Geoscience Reference
In-Depth Information
Over all, a comparison of observed and RegCM3 simulated precipita-
tion shows that the model captured well the spatial distribution of seasonal
precipitation and belts of maximum precipitation. Validation against KMA
data sets show that the precipitation simulated with the EMU convection
scheme is more realistic compared with other schemes. The RegCM3 sim-
ulation locates the maximum precipitation belt over south and east coast
and minimum over northwest coast of South Korea, is good agreement with
observation (KMA).
4. Wavelet Analysis
Wavelet analysis or wavelet transform (WT) is a powerful mathemati-
cal tool well suited for the study of multi-scale non-stationary processes
occurring over finite spatial and temporal domains. A detail computa-
tional procedure for WT is well described in the website http://ion.
researcgsystem.com/IONScript.
The daily rainfall fluctuation and wavelet spectrum are computed from
daily precipitation during 2002. Wavelet spectrum for KMA shows the max-
imum variance after first week of August. It also illustrates that the period
of oscillation increases after first week of August, with maximum variance
centered around days 65 in 60-80 days time band in KMR (not shown).
While model simulated maximum variance is also centered around 65 days in
60-80 days. These maximum values are well reflected in daily precipitation
series between 60 and 80 days band for KMA and model (not shown). This
analysis shows that EMU convective scheme in RegCM3 are more close to
the KMA compared to AS convective schemes.
For examining the intensity of 10-20 and 30-60 days oscillations,
Butterworth band pass filter is applied
9
to the precipitation series. The
variances in 10-20 and 30-60 days band for KMA are compared with
model simulated variance. The model simulation shows that the variance in
10-20 days is approximately twice than those in 30-60 days, agrees well with
KMA. The variance obtained to 10-20 and 30-60 days oscillations are 8.7
and 3.0 for KMA, 9 and 4 for model with EMU convective scheme and 13
and 6 for AS scheme in RegCM3, respectively.
Similar patterns of variance for 10-20 and 30-60 days are also obtained
3
on the basis of 23 years (1978-2000) observed data for 20 uniformly dis-
tributed stations from KMA. A comparison of model simulated variance in
10-20 and 30-60 days with observed (KMA) shows that the variance simu-
lated with EMU convective schemes in RegCM3 is more close than others
schemes.
