174
Epilithic algal abundance in relation to anthropogenic changes in phosphorus bioavailability and limitation in mountain rivers Michelle F. Bowman, Patricia A. Chambers, and David W. Schindler
Abstract: Low-level cultural eutrophication (0.1–3.8 µg·L–1 increase in total phosphorus (TP)) of oligotrophic mountain rivers resulted in 4- to 30-fold increases in benthic algal abundance. Because anthropogenic P was more bioavailable than naturally occurring P, there were higher algal abundances downstream relative to upstream of nutrient point sources at a given P concentration. Neither TP nor soluble reactive P concentrations were indicative of P bioavailability. Of the measures studied, epilithic alkaline phosphatase activity was most strongly correlated with algal abundance, most indicative of P bioavailability and thus the most precise indicator of P limitation. Although changes in dissolved inorganic nitrogen (DIN) to P ratios in river water and carbon (C) to P ratios in epilithon were consistent with changes in algal abundance and nutrient limitation, published water DIN to TP and tissue C to P ratio thresholds did not always yield accurate predictions of the type or degree of nutrient limitation. Epilithic N to P ratios and algal growth on nutrient-diffusing substrates were also inexact measures of epilithic nutrient limitation but, unlike other measures, were not strongly correlated with algal abundance. Thus, the predictability of the benthic algal response to anthropogenic nutrient additions in oligotrophic rivers will be improved by using measures indicative of both nutrient limitation and bioavailability. Résumé : Une eutrophisation culturelle de faible intensité (augmentation de 0,1–3,8 µg·L–1 du phosphore total, TP) de rivières oligotrophes de montagne entraîne des accroissements par un facteur de 4–30 de l’abondance des algues benthiques. Parce que le P d’origine anthropique est plus biodisponible que le P naturel, il y a de plus fortes densités d’algues en aval qu’en amont des points d’origine des nutriments, pour une même concentration de P. Ni les concentrations de TP, ni celles du P réactif soluble ne reflètent la biodisponibilité de P. Des variables étudiées, c’est l’activité de la phosphatase alcaline dans l’épilithon qui possède la plus forte corrélation avec la biomasse des algues, qui reflète le mieux la biodisponibilité de P et qui est ainsi le meilleur indicateur de la limitation en P. Bien que les changements du rapport de l’azote inorganique dissous (DIN) sur P dans l’eau de rivière et du rapport carbone (C) sur P dans l’épilithon concordent avec les variations de l’abondance des algues et la limitation en nutriments, les valeurs publiées des seuils de DIN:TP dans l’eau et de C:P dans les tissus ne permettent pas toujours de prédire avec exactitude le type ou l’importance de la limitation en nutriments. Le rapport N:P dans l’épilithon et la croissance des algues sur des substrats à diffusion de nutriments sont aussi des mesures imprécises de la limitation en nutriments dans l’épilithon, mais, contrairement à d’autres variables, ils ne sont pas fortement corrélés à l’abondance des algues. Ainsi, l’utilisation de variables qui reflètent à la fois la limitation en nutriments et la biodisponibilité augmentera la possibilité de prédire la réaction des algues benthiques à l’addition de nutriments d’origine anthropique dans les rivières oligotrophes. [Traduit par la Rédaction]
Bowman et al.
184
Introduction Cultural eutrophication, resulting from excess inputs of nitrogen (N) and phosphorus (P) to aquatic systems, commonly leads to algal abundance that exceeds that desired for aesthetic reasons or for maintenance of ecological integrity. The ability to predict, explain, and mitigate changes in periphyton abundance in rivers is inexact relative to capabilities
for phytoplankton in lake ecosystems. Nutrient supply may be more closely linked to phytoplankton abundances than to periphyton abundances because of diffusion limitations associated with benthic communities in lakes (e.g., Turner et al. 1994) and physical stressors (e.g., spates) in rivers (e.g., Biggs et al. 1998). However, in the absence of overriding physical or biological controls, N or P or both generally limit the abundance of benthic algae in rivers (Borchardt
Received 30 October 2003. Accepted 16 September 2004. Published on the NRC Research Press Web site at http://cjfas.nrc.ca on 5 March 2005. J17823 M.F. Bowman1,2 and D.W. Schindler. Department of Biological Sciences, University of Alberta, Edmonton, AB T6G 2E9, Canada. P.A. Chambers. National Water Research Institute, Burlington, ON L7R 4A6, Canada. 1 2
Corresponding author (e-mail:
[email protected]). Present address: Dorset Environmental Science Centre, 1026 Bellwood Acres Road, P.O. Box 39, Dorset, ON P0A 1E0, Canada.
Can. J. Fish. Aquat. Sci. 62: 174–184 (2005)
doi: 10.1139/F04-244
© 2005 NRC Canada
Bowman et al.
175
Table 1. Main advantages and potential considerations when using ratios of nitrogen to phosphorus concentrations (N:P) in water and algal tissue, alkaline phosphatase activity (APA), or nutrient-diffusing substrates (NDS) to determine nutrient-limiting benthic algal abundance. Measure
Main advantages
Considerations
Water N:P
Predictable hyperbolic response of algal growth potential to changes in N:P (e.g., Schanz and Juon 1983)
Tissue N:P
Robust community optimum (Redfield 1958); reflects food quality for higher trophic levels (e.g., Sterner and Hessen 1994; Stelzer and Lamberti 2002) Inverse function exists between APA and PO4, total P (TP) and cellular TP (see Jansson et al. 1988)
Not relevant when nutrient-replete, may not reflect cellular nutrient ratios, small changes undetectable (Borchardt 1996); does not reflect tight periphyton–substrate nutrient cycling (Pringle 1987); can be only weakly predictive (Francoeur et al. 1999) Optimum is species-specific (Rhee and Gothman 1980); grazing can change ratios (Hunter 1980); growth-rate-dependent (McMahon et al. 1974); can be only weakly predictive (Francoeur et al. 1999) Generally measures maximum potential APA production, difficult to completely separate from APA produced by bacteria and lysed cells, variable among species (but see Rengefors et al. 2003), also occurs in response to starvation for pyrimidines or for guanine, diurnal cycles can occur (Jansson et al. 1998) Diffusion low and inconsistent, pores may clog, sorb P, contain P-binding agents (Pringle 1987); difficult to partition species according to nutrient-supply ratios (Borchardt 1996); a qualitative not a quantitative assessment
APA
NDS
Used successfully to address diverse theoretical and applied questions in benthology (Scrimgeour and Chambers 1997); early pulse mimics allochthonous decay (Pringle and Bowers 1984; Pringle 1987)
1996). Yet despite strong evidence for nutrient control of benthic algal growth in many rivers, it is not clear how much variation can be explained by nutrient availability, and which nutrient(s) will limit growth in a given location (Dodds et al. 2002). Indicators of nutrient limitation may poorly predict the response of periphyton to changes in nutrient availability in rivers because concentrations and forms of N and P in rivers are highly variable in space and time, are highly correlated, and can colimit algal growth (Francoeur 2001). Methods used to identify and quantify the nutrient status of algae include measurements of nutrient concentrations in stream water or algal tissue, physiological indicators or growth rates, and artificial enrichment of water or substrate. Each method has advantages and potential drawbacks (Table 1). For example, N to P ratios (N:P) in water or algae are often compared with the theoretical ideal molar ratio of 16:1 (e.g., Redfield 1958), but precise ratios have been shown to be speciesspecific (e.g., Kahlert 1998). Alkaline phosphatase activity (APA) measures the ability of organisms (including bacteria and algae) to cleave organically bound P and is usually indicative of the degree of P limitation, but being starved of certain proteins can elicit an APA response (e.g., Jansson et al. 1988). Relative algal growth on nutrient-diffusing substrates (NDS) filled with agar amended with nutrients and incubated in situ is often used to determine the type (but not the degree) of nutrient limitation in a given location (e.g., Pringle 1987). In addition to potential problems in interpreting results from the use of a single indicator, the types or degree of limitation may vary because of differences in the scale of limitation or nutrient sources. For example, APA reflects cellular nutrient limitation, whereas NDS experiments infer community limitation. The P concentration found to saturate thin biofilms was two orders of magnitude lower than the concentration required to saturate benthic diatom mats
(Bothwell 1988, 1989). Also, nutrient-limitation indicators may reflect nutrient availability in overlying water (water N:P), substrate (NDS), or both (APA, tissue N:P). Periphyton can respond to variations in nutrient supply in the water column, in the substrate alone, or in both (Pringle 1987). A comparison of methods is needed to identify practical limitations and choose appropriate methods for various situations and study objectives. In P-limited systems, the relationship between P and algal abundance is particularly weak because there is a strong selective advantage for organisms that can use organically bound P, and tight P spiraling. For example, assessment of algal nutrient status from concentrations of total P (TP) and total N in water can overestimate nutrient availability because measures include organically bound fractions that cannot be used by biota. Conversely, the use of soluble-nutrient forms (soluble reactive phosphorus (SRP), dissolved inorganic nitrogen (DIN)) can underestimate bioavailability because certain organic nutrients can be utilized. Furthermore, available limiting nutrient(s) are quickly assimilated by organisms in oligotrophic systems, and therefore may not be detected in measures of river nutrient concentrations. Paul et al. (1991) concluded that because epilithon in rivers with the least bioavailable P reassimilated the P metabolites most quickly, empirical P-abundance relationships would underestimate algal development in the most oligotrophic rivers. In this study, we assessed the efficacy of various indicators of nutrient limitation in explaining patterns of epilithic abundance in oligotrophic mid-sized (4th–5th order at 1 : 250 000 scale) mountain rivers upstream and downstream of four anthropogenic nutrient point sources. The amount of nutrient loading relative to river discharge varied among rivers and point sources and therefore provided both natural and anthropogenic gradients of nutrient loading to test nutrientlimitation measures. The goals of this study were to (i) compare the types and magnitudes of epilithon nutrient limita© 2005 NRC Canada
176
Can. J. Fish. Aquat. Sci. Vol. 62, 2005
Fig. 1. Map of the Athabasca River within Jasper National Park boundaries, the Bow River within Banff National Park boundaries, and the Kicking Horse River within Yoho National Park boundaries, and the locations of Jasper, Lake Louise, Banff, and Field townsites in Alberta and British Columbia, Canada (51°10′N–53°54′N, 115°35′W–118°05′W; 1000–1500 m above sea level).
tion inferred from C:P and N:P in water and in epilithon tissue, APA, and NDS, (ii) determine whether the type or degree of limitation changes as a result of low-level (parts per billion) eutrophication, and (iii) determine whether nutrient-limitation measures are related to the gradient in anthropogenic nutrient additions and epilithic abundance.
Methods Study sites Methods to determine the type of nutrient (N or P) potentially limiting epilithon growth were used at sites upstream and downstream of discharges from municipal wastewater treatment plants (MWWTP) to the Athabasca, Bow, and Kicking Horse rivers in the National Parks in the Rocky Mountains of Alberta and British Columbia (Fig. 1). Detailed descriptions of the rivers, MWWTPs, subalpine and montane ecoregions, and sampling protocols are given in Bowman (2004). Briefly, the Athabasca River receives wastewater from the Town of Jasper, the Bow River receives wastewater from the towns of Lake Louise and Banff, and the Kicking Horse River receives wastewater directly from the Town of Field and indirectly from the Emerald Lake resort, which discharges wastewater into the Emerald River tributary. Sampling sites were located approximately 2 km upstream and 3 km downstream of each MWWTP effluent discharge. In the case of the Kicking Horse River, downstream samples were collected below the confluence of the Emerald River and reflect impacts from both the Field and Emerald Lake MWWTPs, except for APA, which was sampled above the confluence. All sampling sites contained at least three shallow (30 cm), fast-flowing (0.5 m·s–1) riffles within a 500-m reach, and substrate dominated by cobble (10–20 cm diameter). Samples were collected in late September (in the Bow River) and early October (in the Athabasca and Kicking Horse rivers) in 2000, when reduced light from riparian shading or high turbidity did not limit algal growth. Nutrient-chemistry measurements were collected in 1998–2000, NDS experiments were performed in the
Athabasca and Bow rivers in October of 1998 and 1999, and all other samples were collected in 2000. Epilithic communities were dominated by diatoms both upstream and downstream of MWWTP discharges, whereas the proportional abundance of the dominant primary consumer, mayflies, declined, owing to an increase in chironomids downstream of effluent discharges (Bowman 2004). Taxonomic analysis of benthic algal samples showed that assemblages were composed mainly of diatoms at all sites (both upstream and downstream of discharges) and on all sampling dates, but cyanophytes and chlorophytes were also present at most sites. Epilithic communities on NDS were similar to natural communities. At sites upstream of MWWTPs, the most common scrapers were mayflies (Heptageniidae, Ephemerellidae, and Baetidae), but there were also numerous caddisfly (Brachycentridae) and stonefly (Taeniopterygidae) scrapers. The abundance of mayflies, caddisflies, and stoneflies declined downstream of the Jasper MWWTP, and their proportional abundance declined downstream of all MWWTPs. Water chemistry Water samples were collected monthly from August to October and analyzed for TP, SRP, nitrate+nitrite-N, and ammonia-N following modified standard methods (University of Alberta Limnology Laboratory, Department of Biological Sciences, University of Alberta, Edmonton, AB T6G 2E9). Concentrations of TP (unfiltered) and SRP (filtered; Murphy and Riley 1962) were determined colorimetrically with a Varioan Cary 50 probe and a Milton Roy Model 1001 Plus spectrophotometer. Ammonia-N (unfiltered) and nitrate+nitrite-N (filtered) were determined colorimetrically with a TechniconTM AutoanalyzerTM II using automated Berthelot and cadmium reaction methods, respectively. Concentrations of nitrate+nitrite-N and ammonia-N were summed to give DIN. Estimated water-chemistry concentrations for sites downstream of MWWTPs were derived by multiplying mean monthly (August–October) concentrations in effluent by the effluent to river discharge ratio and © 2005 NRC Canada
Bowman et al.
adding upstream river concentrations. Estimated downstream nutrient concentrations are more indicative of nutrient availability than measured concentrations because historical daily records show that concentrations at upstream sites and in MWWTP effluents are consistent over time but concentrations at downstream sites are highly variable (Bowman 2004). Measured and expected molar N:P ratios in river water were compared with previously published nutrientlimitation thresholds. Summaries of the results of numerous studies on benthic algal nutrient enrichment show that the transition to P limitation occurs over a broad N:P range, from >10 to 30, and N limitation generally occurs at N:P < 10 (Allan 1995; Borchardt 1996). Algal abundance To collect epiphyton for chlorophyll a analyses, a template was used to demarcate a 9.6-cm2 area of maximum epiphyton abundance (generally the downstream end of the cobble), and epiphyton was removed with a scalpel; this procedure was repeated on the same cobble to obtain a sample for determining algal biovolume. Epilithon collected from each of three cobbles was combined into a composite sample (one composite for chlorophyll a and another for algal-biovolume analyses). Three composite samples for chlorophyll a and one composite sample for algal biovolume were collected from each site. Samples for determining algal biovolume were immediately preserved in 5% formalin. Methods used to estimate algal biovolume are described in detail by Findlay and Kling (2004). Samples for measuring chlorophyll a were immediately frozen on dry ice and stored in the dark at –20 °C until analyzed. Samples for measuring chlorophyll a were blended with double-distilled water for 10 s, filtered onto GF/F filters, and extracted in boiling 90% ethanol for 7 min. Chlorophyll a concentrations were measured with a fluorometer (Model 10, Turner Designs, Sunnyvale, California). Elemental composition Composite 20-mL samples of epilithon were removed with a scalpel from three randomly chosen rocks in a riffle, at three riffles per site (for a total of three samples per site). Bryophytes were abundant only at the site downstream of Lake Louise, interfered with APA analysis, and were therefore excluded from samples. Epilithon samples for tissue nutrient content were freeze-dried at –60 °C (Edwards Super Modulyo Freeze Drier, Crawley, West Sussex, UK), ground to a uniform powder with a mortar and pestle, dried at 90 °C to a constant weight, and stored in a desiccator. Between 0.1 and 0.2 g (±0.0001 g) of epilithon sample was weighed for analysis of percent P and 1–4 µg (±0.0001 µg) of epilithon sample was used for analyses of percent N and C. P was extracted using persulfate oxidation followed by analysis using the acid molybdate technique (American Public Health Association 1992). The amount of C and N in samples was measured using a Model 440 elemental analyzer (Control Equipment Corporation, Cazenovia, New York). Benzoic acid and SRM NIST 1575 apple leaves (US National Institute of Standards and Technology) were used as qualitycontrol standards for elemental analyses. All C:N:P ratios are reported on a molar basis. We compared the nutrient
177
content of epilithon to P-limitation thresholds of C:P > 369 and N:P > 32 derived by Kahlert (1998). Alkaline phosphatase activity A composite sample of epilithon taken from a 9.6-cm2 area on each of three cobbles was used to measure APA and chlorophyll a concentration; as with samples for epilithic nutrient content, three samples per site were collected and bryophytes were excluded. APA was measured following the phytoplankton method of Waiser and Bothwell (National Water Research Centre, 11 Innovation Boulevard, Saskatoon, SK S7N 3H5, unpublished data) modified for use with epilithon samples, using information in the literature (e.g., Healey and Hendzel 1979; Jansson et al. 1988). The epilithon sample was suspended in sterile filtered river water, refrigerated, and analyzed within 6 h. Samples were slowly agitated with a syringe to break up clumps, and filtered on GF/C filters (to collect bacteria; M.F. Bowman, unpublished data). Next, one half of the sample was resuspended in sterile filtered river water and heated to 35 °C in a water bath. Epilithon samples and blanks of filtered river water were saturated with 10 µmol·L–1 4-methylumbelliferyl phosphate. Fluorescence occurs when phosphate is enzymatically cleaved from 4-methylumbelliferyl phosphate. Fluorescence was measured at increasing time intervals with a fluorometer (Model 10-AU-005, Turner Designs) until APA production stabilized over time. The fluorometer was set up with stacked emission filters (top: dominant λ 479.8; middle: dominant λ 569.5; bottom: 10% neutral density) and the ultraviolet lamp was set up with an excitation filter and a reference filter. The amount of chlorophyll a on the remaining half of the composite epiphyton sample was used to standardize enzyme activity. Based on comparisons between Psufficient and P-deficient phytoplankton assemblages, Healey and Hendzel (1979) concluded that APA (nmol·µg chlorophyll a–1·h–1) >5 was indicative of severe P deficiency, 3–5 was indicative of slight deficiency, and 10–30; Allan 1995; Borchardt 1996), indicated that epilithon at all upstream sites (with one exception upstream of Jasper) would be limited by unavailability of P (Figs. 2c, 2d; Table 2). Both N:P ratios also indicated that epilithon at sites downstream of Field and Lake Louise would likely remain P-limited. Downstream of Banff and Jasper, average DIN:TP indicated that epilithon would not be P-limited, whereas mean DIN:SRP indicated that epilithon would be P-limited. As expected, there were significant (p < 0.05) increases in chlorophyll a concentration downstream of all MWWTP effluent discharges (Fig. 3a). With the exception of the site downstream of Field, there were also increases in algal biovolume at all downstream sites (Fig. 3a). However, at a given estimated TP or SRP concentration, downstream sites always had higher chlorophyll a concentrations (or algal biovolume) than upstream sites (Figs. 3b–3e). ANCOVA showed that the slopes of chlorophyll a – TP and algal biovolume – TP relationships were significantly different (p = 0.04 and p = 0.001, respectively) upstream and downstream of MWWTPs. It was inappropriate to use ANCOVA for algal abundance – SRP relationships because of nonlinearity. © 2005 NRC Canada
Bowman et al.
179
Table 2. The type of epilithic nutrient limitation (phosphorus (P), nitrogen (N), colimitation, or nutrient saturation) concluded from experiments with molar ratios of dissolved inorganic nitrogen (DIN) to total phosphorus (TP) or soluble reactive phosphorus (SRP) in water, molar C:P and N:P ratios in epilithic tissue, epilithic alkaline phosphatase activity (APA), and nutrient-diffusing substrates (NDS) upstream and downstream of municipal wastewater treatment plant discharges. P-limitation threshold
Field Upstream Downstream Lake Louise Upstream Downstream Banff Upstream Downstream Jasper Upstream Downstream
>10–30
>10–30
C:P > 369
N:P > 32
>3 (slight); >5 (severe)
DIN:TPa
DIN:SRPa
Tissue C:Pb
Tissue N:Pb
APAc
NDSd
P P
P P
P P
P P
Severe P Severe P
Saturation P
P P
P P
P P
P P
Severe P Slight P
P Colimitation
P Saturation
P P
P P
P Saturation
Severe P P saturation
Colimitation Saturation
Saturation Saturation
P P
P Saturation
Saturation P
Severe P P saturation
P Saturation
Note: Discrepancies are indicated in boldface type. a Allan (1995); Borchardt (1996). b Kahlert (1998). c Healey and Hendzel (1979). d Significant in Tukey’s test.
Internal measures of nutrient limitation (tissue C:N:P, APA) also suggested that epilithon was more limited by P availability at sites upstream of MWWTPs than at downstream sites (Table 2). The C:P and N:P contents of epilithon were significantly higher at sites upstream than downstream of MWWTPs (except near Banff and Jasper) (Figs. 4a, 4b). In the case of Jasper, nutrient ratios in epilithon collected from cobble and NDS were consistent; C:P and N:P were four- to six-fold lower for epilithon collected from substrates emitting P and N+P than for epilithon collected from control NDS, and were lower downstream than upstream of MWWTPs in a given NDS treatment (Figs. 4c, 4d). Comparison of tissue C:P and N:P with reported P-limitation thresholds for periphyton (C:P > 369, N:P > 32; Kahlert 1998) suggested that all upstream sites except the site near Jasper (based on N:P) would be P-limited (Figs. 4a, 4b, Table 2). Epilithic nutrient content also suggested that sites downstream of Field and Lake Louise would remain Plimited. However, contrary to other measures, tissue C:P downstream of Banff was just above the P-limitation threshold, and tissue N:P was below the P-limitation threshold upstream but above the threshold downstream. Average APA of epilithon at upstream sites was significantly higher than at sites downstream of the Lake Louise, Banff, and Jasper MWWTPs (Fig. 5a). APA of epilithon from all upstream sites and the site downstream of Field was >5 nmol·µg chlorophyll a–1·h–1, indicative of severe P limitation (Table 2; Healey and Hendzel 1979). APA downstream of Lake Louise was in the 3–5 nmol·µg chlorophyll a–1·h–1 range indicative of slight P limitation, while APA at sites downstream of Banff and Jasper was 0.05; M.F. Bowman, unpublished data). Measures of algal abundance (chlorophyll a concentration and biovolume) were highly correlated (r > 0.7) with estimated DIN:SRP and DIN:TP, C:P and C:N of epilithon, and © 2005 NRC Canada
180 Fig. 3. Epilithic chlorophyll a concentration (mean ± 1 standard error) and algal biovolume (䊉) at sites upstream (open bars) and downstream (shaded bars) of wastewater discharges in Field, Lake Louise, Banff, and Jasper in autumn 2000 (an asterisk indicates a significant difference (p < 0.05) between upstream and downstream sites in a one-tailed t test) (a), and chlorophyll a concentration versus estimated TP (b) and SRP (c) and algal biovolume versus estimated TP (d) and SRP (e) at upstream (open symbols) and downstream (shaded symbols) sites.
APA (Table 3). Within categories (algal-abundance measures, nutrient concentrations, nutrient-limitation measures), there were high correlations between (i) the two measures of algal abundance, (ii) SRP and other water-chemistry measures, and (iii) tissue C:P and N:P, NDS C/P and C/NP, and APA and tissue C:P and NDS C/P.
Discussion Epilithic abundance increased and nutrient limitation decreased as a result of low-level anthropogenic eutrophication of nutrient-poor mountain rivers. Nutrient concentrations in river water and epilithic enzyme activity, nutrient content, and abundance on artificially enriched substrates showed that epilithon in the three oligotrophic mountain rivers was constrained by insufficient P at sites upstream of significant pollution sources. In contrast, sites downstream of MWWTPs showed a variable response, from severe P limitation to P saturation. The observation that epilithic abundance increased and nutrient limitation decreased as a result of low-level eutrophication was consistent with earlier findings (e.g., Bothwell 1989; Scrimgeour and Chambers 2000). However,
Can. J. Fish. Aquat. Sci. Vol. 62, 2005 Fig. 4. Molar ratios of carbon to phosphorus (C:P) (a) and nitrogen to phosphorus (N:P) (b) (mean ± 1 standard error) in epilithon sampled upstream (open bars) and downstream (shaded bars) of wastewater discharges, on natural substrates in autumn 2000 (horizontal lines indicate the P-limitation thresholds of Kahlert (1998); an asterisk indicates a significant difference (p < 0.05) between upstream and downstream sites in a one-tailed t test). C:P (c) and N:P (d) molar ratios of epilithon grown on a subset of P and N+P nutrient-diffusing substrate (NDS) treatments incubated upstream (open bars) and downstream (shaded bars) of the Jasper wastewater discharge in autumn 1999.
contrary to expectations, at a given P concentration epilithic abundance was higher and the degree of nutrient limitation lower at sites downstream than at sites upstream of MWWTPs; this suggests that neither TP nor SRP concentrations precisely reflect P bioavailability. Although nutrient concentrations in river water and their ratios were an imprecise reflection of nutrient bioavailability, the use of mean DIN:TP ratios resulted in accurate predictions (with one exception) of whether epilithon would be Plimited or P-saturated. Conversely, DIN:SRP ratios suggested that benthic algae in the most nutrient-rich sites (downstream of Banff and Jasper) were P-limited, when three other measures of nutrient limitation suggested nutrient saturation. Borchardt (1996) concluded that water nutrient ratios are indicative of nutrient limitation provided absolute nutrient concentrations are lower than growth-limiting levels. Although absolute concentrations of SRP estimated at all sites (mean = 1.6 µg·L–1, standard error = 1.3 µg·L–1) at all sites were considerably less than concentrations found by others to saturate benthic periphyton mats (>30–50 µg·L–1, Bothwell 1989; >25–35 µg·L–1, Chambers et al. 2000), estimated DIN concentrations at all sites (mean = 54–111 µg·L–1) were in excess of those found to saturate periphyton in nearby river reaches (DIN > 50 µg·L–1; Chambers et al. 2000). Because high N concentrations would cause elevated N:P at all sites, the ability to detect a shift between P limitation and saturation using N:P would not be precise in Nreplete systems. However, consistent with previous findings © 2005 NRC Canada
Bowman et al. Fig. 5. Alkaline phosphatase activity (APA) of epilithon on natural substrates (mean ± 1 standard error) sampled upstream (open bars) and downstream (shaded bars) of wastewater discharges in autumn 2000 (the horizontal line is the P-limitation threshold derived by Healey and Hendzel (1979); an asterisk indicates a significant difference (p < 0.05) between upstream and downstream sites in a one-tailed t test) (a) and epilithon grown on P and N+P NDS treatments incubated upstream of Field (b). APA versus estimated TP (c) and SRP (d) at sites upstream (open symbols) and downstream (shaded symbols) of wastewater discharges and downstream of the Field wastewater discharge but upstream of the Emerald Lake wastewater discharge (solid symbol).
for lake ecosystems (Morris and Lewis 1988; Axler et al. 1994), we found that nutrient-limitation predictions based on DIN:TP generally agreed with other measures of nutrient limitation, and may be more indicative of bioavailable fractions of P than DIN:SRP in P-limited systems. APA was the only measure of nutrient limitation studied that directly measured the propensity of benthic organisms to cleave organically bound P, and accurately discriminated among various degrees of P limitation. Although most knowledge of APA as a P-deficiency indicator comes from work on phytoplankton in lakes and oceans (e.g., Healey and Hendzel 1979), it has been shown experimentally that APA increases the P supply to algae in streams (Klotz 1985, 1991), and APA of epilithon is directly related to streamwater N:P (Klotz 1992). Consistent with a general review of APA as a P-deficiency indicator (Jansson et al. 1988), our results showed that APA was inversely related to TP and SRP concentrations and the cellular P level. In contrast with algal nutrient status identified by APA and NDS, predictions based on epilithic C:P and N:P at the three more-P-rich sites (average molar DIN:SRP ratio in river water = 56–115) were not consistent with other measures. Francoeur et al. (1999) also found that the cellular nutrient content of periphyton was of limited use for predicting which nutrient limited NDS bioassays in relatively Prich rivers (DIN:SRP = 15–126). However, results obtained
181 Fig. 6. Chlorophyll a concentrations (mean + 1 standard error) on control (C), phosphorus (P), nitrogen (N), and P+N (NP) NDS incubated in rivers upstream (open bars) and downstream (shaded bars) of wastewater discharges in 1998–2000 (bars with the same letters were statistically equivalent in a Tukey’s test).
at our most P-limited sites (DIN:SRP = 126–216) are consistent with the conclusion from experiments using river water with relatively low P concentration (DIN:SRP = 100–231) that the degree of P limitation based on periphytic nutrient content was consistent with other measures of nutrient limitation, including APA (Bothwell 1985). In contrast with epilithic N:P, epilithic C:P was highly correlated with algal abundance and epilithic APA. C:P ratios of benthic periphyton identified as P-replete were higher in the mountain rivers (C:P < 440) than those of benthic algae from other systems (C:P < 158; Kahlert 1998), but C:P ratios of epilithon identified as P-limited (C:P > 720) were not inconsistent with the C:P ratio of phytoplankton collected in a highly oligotrophic lake (250–1750; Elser and George 1993). If P concentrations are limiting, tissue C:P and N:P would be expected to increase because algae make more efficient use of P incorporated into cells (e.g., Hecky et al. 1993). Thus, algal communities adapted to lower nutrient concentrations will have sufficient nutrient supplies at higher C:P ratios than algal communities in more P-replete systems. The amount of epilithon on control relative to treatment NDS was not always consistent with other measures of nutrient limitation, and was not highly correlated with algal © 2005 NRC Canada
0.70 0.65 0.53 –0.72 –0.86 –0.85 –0.76 –0.40 0.50 0.32 –0.91
Estimated for river water TP + 0.56 SRP + 0.65 DIN + 0.46 DIN:SRP – –0.75 DIN:TP – –0.82
Tissue nutrient content C:P – –0.76 C:N – –0.71 N:P – –0.33
Nutrient-diffusing substrates C/P + 0.61 C/NP + 0.49
Alkaline phosphatase activity APA – –0.90 –0.85
0.57 0.37
–0.80 –0.36 –0.64
0.88 0.88 0.68 –0.88 –0.88
0.77 0.52 –0.88 –0.93
TP +
–0.76
0.74 0.61
–0.56 –0.45 –0.30
0.39 0.96 0.83 –0.88 –0.56
0.88 –0.92 –0.75
SRP +
–0.51
0.48 0.41
–0.53 –0.32 –0.37
0.11 0.78 0.84 –0.64 –0.25
–0.64 –0.39
DIN +
0.88
–0.82 –0.65
0.60 0.55 0.26
–0.62 –0.96 –0.70 0.96 0.79
0.91
DIN: SRP –
0.79
–0.73 –0.57
0.55 0.31 0.37
–0.81 –0.86 –0.54 0.89 0.85
DIN:TP –
–0.70
0.35 0.18
–0.76 –0.28 –0.66
0.6 0.37 –0.68 –0.90
TP +
–0.82
0.72 0.55
–0.68 –0.49 –0.40
0.84 –0.95 –0.70
SRP +
Estimated
–0.66
0.51 0.36
–0.69 –0.29 –0.55
–0.63 –0.38
DIN +
0.84
–0.73 –0.58
0.66 0.60 0.31
0.84
DIN:SRP –
0.88
–0.59 –0.42
0.75 0.54 0.45
DIN:TP –
0.77
–0.27 –0.08
0.50 0.78
C:P –
Tissue
0.59
–0.32 –0.16
–0.14
C:N –
0.42
–0.01 0.10
N:P –
–0.77
0.96
C/P +
NDS
–0.59
C/NP +
Note: Correlations greater than 0.70, shown in boldface type, are statistically significant (p < 0.05); to maintain an overall p value of 0.05, a Bonferroni correction requires a correlation of greater than 0.92 (underlined). The sign (+ or –) indicates an expected relationship with the eutrophication gradient; multiplying row and column signs gives the expected sign of the relationship between parameters.
0.72 0.54 0.36 –0.70 –0.59
+
Algal biovolume +
Measured in river water TP + 0.66 SRP + 0.63 DIN + 0.49 DIN:SRP – –0.71 DIN:TP – –0.56
Algal biovolume
Chlorophyll a + 0.92
Measured
Table 3. Correlations between measures of algal abundance (chlorophyll a concentration, algal biovolume), measured and estimated water nutrient concentrations (TP, SRP, DIN), elemental composition of epilithon (C:P, C:N, N:P), amount of chlorophyll a on the control divided by treatment nutrient-diffusing substrates (C/P, C/NP), and alkaline phosphatase activity.
182 Can. J. Fish. Aquat. Sci. Vol. 62, 2005
© 2005 NRC Canada
Bowman et al.
abundance. Discrepancies occurred at sites downstream of Lake Louise and upstream of Banff, when NDS experiments suggested colimitation by P and N but all other nutrientlimitation measures indicated P limitation. These sites were midway along the gradient of river P concentrations, representing the transition between P limitation and P saturation. The fact that the site downstream of Lake Louise was slightly P-limited (based on APA) supports the idea that this site was closer to the P-limitation threshold than other sites. Therefore, like N:P (Borchardt 1996), NDS near nutrient-limitation thresholds may not be as precise as other measures of nutrient limitation. However, growth of epilithon on NDS did highlight the fact that the response of multispecies communities to enrichment with a single nutrient is less frequent and detectable than the response to enrichment by both N and P (Elser et al. 1990; Francoeur 2001). In conclusion, changes in epilithic P limitation were most accurately identified using APA, and patterns in APA and river-water DIN:TP and epilithic C:P were consistent with patterns in algal abundance and nutrient limitation. The amount of P bioavailable to epilithic organisms is reflected in the APA, DIN:TP, and epilithic C:P measures of nutrient limitation. Given the low concentrations of bioavailable P that naturally occur in mountain rivers (
