Geoscience Reference
In-Depth Information
table 19.2
Regression results at national level
Potato
Barley
Coefficient
t-stat
Coefficient
t-stat
National—GDD
0.864
6.89
0.002
0.11
National—Precipitation
−0.316
−5.19
−0.062
−6.53
National—Time trend
0.304
0.28
2.673
15.13
Constant term
Akershus/Oslo
1450.814
8.10
308.209
10.64
Østfold
1408.290
334.091
2.24
−0.55
Hedmark
1569.159
1.45
315.359
0.59
Oppland
1634.395
2.24
298.217
−0.82
Buskerud
1330.441
−1.55
301.974
−0.54
Vestfold
1621.466
2.19
347.441
3.38
Telemark
1179.064
−3.23
295.154
−1.04
Aust-Agder
1242.762
301.376
−2.53
−0.56
Vest-Agder
1300.017
288.074
−1.86
−1.67
Rogaland
1870.702
5.24
366.499
4.88
Hordaland
1626.031
1.68
364.073
3.40
Sogn & Fjordane
1717.985
2.65
377.930
4.38
Møre & Romsdal
1702.704
2.80
273.754
−2.53
Sør-Trøndelag
1491.112
0.48
276.644
−2.51
Nord-Trøndelag
1764.810
3.88
268.485
−3.28
Nordland
1238.523
229.810
−2.24
−5.27
Troms
1203.093
N/A
N/A
−2.33
Finnmark
904.066
−4.89
N/A
N/A
Note:
Potato: 733 observations (R
2
= 0.50). Barley: 660 observations (R
2
= 0.48).
predictions are presented, which we calculated to ±20% (at
95% interval levels). In these cases, the predicted yield is higher
than in the reference situation, which is based on the model's
estimated yield for average GDD and average annual precipi-
tation in the period 1980-2000. However, in many cases the
yield increase is small, and in some cases yield is reduced. The
largest effect is found in Northern Norway, where the predicted
yield increase for potatoes is between 30% and 35%. Other
cases where the yield increase is more than 20% is potatoes in
Aust-Agder (1989-2001), potatoes in Vest-Agder (1958-1973)
and barley in Sogn & Fjordane (1958-1973). In the remaining
cases, the change is less than 20% and not considered robust.
The relative large prediction intervals reflect that the model
can only explain part of the year-to-year variation in yield per
