This paper presents the use of CLIK technique to establish the statistical
relationship between station rainfall in the Philippines under the Type 1 climate and
the larger-scale atmospheric circulation from GCM outputs for 4 target seasons
(MAM, JJA, SON, DJF). The downscaling is carried out for each station in the Type 1
climate for each model in a cross-validation manner. This downscaling technique is
based on linear regression models and makes use of long-term GCM hindcasts data
to derive the dynamic relationship between observations and model outputs. The
purpose of downscaling is to obtain high-resolution detail as accurately as possible
over the area of interest. The CLIK was designed to take only significantly
correlated areas which are taken into account in the downscaling procedure.
Analysis showed that downscaled GCM model outputs have significant skill in most
of the target seasons evaluated. The improvement gained when using the statistical
downscaling tool is significant Downscaled MME distinctly showed to have better
prediction skill than the original one for all the target seasons. The results indicate
that the large-scale circulation pattern from current GCMs muti model outputs have
the potential in predicting station-scale precipitation in the Philippines using CLIK