=====================
I had sent this message to the DSSAT listserver last week, but it bounced
back, because the listserver is configured not to accept attachments (virus
protection). It did not work either when I incorporated the figures in the
email body.  So those who want the PowerPoint file can send me an email.
=====================

Hi Derek,

There are a great number of methods that estimate critical
soil-water-retention parameters (saturation, field capacity, wilting point)
from soil characteristics that are easily measured, such as texture,
soil-organic-matter content and bulk density; not all methods use all of
these input data. Recently I have been comparing eight of such methods for
their use in crop models. At the crop modeling symposium in Florence last
July, we presented a poster on this (attached as PowerPoint file) and the
article is pending. The results were not very promising, as estimates varied
enormously (see Figure 1+2 of our poster), and the DSSAT model responded in
a very strong way also (Figure 3).

A problem with using these estimated data as input for a crop model is that
almost all methods base their estimate on lab-measured data, which may be
very different from what the plant experiences in the field (Figure 4).
Field capacity can be defined in large as a parameter that is defined by
soil characteristics and which is (almost) the same for most crops, but
wilting point can vary greatly by crop and is not so much a soil-specific
parameter but rather a crop-specific one. Using a soil-dependent estimate as
input to a crop model thus makes not much sense.
In a shrinking-swelling clay, lab-measured data may also not be
representative at all.

So what is the alternative? Measuring your water-retention characteristics
under field conditions is a lot of work, should be done for each crop
species anew (at least ideally) and is quite time-consuming (waiting until
your test crop reaches wilting point). Few researchers have done so, but see
Ratliff et al. (1983; Soil Sci. Soc. Am. J. 47:770-775) and Ritchie et al.
(1987; ARS Technical Bulletin), who coordinated a set of such measurements
across the USA. It is surely recommendable to measure your water-retention
characteristics under field conditions.

The Australian APSRU group (from the APSIM model) realized this also and
suggests in their soil manual to measure water retention in the field. It is
worth giving this a look:
For field-measured water limits:
http://www.farmscape.tag.csiro.au/SoilMatters/Mod4/base.htm
<http://www.farmscape.tag.csiro.au/SoilMatters/Mod4/base.htm>
For shrink-swell soils:
http://www.farmscape.tag.csiro.au/SoilMatters/Mod4/4_02.htm
<http://www.farmscape.tag.csiro.au/SoilMatters/Mod4/4_02.htm>

When using field-measured data is not an option and one has to go with
estimated data, be sure not to use the soil classification as a short-cut to
obtain the estimates. Some authors give an average value for each soil class
(e.g. clay, silty clay loam, or loamy sand, etc.), but since a soil class
incorporates a range of textures, both the field capacity and wilting point
may vary by up to 0.15 cm3[H2O]/cm3[soil]; see Figure 1+2 of our poster. For
crop model purposes such uncertainty is deadly. You will need the real
texture, SOM% and bulk density!

A point that is not very promising is that none of these methods applies to
any texture, so if you have a very sandy or very clayey soil it may be hard
to find a method. Moreover, several of the methods have errors, which I
describe in our manuscript. We found it difficult to make convincing
recommendations on which method to use, but Rawls et al. (1982; Transactions
ASAE 25:1316-1328) and Saxton et al. (1986; Soil Sci. Soc. Am. J.
50:1031-1036) seem the best for not-too-extreme textures.

Just one additional point on soil-water retention for the tropics:
I was discussing with our soil physicist, Edgar Amézquita, the question in
how far a wilting point of pF=4.2 is realistic. He mentioned that under
tropical conditions, the temperature inversion in the soil profile (Day:
topsoil = warm, deep soil = cold. Night: topsoil = cold, deep soil = warm)
may lead to an upward water transport in the profile at night and that
plants often survive in this way under conditions that on the basis of
potted-plant experiments in a greenhouse are seen as mortal. He did his PhD
thesis on this at IITA in Nigeria. I don't know whether such processes are
included in the DSSAT soil water module, but he surely sees them as
important for the tropics.

A few articles in which some methods have been compared (all with
lab-measured data!):
# Kern,J.S. 1995. Evaluation of soil water retention models based on basic
soil physical properties. Soil Sci. Soc. Am. J. 59:1134-1141
# Timlin, D.J., Pachepsky, Ya.A., Acock, B. and Whisler, F. 1996. Indirect
estimation of soil hydraulic properties to predict soybean yield using
GLYCIM. Agricultural Systems 52:331-353.
# Williams,R.D., Ahuja,L.R. and Naney,J.W. 1992. Comparison of methods to
estimate soil water characteristics from soil texture, bulk density and
limited data. Soil Science 153:172-184.


A response to Ken Boote's email:
Though I have never seen the code for the water utility of the DSSAT shell,
I am pretty sure that it is Ritchie et al. (1987). I disagree that this is
much the same as Ritchie 1999, which generalizes much more by giving every
soil, besides the very sandy ones, a water-holding capacity of 0.132
cm3/cm3, only varying with organic-carbon concentration

Not unimportant is that the articles related to Ritchie1987 have severe
errors in the equations, making that soils that differ significantly from
those that were used for deriving the method, get wrong estimates for the
soil-water parameters. This not necessarily means that the code as used for
DSSAT has those errors also, but it is likely. I brought this up with Joe
Ritchie a few months ago, and he acknowledged that his article equations
were wrong (my compliments, Joe!).


Arjan Gijsman
Univ. Florida, Ag. & Bio. Eng. /
  Centro Internacional de Agricultura Tropical (CIAT), Colombia