Geoscience Reference
In-Depth Information
geopotential heights and specific humidities at all levels. The coecients of
determination (
R
2
) after calibration of the SDSM for
t
max
and
t
min
were
very high, from 0.714 to 0.785, respectively, while the value for precipita-
tion was very low, ranging from 0.062 to 0.098. This would indicate the
diculty in finding significant climate variables from the NCEP data that
could explain well the variability of daily precipitation. Similar diculty
was found for the LARS-WG in the simulation of daily precipitation char-
acteristics even though the bias values of different indices for LARS-WG
are generally smaller than those given by the SDSM. Tables 3 and 4 show
the total bias scores produced by the SDSM and LARS-WG models for
daily precipitation and for daily temperature extremes, respectively. It was
found that the LARS-WG model was able to produce daily precipitation
statistics in closer agreement with those of the observed data, while SDSM
Table 3. Total bias scores for daily precipitation
simulation for 1961-1975 calibration period.
Evaluation indices
SDSM
LARS-WG
SDII
13
3
Prec90p
10
6
Prcp1
15
1
CDD
10
1
R3days
12
3
Prec-moy
33
11
Prec-std
38
9
Total
131
34
Table 4. Total bias scores for extreme daily temper-
ature simulation for 1961-1975 calibration period.
Evaluation indices
SDSM
LARS-WG
Tmax-moy
2
46
Tmax-std
10
38
Tmax90p
11
5
Tmin10p
5
10
Tmin-moy
2
45
Tmin-std
29
19
Tmean-moy
1
47
Tmean-std
21
27
DTR
7
3
Fr-Th
11
11
FSLs
3
1
GSL
2
2
Total
104
254
