SEARCH

SEARCH BY CITATION

Keywords:

  • Modeling;
  • Local effect time;
  • Residence time;
  • Flushing time;
  • Embayment

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References

Residence times are defined classically by the physical and chemical aspects of water bodies rather than by their ecological implications. Therefore, a more clear and direct connection between the residence times and ecological effects is necessary to relate these timescales quantitatively to ecology. The concept of local effect time (LET) is proposed in this paper as a timescale with spatial resolution that relates to ecological components and their spatial distribution within an embayment. The LET can predict the susceptibility of real-world ecological components to change from one condition to another. It can provide an efficient way to allow managers and agencies to evaluate the degree of stress or relief from current or projected changes in the loading of contaminants or nutrients. The steps for calculating LETs and defining their correlation with the existing ecological components in an embayment are presented along with illustrative applications to loading from riverine inflow and a wastewater treatment plant. The LET successfully identified the areas within the water body that could be prone to ecological changes due to perturbations in the loading rate of riverine water and its constituents. An example is given that shows how the LET method can be used to delineate the distribution and duration of high levels of coliform bacteria due to a pulsed effluent from the treatment plant.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References

The concept of residence time has been used in aquatic studies for decades. When applied to estuaries and coastal embayments, it typically has been equated with flushing time, or the time it takes to exchange all the water in a system with external water. More recently, 3 related measures of this time have entered the aquatic literature, i.e., flushing time, age, and residence time (Monsen et al. 2002). According to Monsen et al. (2002), flushing time is the ratio of the total mass of a constituent (or water) in a water body to its overall rate of renewal; age is the time a water parcel at a specified location has spent in the water body since entering it; and residence time is the time until a water parcel at a specified location leaves the water body. All these timescales originated from physical properties of the water body (e.g., volume, tidal forcing, freshwater inflow, mixing, exchange) and chemical properties of constituents (e.g., decay, partitioning). They focus on the mixing and movement of water and its constituents, not on how it relates to the ecology of a system and its heterogeneity.

Abdelrhman (2002) offered an additional characteristic time, the local residence time, to deal with spatial heterogeneities within an embayment and applied it to 42 embayments in southern New England, USA (Abdelrhman 2005). The concept behind a local residence time has been in the literature for 20 y (e.g., Takeoka 1984; Zimmerman 1988), however, it never included ecological derivatives. For example, Miller and McPherson (1991) used concentrations of constituents and calculated residence times based on various percentages of the original mass left in an estuary. This method also was used to study residence times of pulses of a conservative constituent released at various locations within the estuary. Also, Oliveira and Baptista (1997) used particle-tracking methods to produce maps of the distribution of residence times (times until leaving the embayment) for suites of particles released at specific locations. However, as with the other residence time metrics, these local residence times only were based on physical and chemical properties.

All the above-mentioned timescales often have been used for ecological purposes even though they are not derived from ecology; therefore, they could only be used empirically and qualitatively in this regard. For example, when studying eutrophication in lakes, Vollenweider (1975) normalized rates of phosphorus loading to residence times to obtain empirical relationships for phosphorus in suites of lakes. Dettmann (2001) used the same approach to normalize nitrogen in estuaries to freshwater residence time and obtain analogous empirical relationships for estuaries. Furthermore, any relationship these residence times might have with the ecological parameters of a system was undermined by the assumption of steady state (or quasi-steady state) conditions; ecological effects require that perturbations be considered.

The ecological components of embayments respond to local changes in concentrations and associated local timescales, not the travel time of water parcels in or out of the embayment. For example, Ulva will respond to nitrogen (N) in the water passing by it, not to any subsequent changes in that water. Thus, the above approaches cannot provide quantitative effects of specific constituents on individual ecological components at a specific location within a water body. Those approaches cannot answer questions such as the following: Which ecological component will be affected? Which locations within the embayment will be affected? How long would it take to affect them? Which locations are most vulnerable? Which systems are most vulnerable?

To best relate water movement and its associated constituents, to the ecology of a system, relevant local timescales need to be derived. This paper offers a way to do that: It presents and uses the local effect time (LET), where effect relates to both the constituent and the ecological component of interest. The constituent can be any property of the water parcel (e.g., temperature or salinity) or chemicals carried by it (i.e., contaminants or nutrients). The resulting LETs are more appropriate and relevant to ecological components with limited or no mobility (e.g., sea grass or benthic organisms) and can be used quantitatively to define areas susceptible to changes within an embayment. Here, susceptibility is characterized by a threshold level and the time period to reach that level. For example, application of LET to a pulsed effluent from a treatment plant (see Discussion section) revealed the distribution and duration of levels of coliform higher than the human health criteria.

THE LOCAL EFFECT TIME

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References

The LET is defined as that time period between when the loading of a constituent to an embayment changes (increases or decreases) and the time when its level in the water at a specific location reaches a value crucial to the condition of an ecological component at that location. The change may be either favorable for the ecological component (a contaminant being flushed out) or unfavorable (a needed nutrient being lost). For the former case, areas with shorter retention of the contaminant (quicker flushing) will be healthier than those with longer retention, and vice versa for the nutrient. Excess nutrients, however, may become undesirable for some ecological components. For the Ulva example above (see Introduction section), a drop in N below the threshold needed for optimal growth will produce signs of stress. In contrast, stress in eelgrass (Zostera marina) may be produced by an excess of N, which increases the density of phytoplankton and, in turn, increases turbidity and limits the light that can reach the eelgrass.

The LET obviously is impacted in part by the geometry of the embayment. It also is affected by at least 4 other ecology-related factors: The constituent, the ecological component, the threshold value of the constituent, and the physical mechanism that changes the loading of the constituent (tidal action, freshwater inflow, mitigation measure, etc.). Three of these 4 connections to ecology are direct (ecological component, constituent of interest and threshold value) and 1 is indirect (loading mechanism). The methods presented here apply to systems that change from one equilibrium state to another. Steady state conditions do not need to be investigated because they result in equilibrated systems.

GENERAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References

The following methodology uses readily available models to simulate hydrodynamics and transport patterns in an embayment.

Step 1—establishing initial (existing) conditions. Define the constituent of interest, the spatial distribution of the ecological component of concern, and the constituent's threshold value (α) that may initiate adverse or favorable changes in the ecological component. (Useful comments on the choice of the threshold value are presented in the Discussion section.) Then, define all sources and loading rates of the constituent from point sources (e.g., rivers, wastewater treatment plants, municipal and industrial outfalls), nonpoint sources (i.e., atmospheric deposition), and boundary sources (e.g., groundwater and seawater). Also, identify all sinks, decay factors, and processes that consume the constituent. Next, generate the flow field and run the transport simulation with all identified sources and sinks until the constituent's initial local values reach steady state or quasi-steady state, Co(x,y), throughout the embayment. Interpret the ecological effect from the concentration distribution, which may fall into 1 of 3 cases:

  • I.
    Co(x,y) > α, for all x,y—ecological component is susceptible everywhere.
  • II.
    Co(x,y) < α, for all x,y—ecological component is not susceptible anywhere.
  • III.
    Co(x,y) > α, for some x,y and < α for the other locations—ecological component is susceptible where Co(x,y) > α.

Step 2—perturbing the system. Based on the question asked, change the constituent's sources or sinks, or the flow field, and rerun the simulation to calculate the new equilibrium values of the constituent, C(x,y), at all x,y, and reinterpret the results, which can be as follows:

  • I′.
    C(x,y) > α, for all x,y—ecological component is susceptible everywhere.
  • II′.
    C(x,y) < α, for all x,y—ecological component is not susceptible anywhere.
  • III′.
    C(x,y) > α, for some x,y and < α, for the other locations—ecological component is susceptible where C(x,y) α

For cases I and II, the ecological component at a location (x,y) is not expected to show adverse (or favorable) signs when perturbed unless the change causes a shift to another case (e.g., from case I to case II′ or III′; or from case II to case I′ or III′). In case III, signs of change on the ecological component will appear with any of the 3 new cases I′, II′, and III′. For III′, changes will appear only at (x,y) locations that experience a change in concentration from <α, to >α, or vice versa.

Step 3—defining LET and interpreting results. Calculate the time needed for the initial local equilibrium value Co(x,y) to reach its threshold value α as it changes to its new equilibrium value(x,y). This is the LET at this location. The LET can be much shorter than the time required to reach the new equilibrium value. The distribution of the LET identifies areas where the ecological component is becoming susceptible to the stated perturbation; i.e., areas that will experience adverse (or favorable) conditions after the periods indicated by the LET values.

To illustrate this procedure, the above steps are applied to 2 cases of a constituent loaded through riverine inflow (see Application to riverine loading section). The special case for a pulsed fecal coliform effluent from a wastewater treatment plant is presented in the Discussion section. This case illustrates the application of LET when the perturbed system returns back to its initial condition.

APPLICATION TO RIVERINE LOADING

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References

The above procedure was applied to New Bedford Harbor, Massachusetts, USA (Figure 1), to identify the LET for freshwater or for a constituent loaded to the system with riverine inflow from the Acushnet River at the northern boundary. The hydrodynamics and transport patterns were simulated with 2-dimensional, depth-averaged, finite-element models (RMA2 and RMA4, respectively). Modeling details, calibration, and the main equations for the model are those presented by Abdelrhman (2002), who used the models to study circulation, transport, and residence time in the harbor.

thumbnail image

Figure Figure 1.. Map of New Bedford Harbor, Massachusetts, USA (MassGIS 1993). The star marks the approximate location of the effluent from the Fairhaven wastewater treatment plant (WWTP).

Download figure to PowerPoint

When a conservative tracer is loaded at a rate L (g s−1) with the discharge Q (m3s−1) from the Acushnet River, its total mass will increase until it reaches a quasi-steady state balance with tidal flushing (Abdelrhman 2002). If the loading rate increases (or decreases), this mass will assume a new quasisteady level. The ratio between L and Q defines the concentration of the tracer in riverine water (g m−3), and the simulated distribution of the tracer defines the mixing of the riverine water (and its constituents) with the existing local water. In general, freshwater and its associated constituents will have high concentrations near the river mouth and will decrease southward toward the sea (Abdelrhman 2002). Mixing of actual constituents entering with the riverine water can be defined by scaling the simulated distribution of the tracer; i.e., by multiplying by the ratio of the loading rates, Lconstituent/L.

To simulate the distribution of a tracer, freshwater was introduced at its high rate (1.7 m3s−1) at the northern boundary of the hydrodynamic model. A surrogate tracer was loaded at a similar rate (i.e., 1.7 g s−1) to produce a unit concentration of 1 g m−3 (1,000 μg L−1) in the inflow. The concentration of tracer in Buzzards Bay and at the southern boundary of the model domain was assumed to be zero. The same procedure was applied at the mean flow (0.54 m3s−1)—the tracer was loaded at 0.54 g s−1 to produce a unit concentration of 1 g m−3 in the inflow. The calibrated transport models of Abdelrhman (2002) were used to simulate the distribution of the tracer in the estuary for each loading rate until quasi-steady state distributions were reached. The LETs were determined for 2 scenarios. In the 1st scenario, the flow changed from its mean value to its high value, which builds up riverine water within the estuary. The 2nd scenario is the reverse, with flow changing from its high value back to the mean, which reduces riverine water in the embayment. The change in flow was assumed to take place instantaneously. The simulation for each scenario was continued until concentrations throughout the embayment became quasi-steady. A benthic organism susceptible at α = 40 μg L−1 was assumed to exist throughout the New Bedford harbor.

thumbnail image

Figure Figure 2.. Quasi-steady state distribution of riverine water as marked by a tracer loaded at a rate designed to produce a concentration of 1 g m−3 in the river inflow (a) for the mean flow rate (0.54 m3s−1) and (b) for the high flow rate (1.7 m3s−1). Point A represents region R1 that will have concentrations above α (40 μg L−1) in the two scenarios (see Figure 3) and, thus, will not have local effect times (LETs). Point C represents region R3 that will have concentrations below α in the two scenarios and, thus, will not have LETs. Point B represents the region R2 that will have LETs.

Download figure to PowerPoint

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References

Figure 2 presents the spatial distribution of the concentration field for each flow rate and Figure 3 shows its longitudinal distribution. The maximum concentration near the mouth was ˜200 μg L−1. For the mean flow rate (Figure 2a), concentrations above α are limited to a very short region, R1, near the mouth of the river. At the higher flow rate (Figure 2b), concentrations extend to south of the bridges. Region R1 concentrations are above α in both cases (see also Figure 3). Thus, the ecological condition in this region does not change and there will be no LET for it. Similarly, region R3 with concentrations lower than α in Figure 2b also has concentrations lower than α in Figure 2a (see also Figure 3) and also will not have an LET. Only region R2 is expected to have LETs, because its concentrations can cross α in the positive direction when the inflow increases or in the negative direction when it decreases (Figure 3).

thumbnail image

Figure Figure 3.. Quasi-steady state longitudinal distribution of riverine water for mean flow (0.54 m3s−1, thin line) and high flow (1.7 m3s−1, bold line). The vertical broken lines mark regions R1, R2, and R3.

Download figure to PowerPoint

Temporal variations in concentration for the 1st scenario (flow rate changing from low to high) at locations A, B, and C (Figure 2), which represent the above-mentioned 3 regions, are shown in Figure 4a. At location A, the concentration exceeds α during both flows. At location C, the concentration stays below α Only at location B does the concentration change from values below α to values above it. The LET at B starts at the onset of perturbed flow (point S on the time axis) and ends when concentration markedly exceeds α (shown by the arrow). Locations north of B will have temporal traces of concentrations that fall above curve B, and thus will have shorter LETs. Conversely, locations south of B will have longer LETs.

thumbnail image

Figure Figure 4.. Temporal variation of concentrations at locations A, B, and C (Figure 2) for (a) the 1st scenario, when riverine inflow changes from the quasisteady state conditions at the mean flow rate (0.54 m3s−1) to the quasisteady state conditions at the high flow rate (1.7 m3s−1), and (b) the 2nd scenario, when riverine inflow changes from the quasi-steady state conditions for the high flow rate (1.7 m3s−1) to the quasi-steady state conditions at the mean flow rate (0.54 m3s−1). Point S on the time axis marks the beginning of the local effect time (LET) at location B. The arrow marks the end of the LET, when the concentration reaches the threshold value (α) during its change. Notice that LET can be much shorter than the time required to reach steady state after the perturbation.

Download figure to PowerPoint

Temporal variations at the same 3 locations for the 2nd scenario (flow rate changing from high to low) are shown in Figure 4b. As with the 1st scenario, there is no LET at location A because the concentration remains above α during both flows, or at C because it remains below α for both flows. Only at location B does the concentration cross α, this time in the negative direction. The LET at B starts at the onset of perturbed flow (point S) and ends when the concentration permanently drops below α (the arrow). Locations north of B will have temporal traces of concentrations that fall above curve B and thus have longer LETs. Conversely, locations south of B will have shorter LETs.

The spatial distribution of LETs for the 2 scenarios are presented in Figure 5. When the inflow increases, the LET is shorter near the river mouth and increases away from it (Figure 5a). This is logical because, as the additional river water proceeds southward, it will be mixed and flushed away by seawater. The farther from the mouth of the river a location is, the greater will be the mixing with seawater and the greater the time needed to reach α. Conversely, these locations farther from the mouth of the river will have the shortest LETs when the high flow recedes to the lower value (Figure 5b), because they will be flushed out faster by the tides. For example, the 9-d contour south of the bridges in Figure 5a indicates the locations and the period for the effect to appear when the flow rate changes from its mean value to the higher rate. In contrast, a contour with similar LET value (9 d) appears north of the bridges when the flow changes from high to mean rates (Figure 5b).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References

This work identifies the local effect time as more relevant than flushing time or residence time for predicting susceptibility of ecological components. Although flushing and residence times are qualitative integrative measures of transport timescales that do not originate from ecology, the LET directly relates to an ecological component and its spatial distribution within an embayment. For example, knowing that the flushing time in New Bedford Harbor is ˜3 d (Applied Science Associates 1987; Abdelrhman 2002) does not clarify the susceptibility of any organism at a location (e.g., near the bridges or elsewhere in the harbor). On the other hand, the LET explicitly deals with local areas within heterogeneous systems that have differential transport times and the consequent contrasting ecological responses within that embayment (e.g., Figure 5). In so doing, the LET increases the relevance of field-sampling schemes with multiple stations within an embayment and makes it possible to relate local measurements to a suite of LETs instead of to the commonly used integrative flushing or residence times. This improves the analysis and understanding of embayments and their heterogeneous ecology. Unlike flushing and residence times, LET is not a property of the system or even a location within the system. The LET depends on the loading mechanism, the perturbation, and the threshold criteria for the ecological component of interest, as presented in the wastewater treatment plant example below.

thumbnail image

Figure Figure 5.. Spatial distribution of the local effect time (LET) values (day) for (a) the 1st scenario, when the riverine inflow changes from the quasi-steady state conditions at the mean flow rate (0.54 m3s−1) to the quasi-steady state conditions at high flow rate (1.7 m3s−1), and (b) the 2nd scenario, when the riverine inflow changes from the quasi-steady state conditions at the high flow rate (1.7 m3s−1) to the quasi-steady state conditions at the mean flow rate (0.54 m3s−1).

Download figure to PowerPoint

The threshold value should relate to the well being of the ecological component, not just the appearance of adverse effects (e.g., as given by the lethal concentration at which 50% mortality occurs [LC50]). Various examples exist of threshold values that regulators have considered in marine and freshwater quality criteria and standards (e.g., USEPA [1986] for dissolved oxygen; National Academy of Science and National Academy of Engineering [1972] for temperature; Federal Register [1984] and Buikema et al. [1982] for chemicals; Salas [1985] for bacteria). The threshold value must have the units of the constituent or parameter of interest. Situations with undefined threshold values cannot be characterized by LET.

The following example illustrates the application of LET to fecal coliform discharge from the combined sewer overflow (CSO) at a wastewater treatment plant (Figure 1). The presence of fecal contamination is an indicator that potential human health risks exist for pathogenic diseases including ear infections, dysentery, typhoid fever, viral and bacterial gastroenteritis, and hepatitis A. Massachusetts, USA, regulations indicate that waters used for class 1 primary contact (including such activities as swimming, rafting, and kayaking) should not have fecal concentrations above 200 fecal coliform colony forming units (cfu) per 100 ml (2 × 106 cfu m−3). (This example is not an attempt to present a full-scale modeling exercise of fecal coliform, e.g., [Canale et al. 1993] nor a study of the effect of various point and nonpoint sources and CSOs around the harbor. This example only serves as a practical application for the concept of LET.)

In New Bedford Harbor, the CSO may load the system with 100,000 to 1,000,000 cfu per 100 ml of the effluent discharge during high rain events (B. Pitt, USEPA, Boston, MA, USA, personal communication). The overall loss rate of fecal coliform in seawater (at 20°C) is k = 1.4 d−1 (Thomann and Mueller 1987). I assumed a 1st order loss relation (i.e., c = coe−kt, where c is the concentration of fecal coliform (cfu m−3) at time t (days) and co is its initial concentration), which accounts for losses due to death from exposure to sunlight, temperature, predation, and salinity, as well as settling loss. On a typical rain event such as 2004 May 28 (0.92 inches of rain), the discharge rate to the CSO was 34.2 mgd (1.5 m3s−1), and the reported concentration of fecal coliform in this effluent was 870 cfu per 100 ml (V. Furtado, City of Bedford, Department of Public Works—Wastewater Division, New Bedford, MA, USA, personal communication). The calculated total amount released on that day was 1.1262 × 1012 cfu. Assuming an instantaneous release (pulse), the calculated LET distribution until fecal coliform levels drop to the criteria threshold value (200 cfu per 100 ml) is presented in Figure 6. Contours with longer LET values exist close to the wastewater treatment plant and those with shorter values are further away, up to the mouth of the Acushnet River. In other words, concentrations above the health criteria existed throughout the harbor, but lasted longer in some parts of the harbor than others. This analysis may be used in the siting of recreational facilities or permitting of fishing and shell fishing. This sort of LET analysis also could be used to inform the siting of another CSO, or the relocation of the existing CSO, compare impacts at a different discharge point with those from the present location, or address other management questions.

thumbnail image

Figure Figure 6.. Local effect time (LET) distribution (h) for a pulsed effluent of fecal coliform from the combined sewer overflow at the wastewater treatment plant near Fairhaven (MA, USA) on May 28, 2004, after a rainfall event (0.92 inch) with the threshold for human health criteria as 200 cfu per 100 ml.

Download figure to PowerPoint

It is important to remember that the timescales found here were based on the time needed for the concentration to reach an ecologically based threshold value, α, which can be a controversial parameter in this method. Although managers and modelers would seek a real design value of this threshold (as in the above-mentioned example of the wastewater treatment plant), ecologists and biologists may point to difficulties in defining this value in real-world situations. However, for design purposes, threshold value(s) still may be identified and used by managers and modelers while accommodating ecological concerns. One concern may be that the ecological effects of a perturbation typically occur through some process of increasing accommodation by the organisms (i.e., adaptation), which eventually becomes either energetically expensive or insufficient. The adaptation of an organism can be accounted for to the point when the process becomes energetically expensive; this can mark the required threshold value. Another concern is that the effects may correspond to a range (not a single value) of the change in the constituent or parameter. For such cases, both upper and lower limits of this range have to be included in the LET analysis as threshold values to define susceptible areas within (or outside) this range. For example, a species may function well within the range α1 to α2 and succumb outside it. Figure 7 illustrates this case for a hypothetical range with α1 = 0.04 g m−3 and α2 = 0.025 gm−3. Regions (R1)α1, (R2)α1, (R2)α2, and (R3)α2 are analogous to their respective regions R1, R2, and R3 in Figure 3, but with LETs based on the threshold values indicated by their subscripts. Region R4 will have 2 LET values: One for entering and the other for exiting the specified range, based on the direction of change (e.g., from low to high flow or vice versa). The time period between these 2 LET values indicates the window for susceptibility.

thumbnail image

Figure Figure 7.. Example of a threshold range between α1 = 0.04 g m−3 and α2 = 0.025 g m−3. Regions (R1)α1, (R2)α1, (R2)α2, and (R3)α2 are analogous to their respective regions R1, R2, and R3 in Figure 3, but with the subscript threshold values α1 and α2 used (instead of α) to mark the local effect time (LET). Region R4 has a window of susceptibility marked by 2 LET values: One for entering and the other for exiting the specified range.

Download figure to PowerPoint

At this initial stage for developing and introducing LET, standardized scenarios were presented. The sharp start of the change (point S in Figure 4) is valid when the time period of the perturbation is short relative to LET. Otherwise, the location of point S on the time axis will depend on the nature of the change. Ecological constituents that already are experiencing a change can be the subjects of future analysis by LET to define retroactively the start of that phase (point S) and predict its end-using LET.

Ecologists and biologists are encouraged to follow this introduction of the LET by further developing the concept and finding threshold values for embayments, constituents, and ecological components of interest. One important refinement would be to add the time required for the metabolism of the ecological component to respond to the threshold level, thus providing the total period for effects to appear. The presented concept and methodology for LET are not tied to any specific computer code or model. Based on the availability of resources, more accurate modeling analysis and results always can be achieved with more advanced, sophisticated, and complex models.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References

The author thanks the reviewers of this manuscript, including D. Campbell, W. Nelson, and G. Cicchetti (USEPA-Atlantic Ecology Division [AED]), and 2 anonymous reviewers for their technical reviews, insights, and constructive comments. Special thanks are due to W. Berry (USEPA-AED) for very valuable comments and suggestions and also K. Rahn (Computer Sciences Corporation) for technical and editorial comments on the manuscript. Although the research described here has been funded by the USEPA, it has not been subject to Agency-level review and therefore does not necessarily reflect the views of the Agency, nor does mentioning trade names or commercial products endorse or recommend them. This manuscript is contribution AED-04–071 of USEPA Office of Research and Development, National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. THE LOCAL EFFECT TIME
  5. GENERAL PROCEDURES
  6. APPLICATION TO RIVERINE LOADING
  7. RESULTS
  8. DISCUSSION
  9. Acknowledgements
  10. References
  • Abdelrhman MA. 2005. Simplified modeling of flushing and residence times in 42 embayments in New England, USA, with special attention to Greenwich Bay, Rhode Island. Estuarine, Coastal and Shelf Science, 62 (1–2): 339351.
  • Abdelrhman MA. 2002. Modeling how a hurricane barrier in New Bedford Harbor, Massachusetts, affects the hydrodynamics and residence times. Estuaries 25: 177196.
  • Applied Science Associates. 1987. Selected studies of PCB transport in New Bedford Harbor. Technical report ASA 86–18. Submitted to Ropes and Gray.
  • Buikema AL, Jr., Neiderlehner BR, Cairns J, Jr. 1982. Biological monitoring. Part IV—toxicity testing. Water Res 16: 239262.
  • Canale RP, Auer MT, Owens EM, Heidtke TM, Effler SW. 1993. Modeling fecal coliform bacteria—II. Model development and application. Wat Res 27: 703714.
  • Dettmann EH. 2001. Effect of water residence time on annual export and denitrification of nitrogen in estuaries: A model analysis. Estuaries 24: 481490.
  • MassGIS. 1993. MassGIS data layer descriptions and guide to user services. Boston (MA): Massachusetts Geographic Information System Executive Office of Environmental Affairs.
  • Miller RL, McPherson BF. 1991. Estimating estuarine flushing and residence times in Charlot Harbor, Florida, via salt balance and box model. Limnol Oceanogr 36: 602612.
  • Monsen NE, Cloern JE, Lucas V. 2002. A comment on the use of flushing time, residence time, and age as transport timescales. Limnol Oceanogr 47: 15451553.
  • National Academy of Science and National Academy of Engineering. 1972. Water quality criteria, Washington DC. p 594.
  • Oliveira A, Baptista AM. 1997. Diagnostic modeling of residence times in estuaries. Water Resour Res 33: 19351946.
  • Salas H. 1985. Historia y applicacion de normas microbiologicas de calidad de agua en el medio marino. In: Hojas de divulgation tecnica, CEPIS, OPS/OMS, 29. Lima, Peru. p 15.
  • Takeoka H. 1984. Fundamental concepts of exchange and transport timescales in a coastal sea. Continental Shelf Research 3: 311326.
  • Thomann RV, Mueller JA. 1987. Principals of surface water quality modeling and control. New York (NY): HarperCollins. p 235.
  • [USEPA] U.S. Environmental Protection Agency. 1986. Ambient water quality criteria for dissolved oxygen. Washington DC: USEPA. EPA/440/5–86–003.
  • Vollenweider RA. 1975. Input-output models with special reference to phosphorous loading in limnology. Schweiz Z Hydrol 37: 5384.
  • Zimmerman JTF. 1988. Estuarine residence times. In: KjerfvB, editor. Hydrodynamics of estuaries. Vol 1. Boca Raton (FL): CRC. p 7584.