Hello Tesfalidet,

here is an answer with some theoretical background,
which I hope will help you further in finding solutions
to your problem:

A pragmatic way to characterize soil temperature T
at depth z and time t is the following :

T(z,t)= Tsavg + A(0)*exp^(-z/D) sin (omega*t - z/D)

where Tsavg is the average temperature of the soil
surface, A(0) the amplitude of the surface fluctuation,
D the damping depth and omega the angular frequency of
the oscillation :

omega = 2pi/tau

where tau is the period of oscillation.

To find D, one needs to know the thermal diffusivity
of the soil k:

D = sqrt(2*k/omega)

k is the ratio of thermal conductivity
to volumetric specific heat. It is
affected by soil composition (mineral,
organic, air), bulk density and soil
wetness.

Volumetric specific heat is a linear
function of the volumetric water fraction.
Differences (i.e. the intercept of the
linear function) between different soil fractions
are due to the corresponding differences in bulk
densities.

Soil thermal conductivity has to be determined
from the different conductivities of its
constitutents (water, gas, mineral and organic
fractions). You will find the methodology in
a paper of de Vries (1963) :

de Vries, D.A. (1963) Thermal properties of
the soil. In Physic of plant environment.
van Wick, W. R. (ed.) North Holland Pub. Co.,
Amsterdam, pp. 210 - 235

which is cited in standard textbooks like
Hillel, D. (1998) Environmental soil physics.
Academic Press or Campbell G.S. and Norman J.M.
(1998) Environmental Biophysics. Springer.

Now comming to your question, you will
find a detailed description of how to establish
a relation between soil texture and thermal conductivity
in the mentioned book of  Campbell and Norman.
They use shape factors to account for the different
effects of soil texture (see pp. 119 - 126).

You could also use information from charts
(vol. water fraction vers. thermal diffusivity)
to charterize soil temperature distribution,
which you will also find in the mentioned books
(Hillel p. 319, Campbell and Norman p. 124)
There is another version in a previous edition
of Campbell's book (Campbell, 1977 p. 18).

I am not so familiar with the approach being
implemented in the CERES/CROPGRO models.
Perhaps, someone of the developers could
respond to this part of your question.

Good luck.

Matthias



--
Matthias Langensiepen, PhD
Hannover , Germany

Internet : www.langensiepen.net
Mailbox : [log in to unmask]