Lake Trout Ecosystems in a Changing Environment - Chapter 7 ppt

Evans RODA Environmental Research Limited Contents Introduction Nutrient–phytoplankton–fish linkages Nutrient–biomass relationships Characterization of Ontario lake trout lakes with res

chapter seven

The effects of phosphorus

and nitrogen on lake trout

(Salvelinus namaycush)

production and habitat

Peter J Dillon

Environmental and Resource Studies, Trent University

Bev J Clark

Dorset Environmental Science Centre, Ontario Ministry of the Environment

Hayla E Evans

RODA Environmental Research Limited

Contents

Introduction

Nutrient–phytoplankton–fish linkages

Nutrient–biomass relationships

Characterization of Ontario lake trout lakes with respect to phosphorus

and nitrogen

Characterization of Ontario lake trout lakes with respect to morphometry

TP–morphometry–oxygen linkages

Summary

References

Introduction

The relationships between the levels of nutrients, particularly phosphorus (TP), in lakes

and lake trout Salvelinus namaycush are complex This is a consequence of the fact that

phosphorus may affect lake trout in two opposite but interrelated ways On one hand, phosphorus is the nutrient that controls algal biomass in almost all lakes in the Boreal ecozone of Canada and most lakes directly to the south in the Laurentian Great Lakes region; i.e., in almost all lakes in North America that are inhabited by lake trout (Dillon and Rigler, 1974; Schindler, 1977) Phytoplankton is a very important component of the

food web in all lakes, and algal abundance may strongly influence the biomass and productivity of higher trophic levels, including fish On the other hand, high algal biomass

is associated with oxygen depletion in the hypolimnia of lakes; an adverse effect on the amount of suitable habitat available to species including lake trout that prefer or require cold, well-oxygenated waters is therefore expected when algal biomass is high

In this chapter, we discuss the two conflicting roles that the limiting nutrient, phos-phorus, plays in lake trout biology Considerably less information is available with respect

to the role of nitrogen, but we have included as much as is available We first review the literature with respect to relationships between nutrients and fish production in general and then review those relationships pertaining to phosphorus and salmonids in particular

We then characterize known lake trout lakes by their phosphorus and nitrogen regimes

We discuss the role of morphometric factors in determining lake trout habitat and then very briefly consider how a relatively new model that can be used to predict lake trout habitat is affected by the combination of phosphorus concentrations and morphometric factors The question of nutrient–morphometry–habitat relationships is discussed in detail

by Clark et al (Chapter 6, this volume)

Nutrient–phytoplankton–fish linkages

The relationship between TP in lakes and biological measurements of primary standing

stock, such as chlorophyll a concentration, has been studied extensively (Sakamoto, 1966;

Dillon and Rigler, 1974; Nicholls and Dillon, 1978; Prepas and Trew, 1983; Stockner and Shortreed, 1985; Ostrofsky and Rigler, 1987; Dillon et al., 1988; Molot and Dillon, 1991) Less well documented, however, is the association of other nutrients, such as nitrogen, with primary production (e.g., Sakamoto, 1966; Stockner and Shortreed, 1985; Prepas and Trew, 1983; Molot and Dillon, 1991) and also the relationship between TP and higher levels (e.g., zooplankton and fish) of the aquatic food web (Moyle, 1956; Hanson and Leggett, 1982; Jones and Hoyer, 1982; Downing et al., 1990)

As mentioned above there are two reasons TP, and perhaps other nutrients, should

be related to fish production, either directly or indirectly First, the ecological concept of food webs suggests that there should be a direct relation between fish production and secondary production and between secondary productivity and primary productivity (McQueen et al., 1986), even if there are allochthonous inputs to the food chain that contribute significantly to fish diets (e.g., France and Steedman, 1996) If phosphorus is controlling primary production, it is reasonable to expect a relationship between TP and fish production Secondly, we know that TP controls the trophic status of lakes (see above), and we know that phosphorus and trophic status can affect oxygen concentrations in thermally stratified lakes (Cornett and Rigler, 1979; Welch and Perkins, 1979; Molot et al., 1991) Because minimum O2 levels can define the habitat for lake trout (Ryan and Marshall, 1994; Ranta and Lindstrom, 1998; Sellers et al., 1998) or other cold-water species, then TP should be related to cold-water fish production

Nutrient–biomass relationships

Perhaps the best-known model used to predict fish production from a knowledge of nutrient concentrations in the lake was developed by Ryder (1965) Ryder’s morphoe-daphic index (MEI) was an empirically derived formula that related lake mean depth (zmean) and total dissolved solids (TDS), a surrogate of phosphorus (intended as an indi-cator of the lake’s fertility or productivity), to potential fish yield or harvest For a number

of years after its publication, Ryder’s MEI spawned a plethora of papers in which the

ability of the MEI to predict fish yield, fish harvest, fish production, and fish biomass was examined in North American lakes (Carlander, 1977; Adams and Olver, 1977; Oglesby, 1977; Ryan and Harvey, 1977; Matuszek, 1978; Prepas, 1983) and reservoirs (Jenkins, 1967, 1982; Henderson et al., 1973) and also in tropical systems (Regier et al., 1971; Henderson

et al., 1973; Toews and Griffith, 1979) During the same period of time, methods were developed to improve the index, for example, by the introduction of scaling factors for latitude (Henderson et al., 1973; Ryder et al., 1974, Schlesinger and Regier, 1982; Kalff, 1991) and also by attempts to find better predictors for fish harvest than TDS/zmean These predictors included biotic variables such as photosynthesis (McDonnell et al., 1977),

chlo-rophyll a (Jones and Hoyer, 1982), primary production (Oglesby, 1977) and benthic biomass (Matuszek, 1978; Hanson and Leggett, 1982), and also abiotic factors Because it is generally

understood that any correlate of TDS can be used as a substitute for that variable (Hend-erson et al., 1973), conductivity and alkalinity were substituted sometimes for TDS in the MEI (e.g., Vighi and Chiaudani,1985)

However, evidence that nutrients, and in particular TP, may ultimately control the rate of fish production in lakes was increasing (Colby et al., 1972) and, in fact, had been provided even earlier than the publication of the MEI (Rawson, 1951, 1952; Moyle, 1956)

In addition, there was a noticeable congruence of the MEI with other models incorporating morphometric and/or edaphic parameters such as TP Ryder et al (1974) argued that the MEI model could be related to Vollenweider’s (1968) method for determining admissible and dangerous levels of phosphorus loading in lakes and also to Schindler’s (1971) index for predicting TP loading to lakes from lake area, catchment area, and lake volume Oglesby (1982) also pointed out the similarity between the MEI and Vollenweider’s (1976,

as cited by Oglesby, 1982) model relating lake trophic status to phosphorus loading, i.e.,

where

Chla = chlorophyll a concentration (mg/m3)

Lp = areal P loading (mg P/m2/yr)

Τw = hydraulic retention time (years)

Assuming that Chl a is proportional to fish yield (Jones and Hoyer, 1982) and that Lp is a surrogate for TDS, which incorporates both allochthonous and autochthonous inputs, then the resemblance of this equation to the MEI is evident

Hanson and Leggett (1982) were perhaps the first to demonstrate conclusively that fish yields could be predicted empirically from TP concentrations in lakes They found that the MEI consistently performed poorly when compared to other indices of fish yield/biomass As they state, total phosphorus concentration and macrobenthos standing crop/mean depth were superior to TDS, mean depth, and the MEI as predictors of fish yield (and biomass) when comparisons were based on the same data set Total phosphorus was the best univariate predictor of fish yield and fish biomass in two of the four data sets they examined Furthermore, although the best multivariate predictor of fish yield in one of these data sets included zmean, P, and TDS, removal of TDS in the equation did not significantly reduce the predictive efficiency of the model They suggested that this resulted from cross correlation of TDS with phosphorus concentration (see also Vighi and Chiaudani, 1985; Chow-Fraser, 1991), which is in agreement with Henderson and co-workers’ (1973) contention that any correlate of TDS should be a suitable substitute for it

in the MEI

Vighi and Chiaudani (1985) used a slightly different approach to utilize the apparent relationship between TP concentration and fish production First they found a significant correlation between TP concentration and the MEI in 53 lakes having negligible phospho-rus load as a result of anthropogenic activities Then they plotted TP concentration versus MEI in lakes subject to cultural eutrophication and found they fell above the predicted line developed from the 53 pristine lakes From this observation they argued that because the MEI was developed for use on unpolluted lakes (i.e., with respect to nutrient inputs), then differences between MEI-predicted fish yields and actual fish yields could be used

to estimate the degree of phosphorus inputs from anthropogenic activities Thus, indirectly, Vighi and Chiaudani used the relationship between phosphorus and fish yield to predict natural phosphorus loadings in these lakes A similar approach was used more recently

by Chow-Fraser (1991) and also by Koussouris et al (1992)

The relationship between phosphorus concentration and fish production was explored further in a series of papers published by Downing and co-workers Downing et al (1990), using data collected from the literature for entire lake fish communities, demonstrated that fish production was closely correlated with mean total phosphorus concentration (r2

= 0.67) in addition to annual phytoplankton production (r2 = 0.79) and annual average fish standing stock (r2 = 0.67) Similar to Hanson and Leggett (1982), they also found no correlation between fish production and the MEI A few years later, Downing and Plante (1993) reported that the residuals from a multivariate equation (which related annual fish production to annual mean standing biomass and maximum individual biomass) indi-cated that fish production was positively correlated with total phosphorus concentration (in µg/L) in addition to temperature, phytoplankton production, chl a, and lakewater pH (see their Figure 4) Once again, the MEI was not a good predictor of fish production Later the same year, Plante and Downing (1993) tested the hypothesis that there was a general positive relationship between lake trophic status (i.e., phosphorus concentration) and salmonine production Their relationship:

log Production = 0.95 log TP - 0.47 (n = 10, r2 = 0.61) (7.2) where

Production = salmonine production (kg/ha/yr)

TP = total lake phosphorus concentration (µg/L)

This is very similar to the relation published by Hanson and Leggett (1982) over a decade earlier, i.e.:

log FY = 1.021 log TP − 1.148 (n = 21, r2 = 0.87) (7.3) where

FY = fish yield (kg/ha)

TP = total phosphorus concentration (mg/m3)

This is surprising because Hanson and Leggett’s (1982) relationship was produced using many fish species, whereas Plante and Downing’s (1993) relationship involves only salmo-nine species Plante and Downing suggest that the similarity may indicate that the link between (abiotic) factors (such as phosphorus), which influence primary production, and salmonine populations is stronger for salmonines than for other fish species because food chains tend to be simpler in lakes in which salmonines dominate Thus, the relationship

between phosphorus and salmonine production is similar to that between phosphorus and fish community production, simply because the food chain is shorter in salmonine lakes

The mounting evidence that TP is the nutrient limiting primary productivity and also fish (salmonine) production in lakes, together with the economic significance of salmonine fisheries, prompted some scientists/biologists to investigate the utility of adding phosphorus to oligotrophic lakes as a means of increasing salmonine production For example, in the early 1970s (1970 to 1973 inclusive), Great Central Lake, British Columbia, was treated with ammonium nitrate and ammonium phosphate in an attempt to test the hypothesis that increasing the supply of inorganic nutrients in the lake would increase production at succeeding trophic levels (LeBrasseur et al., 1978) LeBrasseur and co-work-ers found that during the period when the lake was being enriched, mean summer primary production increased fivefold, zooplankton standing stock increased nine times, and the growth of age 2+ smolts, the survival of age 0+ sockeye, and the mean stock size of adult

sockeye salmon Oncorhynchus nerka increased from <50,000 to >360,000 fish Several years

later, Hyatt and Stockner (1985) reported on the results of a more expanded fertilization experiment in which ammonium nitrate and ammonium phosphate were added to as many as 17 lakes along the British Columbia coast Similar to the results of LeBrasseur

et al (1978), Hyatt and Stockner (1985) found that increased autotrophic and heterotrophic production resulted in larger standing stocks of zooplankton and increased in-lake growth

of juvenile sockeye salmon They suggest that the changes that occurred in the fertilized lakes may lead to increases in the harvestable surplus of sockeye adults In Norway, Johannessen et al (1984) also fertilized six small mountain lakes in Telemark, with ammo-nium phosphate, ammoammo-nium nitrate, and phosphoric acid They found that in Lake

Kanontjern, the length and weight of brown trout Salmo trutta increased during the three

seasons of fertilization Similarly, Johnston et al (1990) found that whole-river fertilization

of the Keogh River, British Columbia, with nitrogen and phosphorus increased the size

of steelhead trout Oncorhynchus mykiss and coho salmon Oncorhynchus kisutch fry More

recently, Ashley et al (1997) reported that phosphorus and nitrogen additions (1992–1994)

to the North Arm of Kootenay Lake, British Columbia, increased the biomass of

phy-toplankton, zooplankton, and kokanee salmon Oncorhynchus nerka in the lake.

In summary, given the key role of TP in the production dynamics of freshwater ecosystems and the strong predictive relationships that have been developed previously (e.g., Hanson and Leggett, 1982; Plante and Downing, 1993), it is important that models incorporating TP and fish production continue to be produced and refined

Characterization of Ontario lake trout lakes with respect to phosphorus and nitrogen

Because of their requirement for cold, well-oxygenated water, the classic picture of a lake suitable for lake trout is one that is oligotrophic and deep, although exceptions to this generalization are known to occur (e.g., in monomictic Pedro Lake, Ontario; see Snucins and Gunn, 1995) Notwithstanding, this notion is based on a conceptual model in which the benefits of the protection of oxygen concentration (i.e., of maintaining habitat size), derived from having low nutrient levels, exceed the advantages of having high nutrient levels and concomitant increased productivity Survey data collected on a substantial number of lakes in Ontario support this concept Of 1220 lakes in the province with TP data available, 293 also were classified as lake trout lakes (both native and introduced) These are distributed throughout the range of lake trout lakes in the province (Figure 7.1) The distributions of TP concentrations in lakes with and without lake trout in the data set

are shown in Figure 7.2 A total of 61% of the lake trout lakes surveyed had TP concen-trations less than 6 µg/L, and 85% had <10 µg TP/L, the concentration that is sometimes used as a criterion for classifying lakes as oligotrophic (Dillon and Rigler, 1975) In com-parison, 27% of non–lake trout lakes had TP < 6 µg/L and 62% had <10 µg TP/L Similarly, the mean (6.9 µg/L) and median TP concentrations (6.0 µg/L) in lake trout lakes were lower than those (10.3 and 8.9 µg/L, respectively) of the non–lake trout lake set This comparison indicates that lake trout lakes, on average, have lower TP levels than other lakes in Ontario In fact, the data set used here may be biased to make this difference appear less than it is, as the surveys that have been compiled here into this database explicitly included almost all of the known lake trout lakes in the southernmost portion

of the province where it may be expected that TP levels are higher because of anthropo-genic contributions to the nutrient budgets, e.g., more shoreline development, higher atmospheric deposition of TP

Figure 7.1 Locations of (A) 1916 known lake trout lakes in Ontario and of (B) the subset of 293 lake trout lakes from the overall set with at least one TP concentration measurement The data were collected using a range of sampling methods and were collected at different times of the year, although almost all samples were either epilimnetic composite samples taken in the summer or spring or fall overturn samples All analyses were carried out using identical methods at the same laboratory

Figure 7.2 Total phosphorus concentration distributions in 293 lake trout lakes and 927 lakes with

no lake trout in Ontario

Lake Trout Lakes Non-Lake Trout Lakes

0-2 4-6 8-10 12-14 16-18 20-22 24-26 28-30

Total Phosphorus ( µg/L)

40

30

20

10

0

The distribution of TP in lake trout lakes in Quebec (Figure 7.3) was very similar to that in Ontario (Prairie, 1994) Of 125 lakes in Quebec with TP data available, 49% had

TP < 6 µg/L, and 86% had TP < 10 µg/L The mean (7.1 µg/L) and median (6.2 µg/L)

TP concentrations in Quebec lake trout lakes were virtually identical to those in Ontario lakes (6.9 and 6.0 µg/L, above) Ratios of total nitrogen to total phosphorus (TN–TP) greater than about 16 (mole ratio, equivalent to a ratio of 7 by weight) generally indicate phosphorus limitation (Wetzel, 1975); this ratio is exceeded in all but a few cases in eastern North America As a result, nitrogen does not control trophic status and/or algal biomass

in the great majority of lakes in the Boreal ecozone or the St Lawrence Lowlands However, nitrogen concentrations may influence species composition because lower TN-TP ratios favor nitrogen-fixing species In Ontario lake trout lakes, total nitrogen concentrations, like TP concentrations, were lower than in non–lake trout lakes (Figure 7.4) The mean and median TN concentrations (330 and 331 µg/L, respectively) were substantially lower than those in non–lake trout lakes (495 and 394 µg/L) The median TN−TP ratio in the lake trout lakes was 56 (by weight), well above the ratio at which nitrogen may control algal biomass and higher than that in non–lake trout lakes (44 by weight)

Figure 7.3 Total phosphorus concentration distributions in 125 Quebec lake trout lakes (from Prairie, 1994)

Figure 7.4 Total nitrogen concentration distributions in 102 lake trout lakes and 394 lakes without lake trout in Ontario

30

20

10

0

0-2 4-6 8-10 12-14 16-18 20-22 24-26 28-30

Total Phosphorus ( µg/L)

25

20

15

10

5

0

Lake Trout lakes Non-Lake Trout Lakes

100-150 150-200 200-250 250-300 300-350 350-400 400-450 450-500 500-550 550-600

Total Nitrogen (µg/L)

It should be noted that the traditional view of a lake trout lake might change as the impacts of species invasions (i.e., the accidental or intentional introduction of nonnative species) into lake trout lakes are more fully elucidated For example, it has been reported

(Vander Zanden et al., 1999) that introduction of smallmouth bass Micropterus dolomieu and rock bass Ambloplites rupestris into some Canadian lakes resulted in a change in the

food-web structure including both a decline in the littoral prey-fish abundance and in the trophic position of the trout Lake trout in the invaded lakes shifted their diet toward pelagic zooplankton and reduced their dependence on littoral fish

However, in general, the available information supports the hypothesis that lake trout are found in lakes with low TP levels and TN–TP ratios Only a few lake trout lakes have

TP concentrations that would be considered indicative of mesotrophic conditions This suggests that the need for relatively low TP concentrations to ensure high concentrations

of oxygen in the cold, deep water supersedes the benefits of increased nutrient levels that would result in increased food resources

Characterization of Ontario lake trout lakes with respect

to morphometry

In addition to nutrient levels, lake trout habitat is controlled by morphometry The require-ment for cold, well-oxygenated water means that thermally stratified lakes with large hypolimnetic volumes should be preferred As the proportion of a lake that is composed

of hypolimnetic water increases, not only is the volume of the water with suitable cold temperatures that are needed increased, but the impact of the algal nutrients and subse-quent production in the trophogenic zone, which leads to oxygen-consuming degradation processes in the hypolimnion, diminishes Thus, any consideration of the role of nutrients

in lake trout biology is inextricably linked with morphometry because of the nutrient–oxy-gen relationship

A comparison of the size of lake trout lakes with other lakes in Ontario is shown in Figure 7.5 There is a much lower proportion (5%) of lake trout lakes in the smallest size category (<25 ha) than is the case for all lakes in the database (34%) More than 34% of the known lake trout lakes are <100 ha in area, about half the total (63%) of all those lakes

Figure 7.5 Distribution of lake surface areas for 1916 Ontario lake trout lakes and 6824 other lakes

in Ontario

40

30

20

10

0

All Lakes Lake Trout Lakes

100-125 150-175 200-225 250-275 300-325 350-375 400-425 450-475

Lake Size (ha)

in the database However, because the database includes very few of the small lakes (<10 ha) in Ontario, the difference is certainly considerably larger than these numbers indicate Because lake size is not normally distributed, we consider the median areas rather than the means; for lake trout lakes, the median (169 ha) is about three times greater than for all lakes (55 ha)

The requirement for cold temperatures and high oxygen concentrations suggests that, other things being equal, lake trout lakes will be relatively deep in comparison with lakes that do not have lake trout populations This is supported by the information available

in Ontario (Figure 7.6); the average mean and maximum depths of 1892 lakes with lake trout are 10.7 and 33.5 m, respectively This is substantially greater than the comparable mean and maximum depths for 436 non–lake trout lakes with data available (4.7 and 13.0 m, respectively)

In summary, the available data support the widely held view that lake trout are more likely to be found in deep lakes; the preference for larger size (area) is probably a conse-quence of the fact that there is a relationship between lake area and depth This is relevant

in the context of nutrients in that an oxygen model developed for habitat (Clark et al.,

Chapter 6, this volume) utilizes both TP concentration and morphometric data related to hypolimnetic volume

TP–morphometry–oxygen linkages

Recently, a model linking TP, morphometry, and water clarity (dissolved organic carbon

or Secchi depth) to lake trout habitat has been proposed (Dillon et al., in press) The model addresses both the oxygen and temperature requirements of lake trout TP and morphom-etry are used in the oxygen submodel, and water clarity and morphommorphom-etry are utilized

in the thermal submodel The former model can be used to estimate the response of oxygen concentrations in individual hypolimnetic strata in a lake to changes in TP concentration; thus, the effects of changing nutrient levels on the volume of suitable habitat may be ascertained provided criteria for optimum and/or acceptable oxygen concentration can

be established In a similar fashion, the thermal submodel addresses the issue of habitat size from the perspective of optimal and/or acceptable temperatures This approach to modeling optimal habitat is discussed in detail by Clark et al (Chapter 6, this volume)

Summary

Nutrients, particularly phosphorus, play two conflicting roles in lake trout biology Ele-vated nutrient levels increase productivity of at least lower trophic levels, which, in turn, may increase fish productivity However, elevated nutrient levels also result in reduced hypolimnetic oxygen levels, i.e., reduced habitat for lake trout There are a number of relationships in the literature between nutrients and fish production in general, but only

a very few pertaining to phosphorus and salmonids in particular The available nutrient data demonstrate that lake trout lakes are typically oligotrophic, with 85% of those in Ontario having TP concentrations below 10 µg/L Mean and median TP values are about two-thirds of those in lakes without lake trout, and total nitrogen concentrations in lake trout lakes are also lower by about the same proportion Nutrients, morphometric factors, and water clarity (through temperature effects) combine to determine the size of lake trout habitat; a relatively new model can be used to predict how lake trout habitat is affected

by the combination of phosphorus concentrations and morphometric factors

Adams, G.F and Olver, C.H., 1977, Yield properties and structure of boreal percid communities in

Ontario, Journal of the Fisheries Research Board of Canada 34:1613–1625.

Ashley, K., Thompson, L.S., Lasenby, D.C., McEachern, L., Smokorowski, K.E., and Sebastian, D.,

1997, Restoration of an interior lake ecosystem: the Kootenay Lake fertilization experiment,

Water Quality Research Journal of Canada 32:295–323.

Carlander, K.D., 1977, Biomass, production, and yields of walleye (Stizostedium vitreum vitreum) and yellow perch (Perca flavescens) in North America lakes, Journal of the Fisheries Research Board

of Canada 34:1602–1612.

Chow-Fraser, P., 1991, Use of the morphoedaphic index to predict nutrient status and algal biomass

in some Canadian lakes, Journal of the Fisheries Research Board of Canada 48:1909–1918.

Colby, P.J., Spangler, G.R., Hurley, D.A., and McCombie, A.M., 1972, Effects of eutrophication on

salmonid communities in oligotrophic lakes, Journal of the Fisheries Research Board of Canada,

29:975–983

(a)

(b)

Figure 7.6 Comparison of (a) maximum and (b) mean depths between 1892 lake trout lakes and 436 other non–lake trout lakes with available data

Lake Trout lakes Non-Lake Trout Lakes

Maximum Depth (m)

10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90

40

30

20

10

0

Lake Trout lakes Non-Lake Trout Lakes

Mean Depth (m)

0-5 5-10

10-15 15-20 20-25 25-30 30-35 35-40 >40

70 60 50 40 30 20 10 0