Geoscience Reference
In-Depth Information
whereas the coefficients for the cereals range between −0.34
and 0.63 (see Table 19.1 and Annex 19.1 for details). There are
two possible explanations for the interesting finding that coef-
ficients have, in some instances, been negative. The first is that
precipitation may become so abundant that it leads to excess
soil moisture. The second could be a result of the positive cor-
relation between increased precipitation and cloud cover. Thus,
increased precipitation means reduced radiation from the sun,
leading to reduced photosynthesis, and thereby reduced yield.
Both explanations go some way towards explaining the nega-
tive correlations between precipitation and yield evident in
Western, Mid, and parts of Northern Norway.
The time trend is positive in most significant cases (overall
37% of instances), with the exception of potatoes in Northern
Norway (and Sør-Trøndelag), where it is negative (which is equiv-
alent to 4% of the cases). The positive trend can be attributed to
long-term productivity gains in agriculture, which can include
structural changes (fewer and larger farms), better crop varieties,
improved farming techniques and equipment, and more efficient
fertiliser use. On the other hand, the negative time trend may
reflect structural changes in agriculture that affect productivity
negatively; these could be related to government policies.
The national level analyses only provided significant results
for potatoes and barley in the model variant allowing for differ-
ent constant terms (but with the same marginal effect of GDD
and precipitation, see Table 19.2). For potatoes the sign of coeffi-
cients is the same as in county-level analyses, though the size of
coefficients is smaller. Instead, the model provides for different
constant yields across counties (i.e. the yield component that is
not influenced by GDD, precipitation or time), where the highest
significant yield is found in Rogaland (1871 kg), and the lowest
in Finnmark (904 kg). For barley, the GDD effect is not signifi-
cant. Instead the significant constant terms vary between 378 kg
in Sogn & Fjordane, and 229 kg in Nordland.
*
The precipitation
coefficient is close to zero, but negative and significant.
predictions
Using the model to give predictions for the RegClim climate
change scenario in 2040, we find that the positive contribution
from increased GDD in most of the significant cases (shown in
Table 19.1) dominates the negative contribution from increased
precipitation. The predictions for potatoes are shown in Table
19.3 (details for all crops are found in Annex 19.2). Only robust
*
There is no barley yield in Finnmark and there are too few observations in
Troms to include in the analysis.
