Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

Model instability and channel connectivity for 2D coastal marsh simulations


Reduced freshwater inflow into a coastal marsh can result in environmental stress through episodic hypersalinity. Hydrodynamic models can be used to evaluate salinity-control strategies when freshwater supplies are constrained by climate or increasing urban demands. However, there remain significant scientific, engineering, and technical barriers to correctly modeling salinity transport in such systems. In particular, the numerical instability at the wetting/drying front caused by strong wind stress and steep surface gradient and the inappropriate representation of the complex channels at practical computational scales are unsolved problems. This study documents recent achievements in modeling the time–space evolution of shallow marsh salinity using the Fine Resolution Environmental Hydrodynamic model (Frehd) applied to the Nueces River Delta (Texas, USA). The 2D depth-integrated model is tested across a variety of bathymetric representations derived from high-resolution lidar data to evaluate the effects of grid refinement and a variety of bathymetry processing methods. Novel treatments are proposed and tested to suppress unrealistic velocities and scalar concentrations caused by rapid wetting/drying and strong wind stress. The model results are compared with the field data collected at 12 spatially-distributed locations across the marsh, yielding good model-data agreements for free surface elevation and reasonable agreements for salinity. Analyses of results indicate that the critical difficulty for capturing salinity transport is in correctly representing connectivity effects (both blocking and channel features) at fine scales on the coarse grid without overestimating fluxes. Modeled water surface elevations are relatively robust to poor representation of connectivity whereas the salinity distribution is strongly affected, particularly at key choke points. This study defines a set of future challenges in developing automated methods for evaluating and preserving geometric connectivity at practical model grid resolution.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12


  1. 1.


  2. 2.


  3. 3.



  1. 1.

    Alexander H, Dunton K (2002) Freshwater inundation effects on emergent vegetation of a hypersaline salt marsh. Estuaries 25(6B):1426–1435

  2. 2.

    Alvarez M, Carol E, Dapena C (2015) The role of evapotranspiration in the groudwater hydrochemistry of an arid coastal wetland (Peninsula Valdes, Argentina). Sci Total Environ 506–507:299–307. https://doi.org/10.1016/j.scitotenv.2014.11.028

  3. 3.

    Bricker J, Inagaki S, Monismith G (2005) Bed drag coefficient variability under wind waves in a tidal estuary. J Hydraul Eng 131(6):497–508. https://doi.org/10.1061/(ASCE)0733-9429(2005)131:6(497)

  4. 4.

    Burchard H, Deleersnijder E, Meister A (2003) A high-order conservative Patankar-type discretisation for stiff systems of production–destruction equations. Appl Numer Math 47:1–30. https://doi.org/10.1016/S0168-9274(03)00101-6

  5. 5.

    Candy AS (2017) An implicit wetting and drying approach for non-hydrostatic baroclinic flows in high aspect ratio domains. Adv Water Resour 102:188–205. https://doi.org/10.1016/j.advwatres.2017.02.004

  6. 6.

    Casulli V (1990) Semi-implicit finite-difference methods for the 2-dimensional shallow-water equations. J Comput Phys 86:56–74

  7. 7.

    Casulli V, Cattani E (1994) Stability, accuracy and efficiency of a semi-implicit method for three-dimensional shallow water flow. Comput Math Appl 27(4):99–112

  8. 8.

    Casulli V (1999) A semi-implicit finite difference method for non-hydrostatic free-surface flows. Int J Numer Methods Fluids 30:425–440

  9. 9.

    Cea L, French JR, Vazquez-Cendon ME (2006) Numerical modeling of tidal flows in complex estuaries including turbulence: an unstructured finite volume solver and experimental validation. Int J Numer Methods Eng 67:1909–1932. https://doi.org/10.1002/nme.1702

  10. 10.

    Cea L, French JR (2012) Bathymetric error estimation for the calibration and validation of estuarine hydrodynamic models. Estuar Coast Shelf Sci 100:124–132. https://doi.org/10.1016/j.ecss.2012.01.004

  11. 11.

    Chao X, Jia Y, Douglas Shields F Jr, Wang SSY, Cooper CM (2008) Three-dimensional numerical modeling of cohesive sediment transport and wind wave impact in a shallow oxbow lake. Adv Water Resour 31:1004–1014. https://doi.org/10.1016/j.advwatres.2008.04.005

  12. 12.

    Cheng RT, Casulli V, Gartner JW (1993) Tidal, residual, intertidal mudflat (TRIM) model and its applications to San Francisco Bay. Calif Estuar Coast Shelf Sci 36:235–280

  13. 13.

    Chou Y, Holleman R, Fringer OB, Stacey MT, Monismith SG, Koseff JR (2015) Three-dimensional wave-coupled hydrodynamics modeling in South San Francisco Bay. Comput Geosci 85:10–21. https://doi.org/10.1016/j.cageo.2015.08.010

  14. 14.

    Chua V, Fringer OB (2011) Sensitivity analysis of three-dimensional salinity simulations in North San Francisco Bay using the unstructured-grid SUNTANS model. Ocean Model 39:332–350. https://doi.org/10.1016/j.ocemod.2011.05.007

  15. 15.

    Gross ES, Koseff JR, Monismith SG (1999) Evaluation of advective schemes for estuarine salinity simulations. J Hydraul Eng 125(1):32–46

  16. 16.

    Gross ES, Bonaventur L, Rosatti G (2002) Consistency with continuity in conservative advection schemes for free-surface models. Int J Numer Methods Fluids 38:307–327

  17. 17.

    Hajduk H, Hodges BR, Aizinger V, Reuter V (2018) Locally filtered transport for computational efficiency in multi-component advection-reaction models. Environ Model Softw 102:185–198

  18. 18.

    Herbert ER, Boon P, Burgin AJ, Neubauer SC, Franklin RB, Ardon M, Hopfensperger KN, Lamers LPM, Gell P (2015) A global perspective on wetland salinization: ecological consequences of a growing threat to freshwater wetlands. Ecosphere 6(10):206. https://doi.org/10.1890/ES14-00534.1

  19. 19.

    Hill EM, Tunnell JW, Nicolau BA (2015) Spatial and temporal effects of the Rincon Bayou Pipiline on hypersaline conditions in the Lower Nueces Delta, Texas, USA. Tex Water J 6(1):11–32

  20. 20.

    Hodges BR (2014) A new approach to the local time stepping problem for scalar transport. Ocean Model 77:1–19. https://doi.org/10.1016/j.ocemod.2014.02.007

  21. 21.

    Hodges BR, Liu F (2014) Rivers and electric networks: crossing disciplines in modeling and simulation. Found Trends Electron Des Autom 8(1):1–116. https://doi.org/10.1561/1000000033

  22. 22.

    Hodges BR (2015) Representing hydrodynamically important blocking features in coastal or riverine lidar topography. Nat Hazards Earth Syst Sci 56:1011–1023. https://doi.org/10.5194/nhessd-3-1427-2015

  23. 23.

    Hodges BR, Imberger J, Saggio A, Winters K (2000) Modeling basin-scale internal waves in a stratified lake. Limnol Oceanogr 47(7):1603–1620

  24. 24.

    Hodges BR (2004) Accuracy order of Crank–Nicolson discretization for hydrostatic free surface flow. ASCE J Eng Mech 130(8):904–910. https://doi.org/10.1061/(ASCE)0733-9399(2004)130:8(904)

  25. 25.

    Hodges BR, Rueda FJ (2008) Semi-implicit two-level predictor-corrector methods for non-linearly coupled, hydrostatic, barotropic/baroclinic flows. Int J Comput Fluid Dyn 22(9):593–607. https://doi.org/10.1080/10618560802353389

  26. 26.

    Inoue M, Park D, Justic D, Wiseman WJ Jr (2008) A high-resolution integrated hydrology-hydrodynamic model of the Barataria Basin system. Environ Model Softw 23:1122–1132. https://doi.org/10.1016/j.envsoft.2008.02.011

  27. 27.

    Ip JTC, Lynch DR, Friedrichs CT (1998) Simulation of estuarine flooding and dewatering with application to Great Bay. New Hamps Estuar Coast Shelf Sci 47:119–141

  28. 28.

    Martins R, Leandro J, Djordjevic S (2018) Wetting and drying numerical treatments for the Roe Riemann scheme. J Hydraul Res 56(2):256–267. https://doi.org/10.1080/00221686.2017.1289256

  29. 29.

    McMahon TA, Peel MC, Lowe L, Srikanthan R, McVicar TR (2013) Estimating actual, potential, reference crop and pan evaporation using standard meteorological data: a pragmatic synthesis. Hydrol Earth Syst Sci 17:1331–1363. https://doi.org/10.5194/hess-17-1331-2013

  30. 30.

    Medeiros SC, Hagen SC (2013) Review of wetting and drying algorithms for numerical tidal flow models. Int J Numer Methods Fluids 71:473–487. https://doi.org/10.1002/fld.3668

  31. 31.

    Murillo J, Garcia-Navarro P, Burguete J (2009) Conservative numerical simulation of multi-component transport in two-dimensional unsteady shallow water flow. J Comput Phys 228:5539–5573. https://doi.org/10.1016/j.jcp.2009.04.039

  32. 32.

    Patankar SV (1980) Numerical heat transfer and fluid flow. McGraw-Hill, New York

  33. 33.

    Penman HL (1948) Natural evaporation from open water, bare soil and grass. Proc R Soc Lond A 193:120–145

  34. 34.

    Ralston DK, Geyer WR, Lerczak JA (2010) Structure, variability, and salt flux in a strongly forced salt wedge estuary. J Geophys Res 115:C06005. https://doi.org/10.1029/2009JC005806

  35. 35.

    Rayson MD, Gross ES, Fringer OB (2015) Modeling the tidal and sub-tidal hydrodynamics in a shallow, micro-tidal estuary. Ocean Model 89:29–44. https://doi.org/10.1016/j.ocemod.2015.02.002

  36. 36.

    Del Rosario EA, Montagna PA (2018) Effects of the Rincon Bayou pipeline on salinity in the upper nueces delta. Tex Water J 9(1):30–49

  37. 37.

    Rueda FJ, Sanmiguel-Rojas E, Hodges BR (2007) Baroclinic stability for a family of two-level, semi-implicit numerical methods for the 3D shallow water equations. Int J Numer Methods Fluids 54(3):237–268. https://doi.org/10.1002/fld.1391

  38. 38.

    Ryan AJ, Hodges BR (2011) Modeling hydrodynamic fluxes in the nueces river delta, center for research in water resources, University of Texas at Austin, CRWR Online Report 11-07, p 92

  39. 39.

    Schoenbaechler C, Negusse S, Guthrie CG (2014) Nueces Delta data collection for calibration and validation of the Nueces Delta Hydrodynamic Model, Technical Report, Bay and Estuaries Program, Surface Water Resources Division, Texas Water Development Board

  40. 40.

    Silva CP, Marti CL, Imberger J (2014) Horizontal transport, mixing and retention in a large, shallow estuary: Rio de la plata. Environ Fluid Mech 14:1173–1197. https://doi.org/10.1007/s10652-013-9330-3

  41. 41.

    Snedden GA, Cretini K, Patton B (2015) Inundation and salinity impacts to above- and belowground productivity in Spartina patens and Spartina alterniflora in the Mississippi River deltaic plain: Implications for using river diversions as restoration tools. Ecol Eng 81:133–139. https://doi.org/10.1016/j.ecoleng.2015.04.035

  42. 42.

    Spitz Y, Klinck J (1998) Estimate of bottom and surface stress during a spring-neap tide cycle by dynamical assimilation of tide gauge observations in the Chesapeake Bay. J Geophys Res 103(C6):12761–12782

  43. 43.

    Stachelek J, Dunton KH (2013) Freshwater inflow requirements for the Nueces Delta, Texas: Spartina alterniflora as an indicator of ecosystem condition. Tex Water J 4(2):62–73

  44. 44.

    Teng J, Jakeman AJ, Vaze J, Croke BFW, Dutta D, Kim S (2017) Flood inundation modelling: a review of methods, recent advances and uncertainty analysis. Environ Model Softw 90:201–216. https://doi.org/10.1016/j.envsoft.2017.01.006

  45. 45.

    Torres R, Styles R (2007) Effects of topographic structure on salt marsh currents. J Geophys Res 112:F02023. https://doi.org/10.1029/2006JF000508

  46. 46.

    Tracy H (1957) Discharge characteristics of broad-crested weir. Technical report, Geological Surface Circular, 397

  47. 47.

    Tunnell JW, Lloyd L (2011) Effects of Rincon Bayou Pipeline inflows on salinity structure within the Nueces Delta, Texas, Corpus Christi (Texas): Coastal Bend Bays & Estuaries Program project 1106, CBBEP-76. p 23

  48. 48.

    Wang B, Fringer OB, Giddings SN, Fong DA (2009) High-resolution simulations of a macrotidal estuary using SUNTANS. Ocean Model 28:167–192. https://doi.org/10.1016/j.ocemod.2008.08.008

  49. 49.

    Wadzuk BM, Hodges BR (2009) Hydrostatic versus nonhydrostatic Euler-equation modeling of nonlinear internal waves. ASCE J Eng Mech 135(4):1069–1080. https://doi.org/10.1061/(ASCE)0733-9399(2009)135:10(1069)

  50. 50.

    Ward GH (1997) Process and trends of circulation within the corpus christi bay national estuary program study area, Publication CCBNEP-21. Coastal Bend Bays & Estuaries Program, Corpus Christi, Texas, p 381. http://cbbep.org/publications/CCBNEP21.pdf

  51. 51.

    Watson EB, Szura K, Wigand C, Raposa KB, Blount K, Cencer M (2016) Sea level rise, drought and the decline of Spartina patens in New England marshes. Biol Conserv 196:173–181. https://doi.org/10.1016/j.biocon.2016.02.011

  52. 52.

    Wunderlich WO (1972) Heat and Mass Transfer Between a Water Surface and the Atmosphere, Water Resources Research Laboratory Report No. 14, Tennessee Valley Authority Report No. 0-6803, Norris Tennessee

  53. 53.

    Yuan D, Lin B, Falconer RA (2007) A modelling study of residence time in a macro-tidal estuary. Estuar Coast Shelf Sci 71:401–411. https://doi.org/10.1016/j.ecss.2006.08.023

  54. 54.

    Zacharias I, Gianni A (2008) Hydrodynamic and dispersion modeling as a tool for restoration of coastal ecosystems. Application to a re-flooded lagoon. Environ Model Softw 23:751–767. https://doi.org/10.1016/j.envsoft.2007.09.007

  55. 55.

    Zhang C, Li L, Lockington D (2014) Numerical study of evaporation-induced salt accumulation and precipitation in bare saline soils: mechanism and feedback. Water Resour Res 50:8084–8106. https://doi.org/10.1002/2013WR015127

  56. 56.

    Zheng L, Chen C, Liu H (2003) A modeling study of the Satilla River Estuary, Georgia: flooding-drying process and water exchange over the salt marsh-estuary-shelf complex. Estuaries 26(3):651–669

Download references


This work has been supported by the Texas Water Development Board under interagency cooperation contracts 1400011719 and 1600011928. Field work was conducted by the TWDB with funding from the US Army Corps of Engineers.

Author information

Correspondence to Z. Li.


Appendix A. Model spin up

The “spin-up” time for a model is the simulation time that it takes for the results to be independent of the initial conditions. Spin-up for a fresh/salt marsh simulation is inherently challenging. Unless the sampling period includes a complete flush of the system, we cannot start from a “clean slate” of zero salinity and expect to reach the actual salinity distribution by some date. In contrast, the hydraulic memory of the velocity and elevation fields is relatively short and can be approximated by the time it takes for an increase in tidal elevation to be seen throughout the domain, which allows uniform conditions to be readily applied as a starting guess. To examine the spin-up behavior of the NDHM, we conducted test simulations starting two weeks apart (from Nov. 15 and Dec. 1, 2012 respectively) to evaluate when the model results are independent of the initial starting date. The start date was chosen based on availability of field data and to ensure sufficient spin-up time prior to the pump operations in summer. Note that during the winter, the secular water elevations in the Nueces and Corpus Christi Bay systems are declining towards a semi-annual low in January [50], and it can be expected that spin-up times during other stages of the secular cycle might be different. For the spin-up simulations, the NDHM was started from quiescent water (\(u_j = 0, j \in \left\{ 1,2 \right\}\)) with a uniform free-surface elevation equal to the tidal boundary condition. The initial condition for salinity was computed using the ordinary kriging method,Footnote 3 in which the salinity field was interpolated based on measurements from the 12 stations at the beginning of the model start date.

We consider an adequate spin-up time as the interval when the residual (difference between 2 simulation results) is less than 2 psu for salinity and 0.002 m for free surface. Using this metric, the spin-up times as well as a final salinity residual after 60 days from Dec. 1, 2012 are listed in Table 5. As expected, the spin-up times for free surface elevation are significantly shorter than that for salinity. The poorest result was at Nueces4 in the upstream tidal flat known as West Lake (Fig. 1) that is poorly flushed during the secular low tidal period in winter. In contrast, the locations in the Rincon Bayou main channel and marsh areas close to the open boundary (e.g. Nueces1, 5, 6, 7, 9, 10 and 11) see more consistent flushing through the winter and hence the spin-up times are much shorter. As a conservative measure, all data analyses herein is based on model results after 60 days of spin-up time.

Table 5 Spin-up time for free surface elevation, salinity and 60-day residual for salinity at each TWDB station

Appendix B. The irrelevance of global drag calibration

Model calibration for a 2D shallow-water model is generally accomplished by adjusting either coefficients of a turbulence model (e.g. \(\nu\)) or drag (\(C_D\)) that control energy dissipation [34, 53]. This is typically conducted using global values: the calibration problem cannot be reasonably constrained if every model grid cell has an independent \(C_D\) and field data is available at limited locations. Our analyses (not shown) indicate the NDHM results are relatively insensitive to the choice of a global \(\nu\) or \(C_D\). The minor variability of results obtained in our calibration exercise does not allow rejection of the hypothesis that our a priori selected baseline values are acceptable. Similar conclusions have been reached by other researchers for simulations with complex bathymetries [9, 35].

Arguably, there are two principal reasons for the insensitivity of the model to calibration: (i) the numerical dissipation associated with our 1st-order upwind advection scheme [14, 15], and (ii) the “topographic” dissipation associated with the convoluted channels in a marsh. We have not seen this latter topic specifically addressed in the literature, but it follows from simple consideration of how momentum turns a corner with the hydrostatic approximation. In the real world, the pressure gradients at a channel bend serve to redirect momentum, i.e. the \(dp/dx_1\) required to slow momentum in the \(x_1\) direction increases the pressure p on the outer edge of the channel, and results in \(dp/dx_2\) that accelerates the flow in the \(x_2\) direction around the bend. Thus, streamwise momentum is not lost around a corner, but is smoothly transferred from \(x_1\) to \(x_2\) coordinate directions through the pressure gradients and nonlinear terms. Indeed, the 1D Saint-Venant equations for channel flow are essentially the mathematical embodiment of this idea [21]. However, when a narrow channel bend is represented by a single set of grid cells in a 2D hydrostatic model, only a small part of the momentum change in the \(x_1\) direction will be recovered in the hydrostatic pressure and returned to the \(x_2\) direction. The fundamental problem is that insufficient grid resolution at the channel scale creates an inability to have smooth transition of pressure gradients and nonlinear terms between coordinate directions. Thus, every bend in a narrow channel causes the flow to stop its streamwise acceleration in the \(x_1\) direction and restart the streamwise acceleration in the \(x_2\) direction. If the marsh system were strongly nonlinear with high channel velocities, then increased grid resolution would be necessary for a reasonable approximation of the fluxes. However, velocities in the marsh channels are slow and only weakly nonlinear, so relatively coarse grid resolution of the channels is an acceptable trade-off for computational efficiency. The main consequence is that the topographic dissipation from channel bends plays a major role in dissipation of energy, which makes \(C_D\), \(\nu\), and drag calibration nearly irrelevant.

It can be argued that a depth-dependent drag model (e.g. Chezy–Manning \(C_D = g{\hat{n}}^2 h^{-1/3}\), where \({\hat{n}}\) is Manning’s n) would be an appropriate baseline model. However, our calibration exercises showed that field data could not adequately discriminate between competing models. Thus, we appeal to Occam’s Razor and use the simplest possible drag model. This model is a baseline uniform \(C_D\) that is only depth-dependent in thin layers (as discussed in Sect. 2.5) where the depth dependency has a clear impact. Nevertheless, we do not consider this the final answer. Results with the automated channel bathymetry treatment (discussed below) indicate that some form of spatially-distributed drag might be useful, although it is not clear that simple depth-dependency such as Chezy–Manning is necessarily the solution. We speculate that a local drag coefficient might be linked to the approximations made in the channel connectivity algorithm and calibrated with some form of global coefficient. This idea remains an area of ongoing research.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Li, Z., Hodges, B.R. Model instability and channel connectivity for 2D coastal marsh simulations. Environ Fluid Mech 19, 1309–1338 (2019). https://doi.org/10.1007/s10652-018-9623-7

Download citation


  • Shallow coastal marsh
  • Numerical modeling
  • Bathymetric error
  • Salinity transport