Transcript Hot extremes in Macao

International Workshop on High Impact Weather Research
2015.1.22
Hot extremes in Macao:
dynamics and predictability
Cheng QIAN (钱诚)1
Wen ZHOU2, Soi Kun FONG3, and Ka Cheng LEONG3
1 Key Laboratory of Regional Climate-Environment for Temperate East Asia & LASG,
Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
2 Guy Carpenter Asia-Pacific Climate Impact Centre, School of Energy and
Environment, City University of Hong Kong, Hong Kong, China
3 Macao Meteorological and Geophysical Bureau, Macao, China
Qian, C., W. Zhou, S. K. Fong, and K. C. Leong, 2015: Two approaches for statistical
prediction of non-Gaussian climate extremes: a case study of Macao hot extremes during
1912−2012. J. Climate, 28(2), 623−636, doi: 10.1175/JCLI-D-14-00159.1
Introduction
• Changes in extreme climate events, especially hot
extremes, could have notable impacts on human
mortality, regional economies, and natural
ecosystems
• Climate change adaptation research requires
spatially fine information
• understanding historical variations and changes in
regional or even local hot extremes and predicting
future changes will be beneficial for human
adaptation to climate change
Introduction
• The Gaussian/normal assumption(正态分布假定)
has been widely used in many previous studies on
climate variability and change that have used
traditional statistical methods (e.g. regression) to
estimate linear trends, diagnose physical
mechanisms, or construct statistical
prediction/downscaling models.
y    x  
m
y     xi  i  
i 1
Introduction
• However, climate extremes sometimes, if not often,
have a non-Gaussian distribution (highly skewed or
kurtotic, or with substantial outliers) (e.g., Klein Tank
et al. 2009), which will distort relationships and
significance tests.
skewed
Kurtotic
outliers
Introduction
• The aim of this study is to propose two approaches
to statistically predict the future occurrence of nonGaussian climate extremes
• the construction of a physically based statistical
prediction/downscaling model
Location of Macao (澳门)
before 1999: a colony of Portugal
Now: a special administrative region of the People's Republic of China
 continuous observations since 1901 and even during World War II
 relatively unaffected by urbanization
 dynamic downscaling is difficult for such a coastal city
Data
• Daily maximum and minimum temperature observations in
Macao during 1912-2012
• the NCEP/NCAR reanalysis data during 1948–2012: sea level
pressure (SLP), winds at 850 hPa (UV850), air temperature at
850 hPa (T850), geopotential height at 500 hPa and 200 hPa
(GHT500 and GHT200), and zonal wind at 200 hPa (U200)
• The monthly NOAA extended reconstructed sea surface
temperature (SST) dataset version 3 (ERSSTv3) for the period
1948–2012
Methods
• Three approaches for normality test: the histogram, QuantileQuantile plotting, and the Jarque-Bera test (Qian and Zhou,
2014)
• Pearson /Spearman correlation coefficient
• effective degrees of freedom (EDOF)
• generalized linear model (GLM)
Hot extreme indices
According to Macao Meteorological
and Geophysical Bureau
hot day (TX>33℃)
– HD33
95% percentile (33.2 ºC)
hot night (TN>28℃)
– HN28
99% percentile (27.8 ºC)
mostly in JJAS
Statistical downscaling for hot days (>33℃)
(1) Weather typing; (2) Weather generators; (3) Regression methods
Distribution of HD33: non-Guassian
x
Solution:
Transform:
After transformation
y  x  0.5
y  x  0.5 to become quasi-Gaussian and use multiple LM
xi  0.5  b0  b1 predictor1 (i)  ...  bp predictorp (i)
Interannual variability
y  x  0.5
Gaussian distribution (a<0.05)
Mostly in JJAS
Associated with the interannual variability of occurrence of hot days at Macao
(1948-2005)
El Niño Modoki
Anomalous more
HD33 year
corresponds to El
Niño Modoki
developing stage.
Macao is located on
the northwest edge of
the cyclonic
circulation system and
thus is controlled by
anomalous northerly
wind, favoring high
temperatures from
mainland China
moving southward
Associated with the interdecadal variability of occurrence of hot days at Macao
(1948-2005)
higher tropospheric temperature in northern Asia
warmer SSTA in North Atlantic Ocean
warmer JJAS mean temperature in Macao
Statistical prediction/downscaling model for HD33
combining the influence factors for the interannual and interdecadal
variability, a physically based multiple linear regression model:
x  0.5  b0  b1MEMI  b2U 850  b3V 850  b4GHT 200  b5 AMO  b6Tx
Schematic diagram
extreme temperature index
transform to normal distribution
find interannual predictors
find interdecadal predictors
training
projection
RCP85
multiple regression model
RCP26
hot nights (HN28)
far from Gaussian
 Multiple linear regression is not appropriate.
 Transform to Gaussian is difficult.
Solution: the non-parametric Spearman's rank correlation coefficient
Generalized Linear Model
Associated with hot nights (HN28) at Macao
non-parametric Spearman’s rank correlation (1948-2005)
Pacific decadal oscillation
(PDO)-like
+
a positive PDO-like
SSTA pattern can
weaken the East Asian
summer monsoon
through weakening
the land-sea thermal
contrast and reduce
JJAS rainfall in Macao,
favoring higher
temperature in Macao
Statistical prediction/downscaling model for HN28
a physically based generalized linear regression model:
log( x  0.25)  b0  b1SSTA  b2 SLP  b3V 850  b4GHT 200
link function is Possion
Schematic diagram
extreme temperature index
using Spearman’s correlation
find predictors
training
projection
RCP85
generalized linear regression model
RCP26
Summary
• Two approaches are proposed to statistically predict/downscaling
non-Gaussian temperature extremes: one uses a multiple linear
regression model after transforming the non-Gaussian predictant to
a quasi-Gaussian variable, and uses Pearson’s correlation test to
identify potential predictors; the other uses a generalized linear
model when the transformation is difficult, and uses a nonparametric Spearman’s correlation test to identify potential
predictors.
• Hot extremes in Macao is associated with the interannual and
interdecadal variability of a coupled El Niño-Southern Oscillation
(ENSO)-East Asian summer monsoon system.
• It is important to test the assumed distribution of climate extremes
and to apply appropriate statistical approaches.
Thank you for your attention!