Accepted author manuscript, 467 KB, PDF document
Available under license: CC BY: Creative Commons Attribution 4.0 International License
Final published version
Licence: CC BY: Creative Commons Attribution 4.0 International License
Research output: Contribution to Journal/Magazine › Journal article › peer-review
Research output: Contribution to Journal/Magazine › Journal article › peer-review
}
TY - JOUR
T1 - A step towards efficient inference for trends in UK extreme temperatures through distributional linkage between observations and climate model data
AU - Tawn, Jonathan Angus
AU - Gabda, Darmesah
AU - Brown, Simon
PY - 2019/9/1
Y1 - 2019/9/1
N2 - The aim of this paper is to set out a strategy for improving the inference for statistical models for the distribution of annual maxima observed temperature data, with a particular focus on past and future trend estimation. The observed data are on a 25 km grid over the UK. The method involves developing a distributional linkage with models for annual maxima temperatures from an ensemble of regional and global climate numerical models. This formulation enables additional information to be incorporated through the longer records, stronger climate change signals, replications over the ensemble and spatial pooling of information over sites. We find evidence for a common trend between the observed data and the average trend over the ensemble with very limited spatial variation in the trends over the UK. The proposed model, that accounts for all the sources of uncertainty, requires a very high dimensional parametric fit, so we develop an operational strategy based on simplifying assumptions and discuss what is required to remove these restrictions. With such simplifications we demonstrate more than an order of magnitude reduction in the local response of extreme temperatures to global mean temperature changes.
AB - The aim of this paper is to set out a strategy for improving the inference for statistical models for the distribution of annual maxima observed temperature data, with a particular focus on past and future trend estimation. The observed data are on a 25 km grid over the UK. The method involves developing a distributional linkage with models for annual maxima temperatures from an ensemble of regional and global climate numerical models. This formulation enables additional information to be incorporated through the longer records, stronger climate change signals, replications over the ensemble and spatial pooling of information over sites. We find evidence for a common trend between the observed data and the average trend over the ensemble with very limited spatial variation in the trends over the UK. The proposed model, that accounts for all the sources of uncertainty, requires a very high dimensional parametric fit, so we develop an operational strategy based on simplifying assumptions and discuss what is required to remove these restrictions. With such simplifications we demonstrate more than an order of magnitude reduction in the local response of extreme temperatures to global mean temperature changes.
KW - climatological data
KW - distributional linkage
KW - generalised extreme value distribution
KW - spatial extremes
KW - temperature data
U2 - 10.1007/s11069-018-3504-8
DO - 10.1007/s11069-018-3504-8
M3 - Journal article
VL - 98
SP - 1135
EP - 1154
JO - Natural Hazards
JF - Natural Hazards
SN - 0921-030X
IS - 3
ER -