Hydrobiologia 448: 27–39, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.

27

Periphyton nutrient limitation and other potential growth-controlling factors in Lake Okeechobee, U.S.A. Andrew J. Rodusky, Alan D. Steinman, Therese L. East, Bruce Sharfstein & Richard H. Meeker Ecosystem Restoration Department, South Florida Water Management District, 3301 Gun Club Road, West Palm Beach, FL 33406, U.S.A. E-mail: [email protected] Received 15 September 1999; in revised form 14 June 2000; accepted 3 November 2000

Key words: periphyton, nutrients, shallow lakes, littoral, light, phytoplankton

Abstract Periphyton nutrient limitation was assessed in Lake Okeechobee, a large, shallow, eutrophic lake in the southeastern U.S.A. Nutrient assays were performed to determine if the same nutrients that limit phytoplankton also limit periphyton growth in the lake. Nutrient diffusing clay substrates containing agar spiked with nitrogen, phosphorus, or both, along with nutrient-free controls, were incubated at four sites in the lake. Three sites were located in a pelagic–littoral interface (ecotone) and one site was located in the interior littoral region. Incubations lasted for 20–26 days, and were repeated on a quarterly basis between 1996 and 1997, to incorporate seasonal variability into the experimental design. The physical and chemical conditions at each site also were measured. Periphyton biomass (chlorophyll a and ash-free dry mass) was highest at the littoral and northern ecotone sites. At the littoral site, nitrogen limited biomass in four of five incubations, although the largest biomass differences between the treatments and controls (≤3 µg cm−2 as chl) were probably not ecologically significant. Periphyton biomass at the western and southern ecotone sites was low compared to the other two sites. Increases in water column depth and associated declines in light penetration strongly correlated with periphyton growth and suggested that they may have limited growth most often at all three ecotone sites. Nitrogen also was found to limit periphyton growth approximately 20% of the time at the ecotone sites and phosphorus was found to limit growth once at the west site.

Introduction The large littoral region and adjoining littoral/pelagic interface (ecotone) of subtropical Lake Okeechobee are areas where periphyton can become abundant (Steinman et al., 1997a), perhaps due to the ability of light to penetrate to the sediments, and to the abundance of macrophytes, which provide a large colonizable surface area. In these areas of the lake, periphyton may be able to compete with phytoplankton for nutrients (Hwang et al., 1998), and may indirectly reduce phytoplankton biomass or bloom frequency via nutrient removal from the water column (Phlips et al., 1993). Periphyton on the sediment surface also may prevent nutrient transport into the water column from the sediments in shallow lakes (Wetzel, 1979; Hansson, 1990; Hansson, 1992; Blumenshine

et al., 1997). Observational data suggest that competition between periphyton and phytoplankton occurs in the littoral and ecotone areas of Lake Okeechobee. Summer periphyton biomass maxima correspond to maximum macrophyte biomass and minimum phytoplankton biomass (Philips et al., 1993; Havens et al., 1996; but see Hwang et al., 1998). While the potential exists for periphyton to remove a large portion of nutrients otherwise available to phytoplankton in Lake Okeechobee and in other shallow lakes, factors that regulate periphyton growth in lakes have received much less study compared to phytoplankton (Lowe, 1996). Phytoplankton studies in Lake Okeechobee commenced in the 1970s (Kratzer & Brezonik, 1984; Schelske, 1989) and growth-controlling factors appear to vary both temporally and spatially (Aldridge

28 et al., 1995). Phytoplankton in this lake is thought to be limited primarily by light (Phlips et al., 1997) and nitrogen (N) (Aldridge et al., 1995; Havens et al., 1995, 1996). However, phytoplankton also has been shown to be limited by phosphorus (P) (Hwang et al., 1998) and co-limited by nitrogen and phosphorus (NP) (Aldridge et al., 1995; Havens et al., 1999). Sampling location and season play a strong role in determining resource limitation in this large, subtropical lake. Factors regulating periphyton growth in Lake Okeechobee, however, have only recently been considered. Steinman et al. (1997a) concluded that both light and water chemistry was important. Havens et al. (1996) found that both periphyton and phytoplankton were either N limited or co-limited by NP during two short-duration experiments and concluded that the potential for competition between the two algal communities existed. Hwang et al. (1998) compared periphyton and phytoplankton P uptake kinetics over a 1-year period, and suggested that P limited both communities year-round in the littoral region, but only during the summer in the ecotone, when phosphorus concentrations were lowest. Zimba (1998) conducted nutrient bioassays on artificial substrates, at a site near the littoral region in the western bay, to test epiphyton nutrient limitation, and observed that N, and sometimes silica, limited growth. Periphyton nutrient limitation in the littoral region also was assessed in mesocosms during a month-long study, and the results suggested that periphyton was co-limited by both N and P (Havens et al., 1999). While these studies have suggested the potential for competition between periphyton and phytoplankton, those that examined nutrient competition were of short duration (e.g. 1 month) or were spatially restricted, and have not been in agreement as to the limiting nutrient. No long-term studies (e.g. >1 yr) have experimentally assessed the spatial and temporal importance of nutrients on periphyton growth. This research was conducted to examine whether periphyton responded to nutrients in the same temporal and spatial patterns as that observed for phytoplankton. Other factors thought to be important as potential limiting factors of periphyton and phytoplankton growth, such as light, also were measured, but were not experimentally manipulated. We used nutrient diffusing substrates to conduct this research, which have been widely used to assess periphyton nutrient limitation in lakes (Fairchild & Lowe, 1984; Fairchild et al., 1985; Fairchild & Sherman, 1992; Havens et al., 1996).

Study site Lake Okeechobee is located in south-central Florida (26◦ 60 N and 80◦ 50 W). It is a large (surface area 1720 km2 ) and shallow (zmean =2.7 m) lake, and the annual water column temperature range (15–32 ◦ C) reflects its subtropical location. The lake is considered eutrophic (Aumen, 1995), and the pelagic region consists of four distinct ecological zones, which differ in nutrient concentrations, turbidity, light availability, underlying sediment composition, macrophyte coverage and chlorophyll a concentrations (Phlips et al., 1993). A large littoral region accounts for approximately 25% of the total lake surface area. Site locations and study dates Nutrient diffusing substrates (NDS) were positioned at four different sites within the lake (Fig. 1) to examine spatial variability in periphyton responses to nutrients. The littoral site was located in an area dominated by the emergent Eleocharis, but also contained floating Nymphaea and submerged Utricularia. This site is located approximately halfway between the pelagic zone and the western shore of the lake. The ecotone sites were located in areas containing scattered Scirpus stands, where the littoral macrophytes interfaced with the pelagic zone. These sites were located in the north, west and southern regions of the lake. NDS were deployed on six occasions at the ecotone sites (2/96, 5/96, 8/96, 11/96, 2/97, 5/97) and on 5 occasions at the littoral site (4/96, 5/96, 8/96, 11/96, 2/97), to examine temporal variability in periphyton growth responses to nutrients. The incubations lasted between 20 and 26 days.

Materials and methods Laboratory diffusion study The release of N and P out of the clay nutrient diffusing substrata (described below) was assessed prior to in situ deployment. A 19-day diffusion study was conducted by placing a NDS of each treatment type (described below) into separate acid-rinsed 20 l carboys, similar to release studies performed by Fairchild et al. (1985). The carboys were filled with 7 l of deionized water and stored at temperatures similar to those in Lake Okeechobee (20–22 ◦ C), to assess diffusion rates. Water was collected three times a week, and analyzed for total and soluble reactive phosphorus

29

Figure 1. Site locations on Lake Okeechobee, FL. Insert shows lake location in Florida, U.S.A.

(TP and SRP), nitrate+nitrite (NOx ), total Kjeldahl nitrogen (TKN) and ammonium (NH4 ). Water in each carboy was changed daily.

Experimental design Nutrient diffusing substrata consisted of small unglazed clay pots (156 cm2 surface area), which were soaked for 24 h in 10% HCl solution and then transferred to deionized water for 48–72 h. The pots were soaked to remove any nutrients or toxic compounds

30 bound to the clay. The pots were filled with approximately 180 ml of nutrient-free agar, spiked with 0.1 M KNO3 (nitrogen–N, at 1400 mg l−1 ), 0.1 M KH2 PO4 (phosphorus–P at 3125 mg l−1 ), both (NP), or neither (control), and sealed as described in Fairchild & Lowe (1984). The pots were attached to 30 cm long wooden dowel rods. The dowel rods were inserted into a rectangular PVC frame filled with sand. Four replicates of each treatment were placed on each frame, for a total of 16 NDS per frame, as described in Havens et al. (1996). Each replicate was placed randomly at a different position on each frame side to minimize position effects. Small Hobo
light and temperature data loggers were placed inside clear polyethylene petri dishes, which were sealed with clear silicone, and attached (one logger of each type) to dowel rods on opposite corners of the frame. The loggers recorded ambient total light (Tot light) and temperature every 24 min. The frames were deployed by slowly lowering into the water column until they rested flat on the sediment surface, so that the pots and data loggers were approximately 0.4 m above the sediments. A PVC collar attached to the middle of each frame was placed over a PVC pipe, which was driven into the sediments, to keep each frame in place. Physical and chemical data were collected weekly, at which time the petri dishes containing the data loggers were cleaned. Incubation periods lasted from 20 to 26 days, at which time the frames were retrieved. Recovery of the NDS replicates was greater than 90% except at the western ecotone site, where it was 88%. At the western ecotone site, all of the NP replicates during 11/96 were lost. Additionally, water column depth at the littoral site during 4/97 was too low to allow the NDS to be incubated. Upon retrieval, all pots were immediately removed from the frames, placed into individual ziploc
bags, and stored on ice in coolers until transport back to the laboratory. The pots were processed randomly, within 24 h of retrieval. Any pots not processed immediately upon returning to the laboratory (always ≤50% of the total) were stored overnight in a large walkin cooler at 2–3◦ C. Processing consisted of gently scraping the periphyton from each pot surface into a polyethylene tray with a toothbrush. The sample was homogenized in a graduated cylinder by vigorous shaking for approximately 30 s, and split into three fractions for chlorophyll a concentration (CHL), ashfree dry mass (AFDM), and cellular carbon to nitrogen molar ratio (C:N) determination. Chlorophyll a concentrations (µg cm−2 ) were determined following a

18 h extraction at 2–3 ◦ C on ground up glass fiber filters (Whatman GF/C) in 90% acetone buffered with Mg CO3 . Pheophytin was corrected for in the extracted samples with 10% HCL. AFDM (mg cm−2 ) was determined by drying the weighed, filtered samples to a constant weight at 105 ◦ C, followed by ignition at 500 ◦ C for 2 h, as described in Steinman & Lamberti (1996). Samples for AFDM on 2/97 were lost prior to AFDM determination. Carbon–nitrogen (C:N) ratios were determined following standard methods (APHA, 1995). Since there never was luxuriant periphyton growth on the NDS, C:N samples were pooled from all four replicates of each treatment to ensure enough material for analysis. Samples were dried to constant weights, and ground to a fine powder and re-dried. Carbon and nitrogen content were determined on a CNS analyzer (C.E. Elantech, U.S.A.). Physical and chemical measurements Temperature, dissolved oxygen (DO), pH and conductivity were measured weekly with a Hydrolab Surveyor III multiparameter sonde. Water column depth was measured using a calibrated line, and Secchi depth was measured with a black and white 20 cm disc. Instantaneous photosynthetically active radiation was measured both above the water column (DECK PAR) and 0.5 m above the sediment surface (BTM PAR), using a LICOR 1000 data logger with a spherical quantum sensor. Nutrient concentrations were measured from water collected weekly with a 2 l Van Dorn bottle, from 0.5 m above the sediments near the center of each frame. Measured water quality variables were the same as those listed for the diffusion study, plus total suspended solids (TSS) and water column chlorophyll samples were collected. Water samples were collected in HCl-rinsed polyethylene bottles. Water column chlorophyll a concentrations (µg l−1 ) were determined spectrophotometrically following a 18 h extraction on ground up glass fiber filters (Whatman GF/C) in 90% acetone buffered with Mg CO3 (APHA, 1995). Water collected for SRP, NOx , and NH4 analyses was filtered in the field through a 1.2 µm glass fiber and 0.45 µm polycarbonate filter. Unfiltered TP and TKN water samples were treated in the field with a 50% H2 SO4 solution to a final sample pH of ≤2. All sample bottles were stored on ice and returned to the laboratory. Phosphorus and N concentrations were determined with a flow-injection autoanalyzer.

31 Sample collection methods followed the South Florida Water Management District Comprehensive QA/QC Plan (SFWMD, 1995), and water quality analyses followed USEPA procedures (1979, 1987) and Standard Methods (APHA, 1995). Statistical analyses Nutrient concentrations from the NDS were regressed log-linearly against time to determine release rates. Separate one-way analysis of variance (ANOVA) tests were run to assess the effects of the nutrient treatments, site locations, and date on periphyton biomass (as CHL and AFDM) and C:N (SAS, 1990). For each ANOVA, two of the factors were blocked (e.g. when nutrient treatment effects were assessed, site location and date were the blocked factors). When assessing site location and date effects, only the control treatment means were used. Scheffe’s multiple contrasts procedure was used to determine if statistically significant (p≤0.05) differences existed between the means. Nutrient limitation is herein defined as a statistically significant (p≤0.05) greater mean biomass on treatments spiked with nutrients, relative to the control treatment. Repeated measures ANOVA also was conducted to assess within and between site variability in measured water quality variables. Pearson product-moment correlation coefficients were calculated between control treatment mean CHL and AFDM, the water quality variables, and water column chlorophyll a concentrations, to identify those factors which may have had a significant (p≤0.05) correspondence with periphyton growth. All data were log transformed prior to statistical analyses to meet homogeneity of variance assumptions. The 4/96 littoral site data were not included in the among-site comparison tests, because the NDS were not incubated at the other sites during the same time period.

Results Diffusion study Nutrient release rates declined between days 3 and 19 in a similar pattern for all three nutrient-spiked treatments. Linear regression lines had the best fit to the ln transformed data for the 16 days and indicated that release rates decreased in a pattern similar to that reported by Fairchild & Lowe (1985). The fit between the linear decay model and the nutrient release rates for NOx (r 2 values −0.88 and 0.94),

TP (r 2 values −0.58 and 0.96), and SRP (r 2 values −0.58 and 0.93) always was statistically significant (p≤0.05). The percentage of N and P released from the pots over the course of the experiment ranged between 68% (N treatment) and 81% (NP treatment) for N, and between 43% (P treatment) and 47% (NP treatment) for P. T-tests also indicated that there were no significant differences (p<0.05) in release rates among the nutrient treatments. Site characteristics – physical and chemical Water quality at the littoral and ecotone sites was significantly different, although there were some similarities between the littoral and northern ecotone site. Water column chlorophyll a, TP, SRP and zmax were lowest at the littoral site and significantly lower (except for zmax ) compared to the other sites (Table 1). Conversely, BTM PAR, total light and Secchi depth:water column depth (percent light penetration in the water column) were highest at the littoral site and significantly higher than at the western and southern ecotone sites. Mean dissolved oxygen and conductivity were significantly lower at the littoral and northern ecotone sites, respectively, compared to the other sites. Seasonal variability of most physical and chemical variables at each site was limited. There were no significant changes in nutrient concentrations or light penetration (always >500 µmol photons m−2 s−1 ) at the littoral site during the study period. In general, nutrient concentrations were highest and light penetration was lowest during the first three experiments at the ecotone sites (<30 µmol m−2 s−1 at the western and southern ecotone sites, and <200 µmol m−2 s−1 at the northern ecotone site). During the remainder of the study, the reverse was observed at these stations; nutrient concentrations decreased somewhat, while light penetration increased (>300 µmol m−2 s−1 at the western and southern ecotone sites, and >500 µmol m−2 s−1 at the northern ecotone site). Water column chlorophyll a concentration between the ecotone sites was similar, and was much lower at the littoral site. There were no statistically significant differences in water column chlorophyll a between incubation dates at any of the sites. Overall, the littoral site was characterized by oligotrophic conditions in the water column, while the western and southern ecotone sites were characterized by eutrophic conditions. The northern ecotone site displayed characteristics of both the littoral (e.g. relatively high light penetration) and ecotone (e.g. high nu-

32 Table 1. Mean (±1 S.D.) and range of the physical, chemical and water column chlorophyll a data for the four sites. (SD: Z (%) is Secchi depth:water column depth, D.O. is dissolved oxygen and Cond is Conductivity). For total light, n=4, except the southern ecotone site, where n=3 Site

Zmax (m) SD: Z (%) BTM PAR (umol m−2 s−1 ) Tot Light (lum ft−2 ) TP (µg l−1 ) SRP (µg l−1 ) NOx (µg l−1 ) NH4 (µg l−1 ) Chl a (µg l−1 ) D.O. (mg l−1 ) Cond (µmhos cm−2 )

Littoral

Western Ecotone

Mean

Range

Mean

Range

0.85±0.19 100 636±384 6448±4006 11±3 5±2 16±34 18±10 3±1 5.2±1.6 514±94

0.48–1.1 100 101–1211 893–9627 6–17 <4–12 <4–160 <9–56 2–5 2.7–8.3 416–712

1.41±0.3 45±28 183±283 2317±2640 83±55 14±13 30±50 30±33 22±11 8.0±2.0 425±51

0.8–1.95 40–100 0.01–775 53–5475 29–227 <4–50 <4–230 <9–171 4–47 2.2–11.8 328–517

Northern Ecotone Zmax (m) SD: Z (%) BTM PAR (umol m−2 s−1 ) Tot Light (lum ft−2 ) TP (µg l−1 ) SRP (µg l−1 ) NOx (µg l−1 ) NH4 (µg l−1 ) Chl a (µg l−1 ) D.O. (mg l−1 ) Cond (µmhos cm−2 )

0.94±0.34 80±20 261±327 5788±5159 72±20 18±13 46±46 35±34 16±7 7.5±2.3 257±109

trient concentrations, higher water column chlorophyll a) regions. Site characteristics – periphyton biomass Chlorophyll a concentration means (CHL) were highest at the littoral and northern ecotone sites, and were significantly lower (p<0.001) at the western and southern ecotone sites (Fig. 2). CHL generally increased during the study period at the ecotone sites, and generally decreased at the littoral site. Nitrogen limited periphyton growth at the littoral site (four of five incubations), the southern ecotone site (three of six incubations), and the northern ecotone site (one of six incubations), but it never limited growth at the western ecotone site. Phosphorus never limited periphyton growth at any site.

0.5–1.5 32–100 6–1055 1203–7910 39–108 <4–44 <4–155 <9–139 7–37 2.4–11.7 122–445

Southern Ecotone 1.08±0.29 50±25 123±175 1710±1872 84±32 12±11 27±30 20±13 21±11 6.6±2.2 459±54

0.49–1.55 15–100 0.06–681 215–3810 24–180 <4–43 <4–111 <9-52 5–54 0.5–10.2 351–569

On the control NDS, CHL displayed the highest variability between incubation dates at the ecotone sites, with highest overall variability at the southern ecotone site, where 80% of all possible among-date comparisons were significantly different. The littoral site had the lowest amount of CHL variability on the control NDS between dates, where 30% of all possible among-date comparisons were significantly different. Control NDS CHL was significantly different in 66% of the among-site comparisons, with the littoral and northern ecotone site having significantly more CHL than the other two ecotone sites. Overall, nutrient limitation had the most significant effect on CHL variability at the littoral site, while date had the most significant effect at the ecotone sites. Location also had a significant effect on CHL on the control NDS.

33

Figure 2. Periphyton biomass means (±1 S.D. as bars) as chl a (µg cm−2 ) at the four sites. Letters above the bars indicate nutrient limitation as determined by ANOVA (N=nitrogen, P=phosphorus, NS=not significant).

Periphyton mean AFDM data showed the same general trends as did chlorophyll, with the highest biomass occurring at the littoral and northern ecotone sites, and a significantly lower biomass (p<0.001) at the southern ecotone site (Fig. 3). Mean AFDM increased somewhat over time at the ecotone sites, and displayed no temporal trend at the littoral site. Nutrient limitation of AFDM was evident only at the western and northern ecotone sites, with N limiting growth once at both sites, and P limiting growth once at the western ecotone site.

There was higher among-date variability on the control NDS AFDM at the ecotone sites relative to the littoral site, and it was highest at the western ecotone site, where 80% of all possible among-date comparisons had significantly different AFDM values. At the littoral site, 20% of the among-date AFDM value comparisons were significantly different. AFDM on the control NDS also was significantly different in 50% of the among-site comparisons, with the littoral, northern and western ecotone sites having significantly higher AFDM than the southern ecotone site.

34

Figure 3. Periphyton biomass means (±1 S.D. as bars) as AFDM (mg cm−2 ) at the four sites. Letters above the bars indicate nutrient limitation as determined by ANOVA (N=nitrogen, P=phosphorus, NS=not significant).

While there was more variability among-dates in AFDM than in CHL, the pattern was similar to that observed for CHL, with date being the most significant factor at the ecotone sites. Nutrients and date displayed little effect on AFDM at the littoral site. Location had a more significant effect on AFDM than on CHL. The C:N ratios of periphyton varied little among nutrients, date or location. The highest C:N ratios were observed at the littoral and southern ecotone sites. The C:N ratios ranged between 3:1 and 20:1 at all sites. Ratios of C:N on the control NDS were relatively

stable among incubation dates at the littoral site, but the mean C:N ratio in the control treatment (11:1) was significantly greater (p<0.01) than the mean C:N ratio of 7:1 in the NP treatment. There were no significant differences in control C:N ratios among sites, and C:N ratios varied little by incubation date or among nutrient treatments, at the ecotone sites. Correlations Other measured variables that might limit periphyton growth in the absence of nutrient limitation were cor-

35 Table 2. Significant correlations (p≤0.05) and relationship slopes between C treatment mean biomass and variables for the four sites (N=number of observations, BTM PAR=par 0.5 m above the sediments, DECK PAR=par above the water column, WCD=water column depth, Cond=conductivity, TP=total phosphorus, SRP=soluble reactive phosphorus) Site

Dependent variable

Independent variable

N

Western Ecotone Western Ecotone Western Ecotone Western Ecotone Western Ecotone Western Ecotone Western Ecotone Western Ecotone Western Ecotone Western Ecotone Western Ecotone Northern Ecotone Northern Ecotone Northern Ecotone Littoral Littoral Littoral Southern Ecotone Southern Ecotone

Log C CHL Log C CHL Log C CHL Log C CHL Log C CHL Log C CHL Log C AFDM Log C AFDM Log C AFDM Log C AFDM Log C AFDM Log C CHL Log C AFDM Log C AFDM Log C CHL Log C CHL Log C AFDM Log C CHL Log C CHL

Log Tot Light Log WCD Log BTM PAR Log zmax Log Sdisc Log Cond Log BTM PAR Log Sdisc Log zmax Log TP Log WCD Log SRP Log zmax Log SRP Log DECK PAR Log pH Log DECK PAR Log WCD Log zmax

6 6 6 6 6 6 5 5 5 5 5 6 5 5 5 5 4 6 6

related with control NDS biomass (Table 2). Light penetration into the water column was important at the ecotone sites. Zmax as an index of light penetration, for example, was inversely related (p<0.05) to either mean control CHL or AFDM at all three ecotone sites. At the littoral site, DECK PAR was significantly correlated (p<0.05) with both mean control CHL and AFDM. Of the measured nutrients, only SRP concentrations at the northern ecotone site were significantly correlated (p<0.05) with both CHL and AFDM, while TP was inversely related (p<0.05) with AFDM at the western ecotone site.

Comparison of environmental factors To better assess the physical environment during the incubation periods relative to long-term conditions, mean long-term water column depth (1991–1998) and wind speed (1989–1998) data from weather platforms closest to each sampling site were compared with means from each incubation period. There was little difference in mean water column depths among the platforms, and they all displayed the same pattern, so the combined means are illustrated in Figure 4a. The

r 2 value +0.98 −0.95 −0.90 −0.90 −0.81 +0.68 +0.91 +0.83 −0.72 −0.70 −0.67 −0.80 −0.66 −0.66 +0.86 +0.81 +0.84 −0.85 −0.81

time interval of the long-term data sets matched those for each incubation period from the present study. Mean water column depth among the weather platforms was higher than or approximately the same as the long-term among platforms mean during the first four of five incubation periods at the littoral site, and during the first three of six incubation periods at the ecotone sites. For the remaining incubations at all sites, mean water depth was below the long-term mean. During the last incubation period (5/97), mean water column depth was significantly lower than any other May between 1991 and 1998. There also was a significant and progressive decrease in mean water column depth during the study period, decreasing from approximately 4.9 to 4 m above sea level. There was little difference in wind speeds among the platforms, and they all displayed the same seasonal trends, so the combined means for all platforms are illustrated in Figure 4b. Mean wind speed during the 8/96 incubation period was significantly higher than any other August between 1989 and 1998. Mean wind speeds during the other incubation periods were similar to the 1989–1998 long-term means and ranged between 16 and 21 km h−1 . There were no significant differences in mean monthly wind speeds among sites

36

Figure 4. (a) Mean incubation period and long-term (1991–1998) mean water depth (m). The asterisk indicates a statistically significant lower 5/97 incubation mean relative to the long-term May mean water depth. (b) Mean incubation period and long-term (1989–1998) mean wind speed (Km h−1 ). The asterisk indicates a statistically higher 8/96 incubation mean relative to the corresponding long-term August mean wind speed.

during 1996, but there was a small and significant increase in mean wind speed at the northern ecotone site during the last two incubation periods.

Discussion In subtropical and tropical lakes, nutrients and light are factors that have been shown to limit phytoplankton growth (Henry et al., 1985; Hecky et al., 1993; Aldridge et al., 1995; Havens et al., 1996; Phlips et al., 1997). However, the role of nutrients and light as limiting factors of periphyton is not as well-known in these systems and has only recently received attention in Lake Okeechobee (Havens et al., 1996, 1999; Steinman et al., 1997; Hwang et al., 1998), one of the largest lakes in the U.S.A. To the best

of our knowledge, no studies using NDS to evaluate periphyton nutrient limitation have been conducted in other subtropical or tropical lakes. When comparing our NDS periphyton biomass (as CHL) to that observed in other lakes (almost all temperate), where similar experiments have been conducted, biomass at the littoral and northern ecotone sites appeared to be intermediate of values reported elsewhere. At the western and southern ecotone sites, periphyton biomass was low compared to that in temperate lakes. In temperate lakes, biomass values ranged between approximately 1 and 20 µg cm−2 , depending on the limiting nutrient and water chemistry of each lake (Fairchild & Lowe, 1984; Fairchild & Everett, 1988; Carrick & Lowe, 1989; Hansson, 1990; Fairchild & Sherman, 1992; Niederhauser & Schanz, 1993). A comprehensive comparison of periphyton biomass in Lake Okeechobee and other lakes can be found in Steinman et al. (1997a). All indices used to measure periphyton biomass in the present study suggested that nutrients and light influenced periphyton growth in Lake Okeechobee, and temporal and spatial variability in the importance of these factors was observed. Nitrogen was the most common limiting nutrient of periphyton biomass. Additionally, periphyton biomass was strongly correlated with light at three of the four sites (the northern ecotone site was the exception), suggesting that light may be an important growth-limiting factor rather than nutrients at certain times. Nutrients and light also are the same factors thought to limit phytoplankton growth in Lake Okeechobee (Aldridge et al., 1995; Phlips et al., 1997). While our results suggest that N and factors related to light penetration might play an important role in regulating periphyton growth in Lake Okeechobee, their importance appears to vary between the littoral and ecotone regions. At the littoral site, N usually limited periphyton biomass. There was no significant change in periphyton biomass on the control treatments throughout the study, nutrient concentrations and phytoplankton biomass were always very low, and there was no significant change in variables related to water column depth. At the ecotone sites, N usually was not important as a periphyton growth-limiting factor. Nutrient concentrations at these sites were high, phytoplankton biomass generally was high, and periphyton biomass on all treatments increased when lake stage decreased, possibly in response to increased water column PAR.

37 When water depth decreased to a point where a minimal critical PAR level was found at the sediment surface, nutrient limitation was observed more frequently. This shift toward nutrient limitation under lower water column depth is consistent with the notion that benthic algae in Lake Okeechobee are limited primarily by light (cf. Steinman et al., 1997a). Nutrient limitation may play a greater role in regulating periphyton growth once light limitation is relieved. Periphyton biomass in the littoral region was greater than that at two of the three ecotone sites, and similar to that at the northern ecotone site. Various factors in the littoral region might favor periphyton growth (Hwang et al., 1998): light always penetrates to the sediments, there is a substantial amount of colonizable surface, and phytoplankton biomass and turbidity are always very low. This contrasts with the ecotone, where water depths are greater, water column PAR is often much lower and there is less colonizable surface area for periphyton. Increased turbulence due to wave action, which has been shown in other systems (Eriksson & Weisner, 1996; Maltais & Vincent, 1997), also could account for decreased periphyton biomass in the ecotone. The northern ecotone site may be better described as transitional between the littoral and the ecotone. Both BTM PAR and total light were lower than at the littoral site, but higher than at the western and southern ecotone sites. When the water depth was high, periphyton biomass at the northern ecotone site was similar to that at the other ecotone sites. However, when water depth was at or below the long-term mean, light penetration increased, and periphyton biomass at the northern ecotone site became as high as at the littoral site. Conversely, while periphyton biomass at both the western and southern ecotone sites increased when water depth was below the long-term mean, the increases were relatively small and biomass was never equivalent to that at the other sites. Our finding that when a nutrient limited periphyton growth in Lake Okeechobee, it was almost always N, supports results obtained by previous investigators. Havens et al. (1996, 1999) reported that both phytoplankton and periphyton were limited by N, or co-limited by N and P in the ecotone and littoral regions of Lake Okeechobee. Zimba (1998) observed that N limited epiphyton growth on artificial macrophytes exposed to lake water. However, N may not always be the limiting nutrient of periphyton growth in Lake Okeechobee, as P also has been inferred as a limiting nutrient of periphyton in both the littoral

and ecotone regions (Hwang et al., 1998), and limited growth once during this study. Phosphorus also has been shown to limit periphyton in the Florida Everglades (McCormick et al., 1996), a nearby area that appears to contain many of the same species and a similar trophic status to that of the littoral region in Lake Okeechobee. Zimba (1998) suggested that silica might periodically limit periphyton growth in Lake Okeechobee when diatoms dominate the assemblage. Since N was found to be limiting at the littoral site and nutrient concentrations at the surface of the nutrient-spiked NDS should have been well above ambient concentrations, we expected periphyton biomass values at the littoral site to be similar to the higher values reported by Havens et al. (1996). However, biomass at the littoral site was unexpectedly low and similar to epiphyton biomass found on artificial substrates at oligotrophic locations in the Everglades (McCormick et al., 1996) and on Scirpus in the ecotone of Lake Okeechobee (Steinman et al., 1997a). A possible reason for lower than expected biomass and small differences between treatment biomass at the littoral site was grazing by Gambusia, a small planktivorous fish that is very abundant in the littoral region of the lake (Bull & Warren, 1994). During the weekly trips to the littoral site, Gambusia was always observed grazing on the NDS. Gambusia has been called an opportunistic feeder, and may have been grazing on detritus and periphyton (Daniels & Felley, 1992). Overall, these results suggest that periphyton in the littoral region and ecotone may be primarily limited by the same variables shown to limit phytoplankton in Lake Okeechobee (Aldridge et al., 1995). Limitation of these two algal fractions by N, when there is sufficient light penetration, suggests that nutrient competition between the two assemblages may be occurring. The light regime and amount of colonizable surface in the littoral region appears to enable periphyton to attain sufficient biomass to compete successfully with phytoplankton for nutrients (Hwang et al., 1998). Conversely, lack of attributes that favor abundant periphyton biomass in the ecotone region may result in no sustained competition for nutrients between periphyton and phytoplankton, and little impact on phytoplankton biomass. For example, results from experiments conducted by Hwang et al. (1998) led them to conclude that low light levels in the ecotone resulted in the small amount of P uptake by periphyton (see also Steinman et al., 1997b). The strong correlations between light-related factors and periphyton growth at the ecotone sites also suggest

38 that when water depths are low enough to allow sufficient PAR to reach the sediments, the increased light regime in the water column may enhance periphyton growth. Under this scenario, the ability of periphyton to compete with phytoplankton for nutrients may be enhanced, possibly resulting in decreased bloom frequency in the ecotone. Future research needs to assess nutrient limitation of both periphyton and phytoplankton simultaneously, as well as impacts of other biotic components, such as macrophytes and vertebrate and invertebrate grazers on the periphyton in Lake Okeechobee.

Acknowledgements We are grateful to Mark Brady, Chuck Hanlon, W. Pat Davis and Soon-Jin Hwang for their assistance in the field and laboratory. We also are grateful to Hunter Carrick, Binhe Gu, Karl Havens, Paul McCormick, and two anonymous reviewers for improving earlier versions of this manuscript.

References Aldridge, F. J., E. J. Phlips & C. L. Schelske, 1995. The use of nutrient enrichment bioassays to test for spatial and temporal distribution of limiting factors affecting phytoplankton dynamics in Lake Okeechobee, Florida. Arch. Hydrobiol Beih. Ergebn. Limnol. 128: 177–190. American Public Health Association (APHA), 1995. Standard Methods for the Examination of Water and Wastewater. 19th edn. APHA, Washington, DC. Aumen, N. G., 1995. The history of human impacts, lake management and limnological research of Lake Okeechobee, Florida (U.S.). Arch. Hydrobiol Beih. Ergebn. Limnol. 45: 1–17. Axler, R. P. & J. E. Reuter, 1996. Nitrate uptake by phytoplankton and periphyton: whole-lake enrichments and mesocosm 15 N experiments in an oligotrophic lake. Limnol. Oceanogr. 41: 659–671. Blumenshine, S. C., Y. Vadeboncoeur, D. M. Lodge, K. L. Cottingham & S. E. Knight, 1997. Benthic-pelagic links:responses of benthos to water-column nutrient enrichment. J. n. am. Benthol. Soc. 16: 466–479. Bull, L. A. & G. L. Warren, 1994. Kissimmee River-Lake Okeechobee-Everglades resource evaluation: Florida Freshwater Game and Fish Commission Report. Tallahassee, FL: 177 pp. Carrick, H. J. & R. L. Lowe, 1988. Response of Lake Michigan benthic algae to in situ enrichment with si, n and p. Can. J. Fish. aquat. Sci. 45: 271–279. Carrick, H. J. & R. L. Lowe, 1989. Benthic algal response to n and p enrichment along a pH gradient. Hydrobiologia 179: 119–127. Daniels, G. L. & J. D. Felley, 1992. Life history and foods of Gambusia affinis in two waterways of southwestern Louisiana. Southwest Assoc. of Nat. 37: 157–165. Eriksson, P. G. & S. E. B. Weisner, 1996. Functional differences in epiphytic microbial communities in nutrient-rich freshwater

ecosystems:an assay of denitrifying capacity. Freshwat. Biol. 36: 555–562. Fairchild, G. W. & R. L. Lowe, 1984. Artificial substrates which release nutrients:effects on periphyton and invertebrate succession. Hydrobiologia 114: 29–37. Fairchild, G. W., R. L. Lowe & W. B. Richardson, 1985. Algal periphyton growth on nutrient-diffusing substrates: an in situ bioassay. Ecology 66: 465–472. Fairchild, G. W. & A. C. Everett, 1988. Effects of nutrient (n, p, c) enrichment upon periphyton standing crop, species composition and primary production in an oligotrophic softwater lake. Freshwat. Biol. 19: 57–70. Fairchild, G. W., J. W. Sherman, & F. W. Acker, 1989. Effects of nutrient (n, p, c) enrichment, grazing and depth upon littoral periphyton of a softwater lake. Hydrobiologia 173: 69–83. Fairchild, G. W. & J. W. Sherman, 1992. Linkage between epilithic algal growth and water column nutrients in softwater lakes. Can. J. Fish. aquat. Sci. 49: 1641–1649. James, R. T., V. H. Smith & B. L. Jones, 1995. Historical trends in the Lake Okeechobee ecosystem: III. water quality. Arch. Hydrobiol. Monogr. Beit. 107: 48–69. Hansson, L-A., 1990. Quantifying the impact of periphytic algae on nutrient availability for phytoplankton. Freshwat. Biol. 24: 265– 273. Hansson, L-A., 1992. Factors regulating periphytic algal biomass. Limnol. Oceanogr. 37: 322–328. Havens, K. E., T. E. East, R. H. Meeker, W. P. Davis & A. D. Steinman. 1996. Phytoplankton and periphyton responses to in situ experimental nutrient enrichment in a shallow subtropical lake. J. Plankton Res. 18: 551–566. Havens, K. E, T. E. East, A. J. Rodusky & B. Sharfstein. 1999. Littoral periphyton responses to nitrogen and phosphorus: an experimental study in a subtropical lake. Aquat. Bot. 1295: 1–24. Hecky, R. E., P. Campbell & L. L. Handsale, 1993. The stoichiometry of carbon, nitrogen and phosphorus in particulate matter of lakes and oceans. Limnol. Oceanogr. 38: 709–724. Henry, R., J. G. Tundisi & J. S. B. Ribeiro, 1985. Responses of phytoplankton in Lake Jacaretinga to enrichment with nitrogen and phosphorus in concentrations similar to those of the River Solimoes (Amazon, Brazil). Arch. Hydrobiol Beih. Ergebn. Limnol. 103: 453–477. Hwang, S. J., K. E. Havens & A. D. Steinman, 1998. Phosphorus kinetics of planktonic and benthic assemblages in a shallow subtropical lake. Freshwat. Biol. 40: 1–17. Kjeldsen, K., 1996. Regulation of algal biomass in a small lowland stream: field experiments on the role of invertebrate grazing, phosphorus and irradiance. Freshwat. Biol. 36: 535–546. Kratzer, C. R. & P. L. Brezonik, 1984. Application of nutrient loading models to the analysis of trophic conditions in Lake Okeechobee, FL. Environ. Manag. 8:1–11. Lowe, R. L., 1996. Periphyton patterns in lakes. In: Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology. Academic Press, San Diego, CA: 57–76. Maltais, M. J. & W. F. Vincent, 1997. Periphyton community structure and dynamics in a subarctic lake. Can. J. Bot. 75: 1556–1569. Mazumder, A. & K. E. Havens, 1998. Nutrient-chlorophyll-secchi relationships under contrasting grazer communities of temperate versus subtropical lakes. Can. J. Fish. aquat. Sci. 55: 1652–1662. McCormick, P. V., P. S. Rawlik, K. Lurding, E. P. Smith & F. K. Sklar, 1996. Periphyton-water quality relationships along a nutrient gradient in the northern Florida Everglades. J. n. am. Benthol. Soc. 15: 433–449.

39 McCormick, P. V. & M. B. O’Dell, 1996. Quantifying periphyton responses to phosphorus in the Florida Everglades: a synopticexperimental approach. J. n. am. Benthol. Soc. 15: 450–468. Neiderhauser, P. & F. Schanz, 1993. Effects of nutrient (n, p, c) enrichment upon the littoral diatom community of an oligotrophic high-mountain lake. Hydrobiologia 269: 453–462. Phlips, E. J., F. J. Aldridge, P. Hansen, P. V. Zimba, J. Ihnat, M. Conroy & P. Ritter, 1993. Spatial and temporal variability of trophic state parameters in a shallow subtropical lake (Lake Okeechobee, Florida, U.S.A.). Arch. Hydrobiol Beih. Ergebn. Limnol. 128: 437–458. Phlips, E. J., M. Cichra, K. E. Havens, C. Hanlon, S. Badylak, B. Reuter, M. Randall & P. Hansen, 1997. Relationships between phytoplankton dynamics and the availability of light and nutrients in a shallow subtropical lake. J. Plankton Res. 19: 319–342. SAS, 1990. SAS/STAT users guide, version 6, third edition. SAS Institute Inc., Cary, NC. Schelske, C. L., 1989. Assessment of nutrient effects and nutrient limitation in Lake Okeechobee. Wat. Res. Bull. 25: 1119–1130. Steinman, A. D. & G. A. Lamberti, 1996. Biomass and pigments of benthic algae. In: Hauer, F. R. & G. A. Lamberti (eds), Stream Ecology: Field and Laboratory Exercises. Academic Press, San Diego, CA: 295–313. Steinman, A. D., R. H. Meeker, A. J. Rodusky, W. P Davis & C. D. McIntire, 1997a. Spatial and temporal distribution of algal biomass in a large, subtropical lake. Arch. Hydrobiol. 139: 29– 50.

Steinman, A. D., R. H. Meeker, A. J. Rodusky, W. P. Davis & S-J. Hwang, 1997b. Ecological properties of charophytes in a large subtropical lake. J. n. am. Benthol. Soc. 16: 781–793. South Florida Water Management District (SFWMD), 1995. Comprehensive quality assurance plan. South Florida Water Management District, West Palm Beach, FL. USEPA, 1979. Methods for chemical analysis of water and wastes. United States Environmental Protection Agency, Washington, DC. USEPA, 1987. Handbook of methods for acid deposition studies. United States Environmental Protection Agency, Washington, DC. Wetzel, R. G., 1979. The role of the littoral zone and detritus in lake metabolism. Ergeb. Limnol. 13: 145–161. Winterbourn, M. J., 1990. Interactions among nutrients, algae and invertebrates in a New Zealand mountain stream. Freshwat. Biol. 23: 463–474. Zimba, P. V., 1995. Epiphytic algal biomass of the littoral zone, Lake Okeechobee, Florida (USA). Arch. Hydrobiol Beih. Ergebn. Limnol. 45: 233–240. Zimba, P. V., 1998. The use of nutrient enrichment bioassays to test for limiting factors affecting epiphytic growth in Lake Okeechobee, Florida: confirmation of nitrogen and silica limitation. Arch. Hydrobiol Beih. Ergebn. Limnol. 141: 459–468.