BUASCSDSEC——Uncertainty Assessment of Coupled Classification and Statistical Downscaling Using Gaussian Process Error Coupling
详细信息 查看官网全文
- 英文论文题名：BUASCSDSEC——Uncertainty Assessment of Coupled Classification and Statistical Downscaling Using Gaussian Process Error Coupling
- 论文作者：Queen Suraajini Rajendran ; Sai Hung Cheung
- 英文论文作者：Queen Suraajini Rajendran ; Sai Hung Cheung ; the School of Civil and Environmental Engineering ; Nanyang Technological University
- 年：2015
- 作者机构：the School of Civil and Environmental Engineering,Nanyang Technological University;
- 英文论文关键词：Stochastic process ; gaussian process ; stochastic classification ; statistical downscaling ; uncertainty quantification ; model inadequacy
- 会议召开时间：2015-03-01
- 会议录名称：International Journal of Environmental Science and Development Vol.6 No.3
- 语种：英文
- 分类号：P435
- 学会代码：CDYA
- 学会名称：成都亚昂教育咨询有限公司
- 页数：6
- 文件大小：1183k
- 原文格式：D
摘要
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.
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.
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.
引文
[1]H.J.Fowler,S.Blenkinsop,and C.Tebaldi,“Linking climate change modelling to impacts studies:recent advancs in downscaling techniques for hydrological modelling,”International Journal for Climatology,vol.27,pp.1547-1578,2007.
[2]R.L.Wilby and I.Harris,“A framework for assessing uncertainties in climate change impacts:Low-flow scenarios for the River Thames,UK,”Water Resources Research,vol.42,pp.1-10,2006.
[3]S.Ghosh and C.Misra,“Assessing Hydrological Impacts of Climate Change:Modelling Techniques and Challenges,”The Open Hydrology Journal,vol.4,pp.115-121,2010.
[4]C.Tebaldi,L.O.Mearns,D.Nychka,and L.Smith,“Regional probabilities of precipitation change:A Bayesian analysis of multimodel simulations,”Geophysical Research Letters,vol.31,2004.
[5]C.Tebaldi,“Quantifying uncertainty in projections of regional climate change:A bayesian approach to the analysis multimodel ensembles,”Journal of Climate,vol.18,pp.1524-1539,2004.
[6]M.Z.Hashmi,A.Y.Shamseldin,and B.W.Melville,“Statistical downscaling of precipitation:state-of-the-art and application of bayesian multi-model approach for uncertainty assessment,”Hydrology and Earth System Sceinces,vol.6,pp.6535-6579,2009.
[7]J.Chen,F.B.Brisstte,and R.Leconte,“Uncertainty of downscaling method in quantifying the impact of climate change on hydrology,”Journal of Hydrology,vol.401,pp.190-202,2011.
[8]B.Axel and N.Danieland B.Gerd,“Effects of climate and land-use change on strom runoff generation:present knowledge and modelling capabilities,”Hydrological Processes,vol.16,pp.509-529,2002.
[9]Beck and Katafygiotis,“Updating Models and Their Uncertainties.I:Bayesian Statistical Framework,”Journal of Engineering Mechanics,vol.124,pp.455-461,1998.
[10]S.H.Cheung and J.L.Beck,“Calculation of Posterior Probabilities for Bayesian Model Class Assessment and Averaging from Posterior Samples Based on Dynamic System Data,”Computer-Aided Civil and Infrastructure Engineering,vol.25,pp.304-321,2010.
[11]J.Ching and Y.Chen,“Transitional Markov Chain Monte Carlo Method for Bayesian Model Updating,Model Class Selection,and Model Averaging,”Journal of Engineering Mechanics,vol.133,pp.816-832,2007.
[12]S.H.Cheung,T.A.Oliver,E.E.Prudencio,S.Prudhomme,and R.D.Moser,“Bayesian Uncertainty Analysis with Applications to Turbulence Modeling,”Reliability Engineering&System Safety,vol.vol.96,issue 9,pp.1137-1149,2011.
[13]C.E.Rasmussen and C.K.I.Williams,Gaussian Processes for Machine Learning,the MIT press,2006,ch.2-5,pp.7-128.
[14]NEA(National Environmetal Agency),Weather Wise Singapore,2009.
[15]Outline map of Singapore.(March 6,2014).[Online].Available:http://www.enchantedlearning.com/asia/singapore/outlinemap/
[2]R.L.Wilby and I.Harris,“A framework for assessing uncertainties in climate change impacts:Low-flow scenarios for the River Thames,UK,”Water Resources Research,vol.42,pp.1-10,2006.
[3]S.Ghosh and C.Misra,“Assessing Hydrological Impacts of Climate Change:Modelling Techniques and Challenges,”The Open Hydrology Journal,vol.4,pp.115-121,2010.
[4]C.Tebaldi,L.O.Mearns,D.Nychka,and L.Smith,“Regional probabilities of precipitation change:A Bayesian analysis of multimodel simulations,”Geophysical Research Letters,vol.31,2004.
[5]C.Tebaldi,“Quantifying uncertainty in projections of regional climate change:A bayesian approach to the analysis multimodel ensembles,”Journal of Climate,vol.18,pp.1524-1539,2004.
[6]M.Z.Hashmi,A.Y.Shamseldin,and B.W.Melville,“Statistical downscaling of precipitation:state-of-the-art and application of bayesian multi-model approach for uncertainty assessment,”Hydrology and Earth System Sceinces,vol.6,pp.6535-6579,2009.
[7]J.Chen,F.B.Brisstte,and R.Leconte,“Uncertainty of downscaling method in quantifying the impact of climate change on hydrology,”Journal of Hydrology,vol.401,pp.190-202,2011.
[8]B.Axel and N.Danieland B.Gerd,“Effects of climate and land-use change on strom runoff generation:present knowledge and modelling capabilities,”Hydrological Processes,vol.16,pp.509-529,2002.
[9]Beck and Katafygiotis,“Updating Models and Their Uncertainties.I:Bayesian Statistical Framework,”Journal of Engineering Mechanics,vol.124,pp.455-461,1998.
[10]S.H.Cheung and J.L.Beck,“Calculation of Posterior Probabilities for Bayesian Model Class Assessment and Averaging from Posterior Samples Based on Dynamic System Data,”Computer-Aided Civil and Infrastructure Engineering,vol.25,pp.304-321,2010.
[11]J.Ching and Y.Chen,“Transitional Markov Chain Monte Carlo Method for Bayesian Model Updating,Model Class Selection,and Model Averaging,”Journal of Engineering Mechanics,vol.133,pp.816-832,2007.
[12]S.H.Cheung,T.A.Oliver,E.E.Prudencio,S.Prudhomme,and R.D.Moser,“Bayesian Uncertainty Analysis with Applications to Turbulence Modeling,”Reliability Engineering&System Safety,vol.vol.96,issue 9,pp.1137-1149,2011.
[13]C.E.Rasmussen and C.K.I.Williams,Gaussian Processes for Machine Learning,the MIT press,2006,ch.2-5,pp.7-128.
[14]NEA(National Environmetal Agency),Weather Wise Singapore,2009.
[15]Outline map of Singapore.(March 6,2014).[Online].Available:http://www.enchantedlearning.com/asia/singapore/outlinemap/