Final Report: Investigation of Hydraulic Characteristics and Alternative Model Development of Subsurface Flow Constructed Wetlands

EPA Grant Number: R825427C009
Subproject: this is subproject number 009 , established and managed by the Center Director under grant R825427
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).

Center: Urban Waste Management and Research Center (University New Orleans)
Center Director: McManis, Kenneth
Title: Investigation of Hydraulic Characteristics and Alternative Model Development of Subsurface Flow Constructed Wetlands
Investigators: Cothren, Gianna M.
Institution: University of New Orleans
EPA Project Officer: Lasat, Mitch
Project Period: August 1, 1998 through May 1, 2001
RFA: Urban Waste Management & Research Center (1998) RFA Text |  Recipients Lists
Research Category: Targeted Research


A trend of increasing dispersion with decreasing aspect ratio as a function of width and of increasing dispersion with increasing flow rate were found in previous work using a bench-scale model (Cothren, 1996). The objectives of this research are to validate these results relating dispersion and system characteristics and to perform a complete water balance and hydraulic characteristic evaluation on a selected full-scale system. The results can then be used to model the SFCW system with an alternative to the common simple plug flow approach. Specifically, the objectives of this research are to:

1. Develop a coefficient relating pan evaporation to the actual evaporation from a SFCW as part of the water balance and hydraulic characteristics evaluation.
2. Evaluate the hydraulic characteristics of a selected SFCW by performing a water balance including inflow, outflow, precipitation, evapotranspiration and seepage.
3. Perform experimental tracer studies an operational SFCW system in Louisiana in order to determine hydraulic residence time and dispersion number.
4. Verify the trend of increasing dispersion with decreasing aspect ratio found in previous research.
5. Investigate an alternative model for the SFCW system that does not neglect the effect of dispersion.

Summary/Accomplishments (Outputs/Outcomes):

Constructed wetlands are complex systems encompassing a variety of mechanisms involving hydraulics, biology and water chemistry. In addition, their performance is variable due to such things as seasonal changes and vegetative cycles (Cole, 1998). Further studies of how wetlands work are needed in order to provide engineers with detailed predictive models. Mathematical models provide a means of establishing design criteria that will ensure treatment efficacy.

The currently practiced design of subsurface flow constructed wetlands (SFCW?s) for wastewater and stormwater treatment is based on the plug flow assumption. Since SFCWs are now commonly designed with low aspect ratios to circumvent surfacing problems, plug flow is not ensured; therefore, alternative models which more accurately describe the flow should be considered. However, to solve the alternative models, a value for dispersion is required. The fundamental variables affecting the dispersion number are the interstitial velocity, which is dependent upon flow rate, porosity, and cross-sectional area, and the pore geometry, which is dependent on the media characteristics such as the average grain diameter and permeability. Although other uncertainties exist, if new advances are to be made in the design of SFCWs, dispersion estimates are required in order to model the system more accurately.

The pan evaporation and evapotranspiration with the media data were collected from the Mandeville, LA, site during each of the three seasonal phases of the study. An analysis of variance was performed between the three seasonal data sets that confirmed a significant difference in evapotranspiration according to season with a p-value less than the significance level of 0.05. Linear regression was used to develop equations relating pan evaporation to evapotranspiration. For all seasons, ET was greater than Ep, however that difference varied in magnitude. Fall showed the most deviation from Ep while winter showed the least. Many environmental factors affect water usage by plants, therefore one correlation equation may not be sufficient or feasible. Since ET has a significant effect on wetland hydraulics by lowering the water level and increasing HRT, all seasons and plant cycles should be considered when designing a SFCW cell. At the LTRC site, an analysis of variance was conducted between the three seasonal data sets showing a significant difference in evapotranspiration according to season at a significance level below 0.05. Linear regression equations relating pan evaporation to evaporation from the media were developed for all seasons. For all phases of the study, ET was greater than Ep. Reasons for this may include the heat absorption capability of the media and its ability to retain that heat. Capillary action could serve to wick the water onto the surface of the media above the water line where evaporation could take place more readily. This difference was seasonally dependent although not as much as the ET from the Mandeville media. This could be because vegetation is more susceptible to seasonal variations due to water usage associated with plant processes such as growth.

Evapotranspiration losses from wetlands can be considerable. These losses can result in lowered water level, heightened pollutant concentration and lengthened HRT. Even when SFCW systems are not planted, ET can vary substantially from Ep. These variations should be considered when designing SFCW systems.

Water Balance

Losses were estimated using each of the combinations of pan evaporation observations and relating equations that were developed. The estimated total volumetric losses were calculated and compared with the evaporation losses found from the water balance. Each equation developed yielded far different results, as did the values for pan evaporation from different locations. The closest approximation was achieved using the observed pan evaporation from the LSU site and the equation developed using the data collected during September and October 2000 (Fall equation). This combination provided an estimate that was 5% less than the evaporation calculated from the water balance equation. It was expected that this combination would provide the best estimate since the equation was derived from data that was seasonally the most appropriate and the pan evaporation was obtained from the site nearest the actual study site. Although still an underestimation, the loss found using ET was closer to the loss found from the water budget than was Ep alone. Using Ep recorded at LSU yielded a result 32% less than the loss found using the water budget where the loss estimated using the Fall equation came within 5% of the actual loss.

Tracer Studies

The mean residence times and dimensionless dispersion numbers are given in ranged from 0.04 to 0.335 for various aspect ratios which is consistent with Kadlec and Knight (1996) who quote typical dispersion numbers between 0.07 and .033 for both SFCW and FWSCW systems. Mean residence times varied depending on depth, flow rate, and aspect ratio.

Longitudinal Dispersion and Aspect Ratio

Although technically the aspect ratio was changed, essentially, only the length of the reactor was changed as samples were gathered at different ports along the length of the cell. No relationship was found to exist between aspect ratio and longitudinal dispersion as the system length increased.

Longitudinal Dispersion and Interstitial Velocity

The interstitial velocities associated with each flow rate and depth combination were calculated as 0.0126 and 0.0253 ft/min at a depth of 18 in. at 2 gpm and 4 gpm respectively. At a depth of 12 in. and 4 gpm, the velocity was 0.038 ft/min. Regression equations and coefficients of determination for each of the longitudinal dispersion and interstitial velocity relationships for each aspect ratio were determined. As expected, a relationship of increasing dispersion number with increasing interstitial velocity was found. This trend did not appear to be linear, however. Another possibility is that the relationship intensifies at some critical velocity after which the relationship could be linear or otherwise. The intensification of the relationship occurs somewhere between 0.025 and 0.038 ft/min. This theory is somewhat supported by correlation analysis that suggests no significant relationship with a correlation coefficient of 0.26 between the lower two velocities and dispersion but confirms a positive relationship when the upper two velocities are considered (correlation coefficient = 0.93). This scenario has been suggested in previous research (Cothren, 1996). The critical velocity found there was far lower, however, ranging from 0.007 to 0.008 ft/min. These results were obtained in bench scale studies, however and may be affected by scale. A wider range of velocity is necessary in order to better define a trend and to identify a possible critical velocity where the relationship intensifies.

Longitudinal Dispersion and Reynolds Number

Reynolds number is very sensitive to the kinematic viscosity of the fluid. The kinematic viscosity of water is, in turn very sensitive to the temperature of the water. The temperature of the water was assumed to be 80° F. Although no actual temperature measurements were taken, the experiments were performed in early fall when water temperature in south Louisiana fall in that range. The relationship between longitudinal dispersion Reynolds number was examined. The sudden increase in dispersion may be related to the transition from laminar to turbulent flow. Because there is no consensus on where exactly this occurs, a definite correlation cannot be verified. The transition from laminar to turbulent flow in porous media has been found to occur with Reynolds number somewhere in the range of 1 to 12. Since the increase in dispersion number also occurs in the range, a relationship may exist.

CSTR-in-Series Model

Experimental tracer data was compared to curves generated by the tanks-in series model. The curves seemed to fit the data well for the shorter distances along the reactor, but underestimated the peak as the distance down the cell was increased.


Equations were developed relating evaporation from SFCW media to pan evaporation. These equations were site specific and seasonally dependent. It can be inferred that the geographical location of the wetland cell would also affect the relationship due to climate conditions such as wind, solar radiation, ambient temperature and humidity. The equation developed for the specific media type and for the same season did a reasonably good job of predicting the evaporation from an unplanted cell, estimating the loss calculated from a water balance to within 5% where the estimate using Ep directly was 32 % less than the actual. This method could prove to be a useful tool once enough data is gathered to develop tables of coefficients for design use. Media type, plant type, geographical location and physical characteristics of the specific reactor such as shading and wind blockage must be considered. Seasonal effects are also significant. Even when the reactor is not planted, solar radiation in conjunction with the reflective/ absorptive characteristics and heat retention capability of the media cause seasonal differences.

A water balance was performed assuming that there was no loss due to seepage. The imprecision of the meter used to control inflow, especially at the low flow rate used, may have introduced some error. Since the inflow and outflow were summed over a long period, any variations in flow should have averaged out and thus provided a fairly good estimate of the total. A method to obtain a more accurate estimate of ET is essential since losses from ET can lower the water level and increase the HRT in a SFCW system, especially in southern climates.

Residence time distribution curves were developed from tracer experiments and were used to determine mean residence times and longitudinal dispersion numbers for variety of scenarios. The trend of increasing dispersion with decreasing aspect ratio found in previous research could not be verified by these results. For the aspect ratios investigated here, no significant relationship was found. This does not necessarily refute the previous findings. In the prior work, it was found that the aspect ratio relationship does not become controlling until aspect ratio is lower than 2:1. This research included only one aspect ratio in this category, so a determination of trend in this interval was impossible. Also, in the research conducted at bench scale, the aspect ratio was varied by changing the width. This was not feasible in this study because a variable length was used. Neither investigation held the volume of the cell constant.

A trend of increasing longitudinal dispersion with increasing interstitial velocity was confirmed by correlation analysis, yielding a Pearson's correlation coefficient of 0.93. This trend became more pronounced at higher velocities. There appears to be a critical velocity where this relationship intensifies. This finding agrees with the conclusion of the previous bench scale research. Since the intensification of the relationship between dispersion and velocity appears to occur at nearly the same velocity that the Reynolds number surpasses the accepted Reynolds number where turbulent flow develops, the critical velocity mentioned above is likely to coincide with the transition from laminar to turbulent flow.

The CSTR-in-Series model gave a reasonably good fit for the residence time distributions as the tracer passed the ports at the near end of the cell. As the distance down the cell increased, the model underestimated the peak concentrations.

Supplemental Keywords:

Constructed wetlands, SCFW, natural systems, dispersion, hydraulic characteristics., Scientific Discipline, Geographic Area, Waste, Municipal, Environmental Chemistry, State, Analytical Chemistry, Ecological Risk Assessment, Ecology and Ecosystems, waste minimization, urban runoff, municipal waste, groundwater quality, New Orleans (NO), waste management, constructed wetlands, technology transfer, outreach, urban waste, storm drainage systems

Progress and Final Reports:

Original Abstract
  • 1999
  • 2000

  • Main Center Abstract and Reports:

    R825427    Urban Waste Management and Research Center (University New Orleans)

    Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
    R825427C001 Comprehensive Evaluation of The Dual Trickling Filter Solids Contact Process
    R825427C002 Issues Involving the Vertical Expansion of Landfills
    R825427C003 Deep Foundations on Brownfields Sites
    R825427C004 Ambient Particulate Concentration Model for Traffic Intersections
    R825427C005 Effectiveness of Rehabilitation Approaches for I/I Reduction
    R825427C006 Urban Solid Waste Management Videos
    R825427C007 UWMRC Community Outreach Multimedia Exhibit
    R825427C008 Including New Technology into the Investigation of Inappropriate Pollutant Entries into Storm Drainage Systems - A User's Guide
    R825427C009 Investigation of Hydraulic Characteristics and Alternative Model Development of Subsurface Flow Constructed Wetlands
    R825427C010 Beneficial Use Of Urban Runoff For Wetland Enhancement
    R825427C011 Urban Storm and Waste Water Outfall Modeling
    R827933C001 Development of a Model Sediment Control Ordinance for Louisiana
    R827933C002 Inappropriate Discharge to Stormwater Drainage (Demonstration Project)
    R827933C003 Alternate Liner Evaluation Model
    R827933C004 LA DNR - DEQ - Regional Waste Management
    R827933C005 Landfill Design Specifications
    R827933C006 Geosynthetic Clay Liners as Alternative Barrier Systems
    R827933C007 Used Tire Monofill
    R827933C008 A Comparison of Upflow Anaerobic Sludge Bed (USAB) and the Anaerobic Biofilm Fluidized Bed Reactor (ABFBR) for the Treatment of Municipal Wastewater
    R827933C009 Integrated Environmental Management Plan for Shipbuilding Facilities
    R827933C010 Nicaragua
    R827933C011 Louisiana Environmental Education and Resource Program
    R827933C012 Costa Rica - Costa Rican Initiative
    R827933C013 Evaluation of Cr(VI) Exposure Assessment in the Shipbuilding Industry
    R827933C014 LaTAP, Louisiana Technical Assistance Program: Pollution Prevention for Small Businesses
    R827933C015 Louisiana Environmental Leadership Pollution Prevention Program
    R827933C016 Inexpensive Non-Toxic Pigment Substitute for Chromium in Primer for Aluminum Sibstrate
    R827933C017 China - Innovative Waste Composting Plan for the City of Benxi, People's Rupublic of China
    R827933C018 Institutional Control in Brownfields Redevelopment: A Methodology for Community Participation and Sustainability
    R827933C019 Physico-Chemical Assessment for Treatment of Storm Water From Impervious Urban Watersheds Typical of the Gulf Coast
    R827933C020 Influence of Cyclic Interfacial Redox Conditions on the Structure and Integrity of Clay Liners for Landfills Subject to Variable High Groundwater Conditions in the Gulf Coast Region
    R827933C021 Characterizing Moisture Content Within Landfills
    R827933C022 Bioreactor Landfill Moisture Management
    R827933C023 Urban Water Issues: A Video Series
    R827933C024 Water Quality Modeling in Urban Storm Water Systems
    R827933C025 The Development of a Web Based Instruction (WBI) Program for the UWMRC User's Guide (Investigation of Inappropriate Pollutant Entries Into Storm Drainage Systems)
    R827933C027 Legal Issues of SSO's: Private Property Sources and Non-NPDES Entities
    R827933C028 Brownfields Issues: A Video Series
    R827933C029 Facultative Landfill Bioreactors (FLB): A Pilot-Scale Study of Waste Stabilization, Landfill Gas Emissions, Leachate Treatment, and Landfill Geotechnical Properties
    R827933C030 Advances in Municipal Wastewater Treatment
    R827933C031 Design Criteria for Sanitary Sewer System Rehabilitation
    R827933C032 Deep Foundations in Brownfield Areas: Continuing Investigation
    R827933C033 Gradation-Based Transport, Kinetics, Coagulation, and Flocculation of Urban Watershed Rainfall-Runoff Particulate Matter
    R827933C034 Leaching and Stabilization of Solid-Phase Residuals Separated by Storm Water BMPs Capturing Urban Runoff Impacted by Transportation Activities and Infrastructure
    R827933C035 Fate of Pathogens in Storm Water Runoff
    R87933C020 Influence of Cyclic Interfacial Redox Conditions on the Structure and Integrity of Clay Liners for Landfills Subject to Variable High Groundwater Conditions in the Gulf Coast Region