Agriculture Reference
In-Depth Information
where coefficients
a
u
,
b
u
,
a
w
and
b
w
are functions of velocity and position, and
d
ξ
u
and
d
ξ
w
are random numbers selected from a Gaussian distribution with mean 0 and
variance
dt.
The variable
v
s,
the fall speed of the spore, is included to account for
gravitational settling. The first term in the expressions for
du
and
dw
contains
information on the velocity 'memory' of the air parcel and the second term
represents fluctuations caused by turbulence. The formulation of the coefficients
depends on the nature of the flow (Legg 1983; Wilson and Sawford, 1996; Wilson
and Flesch, 1997).
LS models are particularly useful for estimating the dispersal of spores close to
sources within plant canopies or close to the ground (Aylor, 1989, 1999).
Consequently they have been used to estimate the escape of
Venturia inaequalis
(cause of apple scab) ascospores from ground cover (Aylor and Flesch, 2001) and
Phytophthora infestans
(cause of potato late blight) sporangia from a potato crop
(Aylor
et al.,
2001). More recently LS models have been used to investigate the
dispersal of pollen from maize crops (Aylor
et al.
, 2003; Jarosz
et al.
, 2004) to
assess the risk of gene flow from transgenic to conventional crops. Because LS
models simulate the flight of 'individual' spores, they have the potential to account
for the effects of wind gusts on spore dispersal. LS model predictions suggest that,
within crops, deposition by impaction could be enhanced by gust release (Legg,
1983) and above crops the concentration and deposition curves will be displaced
down-wind by an amount proportional to the speed of the gusts (McCartney, 1990b).
The success of LS models depends on the accurate parameterisation of air flow and
turbulence, which can be difficult in complex air flows such as at crop boundaries.
Jarosz
et al.
(2004) found that an LS model tended to underestimate maize pollen
deposition within 10 m of the boundary between the crop and bare soil. They
attributed this to an incorrect parameterisation of turbulence in the transition
between the crop and the surroundings. Since LS models used for spores and pollen
ignore effects of particle inertia, they are applicable only to particles less than c. 300
µm in diameter (Walklate, 1987; Wilson, 2000). However, they have been adapted
to describe the application of agricultural sprays in orchards (Walklate, 1992; Xu
et al.,
1998). There has recently been interest in use of LS models for pollutant
dispersal in the convective boundary layer (Raza
et al.,
2001; Oettl
et al.,
2001;
Franzes, 2003) and as understanding of atmospheric flow improves the accuracy and
applicability of LS based models should increase.
Atmospheric dispersal models are becoming more sophisticated as understanding
of mechanisms of atmospheric flow increases. Atmospheric dispersal models are
being developed that can account for not only dispersal processes but also changes
in topography and surface characteristics (e.g. Aloyan, 2004; Wang and Ostoja-
Starzewski, 2004). Such models, although complex, have the potential to enhance
understanding of spore and pollen dispersal within realistic landscapes. This
information is needed to understand gene flow in fungal pathogen communities, for
example, the movement of fungicide resistance or virulence genes. Computational
fluid dynamic systems, developed to calculate air flows in complex terrains such as
around buildings, are now being coupled with dispersal models to investigate
pollutant dispersal in urban or industrial landscapes (Riddle
et al.
, 2004). Such
