Modeling Locational Differences And Prediction Of Temporal Concentration Of Pm10 Using Time Series Analysis
The aim for this research is to model and predict the PM10 concentrations using the probability distributions and time series models to help curb the adverse impact of PM10 on human health. Ten monitoring stations with five years PM10 monitoring records from 2000 to 2004 were used in this researc...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2010
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/41941/1/NURULILYANA_SANSUDDIN_HJ.pdf |
| _version_ | 1797010732604194816 |
|---|---|
| author | Sansuddin, Nurulilyana |
| author_facet | Sansuddin, Nurulilyana |
| author_sort | Sansuddin, Nurulilyana |
| collection | USM |
| description | The aim for this research is to model and predict the PM10 concentrations using the
probability distributions and time series models to help curb the adverse impact of PM10
on human health. Ten monitoring stations with five years PM10 monitoring records from
2000 to 2004 were used in this research. Four distributions namely gamma, log-normal,
Weibull and inverse Gaussian distributions were used to fit hourly average of PM10
observation records. Based on the five types of performance indicator values, the
gamma distribution is chosen as the best distribution to fitting Johor Bharu, Jerantut,
Kangar and Nilai while, log-normal distribution was fitted to Kota Kinabalu, Kuantan,
Kuching, Manjung, Melaka and Seberang Perai. Predicted PM10 concentrations which
exceeds the threshold limit in unit of days were estimated using the best distributions
and were compared to the actual monitoring records. In order to calibrate the
monitoring records from E-sampler and Beta Attenuation Mass (BAM), the most
appropriate k-factor given by Kuching station was used. In addition, the daily average
of PM10 concentrations was used to find the best time series model. Three types of time
series models were used named autoregressive (AR), moving-average (MA) and
autoregressive moving-average (ARMA). The AR(1) is identified as the best model to
represent all stations except for Jerantut which is represented by the ARMA(1, 1). |
| first_indexed | 2024-03-06T15:24:03Z |
| format | Thesis |
| id | usm.eprints-41941 |
| institution | Universiti Sains Malaysia |
| language | English |
| last_indexed | 2024-03-06T15:24:03Z |
| publishDate | 2010 |
| record_format | dspace |
| spelling | usm.eprints-419412019-04-12T05:26:47Z http://eprints.usm.my/41941/ Modeling Locational Differences And Prediction Of Temporal Concentration Of Pm10 Using Time Series Analysis Sansuddin, Nurulilyana TA1-2040 Engineering (General). Civil engineering (General) The aim for this research is to model and predict the PM10 concentrations using the probability distributions and time series models to help curb the adverse impact of PM10 on human health. Ten monitoring stations with five years PM10 monitoring records from 2000 to 2004 were used in this research. Four distributions namely gamma, log-normal, Weibull and inverse Gaussian distributions were used to fit hourly average of PM10 observation records. Based on the five types of performance indicator values, the gamma distribution is chosen as the best distribution to fitting Johor Bharu, Jerantut, Kangar and Nilai while, log-normal distribution was fitted to Kota Kinabalu, Kuantan, Kuching, Manjung, Melaka and Seberang Perai. Predicted PM10 concentrations which exceeds the threshold limit in unit of days were estimated using the best distributions and were compared to the actual monitoring records. In order to calibrate the monitoring records from E-sampler and Beta Attenuation Mass (BAM), the most appropriate k-factor given by Kuching station was used. In addition, the daily average of PM10 concentrations was used to find the best time series model. Three types of time series models were used named autoregressive (AR), moving-average (MA) and autoregressive moving-average (ARMA). The AR(1) is identified as the best model to represent all stations except for Jerantut which is represented by the ARMA(1, 1). 2010-10 Thesis NonPeerReviewed application/pdf en http://eprints.usm.my/41941/1/NURULILYANA_SANSUDDIN_HJ.pdf Sansuddin, Nurulilyana (2010) Modeling Locational Differences And Prediction Of Temporal Concentration Of Pm10 Using Time Series Analysis. PhD thesis, Universiti Sains Malaysia. |
| spellingShingle | TA1-2040 Engineering (General). Civil engineering (General) Sansuddin, Nurulilyana Modeling Locational Differences And Prediction Of Temporal Concentration Of Pm10 Using Time Series Analysis |
| title | Modeling Locational Differences And
Prediction Of Temporal Concentration
Of Pm10 Using Time Series Analysis
|
| title_full | Modeling Locational Differences And
Prediction Of Temporal Concentration
Of Pm10 Using Time Series Analysis
|
| title_fullStr | Modeling Locational Differences And
Prediction Of Temporal Concentration
Of Pm10 Using Time Series Analysis
|
| title_full_unstemmed | Modeling Locational Differences And
Prediction Of Temporal Concentration
Of Pm10 Using Time Series Analysis
|
| title_short | Modeling Locational Differences And
Prediction Of Temporal Concentration
Of Pm10 Using Time Series Analysis
|
| title_sort | modeling locational differences and prediction of temporal concentration of pm10 using time series analysis |
| topic | TA1-2040 Engineering (General). Civil engineering (General) |
| url | http://eprints.usm.my/41941/1/NURULILYANA_SANSUDDIN_HJ.pdf |
| work_keys_str_mv | AT sansuddinnurulilyana modelinglocationaldifferencesandpredictionoftemporalconcentrationofpm10usingtimeseriesanalysis |