General Information
    • ISSN: 2010-0264
    • Frequency: Bimonthly (2010-2014); Monthly (Since 2015)
    • DOI: 10.18178/IJESD
    • Editor-in-Chief: Prof. Richard Haynes
    • Executive Editor: Ms. Nancy Y. Liu
    • Abstracting/ Indexing: Chemical Abstracts Services (CAS), CABI, DOAJ, Ulrich Periodicals Directory, Engineering & Technology Digital Library, Electronic Journals Library, Crossref, ProQuest.
    • E-mail:
  • Jan 09, 2017 News! Vol. 8, No. 2 has been indexed by Crossref.
  • Jan 03, 2017 News! Vol.8, No.2 has been published with online version. 16 peer reviewed articles are published in this issue.   [Click]
The University of Queensland, Australia
It is my honor to be the editor-in-chief of IJESD. The journal publishes good papers in the field of environmental science and development.
IJESD 2015 Vol.6(3): 211-216 ISSN: 2010-0264
DOI: 10.7763/IJESD.2015.V6.592

BUASCSDSEC — Uncertainty Assessment of Coupled Classification and Statistical Downscaling Using Gaussian Process Error Coupling

Queen Suraajini Rajendran and Sai Hung Cheung
Abstract—The statistical downscaling models which are used as a bridging model to connect the global climate model output and the local weather variables have uncertainty associated with it. The uncertainty present in the model as well as in the results should be quantified for reliable climate change impact studies. The sources of uncertainty include natural variability, uncertainty in the climate model(s), downscaling model, model inadequacy and in the predicted results. Uncertainty analysis and quantification in the models is a promising approach for climate change impact studies. In this paper, a new approach called BUASCSDSEC (Bayesian uncertainty analysis for stochastic classification and statistical downscaling with stochastic dependent error coupling) is proposed. It is a robust Bayesian uncertainty analysis methodology and tools for combined classification (to predict the occurrence of rainfall) and statistical downscaling. It is based on coupling dependent modelling error which is viewed as a function modelled as a stochastic process with classification and statistical downscaling models in a way that the dependency among modelling errors will impact the result of the classification and statistical downscaling model calibration and uncertainty analysis for future prediction. Gaussian Process is considered in the error modelling. Singapore data are used and the uncertainty and prediction results are obtained for the validation period (1995-2000). It is observed that the CDFs of the daily predicted samples are consistent with the observed CDF of precipitation. The uncertainty is smaller for the extreme rainfall and the uncertainty for smaller amount of rainfall is more compared to that for the extreme rainfall. From the results obtained, ongoing research for improvement is briefly presented.

Index Terms—Stochastic process, gaussian process, stochastic classification, statistical downscaling, uncertainty quantification, model inadequacy.

The authors are with the School of Civil and Environmental Engineering, Nanyang Technological University, Singapore, 639798 (e-mail:,


Cite:Queen Suraajini Rajendran and Sai Hung Cheung, "BUASCSDSEC — Uncertainty Assessment of Coupled Classification and Statistical Downscaling Using Gaussian Process Error Coupling," International Journal of Environmental Science and Development vol. 6, no. 3, pp. 211-216, 2015.

Copyright © 2008-2016. International Journal of Environmental Science and Development. All rights reserved.