This document discusses climate modeling and weather modification in Indonesia. It outlines computing intensive climate models and introduces soft computing approaches like adaptive neuro-fuzzy inference systems (ANFIS) and fuzzy clustering for climate forecasting. It analyzes the relationship between sunspot numbers and rainfall in several Indonesian regions and finds correlations. The document concludes that numerical climate models need modifications for Indonesian regions and that solar activity is the main factor determining Indonesia's climate.
6. GCM & DARLAM
• CO2 double in 100 years :
• Temperature increase : 0.5 --- 1.5 in 50 years
• Verification based on weather mainly on station
at cities : urban warming instead of global
warming.
• Arakawa cloud formation scheme produced poor
rainfall forecast : 0.46 for rainfall in Bandung
• Sensitivity on initial and boundary conditions
• Needs better data and better model
• Forcing ??
7. Soft Computing : Adaptive Neuro
Fuzzy Inference System
A1
A2
B2
B1 N
N
∏
∏
layer1
layer2
layer3
layer4
layer5
x y
1w
2w
1w
2w
x y
8. ANFIS
•
Layer 1 :
•
• x and y are input of ode -i and O1,i is
membership function of fuzzy set A=(A1,A2)
and B=(B1 ,B2 ) with membership function
A is :
•
•
• ai,bi, and ci are parameters
• Layer 2 : output as the product of input
membership functions :
•
2
1,
1,
( ), 1, 2,
( ), 3,4,
i
i
i A
i B
O x for i or
O y for i
µ
µ −
= =
= =
b2
i
i
A
a
cx
1
1
)x(
−
+
=µ
2,1i)y()x(wO ii BA1i,2 =µµ==
9. • Layer 3 in node -i :
•
• Layer 4 : Node -i is
adaptive node with function
node :
2,1i,
ww
w
wO
21
i
ii,3 =
+
==
)ryqxp(wfwO iiiiiii,4 ++==
10. ANFIS
• Layer 5 : final output :
•
5
i i
i
i i
i i
i
w f
O w f
w
= =
∑
∑
∑
11. • Fuzzy c-means Algorithm
• Fix c (2≤c≤ n) and select a value for parameter m’,
initialize the partition matrix U(0)
, membership
functions and the centers . Each step in this
algorithm will labeled r, where r=0,1,2,..
• Repeat updating the partition matrix for rth step,U (r)
until
•
ε≤−+ )()1( rr
UU
12. • Calculate the new membership functions
1
)1'/(2
1 )(
)(
)1(
−
−
∑
=
=+
m
c
j d
r
jk
d
r
ikr
ik
µ
13. • set r=r+1
• Calculate the new c centers :
∑
=
∑
== n
k
m
ik
n
k kj
xm
ik
ijv
1
'
1
.'
µ
µ
25. Concluding Remarks
• Numerical Climate Model for Indonesian
regions needs modifications
• Long term climate prediction has poor
accuracy due to chaos & forcing
• ANFIS and Fuzzy Clustering can be used
in forecasting of climate in Indonesia
• Solar Activity is the main factor that
determined climate in Indonesian regions