90
G. Langergraber et aJ.
man hyer
OUTLET
--+
~:::: ====~~~=~~~=~=~~~~==
~
dmnage 1ayer
Fig. 2. Construction of the pilot-scale constructed wetland (figures in cm)
For the simulations, only the main layer was considered. The used 2D finite
element mesh consisted of 126 nodes and 200 elements. On top of the main layer,
an atmospheric boundary condition was applied. At the bottom of the main layer,
a constant head boundary condition with a constant head of minus 2 cm was applied. The root uptake was neglected and therefore the sink/source terms in Eqs.
(1) and (6) were zero. The capability of HYDRUS-2D to calculate an inverse
problem was used to fit the van Genuchten model parameters to measured data of
the effluent flow rate, cumulative effluent, water content, and pressure potential
during one feeding interval. The inverse parameter estimation approach in
HYDRUS-2D uses a Marquardt-Levenberg type non-linear parameter optimisation algorithm (Simunek et al. 1999). The results of the parameter estimation are
given in Table 2 and are compared to literature values.
Table 2. Literature and fitted parameter set for the van Genuchten model parameters
Parameter Residual water Saturated water Parameter Parameter Saturated hydraulic
content 6,
content 6.
a.
n
conductivity Ks
Unit
(dm3 w· dm- 3 ,)
(dm' wdm-l,)
(dm;l)
(-)
(dm,.h- I )
Literature 0.045
0.30
1.45
2.68
8.4
Fitted
0.056
0.289
1.26
1.92
8.4
Figures 3 and 4 show the measured and simulated effluent flow rate and cumulative effluent, respectively. The simulation results are presented using the literature parameter set and the parameter data set as given in Table 2. The simulations
using the fitted parameter set showed a better match to the measured data.
Figures 5 and 6 show the measured and simulated water content in a depth of
o to 10 cm and 20 to 30 cm, respectively. The measured water content is a mean
value for an interval of 10 cm. Therefore, for example in Fig. 5, the simulation results for the depths of 10, 15 and 20 cm are given. The simulation results are
shown for the fitted parameter set. For the water content also, simulation results
using the literature parameter set showed a good match to the measured data.
G. Langergraber et aJ.
man hyer
OUTLET
--+
~:::: ====~~~=~~~=~=~~~~==
~
dmnage 1ayer
Fig. 2. Construction of the pilot-scale constructed wetland (figures in cm)
For the simulations, only the main layer was considered. The used 2D finite
element mesh consisted of 126 nodes and 200 elements. On top of the main layer,
an atmospheric boundary condition was applied. At the bottom of the main layer,
a constant head boundary condition with a constant head of minus 2 cm was applied. The root uptake was neglected and therefore the sink/source terms in Eqs.
(1) and (6) were zero. The capability of HYDRUS-2D to calculate an inverse
problem was used to fit the van Genuchten model parameters to measured data of
the effluent flow rate, cumulative effluent, water content, and pressure potential
during one feeding interval. The inverse parameter estimation approach in
HYDRUS-2D uses a Marquardt-Levenberg type non-linear parameter optimisation algorithm (Simunek et al. 1999). The results of the parameter estimation are
given in Table 2 and are compared to literature values.
Table 2. Literature and fitted parameter set for the van Genuchten model parameters
Parameter Residual water Saturated water Parameter Parameter Saturated hydraulic
content 6,
content 6.
a.
n
conductivity Ks
Unit
(dm3 w· dm- 3 ,)
(dm' wdm-l,)
(dm;l)
(-)
(dm,.h- I )
Literature 0.045
0.30
1.45
2.68
8.4
Fitted
0.056
0.289
1.26
1.92
8.4
Figures 3 and 4 show the measured and simulated effluent flow rate and cumulative effluent, respectively. The simulation results are presented using the literature parameter set and the parameter data set as given in Table 2. The simulations
using the fitted parameter set showed a better match to the measured data.
Figures 5 and 6 show the measured and simulated water content in a depth of
o to 10 cm and 20 to 30 cm, respectively. The measured water content is a mean
value for an interval of 10 cm. Therefore, for example in Fig. 5, the simulation results for the depths of 10, 15 and 20 cm are given. The simulation results are
shown for the fitted parameter set. For the water content also, simulation results
using the literature parameter set showed a good match to the measured data.