FAQ 31-40
31. The maximum allowed pressure at the soil surface, hCritS
The hCritS value was in the original DOS SWMS-2D program. It allowed the program to force the surface flux (precipitation) into the profile even when a positive pressure was needed at the soil surface to do that. In the one dimensional program (Hydrus-1D), this value has a very clear meaning. It represents a maximum allowed water layer at the soil surface before the surface runoff is initiated. Since in 2D this would make sense only when the program was coupled with some overland flow program (and thus HYDRUS is ready for that), we have disabled this option and specified it equal to zero. That means that surface runoff is initiated immediately after ponding is reached.
32. The minimum allowed pressure at the soil surface, hCritA
hCritA is the minimum allowed pressure head at the soil surface. This value can be activated only by evaporation. As long as the pressure head at the surface is higher then hCritA, actual evaporation rate is equal to the potential evaporation rate. Once the hCritA value is reached, the actual evaporation rate is decreased from the potential value, because the soil is too dry to deliver this rate. The hCritA value is not used for any other calculations.
The hCritA value is usually specified in the range of -150 m to -1000 m (it may need to be lower for coarse-textured soils - see the discussion below).
hCritA should be selected such so that the corresponding water content is at least 0.005 higher than the residual water content. This may be important especially for coarse-textured soils (sands), which have a very steep retention curve. For coarse-textured soils, small changes in water contents in the dry range lead to large changes in the pressure heads, which can make the numerical solution unstable. It may thus be needed to use a relatively small value of hCritA.
hCritA should also be lower (when negative) than P3 when the root water uptake is considered. When both limits for root water uptake (P3) and evaporation (hCritA) are reached, hCritA>P3 leads to inflow since it controls the flux across the boundary.
hCritA can be calculated from the relative humidity as follows:
Hr=exp[hMg/R/T]
where Hr is the relative humidity, h is the pressure head, M is the molecular weight of water [M/mol] (=0.018015 kg/mol), R is the universal gas constant [ML^2/T^2/K/M] (= 8.314 kg m^2/s^2/K/mol, J/mol/K), and T is the absolute temperature [K].
33. Potential Evapotranspiration
HYDRUS requires user to enter values of potential transpiration and potential evaporation for the atmospheric boundary conditions. HYDRUS then calculates the actual values of transpiration and evaporation based on the availability of water in the profile. The problem is that users usually know the value of the potential evapotranpiration (calculated with Penman, Montheit, FAO, etc.) and not directly its structural parts. Since HYDRUS does not have a crop growing module (does not calculate crop growth, various growth stages, leaf area index, etc) and concentrate only on movement of water and solute in the soil profile, it cannot do this subdivision (into evap and trans) itself. The subdivision is different for each crop and is relatively complex see for example FAO and several other publications suggest using “Crop Coefficients” which can be used to do this subdivision (into evap and trans). We suggest using this procedure to do that.
34. Hydrostatic conditions as initial condition
Necessary steps to specify hydrostatic conditions as initial condition:
- Select a part of the transport domain where you want to specify hydrostatic conditions
- When specifying the actual values, click the radio button “Equilibrium from the lowest located nodal point”.
35. Variable head boundary condition
Only one value of time-variant pressure head boundary conditions could be specified at any given time in Hydrus-2D. In HYDRUS (2D/3D) we presently allow up to four different time-variable pressure head boundary conditions. If, however, the time-variable head BC is specified on an inclined boundary, then the specified value is given to the lowest node and the effect of elevation is automatically added to other nodes with different z coordinate.
36. Moving of concentration and moisture fronts
The moisture front moves at a velocity of roughly q/(WCf-WCi), where q is the Darcian flux (say the imposed boundary flux at the surface), and WCi and WCf are the initial and final water contents. This means that water effectively occupies part (or all) of the initial air phase. The solute front, on the other hand, roughly moves at a velocity of q/WCf, meaning that solute will be distributed over all of the water that is in the soil (initial water and added water). In other words, solute mixes (by advection and diffusion) between the new and old water. The result is that the solute front always lags behind the moisture front, unless WCi is very small.
37. Differences between velocities in v.out and in Boundary.out
There may be some (hopefully relatively small) differences between velocities given in the v.out file and drawn in the Graphics module and those printed to the Boundary.out output file.
- The velocity vectors from the v.out file are calculated based on eq. 5.28 of the manual. I get these fluxes by calculating the arithmetic average of the gradient in a particular node (from all surrounding elements) and by multiplying this gradient by the nodal conductivity. Thus, these velocities are secondary to the numerical solution and do not affect it. Because of the numerical technique applied (linear finite element), some error is always associated with these calculations.
- Velocities printed to the Boundary.out output file are calculated directly from the Q (nodal flux) term in eq. 5.17 (Technical Manual) by dividing this value by the length of the boundary associated with a particular node. Since this term (Q) is part of the primary numerical solution, velocities calculated in this way are more precise than those based on 5.26 (Manual). Boundary fluxes calculated this way are used in HYDRUS (2D/3D) in the mass balance calculations.
Conclusions: For further analyses I would use the values (both fluxes and velocities) from the Boundary.out file since they are more precise.
38. Parameter l in the hydraulic conductivity function
This is the tortuosity factor in the hydraulic conductivity function. It was found to be 0.5 by Mualem (1976), but there is not a strong agreement on this. People at USSL lab (Marcel Schaap) claim, after analyzing huge UNSODA database, that it should be more like -1. So l was made an input parameter in the latest version of HYDRUS codes. Also, all the shape parameters (alpha and n) of the hydraulic conductivity function come directly from the retention curve. One cannot then influence independently the shape of the hydraulic conductivity function; one can only scale it by Ks. By introducing this parameter one has a) this opportunity to influence the shape of K(h), and b) there is one more degree of freedom during the optimization, which is useful for well defined experiments, such as multistep outflow and evaporation methods.
39. Solute transport BC - Pointer to the vector of boundary conditions
For constant head or flux water flow BC this pointer points to the "boundary conditions" vector in the "Reaction parameters for solute" dialog window. Thus if the pointer is specified equal to 3, cBnd3 will be used as concentration on this boundary. Maximum value of this pointer is 6. For time-variable boundary condition this pointer points to columns cValue1, cValue2, and cValue3 in the "Time variable boundary condition" dialog window. Thus the maximum value of the pointer in this case can be 3.
40. Soil hydraulic parameters
In HYDRUS there are multiple ways of entering soil hydraulic parameters:
- One can enter his own hydraulic parameters (hydraulic conductivity, saturated and residual water contents, and shape parameters) for various models (van Genuchten, Brooks and Corey, Kosugi, Durner, etc.).
- One can select soil textural class and the code will provide parameters for this textural class based on paper in WRR by Carsel and Parrish (1988). Carsel R. F., Parrish R. S.: Developing joint probability distributions of soil water retention characteristics, Water Resour. Res., 24, 755-769, 1988.
- Neural Network Predictions: Code can predict the soil hdyraulic parameters using textural characteristics using program Rosetta (Schaap et al., 1998). Schaap, M. G., F. J. Leij & M. Th. van Genuchten, 1998. Neural network analysis for hierarchical prediction of soil water retention and saturated hydraulic conductivity. Soil Sci. Soc. Am. J. 62, 847-855.
Here one can enter textural characteristic, e.g., sand, clay and silt fractions, the bulk density, etc, and parameters are predicted using neural networks trained on large soil databases.