Existing studies on sediment retention ponds (SRPs) have examined the effects of pond layout, inlet and outlet geometry and the installation of baffles on the performance of the SRPs. However, the effects of a temperature difference between the ambient water in the pond and the inflow are often neglected, and the buoyancy forces arising from these differences in temperature can potentially change the flow in the pond and hence its hydraulic performance. This study has experimentally evaluated the effect of these temperature differences on the flow field and residence time in a retention pond for a range of temperature differences. As expected a cold inflow sinks to the bottom of the pond while a hot inflow remains at the surface, but in both cases the inflow flows more rapidly towards the outlet than is the case for isothermal inflow. A counter-current was observed at the bottom or the surface of the pond for colder or hotter influents, respectively. These thermally induced flows significantly reduced the residence time of the pond, reducing the hydraulic performance of the pond and causing severe short-circuiting. The results have also shown that the temperature differences in the pond decrease with time, yet small temperature differences persist with the pond remaining thermally stratified.

Sediment retention ponds (SRPs) have been used for many years as a cost-effective treatment to minimise the adverse impacts of urbanisation on aquatic environments (ARC 2003; Khan et al. 2013; Vymazal 2014). The treatment efficiency of the pond is directly linked to its hydraulic efficiency. Furthermore, providing longer residence time is important to improve the hydraulic efficiency of SRPs. Hence, a plug flow regime – one that leads to uniform velocity and the same residence time for all parcels of water entering the pond – is desirable (von Sperling 2002; Khan et al. 2011). The design of stormwater ponds is currently based primarily on the assumption that plug flow conditions prevail in the pond and that 100% of the pond volume is utilised. This is not the case for most of the ponds because the flow structure is normally composed of zones of recirculation and eddies (Kadlec 1994).

The hydrodynamics of a pond is affected by several factors which can be divided into physical and environmental parameters. The environmental parameters can introduce additional complexities in the current design procedure of retention ponds. Hence, it is important to study the effect of both physical and environmental factors on the behaviour of retention ponds in order to have an efficient and economical design (Watters 1972). Important physical parameters that affect the behaviour of the ponds include the positioning and orientation of pond inlets and outlets, the pond geometry, and the location and orientation of baffles. Environmental parameters of the ponds are external driving forces such as wind and temperature (Shilton 2005). Most of the literature available on the hydraulic performance of SRPs has focused on studying the effects of the physical parameters. Previous research concerning the effect of physical parameters on pond hydraulic efficiency has investigated pond layout (Persson 2000; Su et al. 2009), inlet and outlet design (Pearson et al. 1995; Persson 2000; Bodin et al. 2012), the use of floating treatment wetlands (Khan et al. 2013) and baffles (De Oliveira et al. 2011; Farjood et al. 2015). Variation of density in the vertical direction is also important in analysing the hydrodynamics of the ponds and wetlands. The inflow characteristics and local meteorological conditions are the most important parameters in creating stratification. The most probable causes of stratification are differences in temperature and salinity (Adamsson & Bergdahl 2006). Regarding temperature differences, the results of tracer studies have indicated that temperature variation causes short-circuiting in ponds (Macdonald & Ernst 1986; Pedahzur et al. 1993). Goula et al. (2008) have numerically investigated the effect of variations in influent temperature in a sedimentation tank for potable water treatment and found that ‘only 1 °C difference between influent and tank content is enough to induce a density current’. Few studies have conducted experimental investigations to find the influence of temperature on stormwater ponds. Watters (1972) investigated the effect of influent temperature variations in a waste stabilisation pond. The temperature differences were experimentally modelled by changing the density of the influents using salt water, thereby imposing certain approximations as a method of creating density-stratified flows, such as the dissimilarities in buoyancy and in the tendency of forming a density current. Therefore, a systematic study on the effects of temperature differences is still needed for the better understanding of hydrodynamics of SRPs (Hendi et al. 2017).

The primary aim of this study is to investigate how buoyant, neutrally buoyant or negatively buoyant inflows affect the residence time and flow patterns in a modelled retention pond. The results of this study would also be applicable to the stratification caused by suspended particles or salt.

Tracer studies

In this study, Rhodamine WT was selected as the tracer dye due to its numerous advantages over the other tracers; these advantages include being readily soluble in water, readily detectable by fluorometers, minimally affected by background fluorescence, minimally degradable in short periods of times, periods of harmless in low concentrations (Wilson et al. 1968) and low cost.

Hydraulic performance

The hydraulic performance of ponds can be evaluated by analysing the residence time distribution (RTD) curve. In this study, the hydraulic performance of the pond was assessed using the hydraulic indices as recommended by Farjood et al. (2014), these being a measure of short-circuiting (SC), the moment index (MI), and the Morril index (Mo).

SC is defined as t5, the time for passage of 5% of the added tracer to exit through the outlet, normalised by the nominal residence time tn = pond volume/flow rate.

The MI proposed by Wahl et al. (2010) is a measure of hydraulic efficiency, defined as:
formula
(1)
where , C is the concentration of tracer in pond outflow, C0 is the amount of tracer added to the pond divided by the pond volume, and is the normalised time. The MI for the hydraulic efficiency is within the limits from zero to one, and MI = 1 is the maximum hydraulic efficiency and indicates plug flow.

The Morril index, Mo, is a measure of mixing, defined as t90/t10, where t10 and t90 are the times for 10% and 90%, respectively, of the added tracer to exit the system. Values of Mo close to 1 indicate a flow close to plug flow, and Mo increases with increasing mixing.

Experimental rig

The physical model is a trapezoidal pond made from transparent acrylic sheets fitted on a steel frame with dimensions of 4.1 × 1.6 m and is 0.3 m deep, and has a bank slope of 2:1. The pond is preceded by a rectangular tank of dimensions 0.3 × 1.6 × 0.2 m serving as the sediment forebay (Figure 1). The model pond was designed as a 1:10 scale replica of an existing retention pond situated at the Alpurt B2 Motorway site, north of Auckland, New Zealand (Khan et al. 2011; Farjood et al. 2015). The flow rates used in the experiments were 1 l/s and 2 l/s which represent 316 l/s (treatable field flows) and 632 l/s (a large rainfall event) field flowrates.

The rig was designed so that the temperature differentials could be created using two separate systems and monitored with a thermometer. Before starting the experiments, the water in the tank is pumped to the forebay and over a level spreader into the retention pond. The effluent at this stage is carried by a 40 mm pipe to the waste. After ensuring steady flow conditions in the pond, the effluent in the pond and the water in the tank are recirculated in two different systems. Each system consists of two heater/chiller units to change the temperature of the water. After changing the temperature of the water, the water in the tank is pumped to the pond, and the tracer experiments are conducted by adding Rhodamine WT uniformly across the spread inlet width.

For the outlet, three perforated T-bars are fixed to an outlet riser to model a floating decant dewatering system. The perforated T-bars were constructed from PVC pipes with diameter of 48 mm. Five rows of 6 mm diameter holes on each of the T-bars allow the water to leave the pond. The T-bars are fixed to a 250 mm long, vertical PVC pipe with 200 mm internal diameter, which serves as the outlet riser (Farjood et al. 2015). The T-bars are fixed to the outlet riser at a height of 220 mm from the bottom and are about 80% submerged in the 245 mm water depth during the tests for 1 l/s and fully submerged in the 265 mm water depth for 2 l/s. The tracer concentrations in the outflow are continuously measured using a Cyclops-7TM fluorometer (manufactured by Turner Designs).

Effect of temperature variations on RTDs

To test the effect of temperature variations on the performance of the pond, two conditions were selected. In the first condition, the temperature of water in the pond was hotter than the influent (positive values of ΔT, where ΔT is the difference between the initial temperature of water in the pond and the inflow temperature). In the second condition, the temperature of water in the pond was colder than the influent (negative values of ΔT). Tables 1 and 2 list the temperature differences for each experiment together with the experimental results. The temperature of the inflow for cases 2–6 was lower than the temperature of the pond fluids while the reverse was true for cases 7–11. Typical RTD curves were analysed to assess the hydraulic efficiency of the ponds in the two above-mentioned conditions (see Appendix A, available with the online version of this paper). It can be seen that a temperature difference between the pond water and inlet water can cause significant changes in the RTD curves and consequently can change the hydraulic efficiency of the retention pond. For both positive and negative values of ΔT, as the temperature difference increases, the RTD curves show a sharper peak, and the maximum normalised concentration for the peak is increased. There is also a clear trend showing a decrease in the time taken to reach the peak with increasing temperature difference, indicating a reduced efficiency and the occurrence of short-circuiting in the pond.

The reasons for these differences are given as follows. When ΔT > 0 the inflow is denser than the pond water, and so flows along the bottom of the pond. The restriction of the inflow to a shallow current at the bottom results in a higher velocity for the current.. After running along the length of the pond, the current rises at the far end and exits through the outlet. This results in the tracer taking a shorter time to reach the outlet. The process is similar when ΔT < 0, except that the inflow is concentrated to a thin layer at the top of the pond. Thus, the inflow again reaches the outlet in a shorter time interval than when the inflow and pond temperatures are the same (i.e. ΔT = 0).

The relationships between the RTD curves and temperature differences are shown graphically in Figure 2. As shown in this figure, there is a clear trend showing that all index values decrease as the temperature difference increases, emphasising the importance of temperature effects on the hydraulic behaviour of retention ponds. It is apparent from this figure that even a temperature difference of one degree can significantly reduce the hydraulic efficiency and can cause severe short-circuiting. This finding is consistent with the study conducted by Goula et al. (2008), which concluded that ‘only 1 °C difference between influent and tank content is enough to induce a density current’. The temperature difference also decreases the Mo−1, which indicates decreased mixing levels in the pond for both hot and cold influents. It is also clear that the slope of these trends decreases with increasing temperature difference for both 1 l/s and 2 l/s. For the flowrate of 1 l/s, the hydraulic indices rapidly decrease with ΔT for 1 °C–2 °C temperature difference between the influent and the pond, while the decreasing trend is less marked for 2 °C–8 °C. For the flowrate of 2 l/s, the same trend was observed in the hot influent case while the initial rapid decreasing trend for the hydraulic indices was detected only up to 1 °C temperature difference between the influent and the pond for cold influent.

In order to investigate the importance of inertial and density gravity forces, it is better to compare the hydraulic index values with the corresponding densimetric Froude number. The densimetric Froude number (Frd) – which was proposed by Watters (1972) for constructed ponds – is used to show the differences between hot and cold influent cases for both 1 l/s and 2 l/s.
formula
(2)
where V and d denote the mean flow velocity and the height of the flow at a selected cross-section (in this study the inlet is the selected cross-section), respectively' while denotes the reduced gravity due to the relative density difference:
formula
(3)
where ρin and ρpond denote the density of the influent and the fluid inside the pond, respectively, and β is the thermal expansion coefficient.

The effect of the densimetric Froude number on the results is shown in Figures 35. The important parameters listed in Tables 1 and 2, shown in Figures 35, are plotted against the densimetric Froude number. All the trendlines for both density difference situations and flowrates show the same decreasing trend. The unshaded circles (red circles in the online version of this paper) outside the graphs in Figures 35 illustrate the efficiency of the isothermal case that occurs when . Two horizontal lines have been drawn from the isothermal points that intersect each trend line. These intersections represent an estimate of the minimum densimetric Froude number that is needed to affect the efficiency, short-circuiting and mixing of the pond. The minimum values of efficiencies are 56.9 and 48.1 for 1 l/s and 2 l/s, respectively. These minimum values illustrate the minimum temperature differences needed to change the hydraulic efficiency of the pond; the minimum values 48.1 and 56.9 represent 0.63 °C and 0.46 °C temperature differences between the inflow and the water in the pond, respectively. The minimum values of the short-circuiting index are 46.9, 56.4, and 51.4 for cold influent of 2 l/s, hot influent of 2 l/s, and both influents of 1 l/s, respectively. The minimum values of short-circuiting index illustrate the minimum temperature difference needed to change the short-circuiting index of the pond, given that 46.92, 56.39, and 51.4 represent 0.65, 0.47, and 0.56 °C temperature differences, respectively, between inflow and the water in the pond. Similarly, the minimum values of mixing index are 36.6, 59.1, 49.4, and 58.7 for cold influents of 2 l/s, hot influents of 2 l/s, hot influents of 1 l/s, and cold influents of 2 l/s, respectively. These minimum values illustrate the minimum temperature difference needed to change the mixing index of the pond, given that 36.6, 59.1, 49.35, and 58.7 represent 0.89, 0.42, 0.6, and 0.43 °C temperature difference between inflow and the water in the pond. From the presented results of densimetric Froude numbers, we can see that even a temperature difference of 0.5 °C can change the expected flow pattern. A similar conclusion has been drawn by Crittenden et al. (2012) in a study that showed the temperature differences as small as 0.3 °C can cause density currents.

Effect of temperature on flow patterns

As previously stated, plug flow is the most effective hydraulic condition for SRPs (von Sperling 2002). Hence, the flow patterns in the pond were investigated to shed light on the dissimilarities of the RTDs. The most important parameter that can affect the flow pattern significantly is the buoyancy force while dealing with temperature differences in ponds (Watters 1972). Figure 6 shows photographs of the dye tracer for 1 l/s flowrate as it flows out of the inlet in the case of cold influent, hot influent, and the isothermal case. These photographs visually show the tracer flowing near the bottom or the surface of the pond for the two cases of density-stratified flow.

As can be seen in Figure 6, the dye tracer flows rapidly towards the outlet for the hot and cold influent cases. After 200 seconds (tn = 0.25) from the beginning of the experiments, the majority of the dye tracer appears in the outlet area for the hot and cold influent cases, while for the isothermal case the dye tracer has just reached the outlet. This indicates the short-circuiting in the cold and hot influent cases.

The effect of buoyancy was found to be significant, and the velocity and temperature fields were strongly affected by thermal stratification. The results presented here concern the two flow configurations (hot influent and cold influent) shown in Figures 7 and 8 along with the time-dependent velocity contours that were obtained experimentally. With the increasing temperature difference, the buoyancy effect is more pronounced, and it pushes the streamlines upwards or downwards for hot and cold influent, as expected. This results in a narrower flow region near the surface or bottom and consequently larger flow velocities. The bottom current can cause erosion of sediment resulting from increased bottom velocities. Shear between the inflow water and the standing pond water causes a circulation flow or counter-current in the standing water (the lower region for hot influent and upper region for cold influent); the strength of the circulation increases with increasing temperature differences, thereby leading to larger shear in the upper region for hot influent and lower region for cold influent.

It is apparent from Figures 7 and 8 that time plays a key role in changing the streamlines upwards or downwards, in that in some runs the negative effect of temperature variations becomes weaker after about 600 seconds (tn = 0.75) from the beginning of the simulation. For continuous inflow, as the time elapsed increases, the temperature difference decreases, the velocity decreases and the bottom or top preferential flow becomes weaker, and the flow tends to return to its isothermal form. It is also important to note that even after 600 seconds (tn = 0.75) from the beginning of the simulation the flow velocity is still greater at the bottom or top of the pond for both cold or hot influent cases; this observation is relevant while considering long-term effects of temperature variations.

The following conclusions are drawn from this study.

  • 1.

    Temperature variations in a pond can significantly change the hydraulic performance of the pond. Even 0.5 °C temperature difference between the pond water and inlet lowers the indices representing hydraulic efficiency, short-circuiting, and mixing.

  • 2.

    The decrease in hydraulic efficiency in the pond can be attributed to the temperature-induced flow patterns, which have a significant role in creating short-circuiting. For colder influents, the inflow sinks to the bottom of the pond and moves rapidly towards the outlet, while in the case of hotter influents, the inflow tends to move to the top of the pond. After 200 seconds (tn = 0.25) from the beginning of the experiments, the dye tracer in the isothermal case had just reached the outlet, while the majority of the dye tracer appears in the outlet area for the hot and cold influents. This indicates a high level of short-circuiting in the cold and hot influent cases.

  • 3.

    Inflow particles move very fast in the first 100 seconds (tn = 0.125) from the beginning of the simulations, and this high velocity was more severe for high temperature differences. A circulation flow or counter-current was also observed in the lower region for hot influents and upper region for cold influents, with the strength of this circulation increasing as the temperature difference increases.

  • 4.

    For experiment durations up to 600 seconds (tn = 0.75), the preferential flow pattern evident in some of the runs becomes considerably weaker as the experiment continues. However, the flow velocity is still greater at the bottom or top of the pond in both cases of cold or hot influents.

  • 5.

    Measures to prevent the density flows such as inlet and outlet locations and the effect of baffles should be studied.

ARC
2003
TP10 Design Guideline Manual Stormwater Treatment Devices
.
Auckland Regional Council
,
New Zealand
.
Bodin
H.
,
Mietto
A.
,
Ehde
P. M.
,
Persson
J.
&
Weisner
S. E. B.
2012
Tracer behaviour and analysis of hydraulics in experimental free water surface wetlands
.
Ecological Engineering
49
,
201
211
.
http://dx.doi.org/10.1016/j.ecoleng.2012.07.009
.
Crittenden
J. C.
,
Howe
K. J.
,
Hand
D. W.
,
Tchobanoglous
G.
&
Trussell
R. R.
2012
Principles of Water Treatment
.
Wiley, Hoboken
,
NJ, USA
.
De Oliveira
R.
,
Pearson
H. W.
,
Silva
J. V.
,
Sousa
J. T.
,
Leite
V. D.
&
Lopes
W. S.
2011
Baffled primary facultative ponds with inlets and outlets set at different levels treating domestic wastewater in northeast Brazil
.
Water Science and Technology
63
(
6
),
1183
1187
.
doi: 10.2166/wst.2011.105
.
Farjood
A.
,
Melville
B. W.
&
Shamseldin
A. Y.
2014
A study in hydraulic performance indices for sediment retention ponds
.
Paper presented at the Proceedings of the International Conference on Fluvial Hydraulics, RIVER FLOW 2014
.
Farjood
A.
,
Melville
B. W.
&
Shamseldin
A. Y.
2015
The effect of different baffles on hydraulic performance of a sediment retention pond
.
Ecological Engineering
81
,
228
232
.
http://dx.doi.org/10.1016/j.ecoleng.2015.04.063
.
Goula
A. M.
,
Kostoglou
M.
,
Karapantsios
T. D.
&
Zouboulis
A. I.
2008
The effect of influent temperature variations in a sedimentation tank for potable water treatment – a computational fluid dynamics study
.
Water Research
42
(
13
),
3405
3414
.
http://dx.doi.org/10.1016/j.watres.2008.05.002
.
Hendi
E.
,
Shamseldin
A. Y.
,
Melville
B. W.
&
Norris
S. E.
2017
Experimental investigation of the effect of temperature on flow pattern of a sediment retention pond
.
Paper presented at the Water New Zealand's 2017 Stormwater Conference
,
Auckland, New Zealand
.
Kadlec
R. H.
1994
Detention and mixing in free water wetlands
.
Ecological Engineering
3
(
4
),
345
380
.
http://dx.doi.org/10.1016/0925-8574(94)00007-7
.
Khan
S.
,
Melville
B. W.
&
Shamseldin
A. Y.
2011
Retrofitting a stormwater retention pond using a deflector island
.
Water Science and Technology
63
(
12
),
2867
2872
.
doi: 10.2166/wst.2011.569
.
Khan
S.
,
Melville
B.
&
Shamseldin
A.
2013
Design of storm-water retention ponds with floating treatment wetlands
.
Journal of Environmental Engineering
139
(
11
),
1343
1349
.
doi: 10.1061/(ASCE)EE.1943-7870.0000748
.
Macdonald
R.
&
Ernst
A.
1986
Disinfection efficiency and problems associated with maturation ponds
.
Water Science and Technology
18
(
10
),
19
28
.
Pedahzur
R.
,
Nasser
A. M.
,
Dor
I.
,
Fattal
B.
&
Shuval
H.
1993
The effect of baffle installation on the performance of a single-cell stabilization pond
.
Water Science and Technology
27
(
7–8
),
45
52
.
Persson
J.
2000
The hydraulic performance of ponds of various layouts
.
Urban Water
2
(
3
),
243
250
.
http://dx.doi.org/10.1016/S1462-0758(00)00059-5
.
Shilton
A.
2005
Pond Treatment Technology
.
IWA Publishing
,
London, UK
.
Su
T.-M.
,
Yang
S.-C.
,
Shih
S.-S.
&
Lee
H.-Y.
2009
Optimal design for hydraulic efficiency performance of free-water-surface constructed wetlands
.
Ecological Engineering
35
(
8
),
1200
1207
.
Vymazal
J.
2014
Constructed wetlands for treatment of industrial wastewaters: a review
.
Ecological Engineering
73
,
724
751
.
http://dx.doi.org/10.1016/j.ecoleng.2014.09.034
.
Wahl
M. D.
,
Brown
L. C.
,
Soboyejo
A. O.
,
Martin
J.
&
Dong
B.
2010
Quantifying the hydraulic performance of treatment wetlands using the moment index
.
Ecological Engineering
36
(
12
),
1691
1699
.
http://dx.doi.org/10.1016/j.ecoleng.2010.07.014
.
Watters
G. Z.
1972
The Hydraulics of Waste Stabilization Ponds
.
Utah Water Research Laboratory. https://digitalcommons.usu.edu/water_rep/18
.
Wilson
J. F.
,
Cobb
E. D.
&
Kilpatrick
F. A.
1968
Fluorometric Procedures for Dye Tracing
.
US Government Printing Office
,
Washington, DC, USA
.

Supplementary data