Analysis of constructed treatment wetland hydraulics with the transient storage model OTIS
Introduction
As the use of constructed wetlands for wastewater treatment increases, there is an increased need for predictive tools to estimate expected levels of contaminant removal. In order to achieve accurate prediction of treatment, there must first be an accurate understanding of wetland hydraulics. Due to the varying degrees of short-circuiting experienced in many systems, tracer tests are required to determine the residence times within them. Models of wetland hydraulics may be calibrated against tracer test results in order to develop a simulation model upon which treatment performance can be forecasted.
Short-circuiting of available volume in a constructed wetland reduces the average pollutant residence time and reduces the contact area of the system, leaving large regions of isolated, stagnant water. These isolated areas may exchange water and constituents with more actively flowing regions by diffusion and dispersion. The turnover of water in these isolated areas is relatively slow compared with that in the actively flowing regions so water chemistry and kinetics will likely be different in these two areas.
Much of the constructed treatment wetland design literature assumes ideal plug-flow conditions to estimate expected treatment levels (US EPA, 1988, Water Pollution Control Federation, 1990, Reed et al., 1995). Under plug-flow conditions, an instantaneous spike of tracer injected at the inlet will exit the wetland as an identical spike one nominal residence time (ratio of system volume to volumetric flow through the system) later; no mixing or diffusion is assumed to occur. Models using continuously stirred tank reactors (CSTRs) in series have also been employed and have been recommended by the Water Pollution Control Federation (1990) to model wastewater lagoons. Under continuously stirred conditions, an instantaneous spike of tracer is assumed to mix evenly and instantaneously throughout the reactor. The tanks-in-series model has been used to match field tracer responses with some degree of success by Urban, 1990, Stairs, 1993, Kadlec, 1994, Niswander, 1997.
Experimental results from tracer tests conducted on constructed treatment wetlands typically experience a quick effluent peak of tracer followed by extended tailing (Stairs, 1993). The design models mentioned above have not done well in reproducing such highly skewed tracer responses caused by temporary storage of tracer in stagnant regions and by short-circuiting of the treatment volume. This problem has not been unique to treatment wetlands; previous modeling investigations have tried to deal with the non-ideal flow caused by temporary storage and short-circuiting in other systems. A brief synopsis follows.
Addressing ‘dead’ flow regions in chemical reactors, Hovorka (1961) used combinations of completely mixed and plug-flow vessels to define a ‘finite-stage’ model which could also reproduce the two ideal cases (plug-flow and continuously stirred). A ‘dead-flow’ model element was defined as a completely mixed vessel and a ‘live-flow’ element was defined as a combination of a completely mixed and a plug-flow vessel in series. Combined in parallel, these two elements represented a single ‘model’ element where first-order exchange could occur between the completely mixed portions of the ‘live’ and ‘dead’ elements. The ‘model’ element could then be repeated an integral number of times with convective flow occurring between ‘live-flow’ elements in order to describe the physical system.
Conceptually similar models consisting of CSTRs in series that exchange water and solute with adjacent ‘dead’ CSTRs have been proposed for modeling flow in porous media (Deans, 1963, Levich et al., 1967); in chemical reactors (Levenspiel, 1972); and for treatment wetlands (Kadlec, 1994). Kadlec (1994) model used Levenspiel's ‘Model G’ with the addition of a single plug-flow reactor in series to account for the time lag between the injection of the tracer and the arrival of the tracer at the system outlet.
In their ‘differential capacitance’ model Coates and Smith (1964) coupled the advection–dispersion equation with completely mixed storage zones to model transport in porous media. Their work was further expanded to develop the ‘two-region’ model for flow in porous media to account for the effects of dead-end pores, aggregated media, and thin-film covered particles (e.g. van Genuchten and Wierenga, 1976, van Genuchten and Wagenet, 1989). Noting the effects of temporary storage in wetland hydraulics, Dierks (1997) applied the two-region model to surface-water wetland tracer tests conducted in an isolation cell at the University of Florida in an attempt to determine the proper scale of a representative elementary volume for a surface flow wetland. Similarly, Werner and Kadlec (2000) extended Levenspiel's ‘Model G’ and developed the zones of diminished mixing (ZDM) model to simulate wetland tracer responses. The ZDM model consists of an infinite number of plug-flow reactors in series, each exchanging mass with a side CSTR. To make their model more realistic they introduced longitudinal dispersion to the plug-flow reactors, resulting in a model that is very similar to the ‘differential capacitance’ and ‘two-region’ models.
Along the lines of the model developed by Coates and Smith (1964), ‘transient storage’ models that couple the advection–dispersion equation with completely mixed storage zones have been applied to rivers and streams (e.g. Nordin and Troutman, 1980, Bencala and Walters, 1983). Stream and river studies have used transient storage models to investigate pool-and-riffle streams (Bencala and Walters, 1983), solute sorption to sediments (Bencala, 1983, Bencala, 1984), nutrient-periphyton dynamics (DeAngelis et al., 1995), hyporheic exchange (Harvey et al., 1996), equilibrium chemistry (Runkel et al., 1996), dissolved oxygen (Chapra and Runkel, 1999), and in-stream gas transfer (Chapra and Wilcock, 2000).
The objective of the present work is to apply the United States Geological Survey (USGS) transient storage model one-dimensional transport with inflow and storage (OTIS) to wetland hydraulics using experimental tracer results from a constructed treatment wetland. The use of OTIS has the advantage of being well documented, readily available to users via the World Wide Web, and is complemented with a non-linear regression package for parameter determination.
Section snippets
The OTIS model
The transient storage solute transport model OTIS was coded in fortran and documented by Runkel (1998). The model including documentation is free and available via the USGS website (http://www.co.water.usgs.gov/otis). The model was created to quantify the hydraulic parameters that influence temporary storage in rivers and streams.
The term ‘transient storage’ refers to the temporary detainment of water (and any solutes dissolved or other matter suspended therein) in recirculating eddies and dead
Description of site and tracer experiments
The simulations conducted here use experimental results of bromide tracer experiments conducted at the Orlando Easterly Wetland (OEW) located near Christmas, FL (Martinez and Wise, 2003). The system was designed for nutrient polishing of tertiary treated domestic wastewater and consists of 17 surface flow constructed wetland cells covering approximately 1250 acres (Fig. 2). The cells are separated by earthen berms and use adjustable thin-plate weirs to control flow (through culverts) from one
Results
Modeling results of Cells 1–15 of the OEW are presented here. Model simulations were made using lumped composite tracer responses for cells with multiple weirs. Flows were approximately constant during the study period for all wetland cells with the exception of Cell 11. Accordingly the remaining cells were modeled assuming steady-flow conditions, while Cell 11 was modeled with a transient analysis to account for unsteady flow. Since cell outflows were found to be approximately equal to inflows
Discussion
As can be seen in Fig. 4, Fig. 5, Fig. 6, Fig. 7, OTIS modeled the data very well for both the steady and unsteady flow simulations. The response of Cell 1 appears to be bimodal, indicating that two distinct flow-paths, one ‘slow’ and one ‘fast’, exist (Fig. 4). As can be seen the resulting fit is a compromise between the two. This will be addressed further below. Storage zone exchange coefficients were found to vary over four orders of magnitude and the ratio of A/As varied over a three orders
Conclusion
The transient storage model OTIS was found to fit and parameterize experimental tracer responses from treatment cells of the OEW. The model can account for short-circuiting of the treatment volume and temporary storage of tracer in storage zones once calibrated to experimental tracer responses. Short-circuiting and temporary storage in constructed treatment wetlands are likely caused by variable bathymetry, plant communities and densities, and possibly hyporheic exchange. Parameter
Symbols used
A main channel cross-sectional area (L2) As storage zone cross-sectional area (L2) As,i storage zone cross-sectional area of storage zone i (L2) C main channel solute concentration (ML−3) Cs storage zone solute concentration (ML−3) D longitudinal dispersion coefficient (L2T−1) DaI experimental Damköhler number L flow-path length (L) Pe Peclet number Q volumetric flow (L3T−1) qs,i storage zone exchange flux (L2T−1) T time (T) ts,i storage zone retention time (T) first moment of the tracer response curve (T) U water
References (41)
Continuous flow systems: distribution of residence times
Chem. Eng. Sci.
(1953)- et al.
Modelling nutrient-periphyton dynamics in streams: the importance of transient storage zones
Ecol. Model.
(1995) Detention and mixing in free water wetlands
Ecol. Eng.
(1994)- et al.
On hydrodynamic mixing in a model of a porous medium with stagnant zones
Chem. Eng. Sci.
(1967) - et al.
Wetland residence time distribution modeling
Ecol. Eng.
(2000) - et al.
Direct comparison of kinetic and local equilibrium formulations for solute transport affected by surface reactions
Water Resour. Res.
(1987) Simulation of solute transport in a mountain pool-and-riffle stream with a kinetic mass transfer model for sorption
Water Resour. Res.
(1983)Interactions of solutes and streambed sediment 2. A dynamic analysis of coupled hydrologic and chemical processes that determine solute transport
Water Resour. Res.
(1984)- et al.
Simulation of solute transport in a mountain pool-and-riffle stream: a transient storage model
Water Resour. Res.
(1983) - et al.
Modeling impact of storage zones on stream dissolved oxygen
J. Environ. Eng.
(1999)
Transient storage and gas transfer in lowland stream
J. Environ. Eng.
Characterizing multiple timescales of stream and storage zone interaction that affect solute fate and transport in streams
Water Resour. Res.
Dead-end pore volume and dispersion in porous media
Soc. Petrol. Eng. J.
A mathematical model for dispersion in the direction of flow in porous media
Soc. Petrol. Eng. J.
Transient storage and hyporheic flow along the Willamette River, Oregon: field measurements and model estimates
Water Resour. Res.
Relationships between hydraulic parameters in a small stream under varying flow and seasonal conditions
Hydrol. Process.
Evaluating the reliability of the stream tracer approach to characterize stream-subsurface water exchange
Water Resour. Res.
Cited by (63)
Net export of E. coli from a constructed wetland treating agricultural drainage
2023, Ecological EngineeringProcess formulations and controlling factors of pesticide dissipation in artificial ponds: A critical review
2023, Ecological EngineeringCitation Excerpt :Modeling is an efficient, comprehensive tool for assessing and setting up facilities to treat contaminated environments (Warren et al., 2002). There are mainly two types of models that describe pesticide fate in the environment: (i) physical 3D and 2D models that simulate the hydraulic performance of wetlands by considering water velocity profiles and the vegetation impact of the flow (Carleton and Montas, 2010; Luo et al., 2016; Martinez and Wise, 2003a; Tsavdaris et al., 2013) by testing different internal configurations of artificial ponds (i.e., baffles). ( ii) Conceptual models that simulate the biogeochemistry and the dissipation performance of the wetland for a large-scale application (Gobas et al., 2018; Willkommen et al., 2018).
Review of experimental procedures and modelling techniques for flow behaviour and their relation to residence time in constructed wetlands
2021, Journal of Water Process Engineering
- 1
Tel.: +1-352-846-1745; fax: +1-352-392-3076.