Agriculture Reference
In-Depth Information
Equation 2.2 uses only one cardinal tempera-
ture (
T
base
) in calculating thermal time, and as
temperature increases so does the development
rate. Incorporating a second cardinal temperature
for an upper temperature threshold into equation
2.2 recognizes that development rate does not
increase indefi nitely with temperature. Upper
threshold temperatures, if used, are generally set
to at least 30 ºC when used in equation 2.2 and
limit accumulated thermal time to the upper
threshold (Fig. 2.3b). Most wheat production
systems do not often exceed average daily tem-
peratures greater than 30 ºC, and therefore, errors
in setting this upper threshold seem relatively
minor for fi eld predictions (McMaster et al.,
2008).
If another cardinal temperature, related to the
optimum temperature (
T
opt
) for development
rate, is added along with changing the defi nition
of the upper temperature threshold to be the
maximum temperature that development rate is
greater than zero (
T
max
), then a better representa-
tion of the observed normal temperature-response
curve is improved. A two-segmented linear model
can be used with thermal time increasing linearly
from
T
base
to
T
opt
, then linearly decreasing from
T
opt
to
T
max
(Fig. 2.3c). Dividing
T
opt
into a lower
(
T
optl
) and upper (
T
optu
) temperature optimum,
between which development rate is maximum,
results in a three-segmented linear model that
closely approximates the observed temperature
response curve (Fig. 2.3d). The fi nal adjustment
is to use a curvilinear model such as those pro-
posed by Yan and Hunt (1999) and Streck et al.
(2003). If the curvilinear models are parameter-
ized correctly, they should mimic the observed
temperature-response function shown in Fig.
2.3c,d.
Largely because air temperature is more readily
measured and available, it is used in calculating
T
avg
. The assumption is that the relationship
between air temperature and shoot apex tempera-
ture (where many developmental events occur) is
closely associated. Theory would suggest using
either soil temperature at the depth of the shoot
apex (approximately 2-3 cm) when the shoot apex
is located in the soil or plant canopy temperature
when the apex is located in the plant canopy
(Peacock 1975). Experimental support for this
theory is related to a long history of root-shoot
temperature experiments (Hay and Wilson 1982;
Jamieson et al., 1995; Stone et al., 1999; Vinocur
and Ritchie 2001). However, the possible theo-
retical gain of using soil and plant canopy tem-
perature was not always realized in the fi eld for
wheat (McMaster and Wilhelm 1998; McMaster
et al., 2003). This was explained by (i) the shoot
apex and other intercalary meristems such as for
leaf growth are located over a vertical space where
temperatures vary considerably (Skinner and
Nelson 1995), (ii) the whole plant senses tempera-
ture and that infl uences signal and resource move-
ment throughout the plant, and (iii) the relationship
between air temperature and shoot apex tempera-
ture is quite stable in many environments.
All developmental processes show the curvilin-
ear response to temperature shown in Fig. 2.3c,d.
However, the duration of grain fi lling, a critical
component of determining fi nal yield, merits
special mention here. As temperature increases,
both grain fi lling rates and the accumulation of
thermal time are increased. The net effect of these
two responses is that kernel weight usually is less
under higher temperatures than cooler tempera-
tures because the duration of grain fi lling is
decreased more than the rate of growth is increased
(Marcellos and Single 1971; Sofi eld et al., 1974;
Wiegand and Cuellar 1981; Bhullar and Jenner
1983; Herzog 1986; Wardlaw et al., 1989).
Nontemperature environmental factors
Environmental factors in addition to temperature
can infl uence wheat growth and development,
but it is diffi cult to summarize the diversity of
responses, or nonresponses, to various environ-
mental factors. Cultivars can vary in their
responses to similar treatments, and of course, the
ever-present genotype × environment interaction
further complicates understanding the responses.
Many experiments, particularly those in the fi eld,
have not carefully measured the level of the envi-
ronmental factor, the levels change during the
course of the experiment, and an insuffi cient
number of treatments are applied to fully under-
stand and quantify the response surface. This has
