Review by Jon Arah,
Soil and Water Sciences Division, International Rice Research Institute, P.O. Box 933,
Manila 1099, PHILIPPINES
The excellent online help asserts that WaterMod is intended for those interested in studying
the dynamics of soil water in agricultural systems. The price suggests that educational
institutes, rather than interested individuals, are the main target group. Suggested uses
The model incorporates rainfall and irrigation inputs, runoff, transpiration, evaporation from
the soil and from the canopy, infiltration and drainage. It presents its output in
readily-assimilable form (graphs which may be printed and copied as metafiles or bitmaps).
Typically, a user specifies soil, crop, climatic and irrigation parameters using a
purpose-built front-end utility called the WaterMod Builder; climatic inputs can also be
supplied via a met data file. When the model is run, daily moisture content depth profiles
are displayed beside time-courses of two out of the following system properties: precipitation,
irrigation, soil evaporation, canopy evaporation, evapotranspiration, drainage, runoff.
The model runs fast enough (between 5 and 10 simulated days per second on a 66MHz 486 with
8 MB RAM) to encourage enquiry along the lines of the questions it sets itself to address.
As such, it serves its purpose well.
Its treatment of many processes is necessarily simple. Uptake of water by roots occurs
according to the "distributed model" in which available water in each soil layer is withdrawn
in proportion to the root density in that layer multiplied by the total evapotranspirative
demand, the "preferential model" in which surface layers
are exploited first, or the
" compensating model " in which the previous two models
are combined (distributed then
preferential). Runoff is simply the excess of precipitation and irrigation over infiltration
plus a user-defined surface storage. More significantly, water movement throughout the soil
profile is treated according to a (conceptually-incoherent) version of the tipping-bucket
model: specifically, the amount of water that moves down from one layer to the next is equal
to the mass of water in excess of field capacity in the upper layer multiplied by the ratio
of the saturated hydraulic conductivity of the upper layer to the thickness of that layer.
This last is a gross simplification. It may or it may not lead to significant errors in
simulation, depending on the relative importance of other factors (where evapotranspiration
dominates, soil hydrology may not matter much). Nevertheless, I would like to see a version
of the program in which soil-water-movement paradigms other than the tipping-bucket
(however modified) are on offer. I have in mind Richards' equation or Darcy's law, with any
of the various current approaches to and
k-. These more rigorous treatments may be more
computationally-demanding but they are also more accurate: they should be on the menu.
A great deal of work has been directed, successfully, toward making WaterMod 1.6 user-friendly.
I would like to see the next version offer a few more realistic treatments of water movement