A stem spacing-based non-dimensional model for predicting longitudinal dispersion in low-density emergent vegetation
- 196 Downloads
Abstract
Predicting how pollutants disperse in vegetation is necessary to protect natural watercourses. This can be done using the one-dimensional advection dispersion equation, which requires estimates of longitudinal dispersion coefficients in vegetation. Dye tracing was used to obtain longitudinal dispersion coefficients in emergent artificial vegetation of different densities and stem diameters. Based on these results, a simple non-dimensional model, depending on velocity and stem spacing, was developed to predict the longitudinal dispersion coefficient in uniform emergent vegetation at low densities (solid volume fractions < 0.1). Predictions of the longitudinal dispersion coefficient from this simple model were compared with predictions from a more complex expression for a range of experimental data, including real vegetation. The simple model was found to predict correct order of magnitude dispersion coefficients and to perform as well as the more complex expression. The simple model requires fewer parameters and provides a robust engineering approximation.
Keywords
Stem spacing Longitudinal dispersion Solute transport 1D modelling Vegetated flows Cylinder arraysIntroduction
Natural watercourses often contain vegetation (O’Hare 2015). Furthermore, prior to entering natural watercourses, stormwater is often treated within vegetated sustainable drainage systems (SuDS) to reduce the quantity of pollutants in surface runoff (Woods-Ballard et al. 2015). Applying the 1D ADE to predict pollutant transport therefore requires estimates of longitudinal dispersion coefficient (D_{x}) within vegetation. While velocities can be estimated from simple hydraulics or numerical models, reliable estimation of the longitudinal dispersion coefficient based on vegetation characteristics is often problematic (Sonnenwald et al. 2017). This paper investigates the prediction of the longitudinal dispersion coefficient in uniform emergent vegetation at low densities. It presents results from new laboratory measurements of longitudinal dispersion and compares new and existing predictors of longitudinal dispersion coefficient to measured values.
Predicting D _{x} in vegetation
Tanino (2012), in a review of mixing in vegetation, suggested that there are primarily three mixing mechanisms contributing to longitudinal dispersion within emergent vegetation: turbulent diffusion; secondary wake dispersion; and vortex trapping. Turbulent diffusion within vegetation is the result of instantaneous velocity fluctuations caused by eddies generated by stems. Secondary wake dispersion is caused by velocity-field heterogeneity resulting in different travel times for particles (differential advection). Vortex trapping is caused by the temporary entrainment of particles in vortices behind stems.
Of these three processes, only secondary wake dispersion contributes significantly to longitudinal dispersion at low densities (ϕ < 0.1). Turbulent diffusion is the primary transverse mixing mechanism at low densities (Tanino and Nepf 2008). However, as total transverse dispersion is typically an order of magnitude lower than total longitudinal dispersion (Sonnenwald et al. 2017), it does not make a significant contribution to longitudinal dispersion. Similarly, dispersion due to vortex trapping is typically much lower than secondary wake dispersion at low densities (White and Nepf 2003). Therefore, secondary wake dispersion should provide a reasonable approximation of total longitudinal dispersion within vegetation at low densities.
Comparison of Eq. (4) to experimental data suggested that it provides correct order of magnitude predictions of the longitudinal dispersion coefficient in vegetation when using values of C_{D} estimated with Eq. (5) (Sonnenwald et al. 2018a). However, Eq. (4) depends on several simplifying assumptions for values of σ_{u}*^{2} and Sc_{t} and is insensitive to C_{D}. A simplified model has previously been developed by Nepf (2012) for predicting the transverse dispersion coefficient in emergent vegetation based on velocity and stem diameter. This paper aims to explore whether a similar approach could be adopted for predicting the longitudinal dispersion coefficient in emergent vegetation.
Previous studies
Previous studies investigating longitudinal dispersion in uniform emergent vegetation
Description | d (mm) | ϕ | s (mm)^{a} | U (mm s^{−1}) | References |
---|---|---|---|---|---|
Mixed real species | 1–10 | 0.002–0.022 | 7–26 | 14–38 | Huang et al. (2008) |
Artificial random | 6 | 0.010–0.055 | 8–25 | 29–74 | Nepf et al. (1997) |
Carex | 10–55 | 0.002–0.059 | 70–104 | 99–232 | Shucksmith et al. (2010) |
Phragmites australis | 3 | 0.002 | 28–30 | 171–242 | Shucksmith et al. (2010) |
Regular periodic | 4 | 0.005 | 51.9^{b} | 7–50 | Sonnenwald et al. (2016) |
Typha latifolia | 10–19 | 0.013–0.047 | 29–36 | 9–29 | Sonnenwald et al. (2017) |
Artificial random | 6 | 0.010–0.064 | 7–25 | 12–97 | White and Nepf (2003) |
Methodology
A 15-m long, 300-mm-wide recirculating Armfield flume was fitted with uniform artificial emergent vegetation, and dye tracing was conducted. Uniform flow was established at velocities of 7–122 mm s^{−1} by adjusting flume slope and tailgate, and confirmed using point gauges along the length of the flume. Flow depth was set at 150 mm. A diffuser plate was placed directly after the inlet to straighten the flow.
Artificial emergent vegetation characterisation
Description | d (mm) | ϕ | Estimated s (mm)^{a} | s (mm)^{b} | s_{50} (mm) |
---|---|---|---|---|---|
Regular periodic^{c} | 4 | 0.005 | 24.9 | 51.9 | 51.9 |
Pseudo-random | 4 | 0.005 | 24.9 | 29.4 | 23.9 |
Random | 8 | 0.027 | 18.0 | 22.7 | 22.3 |
Rhodamine WT dye tracing was carried out using four mid-channel mid-depth Turner Designs Cyclops 7 fluorometers to record temporal concentration profiles. The four fluorometers formed three 2.5-m test reaches for the 4-mm stems and three 3-m test reaches for the 8-mm stems, giving total experimental lengths of 7.5 m and 9 m, respectively. The first fluorometer was located 4 m downstream of the inlet diffuser plate. The different reach lengths were to accommodate the difference in vegetation pattern repetition. The 4-mm stems covered the full length of the flume. The 8-mm stems covered between 3 m and 14 m from the inlet diffuser plate, with 1 m before and after the first and last fluorometer.
Determining D _{x} from experimental results
Results and discussion
Figure 6b shows predictions of D_{x} made using Eqs. (4) and (5) compared to measured values of D_{x}. Equations (4) and (5) offer a better prediction of D_{x} for data from Nepf et al. (1997), White and Nepf (2003), and Sonnenwald et al. (2017) than Eq. (7). The vegetation from this study and the vegetation of Shucksmith et al. (2010) are less well predicted. Neither Eq. (7) nor Eqs. (4) and (5) predict the Huang et al. (2008) real vegetation well. Predictions with Eqs. (4) and (5) are biased towards underestimates of D_{x}. The scatter of predictions from Eq. (7) may be due to the use of an estimated value of s rather than a measured value.
Normalisation by median stem spacing, rather than stem diameter, appears to be a useful method of incorporating spatial heterogeneity into non-dimensional longitudinal dispersion coefficient, as it is successful in distinguishing between vegetation arrangements. However, it is worth noting that s_{50}, like d, is still a single length-scale characterisation. Most current theory is based on one such characterisation, while in reality multiple length scales are common, as described in Sonnenwald et al. (2017).
Future work should investigate the suitability of Eq. (7) at higher solid volume fractions and for more types of vegetation. Additional work is also needed to investigate how vegetation is described when calculating dispersion coefficient in complex vegetation arrangements, e.g. with varying stem diameter or stem spacing distributions.
Conclusions
Longitudinal dispersion coefficient (D_{x}) values obtained from dye tracing in artificial emergent vegetation fitted the expected trend of a linear increase with velocity, but could not be normalised using stem diameter. Stem spacing is suggested here for the first time as the appropriate length-scale normalisation for D_{x} in vegetation. From this, D_{x} in vegetation can be modelled by a new simple expression dependent on median stem edge-to-edge spacing. This new model showed reasonable performance when applied to other experimental data, including real vegetation. Although a more complex expression from the literature predicts D_{x} in vegetation equally well, it has multiple implicit assumptions. The new expression presented here gives robust correct order of magnitude estimates of longitudinal dispersion coefficient in vegetation suitable for engineering purposes.
Notes
Acknowledgements
The authors thank Ayuk Merchant, Nathan Wilson, Alexandre Delalande, and Zoe Ball who conducted the laboratory work, and Ian Baylis for his technical support at the University of Warwick. This work was supported by the Engineering and Physical Sciences Research Council (EPSRC Grants EP/K024442/1, EP/K025589/1).
References
- Danish Hydraulic Institute (2009) MIKE 11: a modelling system for rivers and channels. Reference manual. DenmarkGoogle Scholar
- Ergun S (1952) Fluid flow through packed columns. Chem Eng Prog 48:89–94Google Scholar
- Fischer HB, List JE, Koh CR, Imberger J, Brooks NH (1979) Mixing in inland and coastal waters. Elsevier, AmsterdamGoogle Scholar
- Huang YH, Saiers JE, Harvey JW, Noe GB, Mylon S (2008) Advection, dispersion, and filtration of fine particles within emergent vegetation of the Florida everglades. Water Resour Res 44:W04408Google Scholar
- Lightbody AF, Nepf HM (2006) Prediction of velocity profiles and longitudinal dispersion in emergent salt marsh vegetation. Limnol Oceanogr 51(1):218–228CrossRefGoogle Scholar
- Nepf HM (1999) Drag, turbulence, and diffusion in flow through emergent vegetation. Water Resour Res 35(2):479–489CrossRefGoogle Scholar
- Nepf HM (2012) Flow and transport in regions with aquatic vegetation. Annu Rev Fluid Mech 44:123–142CrossRefGoogle Scholar
- Nepf HM, Mugnier C, Zavistoski R (1997) The effects of vegetation on longitudinal dispersion. Estuar Coast Shelf Sci 44(6):675–684CrossRefGoogle Scholar
- O’Hare MT (2015) Aquatic vegetation—a primer for hydrodynamic specialists. J Hydraul Res 53(6):687–698CrossRefGoogle Scholar
- Shucksmith J, Boxall J, Guymer I (2010) Effects of emergent and submerged natural vegetation on longitudinal mixing in open channel flow. Water Resour Res 46(4):W04504CrossRefGoogle Scholar
- Sonnenwald F, Guymer I, Marchant A, Wilson N, Golzar M, Stovin V (2016). Estimating stem-scale mixing coefficients in low velocity flows. In: sustainable hydraulics in the era of global change: proceedings of the 4th IAHR Europe congress. CRC PressGoogle Scholar
- Sonnenwald F, Hart J, West P, Stovin V, Guymer I (2017) Transverse and longitudinal mixing in real emergent vegetation at low velocities. Water Resour Res 53(1):961–978CrossRefGoogle Scholar
- Sonnenwald F, Stovin V, Guymer I (2018a) Use of drag coefficient to predict dispersion coefficients in emergent vegetation at low velocities. Paper presented at the 12th international symposium on ecohydraulics, Tokyo, JapanGoogle Scholar
- Sonnenwald F, Stovin V, Guymer I (2018b) Estimating drag coefficient for arrays of rigid cylinders representing vegetation. J Hydraul Res. https://doi.org/10.1080/00221686.2018.1494050 CrossRefGoogle Scholar
- Tanino Y (2012) Flow and mass transport in vegetated surface waters. In: Gualtieri C, Mihailovic DT (eds) Fluid mechanics of environmental interfaces, 2nd edn. Taylor & Francis, Abingdon, pp 369–394Google Scholar
- Tanino Y, Nepf HM (2008) Lateral dispersion in random cylinder arrays at high Reynolds number. J Fluid Mech 600:339–371CrossRefGoogle Scholar
- The MathWorks Inc (2018) MATLAB R2018a, Natick, MAGoogle Scholar
- Tinoco RO, Cowen EA (2013) The direct and indirect measurement of boundary stress and drag on individual and complex arrays of elements. Exp Fluids 54(4):1–16CrossRefGoogle Scholar
- White BL, Nepf HM (2003) Scalar transport in random cylinder arrays at moderate Reynolds number. J Fluid Mech 487:43–79CrossRefGoogle Scholar
- Woods-Ballard B, Wilson S, Udale-Clarke H, Illman S, Scott T, Ashley R, Kellagher R (2015) The SUDS manual, report C753. CIRIA, LondonGoogle Scholar
Copyright information
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.