Agriculture Reference
In-Depth Information
or, equivalently:
V
i
=
fT
(,,)( , )
θ
cg y
µ
(7)
i
i
i
i
where the macron (over-line) denotes integration over the appropriate
probability distribution (area averaging). Equating the two macroscale
Equations
1
and
6-7
reveals
the
following
assumptions
implicit
in
Equation 1:
F
i
(
T
,
Θ
,
C
)
≡
F
i
(
T, c
, θ
)=
fT
i
(,,)
θ
c
≡
∫
f
i
(
T
,
θ
,
c
)
p
(
T
,
θ
,
c
)
d
(
T
,
θ
,
c
) (A1)
and:
K
i
Y
i
≡
K
i
y
i
=
gy
i
(, µ ≡
∫
g
i
(
y
i
,
µ
i
)
p
(
y
i
,
µ
i
)
d
(
y
i
,
µ
i
)
(A2)
i
i
Assumption A1 can be true only where microsite temperature, moisture
content and clay fraction are homogeneous, in which case
F
i
is by definition
equal to
f
i
. Under any other circumstances (e.g. in particular, where
physical aggregation leads to higher clay fractions and moisture contents in
some microsites than in others), there is no simple relationship between
F
i
and
f
i
. Instead, the macroscale function depends on the probability
distributions of the microsite properties
T
,
and
c
. Macroscale functions
F
i
appropriate for one particular distribution of microsite properties may
therefore be inappropriate for another. Macroscale partition functions
θ
Φ
ij
are similarly vulnerable.
The assumption that there is one set of rate modifier functions
F
i
(and partition functions
Φ
ij
) appropriate for all soils under all climatic
and cultivation regimes amounts to an assumption that the probability
distributions of the microsite properties
T
,
θ
and
c
are always and
everywhere roughly the same.
Similar considerations apply with regard to the microscale distribution
of substrate concentration
y
i
in assumption A2. An additional difficulty
concerns the elimination of microbial biomass
µ
i
from the macroscale
formula. Two possible scenarios (pictures of microbial life in soil) can be
invoked to justify this first-order approximation. The first - probably the
most general where the assumption is questioned at all - is of a universal
and diffuse war of all against all; all possible microbial life strategies
are everywhere present, the most efficient inevitably prevails, successful
microbial populations expand until their substrate becomes limiting. The
longer the model time step and the larger the simulated area, the more
credible such a picture becomes.
What if we want to model short-term SOM dynamics on a time scale
of days or weeks (relevant to the crop, but also within the range of
microbial population fluctuations)? Can we then make the assumption
that potential microbial strategies are universally distributed, and that -
basically - whatever makes thermodynamic sense will occur? The evidence
seems to be against it (J.M. Tiedje, Los Baños, Philippines, 1999, personal
Search WWH ::
Custom Search