Water quality assessment and ecoregional comparison of a reservoir in east-central Indiana

Abstract

Popovičová, J. 2009. Water quality assessment and ecoregional comparison of a reservoir in east-central Indiana. Lake Reserv. Manage. 25:155–166.

This study assessed water quality of a reservoir in an agricultural watershed of east-central Indiana, examined the effects of a thermal and oxygen regime on cycling of nutrients, and compared the results to ecoregional data and reference guidelines. Two locations were monitored biweekly from May through September 2007 for pH, temperature, dissolved oxygen, alkalinity, Secchi disk transparency (SD), chlorophyll a, and nutrient concentrations. The reservoir did not stratify during the monitoring season, although both anoxia and reoxygenation of the hypolimnion were observed. These conditions affected nitrogen (N) and phosphorus (P) cycling because nitrification was found to occur in the hypolimnion, and both internal load and water column mixing affected the concentrations and distribution of P. The reservoir was characterized as a eutrophic water body based on SD, total N and chlorophyll a concentrations, while total P concentrations classified the reservoir as slightly hypereutrophic (TSI = 73). This P overload has shifted the system toward N-limiting conditions (molar TN:TP = 18). Comparison of trophic parameters to Ecoregion 55 data placed this reservoir within the 75th percentile, and all parameters exceeded the U.S. Environmental Protection Agency ecoregional reference guidelines. I discuss a potential restoration of this water body to comply with the ecoregional nutrient criteria and to avoid future deterioration associated with N limitation.

Key words:

• ecoregional comparison
• eutrophication
• mixing
• nutrient cycling

The consequences of cultural eutrophication negatively affect designated uses of lakes and reservoirs, and restoration of eutrophic or hypereutrophic conditions will depend on the natural levels of nutrients influenced by regional environmental characteristics (Omernik 1987, 1995, Fulmer and Cooke 1990, Heiskary et al. 1987, Smith et al. 2003). Ecoregions—land areas with similar land cover, land use, soils and natural vegetation (Omernik 1987)—have been employed to assess ecoregional differences in lake and stream water quality, establish reference conditions for nutrients for development of management strategies, and to determine restoration potential for lakes and reservoirs (Heiskary et al. 1987, Hughes and Larsen 1988, Fulmer and Cooke 1990, Smith et al. 2003, Dodds et al. 2006). However, the validity and success of this approach varies. Ecoregional characterization was successfully applied to lakes in Minnesota and reservoirs in Ohio and Kansas (Heiskary et al. 1987, Fulmer and Cooke 1990, Dodds et al. 2006). This approach, however, was not appropriate for classification of water quality distribution among Northeastern and Michigan lakes (Jenerette et al. 2002, Cheruvelil et al. 2008). A local monitoring program was found to be more accurate and useful for setting lake-specific conditions and management than the utilization of regional criteria due to large variations in nutrients and chlorophyll a relationship among lakes and reservoirs, even within the same region (Hatch 2003, Havens 2003).

Prairie Creek Reservoir, Indiana, serves for fisheries, recreation, and as a secondary drinking water supply. The reservoir is located in east-central Indiana in Ecoregion 55, Eastern Corn Belt Plains, where eutrophication problems are among the worst in the state of Indiana (IDEM 2006, Jones and Medrano 2006). About 75% of the land in the ecoregion is used for crops such as corn and soybeans. The ecoregion is characterized by gently rolling till plains, has few natural lakes, and the soil is poorly drained. Until 2005 water quality investigations at the reservoir were limited to one-time sampling of trophic parameters in 1996 and 2002 (Jones and Barnes 2005, Jones and Medrano 2006). In 2006 Popovičová (2008) found high concentrations of phosphorus and anoxic conditions; however, analyses of total nitrogen (TN) and total phosphorus (TP), the key parameters to determine trophic status, were not included. To date, analysis of phytoplankton composition has been almost nonexistent even though some cases of algal blooms had been reported to local laboratories. Jones and Medrano (2006) found that blue-green algae comprised 41% of the total plankton in 2002, and Cylindrospermopsis raciborskii was measured at 84,247 cells/mL although no toxin was detected (Jones and Sauter 2005). The goal of the present study was to assess eutrophication of Prairie Creek Reservoir, compare the results to the ecoregional data, and determine its restoration potential. Specific objectives were to assess the extent of anoxia and its effects on distribution and cycling of nutrients, determine the limiting nutrient, and analyze available data for Ecoregion 55 to compare the reservoir to ecoregional conditions and reference guidelines.

Materials and methods

Prairie Creek Reservoir (5.1 km²) is located in a watershed (44 km²) of the White River in east-central Indiana (Fig. 1). Land use in the watershed is dominated by agriculture (64.7%), and 13.3% is reserved for green space (DMMPC 2007). Soil is drained by subsurface tiles and surface drains. Mean reservoir depth is 4.7 m, maximum depth is 9.1 m and maximum volume of water is 23.9 × 10⁹ m³.

Figure 1 Location of Prairie Creek Reservoir and sampling sites.

Biweekly sampling and monitoring was performed in the north (8.5 m depth) and center (7 m depth) of the reservoir between 10:00 and 15:00 from May through September 2007 (Fig. 1). Samples for nutrients and alkalinity were collected from the epilimnion at 0.5 m depth and from hypolimnion at 0.5 m from the bottom using a horizontal beta sampler. Samples were transported in high-density polyethylene Nalgene^TM bottles on ice, stored at 4°C, and analyzed within 24 hours of collection (APHA 1998).

Temperature, dissolved oxygen (DO), and chlorophyll a were measured in-situ at 0.5-m intervals from surface to bottom using a Hydrolab Sonde DS5 (Hach, Inc., Loveland, CO). Secchi disk (SD) transparency was measured according to Carlson and Simpson (1996), and a pocket weather meter (Kestrel 2000, Nielsen-Kellerman, Boothwyn, PA) was used to measure wind speed at the time of sample collection. Chlorophyll a concentrations were calculated for the euphotic zone (euphotic depth = 1.7 × SD; Scheffer 2004). Alkalinity was analyzed by titration with a 0.02 N sulfuric acid, nitrates (NO₃-N) by the cadmium reduction method 4500-NO3-E, and ammonia (NH₃-N) by the salicylate method (APHA 1998). Samples for analysis of soluble reactive phosphorus (SRP) were filtered through a membrane filter (0.45 μm) and then analyzed by the ascorbic acid method 4500-P-E (U.S. EPA 1997, APHA 1998). Total nitrogen was analyzed by persulfate digestion, and TP was determined using the ascorbic acid method after digestion with acid persulfate. Quality assurance and quality control procedures were employed according to standard methods (APHA 1998).

Descriptive statistics and analysis of variance for determination of differences among sampling sites, months, and depths were determined using MiniTab® 15.1.0.0 (Minitab Inc., State College, PA). To satisfy the requirement of normal distribution and equal variances, data were tested for normality using the Kolmogorov-Smirnov test and the Leven test for homogeneity of variances. Due to lack of normality, all data were log-transformed. Spearman's rho correlation analysis was used to determine relationships between variables. Trophic state index (TSI) for SD, chlorophyll a, and TP was calculated for July–August data using Carlson's equations (Carlson 1977), and the Kratzer and Brezonik (1981) equation was used to calculate TSI based on TN.

Data for Indiana lakes and reservoirs in Ecoregion 55 were obtained from Jones and Medrano (2006; N = 36), and a median and range for each relevant parameter were calculated. The reference nutrient criteria for lakes and reservoirs in Level III Ecoregion 55 and Ecoregion VI were compared to the collected data (U.S. EPA 2000).

Results

Physicochemical water quality parameters

During the entire monitoring period, results for pH and water temperature did not violate Indiana water quality standards (IAC 2008). In situ pH was significantly higher in the epilimnion but no temporal differences were identified (Table 1). Hypolimnetic temperatures in the center location were higher, and the temperature gradient at both locations indicated lack of thermal stratification at the reservoir (dT/dz < 1°C; Stefan et al. 1996; Figs. 2 and 3). Although not statistically different, the center location had lower temperature gradients (0.01°C/m ⩽ dT/dz ⩽ 0.89°C/m) than did the north location (0.03°C/m ⩽ dT/dz ⩽ 0.94°C/m; Fig. 2 and 3).

Table 1 Water quality parameters measured at the reservoir from May to September 2007.

Parameters		N	Min.	Max.	Mean ± Std. Dev.
Alkalinity** (mg/L CaCO3)	E†	20	130.0	146.0	135.5 ± 5.4
	H ^††	20	132.0	178.0	150.0 ± 18.8
Conductivity (μS/cm)	E	22	278.5	364.9	328.6 ± 2.08
	H	22	298.6	467.4	364.6 ± 42.4
pH***	E	22	7.3	8.5	8.1 ± 0.3
	H	22	6.5	7.6	6.9 ± 0.2
DO (mg/L)***	E	22	4.67	18.62	9.74 ± 3.20
	H	22	0.25	5.89	1.14 ± 1.38
Temperature (°C)***	E	22	18.10	27.79	24.28 ± 2.77
	H	22	14.12	24.96	20.87 ± 3.25
NO3-N** (mg/L)	E	22	0.10	0.50	0.27 ± 0.11
	H	22	0.10	2.30	0.58 ± 0.52
NH3-N (mg/L)***	E	22	<0.01	0.66	0.13 ± 0.16
	H	22	0.03	3.80	0.66 ± 0.88
TN (mg/L)*	E	22	0.3	1.6	1.0 ± 0.4
	H	22	0.6	4.6	1.7 ± 1.2
SRP ^+ (mg/L)	E	22	<0.01	0.07	0.03 ± 0.02
	H	22	<0.01	0.14	0.05 ± 0.04
TP (mg/L)***	E	22	0.07	0.16	0.12 ± 0.02
	H	22	0.11	0.71	0.27 ± 0.20
SD (m)	—	22	0.70	0.95	0.85 ± 0.06
Chlorophyll a (μg/L)	EU ^++	22	10.74	31.03	17.70 ± 6.48

Figure 2 Thermal regime at Prairie Creek Reservoir in 2007.

Figure 3 Temperature gradient and anoxic conditions at the reservoir in 2007.

Epilimnetic alkalinity showed a decreasing trend at both locations from May until the end of September indicating photosynthetic activity. Hypolimnetic alkalinity increased significantly (p< 0.01) in July and August from that measured in the epilimnion and was slightly higher at the north location (Fig. 4). At the center location in mid-August the hypolimnetic alkalinity decreased to the original level measured in spring while at the north location high alkalinity persisted until the end of the month (Fig. 4). The results imply that decomposition was most intense in July and August and persisted longer at the north location.

Figure 4 Time series graph for a) alkalinity, b) nitrates c) ammonia-N, d) total nitrogen, e) soluble reactive phosphorus, f) total phosphorus, g) chlorophyll a, and h) chlorophyll a vs. Secchi disk. Symbols in graphs a) through f) represent: ◯ Center epilimnion, Δ North epilimnion. • Center hypolimnion, ▴ North hypolimnion; in graph g) • represents Center and ▴ North location.

The results for DO violated the Indiana water quality standard of 5 mg/L in 95% of the hypolimnion samples and in 5% of epilimnion samples (IAC 2008; Table 1). Epilimnetic DO concentrations did not exhibit significant temporal differences, although they were significantly higher (p< 0.001) than in the hypolimnion (Table 1; Fig. 5). Hypolimnetic DO concentrations measured on 4 June 2007 indicated the establishment of anoxic conditions with a DO concentration of 0.31 mg/L at the north and 0.32 mg/L at the center location. In general, anoxia persisted until the end of September, although several days of hypolimnetic reoxygenation and mixing were recorded (Fig. 5). At the north location DO concentration in the hypolimnion increased to 1.94 mg/L on 30 July, after which anoxia persisted to varying depths until the end of the monitoring season. At the center, more shallow location, hypolimnetic DO concentrations increased on 20 July (1.37 mg/L), 28 August (1.24 mg/L), and 11 September (5.89 mg/L), whereas 25 September was marked by anoxia (0.4 mg/L).

Figure 5 Vertical and temporal variability of dissolved oxygen concentrations at Prairie Creek Reservoir in 2007.

Trophic parameters

The months characterized by anoxic conditions in the reservoir were also characterized by increasing concentrations of nutrients (Figs. 4 and 5). With one exception, NO₃-N concentrations met the Indiana water quality standard for aquatic life (1.6 mg/L, n = 44; Table 1; IAC 2008). Epilimnetic NO₃-N concentrations showed little temporal difference even as chlorophyll a concentration increased (Fig. 4); however, hypolimnetic concentrations were significantly higher (p< 0.05; Table 1) and increased throughout the monitoring period, especially, at the center location where reoxygenation of the hypolimnion occurred more frequently. The maximum NO₃-N concentration was measured on 25 September (2.3 mg/L) at the center following the increase of hypolimnetic DO concentration to 5.89 mg/L (Fig. 5).

Concentrations of NH₃-N in the hypolimnion and epilimnion were significantly different (p < 0.001; Table 1; Fig. 4), although temporal differences were found only in the hypolimnion (p < 0.01). Epilimnetic NH₃-N concentrations increased at the end of the season when mixing of the water column was observed. Increased hypolimnetic NH₃-N concentrations were measured in July and August, and the maximum was reached in August at the north location (3.8 mg/L). On 25 September concentrations approached those found in May, although NH₃-N production in the hypolimnion was still occurring. Concentrations of TN followed the temporal trend of NO₃-N and NH₃-N with a greater temporal variability in the epilimnion. Hypolimnetic TN concentrations were significantly greater (p < 0.05) than in the epilimnion (Table 1; Fig. 4) and, in general, had a trend similar to NH₃-N in that concentrations increased until July, then declined, and the maximum was reached at the end of August at the north location (Fig. 4). Concentrations of TN characterized the reservoir as eutrophic (TSI = 56; Kratzer and Brezonik 1981).

Contrary to the distribution of N compounds, the soluble reactive phosphorus (SRP) concentrations increased in both epilimnion and hypolimnion as summer progressed, and hypolimnetic concentrations exceeded those found in the epilimnion starting mid-July (Table 1; Fig. 4). The increased SRP concentrations in the hypolimnion are typical of anoxic conditions that began in late May and, with the exception of some mixing periods, persisted until the end of August in the center and until September in the north. Concentrations of TP in the hypolimnion were significantly higher than in the epilimnion (p < 0.001; Table 1; Fig. 4), even though only hypolimnetic TP concentrations showed significant temporal differences (p < 0.05) and reached their maximum in July (Table 1; Fig. 4). High concentrations of TP found at the reservoir characterized this water body as slightly hypereutrophic (TSI = 73; Carlson 1977).

Algal biomass, as measured by chlorophyll a concentration, reached its minimum in June when “stratification” (dT/dz) was greatest and increased until the end of the monitoring period, reaching its maximum in September (Fig. 2, 3 and 4). The late spring decline measured in June could be attributed to the clear-water phase that often occurs at the end of spring (Lampert et al. 1986, Scheffer 2004) due to depletion of available nutrients caused by the spring bloom of diatoms. Although not significantly different, chlorophyll a concentrations were slightly higher at the center location, where more mixing occurred and higher SRP concentrations were measured (Fig. 4). Concentrations of chlorophyll a characterized the reservoir as a eutrophic water body (TSI = 58). Secchi disk transparency correlated inversely with chlorophyll a concentrations (r²= 0.67, p< 0.001; Table 1; Fig. 4). Maximum water transparency was measured in June (0. 93 m) and the minimum in August (0.80 m; Fig. 4), although no significant temporal differences were found. Similar to TSI calculated from TN and chlorophyll a, SD results were characteristic of a eutrophic reservoir (TSI = 62).

Discussion

The results of this study demonstrate distribution and cycling of nutrients in a polymictic reservoir where both anoxic periods and intermittent mixing were observed. A 2006 monitoring study found weak stratification of the water column in July (dT/dz = 1.13 and 1.3°C/m in the center and north location, respectively; Popovičová 2008); however, this was not the case in 2007. This type of thermal regime would be expected for the geometry ratio (A^0.25: z_max) of 5.2 (Stefan et al. 1996). Contrary to the findings of Stefan that anoxic conditions never occur in lakes with the ratio >5, however, anoxia was observed in this reservoir. Furthermore, reoxygenation of the hypolimnion was observed, especially at the shallower center location, which can be attributed to lack of thermal stratification, depth, and contribution of additional factors such as wave action and mixing caused by wind and motorboats that control vertical mixing (Scheffer 2004, Kelton and Chow-Fraser 2005). The first two weeks of July at this reservoir experienced especially heavy motorboat traffic (Fig. 5). Local weather data revealed that storm events occurred prior to monitoring days when hypolimnetic reoxygenation was measured. Precipitation was recorded 18 July (5.6 mm), 19 July (16 mm), 16–25 August (8.9–60.7 mm), and 7–9 September (3.8–41.6 mm). Major storms can have significant effects on water column mixing (Herb and Stefan 2004) and, consequently, can induce oxygenation of the hypolimnion (Fig. 5). Depth may play a role as well because the center, more shallow location (∼7 m) was characterized by a lower temperature gradient and stronger winds (mean speed of 8.85 km/h vs. 8.05 km/h) making it more susceptible to mixing and, thus, hypolimnetic reoxygenation (Fig. 3), even though significant differences in DO concentrations were not found between the monitored locations.

The alternating periods of anoxia and mixing affected the cycling and distribution of nutrients throughout the water column. In shallow reservoirs water mixing can enhance exchange of nutrients between the sediment and the water column (Scheffer 2004). Increased epilimnetic concentrations during the monitoring season were observed especially for SRP, although NO₃-N and TN showed some variability as well (Fig. 4). Increased concentrations of NO₃-N in the hypolimnion during the season were the consequence of organic matter decomposition and subsequent nitrification of ammonia due to reoxygenation of the water column. Ammonification of detritus creates suitable conditions for nitrification when sufficient oxygen becomes available at the top of the sediment (Lijklema 1994, Scheffer 2004). In stratified lakes the zones of nitrification move throughout the water column depending on the presence of oxygen and ammonia, and move to the sediment-water interface when stratification ceases to occur (Stewart et al. 1982). In shallower, mixed lakes (e.g., Prairie Creek Reservoir), however, nitrification mainly occurs at the sediment-water interface in the presence of any available oxygen (Stewart et al. 1982, Lijklema 1994, Scheffer 2004). Increased NO₃-N concentrations in the hypolimnion are especially evident in the center, shallower location where several events of reoxygenation were measured (Figs. 4 and 5). Additionally, nitrification can be enhanced by diurnal temperature fluctuations that can induce vertical mixing of the water column, and, consequently, oxygen replenishment (Herb and Stefan 2004, Scheffer 2004). However, the biweekly water column measurements of temperature (T) and DO are not adequate for investigation of diurnal variations at this reservoir. The results of this study demonstrate that nitrification occurred in the hypolimnion as a result of water column mixing and reoxygenation.

High concentrations of NH₃-N in the hypolimnion are expected when anoxic conditions and higher amounts of decaying algae are present as summer progresses (Scheffer 2004). Increased production of NH₃-N due to ammonification in the bottom sediment resulted in subsequent increase in concentrations of NO₃-N throughout the season. Concentrations of NH₃-N increased significantly in July and then decreased as the season progressed with one exception recorded at the north location in late August. Hypolimnetic NO₃-N concentrations, however, increased throughout July and August due to nitrification of ammonia generated when dissolved oxygen was present as a result of mixing. Concentrations of about 1 mg/L of DO were found to support nitrification, while conditions favorable to denitrification are around 0.2 mg/L (Knowles et al. 1981). In this study the concentrations of DO in the hypolimnion increased above 1 mg/L on several occasions, leading to nitrification (Figs. 4 and 5).

Vertical mixing of water and variations in DO concentrations at the sediment-water interface also influence P cycling between the sediment and water column (Herb and Stefan 2004). In shallow lakes this mixing can have two effects on P fate (Scheffer 2004). Periods of reduced mixing and turbulence can lead to anoxic conditions at the sediment surface that cause release of P (Nürnberg and Peters 1984, Nürnberg 1994, Scheffer 2004, Kelton and Chow-Fraser 2005), while during periods of increased mixing P is distributed throughout the water column. Resuspension and oxygenation of sediment is related to lake size and depth, and this reservoir is classified as a water body where both forms of P release occur (Scheffer 2004). This is supported by the present study where internal P release, as well as increased epilimnetic concentrations of SRP, were observed, indicating the contribution of P from the sediment and hypolimnion as a result of anoxic conditions and intermittent mixing (Figs. 2, 4 and 5). The estimated internal P load calculated from the in situ summer increases in water column TP concentration was 1018 mg/m² at the north and 1803 mg/m² at the center location based on the maximum water column concentrations measured in July at both locations (Nürnberg 1987).

Although high summer TP concentrations characterized this reservoir as slightly hypereutrophic, the symptom of excessive algae scum was not observed in the reservoir. The absence of extensive algal growth, which would be expected at the measured TP concentrations, could be attributed to several factors such as lack of stratification, mixing of the water column, reoxygenation of the hypolimnion, nutrient limitations, and nonalgal turbidity. First, in a lake with a mixed water column, epilimnetic concentrations of chlorophyll a are lower than in a stratified lake because algae is not confined to the surface as it is in a stratified lake (Havens 2003). Furthermore, historical data suggest an increasing trend in TP loading at the reservoir (Jones and Barnes 2005) that has caused a shift in nutrient limitation (Table 2). The mean summer TN:TP ratio of 18 indicates a co-limitation of N and P and a shift toward N-limiting conditions (Elser et al. 1990, Guildford and Hecky 2000, Dzialowski et al. 2005). A decrease in N/P ratio develops with increasing eutrophication and can be related to higher P load from the watershed (Lijklema 1994). Other studies found that lakes initially identified as P-limited had shifted to N-limitation due to excessive P loading (Nürnberg and Peters 1984, Elser et al. 1990, Havens 1995, James et al. 1995, Havens et al. 1996). Finally, nonalgal turbidity can play a significant role in phytoplankton production due to light limitation (Scheffer 2004, Carlson and Havens 2005). Indiana lakes and reservoirs were found to be affected by nonalgal turbidity (Jones and Medrano 2006); however, the results for this reservoir showed that nonalgal turbidity does not significantly limit algal growth because the correlation between TSI(SD) and TSI(Chlor) is stronger (r²= 0.65, p < 0.001) than correlation between TSI(SD) and TSI(TP) (r²= 0.26, p = 0.245; Carlson and Havens 2005). Additionally, differences between the TSI results for TN, SD and chlorophyll a fit in the upper left quadrant of a deviation plot, suggesting the effect of nitrogen (Fig. 6.) Therefore, algal growth at the reservoir is most likely limited by nitrogen rather than nonalgal turbidity, although bioassay analyses would be useful to confirm these findings in the future (Carlson and Havens 2005).

Table 2 Historical changes in trophic levels at Prairie Creek Reservoir.

	1996 †	2002 *	2007
TSI (TN) ^+	53	54	56
TSI (TP) ^††	59	68	73
TSI (SD) ^††	50	65	62
TSI (Chlor a) ^††	44	61	58

Figure 6 TSI deviation plot for Prairie Creek Reservoir in 2007. Data for 2002 are from Jones and Medrano (2002).

Comparison to lakes and reference conditions of Ecoregion 55

Compared to other Indiana lakes and reservoirs in Ecoregion 55, the percent of water column under oxic conditions was higher in Prairie Creek Reservoir than in 50% of water bodies of this ecoregion (Table 3). However, these ecoregional conditions only reflect an instantaneous state measured on the day of monitoring that can change depending on weather patterns, wind, boat traffic, and stratification (Scheffer 2004, Kelton and Chow-Fraser 2005, Knowlton and Jones 2006). The lowest DO concentrations at Prairie Creek Reservoir were measured in June when about 40% of the bottom water column was anoxic; similar conditions were also observed in 2006 (Popovičová 2008). Considering the nutrient concentrations in this reservoir, the lack of thermal stratification alleviated anoxia, reduced the extent of the anoxic layer, and prevented an even a greater release of TP from the sediment (Scheffer 2004).

Table 3 Comparison of Prairie Creek Reservoir 2007 data to other Indiana lakes and reservoirs in Ecoregion 55 and the reference values.

	Ecoregion 55 † (N = 36)	Reference	Reference	Prairie Creek Reservoir ^††
Parameters	Median/Mean	conditions *	conditions ^+	Median/Mean
Water column oxic (%)	52 / 57 (13–100)	—	—	81
Secchi Disk (m)	0.9 / 1.4 (0.3–6.2)	0.87	1.36	0.85/0.84
Nitrate (mg/L)	0.022 / 0.589 (0.013-3.853)	0.257**	—	0.30/0.32
Ammonia (mg/L)	0.018 / 0.04 (0.013–0.163)	—	—	0.10/0.156
TP (mg/L)	0.067 / 0.077 (0.012–0.246)	0.035	0.0375	0.123/.117
TN (mg/L)	1.450 / 1.797 (0.437–4.643)	0.782	0.781	1.25/1.08
Chlorophyll a (μg/L)	6.69 / 17.0 (0.18–135.7)	5.27	8.59	13.38/15.95

The Secchi disk transparency in the reservoir was only slightly lower than the median for Ecoregion 55 (Table 3). Concentrations of TN were within the 50th percentile of the ecoregional levels, while NO₃-N, NH₃-N, TP, and chlorophyll a concentrations were within the 75th percentile. A review of the ecoregional data acquired between 1999 and 2003 (Jones and Medrano 2005) showed that in 2002 the reservoir had better water quality in comparison to other lakes and reservoirs of the same ecoregion; nitrates were within the 25th percentile, TN and TP values were within the 50th percentile, and chlorophyll a was in the 75th percentile of the ecoregional data. In 2005 Jones and Barnes reported that the Indiana TSI (ITSI) for this water body, a scoring system developed specifically for lakes and reservoirs in the state, changed from mesotrophic (ITSI = 24) to hypereutrophic status (ITSI = 55; Jones and Barnes 2005, Jones and Medrano 2006).

The SD transparency decreased significantly from 2.7 m in 1996 to 0.7 m in 2002 (Jones and Barnes 2005). These results demonstrate degradation of the reservoir water quality due to eutrophication between 1996 and 2007 that placed the reservoir among 10% of hypereutrophic water bodies in the state (Table 2). This decrease in water quality is a result of nutrient loading, increased load of suspended solids, and absence of adequate land management practices. Historical trends in TSI (Carlson 1977) show that while levels of N remained stable between 1996 and 2007, P load has increased considerably, consequently affecting SD transparency and chlorophyll a concentrations, although SD transparency and chlorophyll a concentrations did not change significantly between 2002 and 2007 (Table 2). These trends also explain the shift toward the N-limited system (Guildford and Hecky 2000, Dzialowski et al. 2005) because N limitation has been found to occur in lakes with excessive P input (Nürnberg and Peters 1984, James et al. 1995, Havens et al. 1996). This finding is of great consequence for the management of this reservoir because N-limiting conditions produce blooms of green and blue-green algae that can generate toxins and obnoxious taste and odor compounds (Havens 1995, Havens et al. 1996, Johnston and Jacoby 2003). Decreasing N inputs, however, might not decrease cyanobacteria growth in N-limited systems (Scheffer 2004), and management of both N and P is necessary to control production of cyanobacteria (Dzialowski et al. 2005). Current management practices in the reservoir watershed are limited to only minimum tillage agriculture; use of other strategies such as buffer strips, livestock fencing, use of constructed wetlands at the reservoir inlets and in-lake management practices have not been implemented.

To determine the reservoir restoration potential, water quality parameters were compared to the reference guidelines and the corresponding TSI proposed by the U.S. EPA (2000) because Indiana does not have sufficient data for development of its own criteria. All parameters measured in 2007 exceeded the reference guidelines for the ecoregion (Tables 2 and 3), and the attainment of the proposed levels would restore the reservoir water quality to the level found in 1996 (U.S. EPA 2000, Jones and Sauter 2005; Tables 2 and 3). Such comparisons, however, could be negatively affected by inadequate long-term data for this water body needed to account for temporal variability. Knowlton and Jones (2006) showed that reservoirs with long-term means for TP and chlorophyll a concentrations in compliance with the reference nutrient criteria often exceeded those criteria when individual values or seasonal averages were considered. Numeric criteria should therefore be applied only to long-term averages; otherwise, compliance with the criteria can be misclassified. Consequently, monitoring at this reservoir should continue to obtain at least four seasons of data to assess typical conditions and variability (Knowlton and Jones 2006).

The suitability of an ecoregional approach to guide restoration and management of Indiana lakes and reservoirs has not been established. Ecoregional comparisons performed in other states have used extensive databases of lakes data spanning several years (Jenerette et al. 2002, Havens 2003, Dodds et al. 2006, Knowlton and Jones 2006, Cheruvelil et al. 2008), while Indiana lakes and reservoirs were monitored once between 1994 and 1998 and once between 1999 and 2003. In both monitoring periods only one measurement was performed per water body (Jones and Sauter 2005, Jones and Medrano 2006). Although Jones and Medrano (2006) attempted to classify lakes and reservoirs according to ecoregions, significant differences among them were not established. Additionally, the results for Ecoregion 55 demonstrate very high variability, especially in SD and chlorophyll a concentrations (Table 3). If reference conditions were based on these data, reference levels for chlorophyll a would be set at 0.87 μg/L (lower 25th percentile) or 0.66 μg/L (trisection method), and reference conditions for SD would be set at 1.8 m (25th percentile) or at 2.7 m (trisection method); such conditions, especially for chlorophyll a, would be unattainable for this reservoir.

Prairie Creek Reservoir showed high temporal variability of nutrients caused by lack of stratification and intermittent reoxygenation of the hypolimnion. Trophic parameters showed worsening of eutrophic conditions and development of N limitation. Comparison of trophic variables to the ecoregional levels placed the reservoir in the 75th percentile. Restoration would be possible if management practices were implemented to prevent further deterioration and to avoid potential cyanobacterial blooms in the future that could threaten its uses for recreation and drinking water (Dodds et al. 2006). However, the comparison analysis was limited by availability of historical data for this reservoir as well as Ecoregion 55. Therefore, further monitoring is necessary to establish ecoregional reference conditions for Indiana lakes and reservoirs and to assess significant differences in water quality among ecoregions. Such assessment should also consider classification of lakes and reservoirs within smaller groups, such as eight-digit hydrologic units (Cheruvelil et al. 2008) or subregions (Hughes and Larsen 1988, Smith et al. 2003) due to large variability of reference data and natural background concentrations (Smith et al. 2003). Additionally, further long-term investigation of this reservoir would be beneficial to assess temporal variability of trophic parameters and determine more accurate lake-specific management strategies. Finally, an investigation of diurnal changes in temperature, dissolved oxygen and nutrients would be valuable to gain a better understanding of nitrification-denitrification processes and nutrient cycling in the reservoir.

Acknowledgments

The author is grateful to Diana Fiallos Celi for her work on field monitoring and laboratory analyses. This project was funded by Ball State University (BSU), Indiana Academy of Science, and a Lilly V grant.

Notes

†Epilimnion.

^††Hypolimnion.

⁺Soluble reactive phosphorus.

⁺⁺Euphotic zone.

*Depth difference significant at p< 0.05.

**Depth differences significant at p< 0.01.

***Depth differences significant at p< 0.001.

†data from Jones and Sauter (2005).

*data from Jones and Medrano (2006).

⁺calculated from Kratzer and Brezonik (1981).

^††calculated from Carlson (1977).

†epilimnetic values from July and August for the ecoregion (data from Jones and Medrano 2006); data in parentheses denote minimum and maximum.

*reference guidelines for Ecoregion 55; median value from 4 seasons—25th percentile % for each season (U.S. EPA 2000).

⁺reference guidelines for Ecoregion VI (U.S. EPA 2000).

^††July and August epilimnetic values.

**nitrate + nitrite.