Malaysian Journal of Analytical Sciences Vol 20 No 3 (2016): 551 - 559

DOI: http://dx.doi.org/10.17576/mjas-2016-2003-13

 

 

 

FITTING STATISTICAL DISTRIBUTIONS FUNCTIONS ON OZONE CONCENTRATION DATA AT COASTAL AREAS

 

(Penyesuaian Fungsi Taburan Statistik pada Data Kepekatan Ozon di Kawasan Pesisiran Pantai)

 

Muhammad Yazid Nasir*, Nurul Adyani Ghazali, Muhammad Izwan Zariq Mokhtar, Norhazlina Suhaimi

 

School of Ocean Engineering,

Universiti Malaysia Terengganu, 21030 Kuala Terengganu,Terengganu, Malaysia

 

*Corresponding author: muhammad.yazidnasir@gmail.com

 

 

Received: 24 February 2015; Accepted: 27 October 2015

 

 

Abstract

Ozone is known as one of the pollutant that contributes to the air pollution problem. Therefore, it is important to carry out the study on ozone. The objective of this study is to find the best statistical distribution for ozone concentration. There are three distributions namely Inverse Gaussian, Weibull and Lognormal were chosen to fit one year hourly average ozone concentration data in 2010 at Port Dickson and Port Klang. Maximum likelihood estimation (MLE) method was used to estimate the parameters to develop the probability density function (PDF) graph and cumulative density function (CDF) graph. Three performance indicators (PI) that are normalized absolute error (NAE), prediction accuracy (PA), and coefficient of determination (R²) were used to determine the goodness-of-fit criteria of the distribution. Result shows that Weibull distribution is the best distribution with the smallest error measure value (NAE) at Port Klang and Port Dickson is 0.08 and 0.31, respectively. The best score for highest adequacy measure (PA: 0.99) with the value of R² is 0.98 (Port Klang) and 0.99 (Port Dickson). These results provide useful information to local authorities for prediction purpose.

 

Keywords: ozone concentration, coastal area, statistical distributions, goodness-of-fit, performance indicator

 

Abstrak

Ozon merupakan salah satu pencemar yang banyak menyumbang kepada masalah pencemaran udara. Maka kajian tentang ozon adalah penting untuk dijalankan. Objektif kajian ini adalah mencari taburan statistik yang terbaik untuk mewakili data kepekatan ozon. Tiga fungsi taburan yang digunakan dalam kajian ini adalah Gaussian songsang, Weibull dan Lognormal telah dipilih bagi menentukan taburan statistik yang terbaik untuk mewakili data taburan ozon per jam pada tahun 2010 di Port Dickson dan Port Klang. Kaedah penganggar kebolehjadian maksimum (MLE) telah digunakan untuk mengira parameter yang membentuk graf fungsi taburan kebarangkalian (PDF) dan fungsi taburan kumulatif (CDF). Tiga penunjuk prestasi (PI) iaitu ralat mutlak dinormalkan (NAE), kejituan ramalan (PA) dan pekali penentuan (R²) telah digunakan untuk menguji prestasi kriteria taburan yang terbaik. Hasil kajian menunjukkan taburan Weibull adalah yang terbaik untuk mewakili data kepekatan ozon dengan nilai ukuran ralat terkecil (NAE) di Port Klang dan Port Dickson masing-masing ialah 0.08 dan 0.031. Skor terbaik juga terhasil untuk pengiraan kejituan tertinggi (PA: 0.99) dengan nilai R² di kedua-dua tempat ialah 0.98 (Port Klang) dan 0.99 (Port Dickson). Hasil kajian ini membekalkan maklumat penting kepada penguatkuasa tempatan untuk tujuan ramalan kepekatan pada masa akan datang.

 

Kata kunci: Kepekatan ozon, pesisir pantai, taburan statistik, penyesuaian terbaik, penunjuk prestasi

 

References

1.       Ghazali N. A., Ramli N. A., Yahaya A. S., Md Yusof N. F. F., Sansuddin N. and Al Madhoun W. A. (2010). Transformation of nitrogen dioxide into ozone and prediction of ozone concentrations using multiple linear regression techniques. Environmental Monitoring and Assessment, 165: 475 – 489.

2.       Munir S, Chen H, and Ropkins K (2012). Modelling the impact of road traffic on ground level ozone concentration using a quantile regression approach. Atmospheric Environment, 60: 283 – 291.

3.       Yang, W.  and  Omaye, S., (2009).  Air  pollutants, oxidative  stress  and  human  health.   Research/Genetic toxicology and Environmental Mutagenesis, 674: 45-54.

4.       Shan W., Zhang J., Huang Z. and You L. (2010). Characterizations of ozone and related compounds under the influence of maritime and continental winds at a coastal site in the Yangtze Delta, nearby  Shanghai. Atmospheric Research, 97: 26 – 34.

5.       Donev E., Zeller K. and Avramov A., (2002). Preliminary background ozone concentration in the mountain and coastal areas of Bulgaria. Environment Pollution, 117: 281 – 286.

6.       Nair P. R.,  Chand D., Lal S., Modh K. S., Naja M., Parameswaran K., Ravindran S. and Venkataramani S. (2002). Temporal variations in surface ozone at Thumba (8.6^oN, 77^oE)- a tropical coastal site in India. Atmospheric Environment, 36:603 – 610.

7.       Song S. K., Shon Z. H., Kim Y. K., Kang Y. H., Oh I. B. and Jung C.H. (2010). Influence of ship emissions on ozone concentrations around coastal areas during summer season. Atmospheric Environment, 44: 713 – 723.

8.       Lu, H. C., (2000). The statistical characters of PM₁₀ concentration in Taiwan area. Atmospheric Environment, 36: 491 – 502.

9.       Lu, H. C., (2003). Estimating the Emission Source Reduction of PM₁₀ in Central Taiwan. Chemosphere, 54: 805-814.

10.    Forbes C., Evan M., Hastings N. and Peacock B., (2011). Statistical Distributions. 4^th Edition. John Wiley & sons. New Jersey.

11.    Zerda I. (2012). An experimental comparison of popular estimation method for the Weibull, Gamma, and Gompertz distributions. Schedae Informaticae, 20: 67 – 82.

12.    Zhibin T., (2009). A new approach to MLE of Weibull distribution with interval data. Reliability Engineering and System Safety, 94: 394 – 403.

13.    Hadley A. and Toumi R., (2003). Assessing changes to the probability distribution of sulphur dioxide in the UK using a lognormal model. Atmospheric Environment, 37: 1461 – 1474.

14.    Chikara R. S. and Folks J. L., (1989). The inverse Gaussian distribution: Theory methodology and applications. Marcel Dekker. New York.

15.    Tweedie M. C. K. (1957). Statistical properties of inverse Gaussian distribution. I. The Annal of Mathematical Statistics, 362 – 367.

16.    Department of Environment, Ministry of Natural Resources and Environment (2012). Malaysia Environmental Quality Report 2012; ISSN 0127-6433.

17.    Wilks D. S. (2011). Statistical methods in the atmospheric sciences. 3^rd Edition. Elsevier. USA.

18.    Alshangiti, A.M., Kayid, M. and Alarfaj, B. (2014). A new family of Marshall–Olkin extended distribution. Journal of Computational and Applied, 271: 369 – 379.

19.    Hamid, H. A., Yahaya, A. S., Ramli N. A., and Ul-Saufie, A. Z. (2013). Finding the best statistical distribution model in PM₁₀ concentration modeling by using lognormal distribution. Journal of  Applied Science, 13 (2): 294 – 300.

20.    Sansuddin, N., Ramli, N. A., Yahaya, A. S., Yusof, N. F. F. M., Ghazali, N. A. and Madhoun, W. A. A. (2011). Statistical analysis of PM₁₀ concentrations at different locations in Malaysia. Environment Monitoring and Assessment, 180: 573 – 588.

21.    Sharma, P., Sharma, P., Jain, S. and Kumar, P., (2013). An intergrated statistical approach for evaluating the exceedance of criteria pollutants in the ambient air of megacity Delhi. Atmospheric Environment, 70: 7 – 17.

22.    Yahaya, A. S. Ramli, N. A. Ul-Saufie, A. Z., Hamid, H. A., Ahmat, H. and Mohtar, Z. A. (2013).      Predicting CO concentrations levels using probability distributions. International Journal of Engineering and Technology, 3: 900 – 905.

23.    Banan N, Latif M. T., Juneng L. and Ahamad F., (2013). Characteristics of surface ozone concentrations at stations with different backgrounds in the Malaysia Peninsula. Aerosol and Air Quality Research, 13: 1090 – 1106.