 2‰ depletion in POM in lakes with no seasonal sampling can obscure the trophic connection between the planktivores and POM. The discrepancy was small (0.6‰) in oligotrophic lakes, but increased dramatically in mesotrophic (3.8‰) and eutrophic lakes (2.6‰) (Fig. 1c). The seasonal variability also increases with latitude. Therefore, intensive sampling, such as bimonthly or monthly sampling, during the open period is desirable for mesotrophic and eutrophic lakes, and especially eutrophic lakes in the high latitude region. Sources of additional variability in
15NPOM The large variation in
15NPOM is attributed to the complexity of nitrogen cycling pathways and the lack of several

429

important variables for the study lakes. These variables include
15NDIN, sources of the external inputs, and the rates of nitrogen Wxation and nitrogen loading. Except for a few cases,
15NDIN is seldom measured due to technical diYculties. The source of external inputs of nitrogen is generally unknown and may inXuence the isotope composition of POM in the surface water. Nitrogen from residential wastewater and agricultural land use is a major source of nutrients to the lakes (Carpenter et al. 1998; Leavitt et al. 2006; Xie et al. 2007) and often carries distinct isotope ratios (Anderson and Cabana 2005; Savage 2005; Vander Zanden et al. 2005). There was also a lack of data for nitrogen Wxation associated with POM collection. As discussed above, nitrogen Wxation tends to dilute the isotope pool of POM in lakes of various trophic states. DiVerences in sampling methods may also contribute to the unexplained variations in
15NPOM and made data interpretation diYcult. Many studies reported POM as those that are Wltered onto GF/F Wlter papers. Other studies deWned POM as those that pass diVerent mesh sizes. It is known that diVerent size categories of suspended particles possess diVerent isotope compositions (Yoshioka et al. 1988; Sarvala et al. 2003; Vuorio et al. 2006). Another variation in sampling technique is the water depth where POM samples were taken. Water samples for POM analysis were typically collected at 0.5 m below the surface in shallow lakes (e.g., Gu et al. 1996, 2006). However, water samples from integrated depth (e.g., Pulido-Villena et al. 2005; Patoine et al. 2006; Vuorio et al. 2006), subsurface layers (Harvey and Kitchell 2000; Mooy et al. 2001; Sarvala et al. 2003), or sediment traps (Hodell and Schelske 1998; Teranes and Bernasconi 2000) were also reported. The POM at diVerent water depths may consist of diVerent proportions of living organisms and detritus with contrasting isotope composition (Yoshioka et al. 1988; Sarvala et al. 2003; Lehmann et al. 2004).

Conclusions Despite the complexity of nitrogen cycling processes, the absence of some critical data, and the variations in POM collection techniques, this analysis revealed important relationships between lake morphometry, trophic state,
15NDIN, nitrogen Wxation and seasonal variability of
15NPOM in the global range of lakes. High
15NPOM is often associated with productive lakes, indicating that higher lake trophic state leads to 15N enrichment in POM. 15N depletion in nitrogen-limited systems that may be unproductive or productive lakes is a strong indication of planktonic nitrogen Wxation. The use of
15N values of suspended or sedimented POM to infer trophic or nutrient status in surface water must also consider other closely related factors such

123

430

as phytoplankton species composition and the nutrient ratios. These results revealed large seasonal variability of
15NPOM that is associated with trophic state and latitudinal locations of lakes. Therefore, a single or a few samples of POM very likely do not represent the isotope characteristics of the surface waters during the growth seasons. This Wnding provides important insights of sampling design for the studies of pelagic food web baseline. Further research is needed to gain a better understanding of the relationships between stable nitrogen isotopes of POM and DIN, the rates of nitrogen Wxation and the allochthonous loading of nitrogen compounds in lakes of diVerent trophic states. Acknowledgments I appreciated the critical comments by two anonymous reviewers and the editorial improvements by Dr. Thomas Dreschel and Dr. Sharon Ewe.

References Anderson C, Cabana G (2005)
15N in riverine food webs: eVects of N inputs from agricultural watersheds. Can J Fish Aquat Sci 62: 333–340 Bade DL, Carpenter SR, Cole JJ, Hanson PC, Hesslein RH (2004) Controls of
13C-DIC in lakes: geochemistry, lake metabolism, and morphometry. Limnol Oceanogr 49:1160–1172 Bootsma HA, Hecky RE, Hesslein RH, Turner GF (1996) Food partitioning among Lake Malawi nearshore Wshes as revealed by stable isotope analyses. Ecology 77:1286–1290 Cabana G, Rasmussen JB (1996) Comparison of aquatic food chains using nitrogen isotopes. Proc Natl Acad Sci USA 93:10844– 10847 Carpenter SR, Caraco NF, Correll DL, Howarth RW, Sharpley AN, Smith VH (1998) Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecol Appl 8:559–568 Chapman AD, Schelske CL (1997) Recent appearance of Cylindrospermopsis (Cyanobacteria) in Wve hypereutrophic Florida lakes. J Phycol 33:191–195 Delwiche CC, Steyn PL (1970) Nitrogen isotope fractionation in soils and microbial reactions. Environ Sci Technol 4:929–935 Elser JJ, Bracken MES, Cleland EE, Gruner DS, Harpole WS, Hillebrand H, Ngai JT, Seabloom EW, Shurin JB, Smith JE (2007) Global analysis of nitrogen and phosphorus limitation of primary producers in freshwater, marine and terrestrial ecosystems. Ecol Lett 10:1135–1142 Estep MLF, Vigg S (1985) Stable carbon and nitrogen isotope tracers of trophic dynamics in natural populations and Wsheries of the Lahontan Lake system, Nevada. Can J Fish Aquat Sci 42:1712– 1719 Goericke R, Montoya JP, Fry B (1994) Physiology of isotopic fractionation in algae and cyanobacteria. In: Lajtha K, Michener RH (eds) Stable isotopes in ecology and environmental science. Blackwell, Oxford, pp 187–221 Goldman CR, Marzolf ER, Elser JJ (1990) Phosphorus and nitrogen limitation of phytoplankton growth in the freshwaters of North America: a review and critique of experimental enrichments. Can J Fish Aquat Sci 47:1468–1477 Gondwe MJ, Guildford SJ, Hecky RE (2008) Planktonic nitrogen Wxation in Lake Malawi/Nyasa. Hydrobiologia 596:251–267 Gu B (1993) Natural abundance of 15N in a subarctic lake and biogeochemical implications to nitrogen cycling. Dissertation, University of Alaska-Fairbanks

123

Oecologia (2009) 160:421–431 Gu B, Alexander V (1993) Estimation of N2 Wxation based on diVerences in the natural abundance of 15N among freshwater N2-Wxing and non-N2-Wxing algae. Oecologia 96:43–48 Gu B, Schell DM, Alexander V (1994) Stable carbon and nitrogen isotopic analysis of the plankton food web in a subarctic lake. Can J Fish Aquat Sci 51:1338–1344 Gu B, Schelske CL, Brenner M (1996) Relationship between sediment and plankton isotope ratios (
13C and
15N) and primary productivity in Florida lakes. Can J Fish Aquat Sci 53:875–883 Gu B, Chapman AD, Schelske CL (2006) Factors controlling seasonal variations in stable isotope composition of particulate organic matter in a soft water eutrophic lake. Limnol Oceanogr 51:2837– 2848 Hadas O, Altabet MA, Agnihotri R (2009) Seasonally varying nitrogen isotope biogeochemistry of particulate organic matter in Lake Kinneret, Israel. Limnol Oceanogr 54:75–85 Harvey CJ, Kitchell JF (2000) A stable isotope evaluation of the structure and spatial heterogeneity of a Lake Superior food web. Can J Fish Aquat Sci 57:1395–1403 Havens KE, Phlips EJ, Cichra MF, Li B (1998) Light availability as a possible regulator of cyanobacteria species composition in a shallow subtropical lake. Freshw Biol 39:547–556 Hodell DA, Schelske CL (1998) Production, sedimentation, and isotopic composition of organic matter in Lake Ontario. Limnol Oceanogr 43:200–214 Hoering TC, Ford HT (1960) The isotope eVect in the Wxation of nitrogen by Azotobacter. J Am Chem Soc 82:376–378 Jäntti H (2007) The spatial and temporal variation of nitrogen Wxation in aquatic environments. Master Thesis, University of Jyväskylä Kling GW, Fry B, O’Brien WJ (1992) Stable isotopes and planktonic trophic structure in Arctic lakes. Ecology 73:561–566 Leavitt PR, Brock CS, Ebel C, Patoine A (2006) Landscape-scale eVects of urban nitrogen on a chain of freshwater lakes in central North America. Limnol Oceanogr 51:2262–2277 Lehmann MF, Bernasconi SM, McKenzie JA, Barbieri A, Simona M, Veronesi M (2004) Seasonal variation of the
13C and
15N of particulate and dissolved carbon and nitrogen in Lake Lugano: constraints on biogeochemical cycling in a eutrophic lake. Limnol Oceanogr 49:415–429 MacGregor BJ, Van Mooy B, Baker BJ, Mellon M, Moisander PH, Paerl HW, Zehr J, Hollander D, Stahl D (2001) Microbiological, molecular biological and stable isotopic evidence for nitrogen Wxation in the open waters of Lake Michigan. Environ Microbiol 3:205–219 McCusker EM, Ostrom PH, Ostrom NE, Jeremiason JD, Baker JE (1999) Seasonal variation in the biogeochemical cycling of seston in Grand Traverse Bay, Lake Michigan. Org Geochem 30:1543–1557 Minagawa M, Wada E (1986) Nitrogen isotope ratios of red tide organisms in the East China Sea: a characterization of biological nitrogen Wxation. Mar Chem 19:245–259 Montoya JP, McCarthy JJ (1995) Isotopic fractionation during nitrate uptake by phytoplankton grown in continuous culture. J Plankton Res 17:439–464 Montoya JP, Carpenter EJ, Capone DG (2002) Nitrogen Wxation and nitrogen isotope abundances in zooplankton of the oligotrophic North Atlantic. Limnol Oceanogr 47:1617–1628 Mooy BV, MacGregor B, Hollander D, Nealson K, Stahl D (2001) Evidence for tight coupling between active bacteria and particulate organic carbon during seasonal stratiWcation of Lake Michigan. Limnol Oceanogr 46:1202–1208 Ostrom NE, Long DT, Bell EM, Beals T (1998) The origin and cycling of particulate and sedimentary organic matter and nitrate in Lake Superior. Chem Geol 152:13–28 Owen JS, Mitchell MJ, Michener RH (1999) Stable nitrogen and carbon isotonic composition of seston and sediment in two adirondack lakes. Can J Fish Aquat Sci 56:2186–2192

Oecologia (2009) 160:421–431 Owens NJP (1987) Natural variations in 15N in the marine environment. Adv Mar Biol 24:389–451 Patoine A, Graham MD, Leavitt PR (2006) Spatial variation of nitrogen Wxation in lakes of the northern Great Plains. Limnol Oceanogr 51:1665–1677 Peterson BJ, Fry B (1987) Stable isotopes in ecosystem studies. Annu Rev Ecol Syst 18:293–320 Post DM (2002) Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83:703–718 Pulido-Villena E, Reche I, Morales-Baquero R (2005) Food web reliance on allochthonous carbon in two high mountain lakes with contrasting catchments: a stable isotope approach. Can J Fish Aquat Sci 62:2640–2648 Sarvala J, Badende S, Chitamwebwa D, Juvonen P, Mwape L, Molsa H, Mulimbwa N, Salonen K, Tarvainen M, Vuorio K (2003) Sizefractionated
15N and
13C isotope ratios elucidate the role of the microbial food web in the pelagial of Lake Tanganyika. Aq Ecosyst Health Manage 6:241–250 Savage C (2005) Tracing the inXuence of sewage nitrogen in a coastal ecosystem using stable nitrogen isotopes. Ambio 34:143–148 Syväranta J, Tiirola M, Jones RI (2008) Seasonality in lake pelagic
15N values: patterns, possible explanations, and implications for food web baseline. Fundam Appl Limnol 172:255–262 Teranes JL, Bernasconi SM (2000) The record of nitrate utilization and productivity limitation provided by
15N values in lake organic matter—a study of sediment trap and core sediments from Baldeggersee, Switzerland. Limnol Oceanogr 45:801–813

431 Vander Zanden MJ, Vadeboncoeur Y, Diebel MW, Jeppensen E (2005) Primary consumer stable nitrogen isotopes as indicators of nutrient source. Environ Sci Technol 39:7509–7515 Vuorio K, Meili M, Sarvala J (2006) Taxon-speciWc variation in the stable isotopic signatures (
15N and
13C) of lake phytoplankton. Freshw Biol 51:807–822 Wada E, Hattori A (1978) Nitrogen isotope eVects in the assimilation of inorganic nitrogenous compounds by marine diatoms [Phaeodactylum tricornutum and Chaetoceros sp.]. Geomicrobiol J 1: 85–101 Wang J, Wu F, Li W, Wang L, Guo J, Fu P, Zhang R (2008) Seasonal variation and vertical characteristics of the
15N of particulate organic matter in Lake Hongfeng and Lake Baihua, Guizhou Province. J Lake Sci 20:571–578 (in Chinese) Waser NAD, Harrison PJ, Nielsen B, Calvert SE, Turpin DH (1998) Nitrogen isotope fractionation during the uptake and assimilation of nitrate, nitrite, ammonium, and urea by a marine diatom. Limnol Oceanogr 43:215–224 Wetzel RG (2001) Limnology: lake and river ecosystems, 3rd edn. Academic Press, San Diego Xie YX, Xiong ZQ, Xing GX, Sun GQ, Zhu ZL (2007) Assessment of nitrogen pollutant sources in surface waters of Taihu Lake region. Pedosphere 17:200–208 Yoshioka T, Wada E, Saijo Y (1988) Isotopic characterization of Lake Kizaki and Lake Suwa. Jpn J Limnol 49:119–128 Yoshioka T, Wada E, Hayashi H (1994) A stable isotope study on seasonal food web dynamics in a eutrophic lake. Ecology 75:835–846

123