Page 70 - Special Topic Session (STS) - Volume 2

STS459 Norshahida S. et al.
To evaluate the performance of a fitted distribution model in comparison
to the empirical distribution, the visualization approach is used. The CDF plots
for the theoretical distribution curve and the empirical distribution curve were
plot together in the same graph.
The performance of the best distribution model obtained was validated
using the coefficient of determination R . High value of R (> 0.8) and a linear
2
2
pattern of the plotted points between the predicted and the observe values
indicate that the fitted model is a good model to represent the random
distribution of Ozone hourly concentrations and the prediction can be done.
Once the best-fit distribution is determined, the cumulative distribution
function (CDF) of the fitted distribution was used to predict exceedances and
the return period. Exceedances is defined as the condition when ozone
concentration level exceeds or greater than the given government standard
(H = 0.1 ppm). In environmental problems, return period is considered as the
average number of days between exceedances. Thus to compute the return
period (T) of exceedances of Ozone concentration (x) using CDF, the formula
is given by:

= 1 (1)
1−( )
where F(xexceedance)is the CDF of the ozone concentration and xexceedence is when
x ≥H .
1 − () = ( > ) (2)
According to Malaysia Air Ambient Quality Guideline (MAAQG), the
standard for exceedances is whenever the hourly ozone concentration exceeds
0.1 ppm.

3. Results
The hourly ozone concentration for Southwest monsoon is reported in
Table 1 shows the descriptive statistics for O3 concentration for Shah Alam and
Putrajaya areas. The statistics minimum O3 for the areas are about the same
with level of 0.000 ppm while the maximum O3 for Shah Alam and Putrajaya
are 0.1410 ppm and 0.1150 ppm respectively.
The results indicate that that Shah Alam has higher O3 concentration
compared to Putrajaya. Besides, the mean values were higher than median
values for the three areas. The result indicates that there exist outlying data.
The statistic standard deviation for Putrajaya is larger than Shah Alam. The
results indicate that Putrajaya has higher variability in the Ozone concentration
level. For all areas, the coefficient of skewness and kurtosis are greater than
zero. The result shows that the Ozone distribution at the locations are
positively skewed, if the model were to be fitted to the Ozone data, a skewed
model is more appropriate. Parameter estimates for four distributions
according to Southwest Monsoon season data are shows in Table 2. These

59 | I S I W S C 2 0 1 9