Geoscience Reference
In-Depth Information
Furthermore, the method is founded on correlation analysis and
not necessarily on causal mechanisms. There may be depen-
dency between explaining variables (multicollinearity), and
relationships between yield, precipitation and temperature may
be non-linear. Moreover, the simple model we have chosen is
not able to account for effects caused by variability in weather
and extreme weather events on yields [7]. Since we are studying
a smaller change in climate (as defined by the RegClim sce-
nario), a linear model is probably an acceptable approximation
even if the relationships are non-linear. In addition, data avail-
ability has put strong restrictions on which variables could be
included in the analysis. One example of an important weather
variable for plant growth that could not be included was sun
radiation, which could be represented through a measure of
cloud cover. Through the chosen approach we were able to link
changes in climate variables at local level (weather stations) to
secondary climate change impacts in terms of changes in agri-
cultural productivity for some crops at county level in Norway.
Some major benefits of the approach are simplicity, limited
data requirements and the ability to get some control over the
significance of various explaining factors. The study is in line
with the call of Zilberman et al. [25] to analyse the impact of
climate change on agriculture within a disaggregated model-
ling framework and a focus on empirical research. The results
should indicate if county level is a suitable aggregation level to
disclose significant effects, or if this is an aggregation level that
only produces moderate effects since more distinct local effects
are averaged out [25].
An overview and assessment of climate change impacts in
Europe, including agriculture, can be found in Parry (2000) [14].
NILF [11] provides a comprehensive survey of climate change
impacts for the agricultural sector in Norway. Based on average
yields in various climate zones, the climate change impact on
agricultural productivity is analysed through a shift in climate
zones leading to increased yields for most crops.
An early application of a statistical model is Warrick [24],
who simulated wheat yields on the US Great Plains, assum-
ing technology as in 1975 and climate conditions as under the
1936 drought. Leemans and Soloman [8] studied the potential
yield changes for spring and winter wheat and other major
crops at a global scale under a warmed climate. Using a crop-
prediction model with geographic information systems (GIS),
they reported that high-latitude regions will be the beneficia-
ries of climate change, enjoying extended growing seasons and
increased productivity. Rötter and Van de Geijn [20] provided
