Geoscience Reference
In-Depth Information
Southland region of New Zealand. With this background,
we expected that a weather event, such as a late spring frost
episode, would likely have negative consequences for yield.
To capture an element of this vulnerability, a dummy vari-
able was introduced in the model, with '1' indicating the
occurrence of one or more 'frost events' during that year
in a given county and a '0' representing the absence of one.
Given the sensitivity of crops to low temperatures during
the early phases of their development, a 'frost event' was
said to have taken place when the minimum air temperature
during one or more days in May was equal to, or fell below
−2°C (or −4°C in a second variant of the model). May was
selected as a key month as grains are commonly sown in
April in Norway.
*,†
In cases where observations from sev-
eral weather stations had been used to compile weather data
for a particular county, the records of all relevant stations
were examined for evidence of frost events. Weather sta-
tions were initially chosen due to their proximity to areas
of agricultural activity in a county; therefore, a frost event
occurring at any one of the stations would be likely to have
some relevance for at least part of the crop area under cul-
tivation in that county. In terms of our results, we found no
evidence to suggest that frost events influence crop yields.
This suggests that the model was not well suited to incorpo-
rate such a variable.
Fertiliser
application to
grain
production
The limited fertiliser use data that was available at the county
level was integrated into the model for the brief period, 1989-
1996. Sample surveys provided figures for the application of
commercial nitrogen and phosphorus fertilisers to grain and
oil seeds in the form of average kilograms per decare for most
counties.
‡
Based on the assumption that farmers used both
nitrogen and phosphorus optimally, the sum of the two was cal-
culated and included as a third independent variable in addition
to the two central climate variables—GDD during the growing
season and annual precipitation. The analyses showed that fer-
tiliser use—for the limited period data was available—did not
have any significant positive effects on yield.
*
Note that if spring arrives late, sowing can be delayed.
†
Thirty percent of wheat in Norway is sown in the autumn.
‡
Resultatkontroll jordbruk, Statistics Norway, 1993, 1995 and 1997. A com-
plete data set for the 8-year period was not available for some Northern and
Western counties, that is, Telemark, Hordaland, Sogn & Fjordane, Nord-
Trøndelag, Nordland, Troms and Finnmark.
