Since the SAR, there has been more attention paid to the analysis of extreme events in climate models. Unfortunately, none of the major intercomparison projects such as AMIP and CMIP have had diagnostic sub-projects that concentrated on analyses of extreme events. Very few coupled models have been subjected to any form of systematic extreme event analysis. Intercomparison of extreme events between models is also made very difficult due to the lack of consistent methodologies amongst the various analyses and also to the lack of access to high-frequency (at least daily) model data. Analysis has also been limited by the comparatively low resolution at which most models are run, this presents difficulties since most extreme events are envisaged to occur at the regional scale and have comparatively short lifetimes. However other forms of extreme event analysis have been developed which use the large-scale fields produced by a climate model and produce various indices of extreme events; such indices include maximum potential intensity of tropical cyclones (Holland, 1997) or maps of 20-year return values of variables such as precipitation or maximum temperature (Zwiers and Kharin, 1998) (a 20-year return value implies that the value given is reached once in every 20 years).

In this chapter we assess the following types of extreme events that can be presented in terms of global patterns; frequency of tropical cyclones, daily maximum and minimum temperature, length of hot or cold spells, and precipitation intensity and frequency (floods and droughts). While it is arguable that extra-tropical cyclones belong to the class of “extreme events” we choose to include them here for consistency with other chapters. Table 8.4 summarises the climate models and the types of extreme events that have been analysed since the SAR. Assessments of extreme events that are purely local or regional are discussed in Chapter 10.

Table 8.4: Analyses of extreme events in GCMs since the SAR. Wherever possible the model names have been made consistent with Table 8.1; however, since much of the analysis has been done with AGCMs alone (and often with comparatively old model versions) there often is no correspondence between these two tables. The references given refer to the particular analysis used, and are not necessarily tied to a specific model description.
Names	References	Characteristics		Extreme events
		AGCM	OGCM	T	Pr	ETC	TC
ARPEGE-C	Royer et al.,1998	T42, L30	no				F,G
CCC2	Zwiers and Kharin, 1998	T32, L10	no	D,L,R	D,L,R	P	P
CCM	Tsutsui and Kasahara, 1996	T42, L18	no				F,G
	Zhang and Wang, 1997	T42, L18	no				F
	Kothavala, 1997	T42L18	no		E
CGCM1	Kharin and Zwiers, 2000	T32, L10	T64, L29	D,L.R	D,L,R	P	P
CSIRO	Watterson et al., 1995	R21, L9	no				F,G
	Walsh and Pittock, 1998	R21, L9	no	E			I,T,W
	Schubert et al., 1998	R21, L9	no	D,E
ECHAM	Bengtsson et al., 1995, 1996, 1999	T106, L19	no				F,I,N
	Lunkeit et al., 1996	ECHAM2	OPYC			M
	Beersma et al., 1997	ECHAM3	no				F,S
	Christoph et al.1997	T42, L19	no				S
	Schubert et al., 1998	T42, L19	LSG				F,I,S
FSU	Krishnamurti et al., 1998	T42, L16	no				F
GFDL	Vitart et al., 1997	T42, L18	no				F
	Haywood et al., 1997	R15, L9	GFDL_R15_a		D
	Knutson et al., 1998	R30, L14	no				I
	Delworth et al., 1999	R15L9	no	H
	Wetherald and Manabe, 1999	R15L9	no		D
HadCM2	Carnell and Senior, 1998	2.5x3.75, L19	2.5x3.75, L20				N,S
HadCM2b	Bhaskaran and Mitchell, 1998	2.5x3.75, L19	2.5x3.75, L20		E
HadAM2	Thorncroft and Rowell, 1998	2.5x3.75,L19	no				L,W
	Durman et al., 2001	2.5x3.75,L19	no		D,E
JMA	Sugi et al., 1997	T106, L21	no				C
	Yoshimura et al., 1999	T106, L21	no				C
JMA/NIED	Matsuura et al., 1999	T106, L21	0.5x1.0, L37				F
PMIP	Kageyama et al., 1999	ECHAM3, LMD, UGAMP, UKMO	no		D		S
UKMO	Gregory and Mitchell1, 1995	2.5x3.75, L11	no	D	D
	Hulme and Viner, 1998	UKTR	no				T
AGCMs	Hennessy et al., 1997	CSIRO, UKHI	no		D
	Henderson-Sellers et al., 1998	GFDL,ECHAM3	no				F,G,I,N,U
	McGuffie et al., 1999	BMRC, CCM	no		E,L,R		E,L,R
	Zhao et al., 2000	IAP, NCC	IAP
Under “Extreme events”, column T denotes extremes in temperature, Pr denotes extremes in precipitation, ETC denotes extra-tropical cyclone, TC denotes tropical cyclone. The model names and characteristics are further explained (where possible) in Table 8.1. GCM analyses have been with different techniques and methods designated as: C for cyclone centres; D for daily variability of temperature or precipitation; E for extreme temperature or precipitation; F for frequency of cyclones; G for Gray’s yearly genesis parameter; H for heat index; I for intensity of cyclone; L for dry/wet spells or hot/cold spells; M for maximum eddy growth rate; N for numbers of cyclones; P for wind speed; R for return value or return period; S for storm track; T for sea surface temperature; U for maximum potential intensity; W for wave activity.

IPCC Homepage