Tweedie models for Malaysia rainfall simulations with seasonal variabilities
This study aims to evaluate the suitability of the Tweedie generalised linear model for characterising monthly rainfall patterns across 18 meteorological stations in Peninsular Malaysia. It incorporates harmonic functions consisting of sine and cosine functions as seasonal pre-dictors and El Niño So...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IWA Publishing
2023
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/104994/1/SuhailaJamaludin2023_TweedieModelsforMalaysiaRainfallSimulations.pdf http://eprints.utm.my/104994/ http://dx.doi.org/10.2166/wcc.2023.275 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This study aims to evaluate the suitability of the Tweedie generalised linear model for characterising monthly rainfall patterns across 18 meteorological stations in Peninsular Malaysia. It incorporates harmonic functions consisting of sine and cosine functions as seasonal pre-dictors and El Niño Southern Oscillation (ENSO) indices as climatic predictors. Results indicate that three harmonic functions are essential to accurately portray rainfall dynamics in the southwestern and northwestern regions, while two suffice for the inland and western regions. However, incorporating four harmonic functions is the most optimal representation of the eastern region. An additional 1-month lag in ENSO indices is introduced to the optimal seasonal predictor model. Based on the findings, the southern oscillation index notably impacts monthly rainfall significantly in eastern and inland areas, while meteorological stations in the western and northwestern areas fit better with the multivariate ENSO index. Strikingly, no substantial impact of climate predictors is observed on the monthly rainfall within the southwestern region. Thus, the influence of climate indices is very much influenced by the geographical locations of the regions. Importantly, generating simulated data through the Tweedie model contributes to a more accurate representation of the statistical properties inherent in rainfall analysis. |
|---|