Fuzzy Clustering of Climate Regions of Indonesian Maritime Continent

Jurnal Matematika dan Sains Vol. 8 No. 2, June 2003, p. 57 – 61

The Houw Liong 1) ,
Bannu 2) ,
P.M. Siregar 3)

1) Department of Physics, ITB
2) Department of Physics, UNHAS.
3) Departement of Geophysics & Meteorology, ITB

Abstract
In general drought in Indonesia can be predicted from intensities of El Niño that can be defined by using time series of sea surface anomaly on Pacific Ocean (SSTA 3.4). It can be shown that when El Niño with strong intensities occur then more than 65% regions in Indonesia the precipitations are below normal (drought in Indonesia). The correlation between strong El Niño intensities and percentages of regions in Indonesia with precipitations below normal are high, but when the intensities are weak the correlations are low. In this case other phenomena such as on Indian Ocean Dipole Mode (IOD) can contribute to drought in Indonesia. Clustering of climatic regions in Indonesia based on monthly rainfall pattern using fuzzy set, fuzzy relations or Kohonen’s neural network will help to clarify drought on these regions. It can be shown that climatic regions in Indonesia can be clustered based on monthly rainfall patterns that are strongly influence by Australian monsoon which is known as North Australia-Indonesian Monsoon (NAIM) and Maritime Continent (MC) which has equatorial precipitation characteristic. The climatic clustering is based on the ground that ENSO and IOD are regional atmospheric dynamic so the clustering should be based on average monthly pattern or geopotential height. The east MC and NAIM will be influence strongly by ENSO and the western MC especially south Sumatra and west Java is influence also by IOD.

Keywords : fuzzy clustering, fuzzy set, fuzzy relations, ENSO, IOD, drought, monsoon, maritime continent