ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on July 25, 2008
ICES Journal of Marine Science: Journal du Conseil 2008 65(8):1469-1474; doi:10.1093/icesjms/fsn121
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
This article appears in the following ICES Journal of Marine Science issue: Marine Environmental Indicators: Utility in Meeting Regulatory Needs [View the issue table of contents]
Development and application of seasonal indices of coastal-zone eutrophication
1 K. C. Irving Environmental Science Centre, Acadia University, Wolfville, NS, Canada B4P 2R6
2 Department of Earth and Environmental Science, Acadia University, Wolfville, NS, Canada B4P 2R6
3 Department of Fisheries and Oceans, Ecosystem Research Division, Bedford Institute of Oceanography, PO Box 1006, Dartmouth, NS, Canada B2Y 4A2
Correspondence to S. A. Ryan: tel: +1 902 5851687; fax: +1 902 5851034; e-mail: 038004r{at}acadiau.ca
Ryan, S. A., Roff, J. C., and Yeats, P. A. 2008. Development and application of seasonal indices of coastal-zone eutrophication. – ICES Journal of Marine Science, 65: 1469–1474.Indices of coastal-zone eutrophication were developed based on combined values of nitrogen, phosphorus, and phytoplankton biomass (measured as Chl a). Measurements of nitrate, ammonia, phosphate, Chl a, and total phosphorous were taken at a series of inlets (bays and estuaries) along the Atlantic coast of Nova Scotia and the north shore of Prince Edward Island in summer 2006. Reference data from offshore sites and measured data are displayed on a N (nitrogen):P (phosphorus) nutrient–space diagram, which clearly indicates thresholds between impacted and unimpacted inlets. The problem of seasonal variability in nutrient levels and phytoplankton concentrations is accounted for by using an amalgamation of the Redfield nutrient ratios and the coastal carbon–Chl a ratio and allowing for sedimentary losses. This amalgamated ratio allows for the inverse seasonal relationship between labile nutrient levels and Chl a, permitting the collection, comparison, and interpretation of seasonally variable data. Measured concentrations of phosphate were adjusted from measured Chl a and were compared with measured levels of total phosphorous for the Nova Scotian sites to assess the accuracy of conversions. Given regional calibration, these indices should be applicable to all coastal waters.
Keywords: chlorophyll, coastal inlets, eutrophication, indicators, Nova Scotia, nutrients, Prince Edward Island, water quality
Received 23 November 2007; accepted 6 May 2008; advance access publication 25 July 2008.
 |
Introduction |
|---|
 Top Introduction MethodsResultsDiscussionReferences |
|---|
As global populations climb and coastlines are inundated with human activity, the problem of coastal marine eutrophication grows. A recent study by the United Nations Environment Programme (UNEP, 2006) identified coastal marine eutrophication as a red-flag issue because of its prevalence along the world's coastlines and the lack of action currently being taken to remedy it. Behind carbon dioxide, anthropogenic nitrogen pollution is one of the major problems facing the world today. Bricker et al. (1999) state that significant signs of eutrophication have been reported globally, and major affected areas include the Baltic, Adriatic, and Black seas, and the coastal and estuarine waters of Japan, China, Australia, and the US.
If the problem of coastal marine eutrophication is to be fully understood by researchers, and perhaps more important by policy-makers, a simple but effective index of eutrophication must be developed that does not rely on labour-intensive methods and that, when possible, makes use of existing data. The index presented here is a broadly applicable diagnostic index of eutrophication for coastal marine environments and is designed to be easy to use and interpret. The index we propose follows the lead of the OSPAR Commission (1997) by establishing reference conditions for nearshore nutrient concentrations using offshore winter data. These reference conditions are initially used in combination with the generalized 16N (nitrogen):1P (phosphorus) Redfield (1958) ratio, to construct boundaries representing the maximum nutrient concentrations expected in unimpacted nearshore waters. These calculated boundaries are then plotted on a total inorganic nitrogen (TIN) vs. phosphate nutrient–space diagram, which can be used as a preliminary index of eutrophication for the nearshore environment.
However, indices of eutrophication that are based solely on inorganic nutrient concentrations tend to be unable to account for seasonal variation in temperate regions, owing to the influence of biological activity, particularly phytoplankton production. Ideally, indices that are developed using labile nutrient concentrations should be limited to data collected during winter, when nutrient levels are at their annual maximum. However, the problems associated with including only winter data are threefold in northern temperate regions: freezing temperatures make sampling within bays and estuaries hazardous; many of the land-based sources of nutrients are substantially reduced at this time of year; and much of the existing data for a region, typically collected during summer, are not reliable for indexing eutrophication. Given these limitations, it is clear that modifications to an index based on nutrients alone are required to allow data from any season to be used for the interpretation of coastal eutrophication.
We address the issue of seasonally variable nutrient concentrations through the addition of stoichiometric Chl a data to the Redfield N and P ratios, then compare these results with the measurements of total P to determine whether or not the Chl a conversion adequately compensates for the presence of non-labile nutrients. Amalgamating the regional carbon–Chl a ratio with the Redfield ratios allows us to estimate the equivalent nutrient content of phytoplankton biomass throughout seasonal cycles, mitigating the effects of seasonal variability reflected in the inverse relationship between nutrient concentrations and biological activity. This should allow the eutrophication status of coastal waters to be more easily determined, irrespective of seasonally variable nutrient and Chl a data and time of assessment.
|
Methods |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
To obtain an offshore reference condition dataset, a query was run on the BioChem database (DFO, 2006) for December, January, and February, the time of year when labile nutrient concentrations are at their maximum. Data were retrieved for the Scotian Shelf on TIN (NO3 + NO2 + negligible NH3) and phosphate (PO4) from depths <60 m. The resulting data contained points collected between 1969 and 2005. The raw data were imported into ESRI ArcGIS Desktop 9.1, and spatially selected using a polygon extending
20 km out from the Nova Scotian coastline (to avoid direct coastal influence) to the edge of the Scotian Shelf (Figure 1). The area offshore of southwest Nova Scotia (NS) was excluded because of upwelling in this region, and associated nutrient enrichment makes it unsuitable for present purposes. The selected polygon was subdivided into four subregions of similar size, serving as reference for further analysis: north (A and B), south (C and D), east (B and D), and west (A and C).
|
To allow for some outliers in the BioChem data, we chose the 90th percentiles of nutrient concentrations for TIN and phosphate within the selected offshore areas, to serve as the appropriate boundaries for plotting on the nutrient–space diagram. The percentile to be used might be increased or decreased depending on the quality of the dataset being used. In this case, choosing the 95th percentile rather than the 90th would result in a 10% increase in the nutrient concentration selected. This relatively small difference indicates that the number of anomalously high concentrations is quite low. From the 90th percentile point for N and P, straight lines were drawn, intersecting vertically with the x-axis and horizontally with the Redfield 16N:1P line, to form a polygon reflecting offshore source-water conditions, within which we expect to find nutrient data from unimpacted coastal sites. Sites where N and P concentrations fall outside this polygon are presumably impacted by nutrients from coastal or land sources and therefore rank as eutrophic. Greater distance from the polygon indicates a greater degree of eutrophication.
To illustrate applications of this methodology, two datasets were examined. First, data from nutrient-monitoring studies in St Margaret's Bay, NS (DFO, 2006), were plotted on the nutrient–space diagram. Second, the initial results from the Strain and Yeats' (1999) eutrophication survey of 34 inlets along the east coast of NS were compared with the results obtained by assessing the same dataset using the nutrient–space index. Strain and Yeats (1999) calculated a numerical eutrophication index based on principal component analysis, which is equal to the sum of the log of the individual concentrations of the following parameters, multiplied by their respective component loadings (in parentheses): silica (0.74), phosphate (0.92), ammonia (0.72), oxygen (–0.85), iron (0.86), and manganese (0.89). For the comparison, we developed a metric for use with the nutrient–space diagram to act as a measure of distance from the polygon of reference conditions. The formula for this metric was [(N – Nr)/16 + (P – Pr)], where N and P are the TIN and phosphate concentrations of the sample, respectively, and Nr and Pr are the boundaries of the offshore reference polygon for the two nutrients. The division by the Redfield ratio (16) was introduced to normalize the scale of N relative to P. Results for the two indices were then ranked, and a Friedman test was conducted to determine whether or not the rankings were statistically different.
Water samples were collected from small boats along the Atlantic coast of NS and the nutrient-laden waters along the north shore of Prince Edward Island (PEI) between 23 March and 19 August 2006. During this period, 17 NS sites were sampled three times each, and single samples were collected from six PEI sites (Figure 1). During each visit to the NS sites, surface (1-m depth) samples were collected from the approximate middle of the inlet, and near-bottom (1 m above sediment) samples were collected as close as possible (subject to weather conditions) to the headlands of the inlets. Only mid-inlet samples were collected from the PEI sites. Water samples were collected using a Wildco 2.2L Horizontal Alpha Water Sampler, and subsamples were taken for the analysis of nutrients, salinity, and at the NS sites, Chl a. Samples for nutrient analysis were stored in 30 ml wide-mouth Nalgene bottles, which had previously been soaked in 10% ACS grade HCl for 2 h and rinsed with deionized water (Milli-Q Gradient). Before being stored in the bottle, water samples collected for inorganic nutrient analysis were filtered through a 25 mm Millipore APFC glassfibre filter. A second subsample of 500–1000 ml was filtered through a 47 mm Millipore APFC glassfibre filter. The filter was wrapped in a large piece of filter paper, and again in tinfoil, to be saved for Chl a analysis. Following collection, all bottles and filters were immediately placed in a 9-V portable cooler and frozen at –20°C within 6 h of collection, until they could be analysed in the laboratory. Salinity was measured in situ using a Hydrolab Minisonde 4a.
All nutrient analyses were performed using a Technicon Autoanalyzer II system. Methods used for analysing nitrate + nitrite and phosphate can be found in Strain and Clement (1996), and those for ammonia in Kérouel and Aminot (1997). Total phosphorous samples were digested following ESS Method 310.2 (USEPA, 1992), except that an Accu-Test C. O. D. Heater Block (Bioscience 163–466) was used instead of an autoclave and analysed using the phosphate manifold. Duplicates of each sample were analysed in runs of 42 with standards at the beginning of runs and blanks placed every six samples. Chl a samples were analysed fluorometrically in the laboratory according to USEPA Method 445.0 (USEPA, 1997).
Once nutrient and Chl a data had been collected, TIN was calculated by summing the molar concentrations of nitrate and ammonia. Concentrations of TIN and PO4 were adjusted to allow for the inverse relationship between nutrients and Chl a, using an amalgamation of the molar Redfield (1958) ratios (106C:16N:1P) and two C–Chl a ratios representing the upper and lower values of the expected range for this area, which were then corrected for phytoplankton biomass loss caused by sedimentation, to determine which is more suitable for use here. The issue of C–Chl a ratios has been considered in many publications, and it is necessary to explain our rationale behind the selection of these two ratios. Although the limitations of determining this ratio for marine phytoplankton have been widely discussed (Zeitzschel, 1970; Banse, 1977; Li et al., 1993), a range of 30:1–75:1 seems to be accepted for the North Atlantic Ocean at a latitude of 47°N, depending on time of year and nutrient conditions (Taylor et al., 1997). Values at the lower end of this range occur during autumn, winter, and spring, when nutrient levels are relatively high and light availability is low, whereas higher ratios occur during summer when nutrients are low and light availability is high. In addition to the conversion from labile to particulate forms, loss of nutrients from the water column in particulate form also occurs during spring and summer through sedimentation. Based on work by Hopcroft et al. (1990), we assumed a rate of loss of phytoplankton biomass (as Chl a) of 25% per year. The carbon component of the upper and lower ends of the range mentioned above (30:1 and 75:1) was divided by 0.75 (1–0.25) to account for this loss in phytoplankton biomass, which resulted in new limit ratios of 40:1 and 100:1. The resulting amalgamated molar ratios and nutrient–Chl a conversion factors are:
- C:Chl a = 40:1–106C:16N:1P:0.036Chl a; N = 444*Chl a;P = 27.8*Chl a
- C:Chl a = 100:1–106C:16N:1P:0.014Chl a; N = 1143*Chl a; P = 71.4*Chl a
Therefore, the molar Chl a concentration for a given sample is multiplied by the N and P conversion factors, and the results are added to the appropriate labile nutrient concentration. Subsequently, the eutrophication status of the inlets was assessed by the N:P nutrient–space diagrams, using values either adjusted or not adjusted for Chl a conversion and losses.
|
Results |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
The initial query of the BioChem database yielded 646 data points within the offshore reference subregions. Because the northern subsections (A and B in Figure 1) of the offshore reference area are most likely to supply source water to the embayments being studied because it is closer to the coastline, these data were used to construct the boundaries on the nutrient–space diagram (Figure 2).
|
Of the year-round nutrient concentrations from St Margaret's Bay (generally regarded as relatively pristine), 98.8% of the 894 points fell within the nutrient–space reference polygon, whereas the remaining 11 points all fell close to the boundaries. Analysis of the Strain and Yeats (1999) dataset using the nutrient–space diagram revealed that sites identified as being minimally impacted in their original study fell within or very close to the reference condition polygon, whereas sites identified as being highly impacted fell outside the polygon. A Friedman test performed on the ranked indices from the two studies revealed no significant difference (p = 0.86) between the two rankings.
Because seasonally comprehensive data on TIN, PO4, and Chl a for the NS nearshore environment do not exist, we used data from the offshore reference area for an analysis of seasonal changes in nutrient and Chl a concentrations, and to illustrate the effects of adjusting for conversion of Chl a and sedimentary losses. The means of all observations containing PO4, NO3, and Chl a concentrations were plotted for each month in the four subsections of the roughly 500 x 200 km offshore reference area (Figure 1, A–D) to determine whether or not trends would differ within the reference area. Because the observed trends were similar across all four subsections (Figure 3), and in the interest of maintaining a sufficiently large dataset, all further Chl a analysis was conducted using mean monthly nutrient and Chl a values for the entire area.
|
Nutrient data adjusted using the 40:1 and 100:1 C–Chl a ratios were plotted along with labile nutrient data to determine if the adjustments were effective in reducing seasonal variability (Figure 4). Evidently, following the spring bloom, nutrients were being lost from the system at a rate higher than anticipated. The decline is by a factor <2 for P (Figure 4a) and by a factor >4 for TIN (Figure 4b), even after adjustments. This phenomenon may be partially caused by the natural loss of nutrients in surface waters, owing to thermal stratification and sedimentation on the shelf, but further exploration is required to determine the reasons for these declines in total nutrients.
|
In view of these results in the offshore data, we decided to focus on the coastal environment, where inlets are well mixed and natural circulation patterns act as nutrient retention mechanisms that may offset the losses observed offshore. Once again faced with the absence of temporally comprehensive nearshore data, we chose to compare the adjusted nutrient data with a limited total phosphorous dataset that had been collected during our initial sampling. After repeating the analyses with carbon–Chl a ratios of 40:1, 70:1, 100:1, and no Chl a adjustment at all, adjustments made using the 40:1 ratio appeared to give the best fit to the total phosphorous data and resulted in fewer adjusted phosphate values being greatly more than the TP data (Figure 5). Although the selection of this ratio may seem overly pragmatic given the small size of the dataset and the lack of a statistically significant relationship, the combination of increased turbidity (decreased light availability), total vertical mixing, and high nutrient levels in coastal inlets compared with offshore waters would tend to favour a relatively low C–Chl a ratio. To determine conclusively which ratio is most suitable for nutrient adjustments, a more extensive dataset would be needed.
|
The results of surveys of NS and PEI coastal inlets, using only labile nutrient data, are shown in the nutrient–space diagram in Figure 6a. All NS sites lie within the boundaries of the reference polygon, except the LaHave River estuary. In contrast, all samples collected in PEI waters fall well outside the polygon, indicating that the area is experiencing eutrophication. This finding is important in validating the results of the eutrophication index, because these waters are well known for being highly eutrophic owing to the combination of naturally sandy soils and heavy fertilizer use by the agriculture industry (Raymond et al., 2002).
|
Although the analysis of labile nutrient data reveals no clear signs of eutrophication at any of the NS sampling locations, the adjusted data using the amalgamated nutrient–Chl a ratios (Figure 6b) indicate two additional locations displaying signs of eutrophication: Barrington Bay and New Harbour. Both of these sites demonstrate low concentrations of labile nutrients (0.42 µM PO4, 0.66 µM TIN, and 0.23 µM PO4, 1.16 µM TIN, respectively) but high measures of Chl a (22.7 and 55.3 µg l–1, respectively). The potential eutrophication problem would not have been detected based solely on inorganic nutrient concentrations!
|
Discussion |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
The nutrient–space eutrophication index method represents a considerable simplification in the assessment of water quality over many of the indices currently being developed. For example, whereas Strain and Yeats (1999) used six factors (silica, phosphate, ammonia, dissolved oxygen, iron, and manganese) to construct their eutrophication index, we obtained similar results from their data using only two factors (N and P), both of which are easily measured and widely available.
All but three NS inlets—Barrington Bay, New Harbour, and the LaHave River (Figure 1)—lie within the boundaries of the calculated reference conditions (Figure 6a and b). Although adjusted nutrient levels in Barrington Bay are high, they do not fall very far outside. This may be an early warning that if eutrophication is not already a problem in the area, it may be developing, and further work should be initiated to detect nutrient sources to avoid serious problems in the future. With TIN and PO4 concentrations approximately twice the boundary values in one of the mid-bay samples, New Harbour falls well outside the reference polygon, suggesting a serious eutrophication problem. This particular sample was collected in a tidal impoundment, which is constrained by a causeway and only linked to the outer harbour by water flowing under a small bridge. The relatively low human population in the area surrounding New Harbour can hardly be the primary cause of a local eutrophication problem. Rather, the low flushing rate resulting from the causeway is making the impoundment highly sensitive to nutrient enrichment. Unfortunately, little can be done to remedy the problem other than removing the causeway altogether. The LaHave River estuary, one of the larger and more heavily used estuaries in the region, represents a special case. The salinity of this sample was quite low (8 psu), and its slight deviation above the 16N:1P Redfield ratio might be the result of the high N–P ratio of fresh-water inputs relative to seawater. For this reason, the prospect of incorporating fresh-water nutrient criteria into future applications of the nutrient–space diagram must be considered to account for low-salinity estuarine water samples.
Assessment of the eutrophication status of coastal inlets using the method proposed here is fast, visually informative, and easily adaptable to different regions of the world. It simultaneously provides researchers with nutrient concentrations that have been adjusted for phytoplankton biomass, and directly illustrates how those values compare with regional expectations. Making such information readily available and easy to understand is essential to researchers for interpreting eutrophication, based on their knowledge of the areas being studied, and to legislators for making educated decisions about eutrophication measures in the coastal-zone.
The next steps in the development of a coastal marine eutrophication framework are to incorporate fresh-water nutrient criteria and estuarine retention effects into the method. If such an expanded nutrient–space eutrophication index were to be used in combination with a measure of eutrophication susceptibility, development of a straightforward and widely applicable framework for the assessment of eutrophication in the coastal-zone should be possible.
|
References |
|---|
|
TopIntroductionMethodsResultsDiscussionReferences |
|---|
-
Banse K. Determining the carbon-to-chlorophyll ratio of natural phytoplankton. Marine Biology (1977) 41:199–212.[CrossRef]
Bricker S. B., Clement C. G., Pirhalla D. E., Orlando S. P., Farrow D. R. G. National Estuarine Eutrophication Assessment: Effects of Nutrient Enrichment in the Nation's Estuaries. NOAA, National Ocean Service, Special Projects Office and the National Centers for Coastal Ocean Science. (1999) MD: Silver Spring. 71.
DFO. BioChem: database of biological and chemical oceanographic data. In: Department of Fisheries and Oceans (2006) Canada. Version 8 (2005) http://www.meds-sdmm.dfo-po.gc.ca/biochem/Biochem_e.htm.
Hopcroft R. R., Roff J. C., Berges J. A. Size-fractioned sedimentation in a tropical neritic ecosystem near Kingston, Jamaica. Continental Shelf Research (1990) 10:795–806.[CrossRef][Web of Science]
Kérouel R., Aminot A. Fluorometric determination of ammonia in sea and estuarine waters by direct segmented flow analysis. Marine Chemistry (1997) 57:265–275.[CrossRef][Web of Science]
Li W. K. W., Dickie P. M., Harrison W. G., Irwin B. D. Biomass and production of bacteria and phytoplankton during the spring bloom in the western North Atlantic Ocean. Deep Sea Research II (1993) 40:307–327.[CrossRef]
OSPAR. The Common Procedure for the Identification of the Eutrophication Status of the OSPAR Maritime Area. OSPAR Agreement 1997–17 (1997).
Raymond B. G., Crane C. S., Cairns D. K. Nutrient and chlorophyll trends in Prince Edward Island estuaries. Canadian Technical Report of Fisheries and Aquatic Sciences (2002) 2408:142–153.
Redfield A. C. The biological control of chemical factors in the environment. American Scientist (1958) 46:205–221.[Web of Science]
Strain P. M., Clement P. M. Nutrient and dissolved oxygen concentrations in the Letang Inlet, New Brunswick, in the summer of 1994. In: Canadian Data Report on Fisheries and Aquatic Sciences, 1004 (1996) 33.
Strain P. M., Yeats P. A. The relationships between chemical measures and potential predictors of the eutrophication status of inlets. Marine Pollution Bulletin (1999) 38:1163–1170.[CrossRef][Web of Science]
Taylor A. H., Geider R. J., Gilbert F. J. H. Seasonal and latitudinal dependencies of phytoplankton carbon-to-chlorophyll a ratios: results of a modeling study. Marine Ecology Progress Series (1997) 152:51–66.[CrossRef][Web of Science]
UNEP. The State of the Marine Environment: trends and processes. (2006) The Hague: United Nations Environment Programme. 28.
USEPA. ESS Method 310.2: Phosphorus, Total, Low Level (Persulfate Digestion). US Environmental Protection Agency, Environmental Sciences Section, Inorganic Chemistry Unit (1992).
USEPA. Method 445.0: In vitro determination of chlorophyll a and pheophytin a in marine and freshwater algae by fluorescence. US Environmental Protection Agency, National Exposure Research Laboratory, Office of Research and Development (1997).
Zeitzschel B. The quantity, composition and distribution of suspended particulate matter in the Gulf of California. Marine Biology (1970) 7:305–318.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||