1. Introduction
a. Observational and dynamical studies of the SPCZ
The South Pacific convergence zone (SPCZ) is a band of low-level atmospheric convergence and precipitation that forms one of the major components of South Pacific climate. The SPCZ is responsible for a large fraction of South Pacific precipitation, especially during austral summer [December–February (DJF)]. The SPCZ was first described in surface observations by Bergeron (1930) and identified in early satellite cloud images by Hubert (1961). Using satellite and surface observations, subsequent studies (e.g., Streten 1973; Trenberth 1976; Meehl 1987; Kiladis et al. 1989; Vincent 1994) describe the SPCZ as a band of convective cloud and precipitation extending northwest–southeast in a diagonal line from near New Guinea (0°, 150°E) to the southeastern Pacific (around 30°S, 120°W), as shown in Fig. 1.
The SPCZ has been identified in a range of fields, including precipitation, surface wind convergence, outgoing longwave radiation (OLR), vertical atmospheric motion, and mean sea level pressure (SLP). Kiladis et al. (1989) found that the SPCZ surface wind convergence maximum lies to the south of the band of maximum precipitation, which in turn lies to the south of maximum sea surface temperature (SST).
The SPCZ may be divided into two portions: the zonally oriented western portion attached to the western Pacific warm pool, which interacts with the Australian–Indonesian summer monsoon; and the diagonally oriented eastern portion, which extends into the subtropics (e.g., Trenberth 1976; Kiladis et al. 1989; Vincent 1994). The western portion is a tropical convergence zone, whereas the eastern portion has mixed tropical and extratropical characteristics (Kiladis et al. 1989). The SPCZ forms in the zone of convergence of northeasterly trade winds from the South Pacific high with the southeasterly circulation ahead of anticyclones from the Australian region (Trenberth 1976; Streten and Zillman 1984). The western tropical portion of the SPCZ lies over the warmest SSTs of the western Pacific warm pool, while the eastern portion is forced by interactions with troughs in the midlatitude circulation, which is thought to contribute to its diagonal orientation (Kiladis et al. 1989; Vincent 1994).
Some studies have suggested that the location and orientation of the eastern, diagonal portion of the SPCZ is controlled by the dry zone in the southeast Pacific, which is in turn driven by dry subsiding airflow over the Andes (e.g., Takahashi and Battisti 2007a,b). The strength of trade winds bringing dry zonal inflow from the southeast Pacific is also thought to influence the eastern margin of the SPCZ on synoptic time scales (Lintner and Neelin 2008). Alternatively, Widlansky et al. (2011) argue that the location and orientation of the SPCZ is explained by a region of negative zonal stretching deformation that slows the propagation of synoptic disturbances to the west of the South Pacific high, forming a so-called graveyard (Trenberth 1976) for frontal systems. Widlansky et al. (2011) find that the diagonal orientation of the SPCZ is forced by the Pacific zonal SST gradient, which determines the location of this zone of wave accumulation.
The location and intensity of the SPCZ, particularly the eastern portion, are observed to vary on a range of time scales, from synoptic to interannual and decadal. On intraseasonal time scales, there is evidence that the Madden–Julian oscillation (MJO) influences the SPCZ, with intense convection in the SPCZ out of phase with convection in the Indian Ocean (Vincent 1994; Matthews et al. 1996). Seasonally, the SPCZ is most clearly defined in the austral summer and weaker and less well defined in the austral winter [June–August (JJA)] (e.g., Vincent 1994). The SPCZ experiences a much larger seasonal cycle in position and intensity than the intertropical convergence zone (ITCZ) (Kiladis and van Loon 1988).
On interannual time scales, the location and intensity of the SPCZ varies with El Niño–Southern Oscillation (ENSO). The response of the SPCZ to ENSO has been documented in numerous studies (e.g., Trenberth 1976; van Loon and Shea 1985; Meehl 1987; van Loon and Shea 1987; Kiladis and van Loon 1988; von Storch et al. 1988; Vincent 1994; Folland et al. 2002; Vincent et al. 2011). Trenberth (1976) described the movement of the SPCZ as south and west during positive Southern Oscillation (La Niña) events and north and east during negative Southern Oscillation (El Niño) events. During El Niño events, the SPCZ and ITCZ tend to fuse in the central equatorial Pacific with enhanced convergence there and in the eastern tropical Pacific (e.g., Kiladis and van Loon 1988; Folland et al. 2002). Vincent et al. (2011) classified the SPCZ orientation into four classes—negative, neutral, positive, and asymmetric—corresponding approximately to La Niña, neutral, El Niño, and strong El Niño states, respectively. The “asymmetric” class consists of a near-zonal SPCZ, merged with the ITCZ and located over the region of maximum SST (Vincent et al. 2011).
Salinger et al. (2001) and Folland et al. (2002) found that the SPCZ varies on multidecadal time scales with the interdecadal Pacific oscillation (IPO; Power et al. 1999). An SPCZ position index (SPI) was defined to measure the seasonal mean location of the SPCZ based on sea level pressure at Suva in Fiji and Apia in Samoa. Both ENSO and IPO were found to influence the SPCZ position in a quasi-independent manner (Salinger et al. 2001; Folland et al. 2002).
b. Coupled model simulations of the SPCZ
The ability of coupled models to simulate a realistic SPCZ has improved with each generation of models. Some early Coupled Model Intercomparison Project phase 2 (CMIP2) models simulated a “double ITCZ” straddling the equator in the eastern Pacific (100°–150°W), while others simulated an ITCZ that migrated across the equator with the seasons following solar heating (Mechoso et al. 1995). The double ITCZ was associated with biases in the mean state, including equatorial cold tongue SSTs that were too cold, too narrow, and extended too far west; and SSTs south of the equator that were too warm (Mechoso et al. 1995).
More recent World Climate Research Programme (WCRP) CMIP3 (Meehl et al. 2007) models have an improved representation of the mean state and seasonal cycle as well as higher resolution and more sophisticated atmosphere and ocean physics (e.g., Randall et al. 2007). These state-of-the-art coupled models have some skill in capturing the mean state and seasonal cycle of tropical Pacific precipitation, SSTs, and atmospheric circulation but errors remain. Studies of the CMIP3-generation models have found that those models without ocean heat flux adjustment often simulate an equatorial cold tongue that also extends too far into the western Pacific, with an associated dry bias on the equator. In these models, the SPCZ tends to be too zonal and to extend too far into the eastern tropical Pacific (e.g., Lin 2007; Zhang et al. 2007).
Other studies focus on the presence of a double ITCZ in the eastern Pacific, where a small secondary precipitation band is found in observations only in March–April (e.g., de Szoeke and Xie 2008; Bellucci et al. 2010). Many nonflux-adjusted models simulate an extension of their SPCZ precipitation band to the far east for part of the year, with an absent or weakened northern ITCZ at the same time. This seasonally “alternating” or “migrating” convergence zone produces an apparent double ITCZ in the annual mean, while a smaller set of models simulate a permanent convergence zone in both hemispheres in all months (de Szoeke and Xie 2008).
To our knowledge, the ability of CMIP3 coupled atmosphere–ocean models to capture the location, intensity, and variability of the SPCZ has not previously been evaluated in detail. The SPCZ has been examined in climate models as part of understanding the double ITCZ error, but such analysis has not focused on the ability of models to capture a realistic SPCZ climatology or its interannual variability. In this study, the climatology and interannual variability of the SPCZ in twentieth-century climate simulations from 24 coupled models are compared with observations. It is hoped that by evaluating the models, we will be able to provide useful guidance in developing reliable climate projections for the Pacific region.
The paper is structured as follows. In section 2, the models and datasets used are described. In section 3, the metrics used to define the SPCZ position and intensity are outlined. In section 4, the seasonal cycle of the SPCZ in the models is evaluated. In section 5, the interannual variability of the simulated SPCZ is compared with corresponding variability in the observational data. In section 6, the results are summarized, and the implications for using models to examine SPCZ changes in future climate projections are discussed.
2. Models and observational data
Dynamical fields (surface winds and wind divergence, vertical motion) are obtained from National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis (Kalnay et al. 1996) for the period 1950–99. Gridded precipitation observations are taken from the Climate Prediction Center Merged Analysis of Precipitation (CMAP; Xie and Arkin 1997) and Global Precipitation Climatology Project (GPCP) version 2 (Adler et al. 2003) for the period 1979–99, as these datasets are only available for the satellite era, that is, from the late 1970s onward. Outgoing longwave radiation (OLR) data are obtained from the National Oceanic and Atmospheric Administration (NOAA)’s interpolated OLR dataset (Liebmann and Smith 1996) for the period 1979–99, excluding years before 1979 for consistency with the precipitation data. All the observational datasets use a regular 2.5° latitude × 2.5° longitude grid.
There are some well-known differences between the CMAP and GPCP precipitation datasets over the tropical Pacific region. In particular, CMAP is known to have higher precipitation rates over the tropical ocean because of the inclusion of rain gauge data from atoll stations to adjust the satellite estimates, while GPCP has higher precipitation over some land areas because of the wind correction of rain gauge data (e.g., Adler et al. 2003; Yin et al. 2004). As the “correct” value of rainfall over the tropical Pacific region is not known, the models are compared with both CMAP and GPCP precipitation in some cases, providing a range of uncertainty for the observations. In cases where a single observation is used, CMAP precipitation is chosen.
The observations are compared with simulations from 24 atmosphere–ocean coupled models from the WCRP CMIP3 multimodel dataset (Meehl et al. 2007). The models and climate simulations are those used for the Fourth Assessment Report (AR4) of the Intergovernmental Panel on Climate Change (IPCC). The model host institutes and model names are listed in Table 1 (details of the model configurations and associated references are available online at http://www-pcmdi.llnl.gov/ipcc/model_documentation/ipcc_model_documentation.php). The twentieth-century (20C3M) model simulations are analyzed in this study. The 20C3M simulations are forced by solar, volcanic, sulfate aerosol direct effects and greenhouse gas forcings to reproduce the climate of the twentieth century.
The seasonal climatologies are calculated over the period 1950–99 for both models and observations, with the exception of precipitation, where the years 1979–99 are used to allow comparison with the observations for the same period. When interannual variability is investigated, a longer (1950–99) period is used for model precipitation to obtain a larger sample of ENSO events. All models are regridded to a 2.5° × 2.5° grid using first-order conservative remapping, which guarantees that the area-integrated field is conserved. Four of the 24 models use ocean heat flux adjustments: CGCM3.1(T47) and CGCM3.1(T63), ECHO-G, and MRI-CGCM2.3.2.
3. Definition of the SPCZ
To evaluate the simulation of the SPCZ in the climate models, we require a definition of the SPCZ that is robust despite model systematic biases and that captures the most important features of the SPCZ. The definition can then be used to calculate parameters such as (i) SPCZ position (horizontal orientation or “slope” and mean latitude), (ii) SPCZ seasonal precipitation, and (iii) SPCZ interannual variability—for example, meridional displacement in response to ENSO.
Previous studies have defined the location of the SPCZ using different observed variables (e.g., Kiladis et al. 1989; Folland et al. 2002; Vincent et al. 2011). Folland et al. (2002) used the convergence of low-level (10 m) winds to calculate a 2-dimensional SPCZ position, while the SPI of pressure difference between Apia and Suva was used to measure the interannual variability of mean SPCZ position. Vincent et al. (2011) defined the SPCZ position in gridded observed precipitation as the latitude of maximum precipitation between the equator and 30°S for grid points with precipitation greater than 6 mm day−1.
A comparison of SPCZ location from CMAP precipitation, NOAA OLR, NCEP near-surface wind and divergence, and omega (vertical velocity of the atmosphere) at 500 hPa is given in Fig. 1. We use the seasonal mean for austral summer, as this is when the SPCZ is most well defined (e.g., Vincent 1994). The SPCZ location can be defined as the line of maximum precipitation and minimum OLR, wind divergence, and omega at 500 hPa, as indicated in Fig. 1. Using this definition, the location of the SPCZ is similar for precipitation, OLR, and omega but slightly further poleward (approximately 5° farther south) for convergence, in agreement with Kiladis et al. (1989). As precipitation, OLR, and omega at 500 hPa produce a similar SPCZ location, we hereafter use maximum precipitation to define the position of the SPCZ. Unlike Vincent et al. (2011), we do not set a minimum precipitation threshold to define the SPCZ but instead use all values in a given longitude range, as precipitation amount may vary widely between models.
To provide a simple metric for SPCZ latitude and orientation, we will identify the latitude of maximum precipitation at each longitude point over the SPCZ region, specified as the domain where the seasonal mean SPCZ is well defined in observations (from 155°E to 140°W and from the equator to 30°S). A linear fit will be calculated to provide an objective estimation of the slope and mean latitude of the SPCZ for comparison between models and observations.
This approach provides a first-order description of the SPCZ that captures the western, tropical component of the SPCZ well. The eastern, extratropical component of the SPCZ extends beyond the 140°W limit of the SPCZ region and is not easily represented with a linear fit. While other studies have used two line sections to characterize the western and eastern SPCZ (Vincent et al. 2011), that approach is less suited to the CMIP3 models, as many of the models do not simulate the extratropical component of the SPCZ. For this reason, we use a single linear fit to define the location of the SPCZ, while noting that this does not completely represent the eastern portion of the SPCZ.
4. Seasonal climatology of the SPCZ in observations and models
a. Precipitation seasonal climatology
In Fig. 2, the DJF seasonal mean precipitation calculated over the period 1979–99 is shown for the 24 CMIP3 models, the flux-adjusted and nonflux-adjusted model means, and observed CMAP precipitation. A band of precipitation extending from the tropical southwest Pacific toward the east is evident in all model simulations; however, the width, intensity, and alignment of the model SPCZ differ greatly from one model to another. We first discuss the broad-scale features of the SPCZ in the models before providing a quantitative evaluation in the next subsection.
Out of the 24 models, 20 simulate an SPCZ that is distinct from the Northern Hemisphere ITCZ, while two models simulate an SPCZ that is merged with the ITCZ (GISS-AOM, GISS-ER) and two simulate an SPCZ that does not extend beyond the western Pacific into the central Pacific [MIROC3.2(medres) and MIROC3.2(hires)]. Of the 20 models with a distinct SPCZ, all the nonflux-adjusted models have dry biases of varying magnitudes on the equator because of the SST cold tongue extending too far into the western Pacific warm pool (e.g., Randall et al. 2007; Cai et al. 2009). The four models using ocean heat flux adjustment simulate a more realistic DJF SPCZ precipitation distribution, which is consistent with their reduced SST biases.
The orientation of the SPCZ in the nonflux-adjusted models is generally too zonal, in agreement with previous studies (e.g., Lin 2007; Zhang et al. 2007). Those models with the largest dry bias in the western equatorial Pacific also have the most zonal SPCZ (e.g., BCCR-BCM2, GISS-EH, INM-CM3.0, and PCM). In these models, the SPCZ does not extend toward the equator and merge with the region of monsoon convergence west of the date line as in observations (e.g., Kiladis and van Loon 1988). Instead, the SPCZ lies south of the Solomon Islands in the western Pacific in these models, with an expanded equatorial “dry zone” separating it from the ITCZ.
The zonal average of DJF precipitation over the range 155°E–140°W for the models and CMAP observations is shown in Fig. 3. The maximum of zonal average precipitation south of the equator provides an estimate of the mean latitude of the SPCZ in this longitude range. For CMAP observations, the peak occurs at 11.25°S, with most models having a peak near this latitude. The exceptions are those models without a distinct SPCZ (GISS-AOM, GISS-ER, and the MIROC models), which have a single maximum extending over the equator. The majority of nonflux-adjusted models are too dry on the equator, with dry biases of up to 5 mm day−1.
The seasonal cycle of precipitation in the SPCZ region is examined using time–latitude plots of zonal mean precipitation over the region 155°E–140°W (Fig. 4). Both CMAP and GPCP precipitation show a Southern Hemisphere convergence zone throughout the year, with the most intense precipitation during austral summer. The dry zone on the equator expands as the ITCZ moves north in boreal summer [June–August (JJA)] and peaks in observations around August–October. The nonflux-adjusted and flux-adjusted model means both capture seasonal variability in the SPCZ and ITCZ, although they show a wider ITCZ than observed, and the northward movement and intensification of the ITCZ occurs later in the nonflux-adjusted model mean than observations. The flux-adjusted model mean seasonal cycle is closer to observations in the Southern Hemisphere, with more precipitation on the equator and a stronger SPCZ seasonal cycle.
A distinct Southern Hemisphere precipitation band is apparent throughout the seasonal cycle in most models, with the exception of GISS-AOM, GISS-ER, and the two MIROC models. The timing of the seasonal intensification of the SPCZ in austral summer is also captured by most models, except for the above models as well as INGV-SXG, INM-CM3.0, and IPSL CM4. The seasonal meridional migration of the ITCZ in the Northern Hemisphere is evident for most models, although some models have a wider-than-observed ITCZ or a mean location that is too far north. The dry bias on the equator persists throughout the seasonal cycle in the majority of models without flux adjustment.
b. Slope and latitude of the SPCZ
The orientation and latitude of the SPCZ in the observations and models is now identified in a quantitative manner using a linear fit to the maximum precipitation within the SPCZ region, as shown in Fig. 5. A line is fitted to the latitudes of maximum precipitation for the DJF seasonal mean at each longitude from 155°E to 140°W and from the equator to 30°S. We calculate the slope (in degrees north per degree east) and the mean latitude (in degrees north) of the SPCZ line. The slope and mean latitude of the SPCZ are listed in Table 2 for CMAP and GPCP and for each model. In most cases, the tropical portion of the model SPCZ precipitation band can be captured with the linear fit technique. The linear fit is poorly defined for the models with an indistinct SPCZ: GISS-ER, MIROC3.2(hires), and MIROC3.2(medres). The method also calculates a weakly positive SPCZ slope for a small number of models that have strong a zonal precipitation band extending into the eastern Pacific: FGOALS-g1.0 and GFDL CM2.1.
While Vincent et al. (2011) use only areas with precipitation above 6 mm day−1 to define an SPCZ line, we use all points within the specified longitude range. For comparison, the model SPCZ lines were also calculated using the 6 mm day−1 threshold (not shown). Only a small number of models do not have precipitation over the 6 mm day−1 threshold within the SPCZ longitude range (see Fig. 5). When the model SPCZ line is fitted only to points above the threshold, the slope of the MIROC3.2(medres) line is much greater, as the line is now fitted only within 155°–170°E. The MIROC3.2(hires) SPCZ line is also shorter, but it has a similar slope. In general, the SPCZ linear fit method is robust for all those models with a distinct SPCZ precipitation band.
The mean latitude versus the slope of the observed and model SPCZ are plotted in Fig. 6. The CMAP and GPCP-observed SPCZ have slopes of −0.29°N/°E and −0.25°N/°E and mean latitudes of 11.2° and 12.2°S, respectively. Only two models (U for MRI-CGCM2.3.2 and X for UKMO-HadGEM1) have an SPCZ slope close to the range of the observations. All other models have an SPCZ slope that is smaller than observed, indicating that the SPCZ orientation is too zonal, in agreement with previous studies (e.g., Lin 2007; Zhang et al. 2007). The model SPCZ mean latitudes span a large range, from around 2°S to around 16°S, although much of this spread is due to a few outliers. The flux-adjusted model mean SPCZ slope (−0.21°N–°E) is close to observations, whereas the nonflux-adjusted model mean slope (−0.05°N–°E) is too zonal. The flux-adjusted and nonflux-adjusted model mean SPCZ latitudes of 11.4° and 12.3°S, respectively, are close to the observed CMAP and GPCP mean latitudes. The interannual variability of the observed and model SPCZ slope and latitude is examined in section 5.
c. Evaluating model simulations of the SPCZ
The skill of models in simulating the seasonal SPCZ precipitation climatology can be represented using a Taylor diagram (Taylor 2001). The Taylor diagram summarizes two statistics in this case: the spatial correlation between model and observed climatology, and the normalized standard deviation of the model spatial precipitation distribution (model standard deviation divided by the standard deviation of the observations). The closer the models are to the reference observations, the better the agreement in the both spatial pattern and amplitude of precipitation variation over the SPCZ region.
The Taylor diagrams of DJF and JJA seasonal mean model precipitation relative to CMAP observations for the SPCZ region are shown in Fig. 7, with models indicated using the letters given in Table 1. The spatial correlation between each model and CMAP seasonal precipitation pattern is shown on the outside of a unit circle; the spatial standard deviation, normalized relative to CMAP, is shown on the radius. The location of GPCP on the Taylor diagram (shown as a star) provides an indication of observational uncertainty, with GPCP having the largest spatial correlation with CMAP (r = 0.98) but a substantially lower standard deviation. The Taylor diagram is plotted relative to CMAP, as the model precipitation amounts are generally much larger than GPCP.
The models have higher spatial correlations with CMAP precipitation in austral summer than in austral winter on average, and a smaller scatter of correlation coefficients in DJF. The relative position of the individual nonflux-adjusted models in the Taylor diagram is quite different in DJF and JJA, indicating that models do not have the same skill in each season. The flux-adjusted models (C, D, I, and U), on the other hand, are among those with the highest correlations in both DJF and JJA (and in the other two seasons, not shown). The majority of models have a larger range of precipitation values (normalized standard deviation) than CMAP in both DJF and JJA, and all models have a larger standard deviation than GPCP.
As noted above, there is large uncertainty in the magnitude of observed precipitation for the region; however, the spatial pattern is robust across the observational products. As such, the spatial correlation between model and CMAP precipitation is a more reliable metric of model SPCZ fidelity than statistics combining both correlation and standard deviation, such as root-mean-square (RMS) error. The spatial correlation between CMAP and model DJF precipitation is calculated over two domains: the SPCZ domain, extending 0°–30°S, 155°E–140°W; and a larger tropical Pacific domain, covering 30°S–20°N, 120°E–140°W (values given in Table 2). The correlation over the larger domain indicates whether the model also captures the spatial pattern of the ITCZ, highlighting models with a good overall simulation of tropical Pacific precipitation. Those models with the highest spatial correlations in both domains are the 4 flux-adjusted models, as well as CSIRO Mk3.5 and GFDL CM2.0. Other models with high correlations (r > 0.70) in the SPCZ region are CSIRO Mk3.0, ECHAM5, FGOALS-g1.0, GFDL CM2.1, IPSL CM4 and UKMO-HadCM3. We now consider whether the models are able to capture the interannual variability of the SPCZ position and intensity.
5. Interannual variability of the SPCZ in observations and models
a. Observed and model SPCZ variability in response to ENSO
As discussed in section 1, many previous studies have identified a relationship between the position and intensity of the SPCZ and ENSO. Here we investigate the variability of the SPCZ position (latitude and slope) in relation to the Niño-3.4 region (5°N–5°S, 120°–170°W) SST variability. The observed Niño-3.4 SST anomaly time series is calculated using the Met Office Hadley Centre Sea Ice and SST (HadISST) data (Rayner et al. 2003). Both observed and model Niño-3.4 SST time series are detrended prior to calculating correlation coefficients.
The correlations between the mean latitudes of the SPCZ lines and the Niño-3.4 SST anomaly for DJF seasonal means are given in Table 2. A strong relationship is found for observed CMAP and GPCP precipitation for the years 1979–99 (r = 0.83 for CMAP, r = 0.86 for GPCP), indicating a northward (southward) displacement of the SPCZ in response to El Niño (La Niña) events. A longer 50-yr period (model years 1950–99) is used for the model comparison to provide a larger sample of model SPCZ responses to ENSO. As many of the CMIP3 models do not capture the observed phase locking of ENSO to the seasonal cycle (e.g., Randall et al. 2007), the relationship between SPCZ position and ENSO may not be strongest in DJF. However, similar results were obtained for November–April mean precipitation and Niño-3.4 SSTs, implying that the results are robust for the entire austral “wet season.”
The mean latitude of the SPCZ line for each DJF seasonal mean is plotted against the Niño-3.4 SST anomaly for the same season in Fig. 8, comparing the relationship calculated from CMAP precipitation with the 24 CMIP3 models. The majority of models capture a positive relationship between SPCZ latitude and Niño-3.4 SST anomaly, with the SPCZ shifting north during El Niño events and south during La Niña events. Out of 24 models, 10 have a correlation coefficient between SPCZ latitude and Niño-3.4 above r = 0.8, while a further five have a correlation above r = 0.6. A weak relationship (r < 0.6) between SPCZ latitude and ENSO is found for CSIRO Mk3.0, INM-CM3.0, MIROC3.2(medres), MIROC3.2(hires), and MRI-CGCM2.3.2; although, the INM-CM3.0 model simulates very little SPCZ variability, and MIROC3.2(hires) and MIROC3.2(medres) do not have a distinct SPCZ. Those models that simulate a very weak relationship or none at all (r < 0.3) between SPCZ latitude and ENSO are GISS-AOM, GISS-EH, GISS-ER, and UKMO-HadGEM1.
The slope of the SPCZ line is plotted against Niño-3.4 SST for CMAP and the CMIP3 models in Fig. 9. The CMAP SPCZ slope does not vary substantially with Niño-3.4 SST, with the exception of the two largest El Niño events, when the magnitude of the slope is reduced to −0.02 in 1982/83 and to +0.03 in 1997/98 compared with the mean value of −0.29°N/°E (Table 2). These two years were identified by Vincent et al. (2011) as “asymmetric” SPCZ years with a reduced SPCZ slope, in agreement with our results (a third asymmetric year identified by Vincent et al. (2011), 1991/92, does not display a reduced SPCZ slope using our method). The combined variability of SPCZ latitude and slope provides an alternative decomposition of SPCZ position to the two principal components of SPCZ location (north–south and east–west) identified by Vincent et al. (2011). However, shifts in the east–west direction are not captured explicitly by this approach, as the longitude range of the SPCZ line is fixed.
As the CMIP3 models generally simulate a shallower SPCZ slope than observed, the models do not show a large reduction in SPCZ slope for strong El Niño events. However, several models do appear to have a near-zero SPCZ slope during a few strong El Niño events, including CNRM-CM3, ECHAM5, ECHO-G, FGOALS-g1.0, GFDL CM2.1, IPSL CM4, MRI-CGCM2.3.2, PCM, and UKMO-HadCM3. In disagreement with CMAP observations, some of these models also show a reduced SPCZ slope during La Niña events. This may be due to the extension of cold equatorial SSTs (and the associated dry bias) to the far western Pacific in model La Niña events, enhancing the overly zonal mean alignment of the SPCZ in these models. Several models, including GISS-AOM and GISS-ER, simulate variability of SPCZ slope despite having a very weak ENSO, while other models, such as GISS-EH and INM-CM3.0, simulate little SPCZ slope variability.
b. Relationship between SPCZ-region precipitation and ENSO
The interannual variability of seasonal mean precipitation in the SPCZ region is examined using composites for El Niño and La Niña years. The composites are constructed for DJF seasonal mean precipitation with a threshold of ±0.75 times the standard deviation of Niño-3.4 SST anomalies, using 50 years (1950–99) of model output and 21 years (1979–99) of CMAP and GPCP data. The average precipitation in the SPCZ region for El Niño years, La Niña years, and all years is calculated for CMAP, GPCP, and CMIP3 models, shown in Fig. 10. The results are similar using higher and lower thresholds to define ENSO events.
Both CMAP and GPCP show increased DJF precipitation in the SPCZ region during El Niño years, consistent with previous studies (e.g., Vincent et al. 2011). The increased precipitation in the SPCZ region in El Niño years is largely due to an increase in the area of the SPCZ (defined by the 6 mm day−1 precipitation threshold), with a merging of the SPCZ and ITCZ regions of tropical convection, especially in the case of strong El Niño events. The observed amount of precipitation in the SPCZ region during La Niña events is similar to the average for all years; although, this result depends on the boundary chosen for the SPCZ region, as precipitation commonly shifts southwest during La Niña events.
The majority of models simulate more precipitation in the SPCZ region in El Niño years than in all years or La Niña years, in agreement with observations. The difference in SPCZ-region precipitation between El Niño years and La Niña years is statistically significant at the 95% confidence level, using a Student’s t test, for all models except BCCR-BCM2.0, CGCM3.1(T47), GISS-AOM, GISS-EH, GISS-ER, INM-CM3, MIROC3.2(medres), and UKMO-HadGEM1. The main cause of the increased model precipitation in the SPCZ region in El Niño years is also an increase in the area of the SPCZ, with some models simulating up to a 70% increase in the number of grid points over 6 mm day−1 within the SPCZ region (not shown). It is encouraging that the majority of CMIP3 models capture the change in precipitation amount with El Niño events, with the exception of those models with extremely weak ENSO or SPCZ variability.
6. Summary and conclusions
The seasonal and interannual characteristics of the SPCZ in 24 CMIP3 coupled climate models were compared with observations to evaluate model skill. A comparison of observed and model austral summer precipitation showed that 20 out of 24 models simulate a distinct SPCZ precipitation band extending from the western Pacific to the central Pacific. The exceptions are the GISS-AOM, GISS-ER, MIROC3.2(medres), and MIROC3.2(hires) models, which simulate a single tropical precipitation band (GISS models) or a precipitation band confined to the west of the date line (MIROC models).
The four models with heat flux adjustment reproduce the observed spatial pattern of SPCZ precipitation most closely, while many of the models without flux adjustment simulate an SPCZ that is too zonal in orientation, failing to capture the diagonal extension of the SPCZ into the extratropics. In the majority of nonflux-adjusted models, the SPCZ meets the ITCZ too far to the west because of model cold SST biases on the equator. Most models are able to capture the seasonal intensification of the SPCZ in austral summer, with the exception of those models without a distinct SPCZ (listed above) as well as INGV-SXG, INM-CM3.0, and IPSL CM4.
The position of the SPCZ was identified using the latitudes of maximum precipitation at each longitude in a defined SPCZ region, with a linear fit applied to obtain a simplified representation of the observed and model SPCZ. The slope and mean latitude of the SPCZ was calculated for the models and compared with the observed estimates from CMAP and GPCP precipitation. The model SPCZ slopes are too weak in all cases, except in MRI-CGCM2.3.2 and UKMO-HadGEM1, confirming the overly zonal SPCZ orientation. The mean latitudes of the flux-adjusted and nonflux-adjusted model mean SPCZ lines are within the range of observations, although the SPCZ line lies too far south in the western tropical Pacific and too far north in the central Pacific in the majority of models.
Model skill in simulating the climatology of the SPCZ was also evaluated using spatial correlations between CMAP and model austral summer precipitation in the SPCZ region. The models with the highest spatial correlations are the four flux-adjusted models—CGCM3.1(T47), CGCM3.1(T63), ECHO-G, and MRI-CGCM2.3.2—as well as the nonflux-adjusted models CSIRO Mk3.5 and GFDL CM2.0. Other models with high spatial correlations for DJF precipitation in the SPCZ region are CSIRO Mk3.0, ECHAM5, FGOALS-g1.0, GFDL CM2.1, IPSL CM4, and UKMO-HadCM3.
Some CMIP3 models are found to have an overly zonal SPCZ that extends to the tropical eastern Pacific, producing the so-called double ITCZ examined in numerous studies (e.g., de Szoeke and Xie 2008; Bellucci et al. 2010). Other models have a zonal SPCZ that is confined to the western and central Pacific, without this eastern band. The absence of a zonal SPCZ or double ITCZ in the four flux-adjusted models suggests that the SST pattern, and zonal SST gradient in particular, is an important control on the position and orientation of the model SPCZ. Studies have also suggested that SST biases may be amplified via feedbacks with atmospheric circulation (e.g., Zhang et al. 2007). There is evidence from models and observations that the easterly trade wind circulation plays a role in setting the position of the SPCZ (e.g., Takahashi and Battisti 2007a; Lintner and Neelin 2008). Therefore, model biases in atmospheric circulation or convection may also contribute to biases in the SPCZ position. Further studies can play a role in exploring the influence of both ocean and atmospheric processes on the SPCZ location.
The response of the SPCZ to ENSO variability was examined using changes in the mean latitude and slope of the SPCZ line, as well as changes in the amount of precipitation in the SPCZ region. A strong positive relationship between the austral summer latitude of the SPCZ line and Niño-3.4 SST is found for observations, with the SPCZ moving north during El Niño events and south during La Niña events. The majority of models capture this relationship. The four flux-adjusted models, which simulate the best seasonal SPCZ climatology, do not uniformly perform well in capturing the SPCZ response to ENSO. Only one of the flux-adjusted models (ECHO-G) has a correlation between SPCZ latitude and Niño-3.4 among the top 10 models.
The observed slope of the SPCZ is found to vary during very strong El Niño events, when the slope becomes small as the SPCZ merges with the ITCZ and precipitation shifts to the central Pacific. A smaller subset of models is able to capture this behavior, including CNRM-CM3, ECHAM5, ECHO-G, FGOALS-g1.0, GFDL CM2.1, IPSL CM4, MRI-CGCM2.3.2, PCM, and UKMO-HadCM3. Some of these models also simulate a near-zero SPCZ slope during La Niña events, which is not consistent with observations. The majority of models capture an increase in precipitation in the SPCZ region during El Niño events, largely because of an increased SPCZ area (within the 6 mm day−1 precipitation contour).
We have found that the majority of CMIP3 models capture the seasonal climatology of the SPCZ and a subset of these models also capture SPCZ variability on interannual time scales. Skill in reproducing the seasonal climatology and interannual variability of the SPCZ may increase confidence in future projections of SPCZ changes, although such skill is only a partial guide to reliability in climate projections (e.g., Whetton et al. 2007). There is a clear distinction between those models that capture the main features of the SPCZ and those with an inadequate representation of seasonal and interannual SPCZ characteristics. This group includes models without a distinct SPCZ: GISS-AOM and GISS-ER; models with an SPCZ confined to the western Pacific: MIROC3.2(hires) and MIROC3.2(medres); and models with an extremely weak SPCZ response to ENSO variability: the GISS models, INM-CM3.0, and UKMO-HadGEM1. Changes in the SPCZ in future climate projections will be outlined in a forthcoming study, using the evaluation of model SPCZ performance presented here to provide guidance regarding model strengths and weaknesses and to inform a discussion of confidence in the future projections.
Acknowledgments
The research discussed in this paper was conducted with the support of the Pacific Climate Change Science Program, a program supported by AusAID, in collaboration with the Department of Climate Change and Energy Efficiency, and delivered by the Bureau of Meteorology and the Commonwealth Scientific and Industrial Research Organisation (CSIRO). Alex Sen Gupta provided useful comments on this work. The comments of George Kiladis, Yangxing Zheng, and one anonymous reviewer greatly improved this manuscript. We acknowledge the modeling groups, the Program for Climate Model Diagnosis and Intercomparison (PCMDI) and the WCRP’s Working Group on Coupled Modelling (WGCM)for their roles in making available the WCRP CMIP3 multimodel dataset. Support of this dataset is provided by the Office of Science, U.S. Department of Energy. More details on model documentation are available at the PCMDI Web site (available at http://www-pcmdi.llnl.gov).
REFERENCES
Adler, R. F., and Coauthors, 2003: The version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4 , 1147–1167.
Bellucci, A., S. Gualdi, and A. Navarra, 2010: The double-ITCZ syndrome in coupled general circulation models: The role of large-scale vertical circulation regimes. J. Climate, 23 , 1127–1145.
Bergeron, T., 1930: Richtlinien einer dynamischen Klimatologie. Meteor. Z., 47 , 246–262.
Cai, W., A. Sullivan, and T. Cowan, 2009: Rainfall teleconnections with Indo-Pacific variability in WCRP CMIP3 models. J. Climate, 22 , 5046–5071.
de Szoeke, S. P., and S.-P. Xie, 2008: The tropical eastern Pacific seasonal cycle: Assessment of errors and mechanisms in IPCC AR4 coupled ocean–atmosphere general circulation models. J. Climate, 21 , 2573–2590.
Folland, C. K., J. A. Renwick, M. J. Salinger, and A. B. Mullen, 2002: Relative influences of the interdecadal Pacific oscillation and ENSO on the South Pacific convergence zone. Geophys. Res. Lett., 29 , 1643. doi:10.1029/2001GL014201.
Hubert, L. F., 1961: A subtropical convergence line of the South Pacific: A case study using meteorological satellite data. J. Geophys. Res., 66 , 797–812.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–471.
Kiladis, G. N., and H. van Loon, 1988: The Southern Oscillation. Part VII: Meteorological anomalies over the Indian and Pacific Ocean sectors associated with extremes of the oscillation. Mon. Wea. Rev., 116 , 120–136.
Kiladis, G. N., H. von Storch, and H. van Loon, 1989: Origin of the South Pacific convergence zone. J. Climate, 2 , 1185–1195.
Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77 , 1275–1277.
Lin, J.-L., 2007: The double-ITCZ problem in IPCC AR4 coupled GCMs: Ocean–atmosphere feedback analysis. J. Climate, 20 , 4497–4525.
Lintner, B. R., and J. D. Neelin, 2008: Eastern margin variability of the South Pacific convergence zone. Geophys. Res. Lett., 35 , L16701. doi:10.1029/2008GL034298.
Matthews, A. J., B. J. Hoskins, J. M. Slingo, and M. Blackburn, 1996: Development of convection along the SPCZ within a Madden-Julian oscillation. Quart. J. Roy. Meteor. Soc., 122 , 669–688.
Mechoso, C. R., and Coauthors, 1995: The seasonal cycle over the tropical Pacific in coupled ocean–atmosphere general circulation models. Mon. Wea. Rev., 123 , 2825–2838.
Meehl, G. A., 1987: The annual cycle and interannual variability in the tropical Pacific and Indian Ocean regions. Mon. Wea. Rev., 115 , 27–50.
Meehl, G. A., C. Covey, T. Delworth, M. Latif, B. McAvaney, J. F. B. Mitchell, R. J. Stouffer, and K. E. Taylor, 2007: The WCRP CMIP3 multimodel dataset: A new era in climate change research. Bull. Amer. Meteor. Soc., 88 , 1383–1394.
Power, S., T. Casey, C. Folland, A. Colman, and V. Mehta, 1999: Inter-decadal modulation of the impact of ENSO on Australia. Climate Dyn., 15 , 319–324.
Randall, D. A., and Coauthors, 2007: Climate models and their evaluation. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 589–662.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108 , 4407. doi:10.1029/2002JD002670.
Salinger, M. J., J. A. Renwick, and A. B. Mullen, 2001: Interdecadal Pacific oscillation and South Pacific climate. Int. J. Climatol., 21 , 1705–1721.
Streten, N. A., 1973: Some characteristics of satellite-observed bands of persistent cloudiness over the Southern Hemisphere. Mon. Wea. Rev., 101 , 486–495.
Streten, N. A., and J. W. Zillman, 1984: Climate of the South Pacific Ocean. Climates of the Oceans, H. van Loon, Ed., World Survey of Climatology, Vol. 15, Elsevier, 263–429.
Takahashi, K., and D. S. Battisti, 2007a: Processes controlling the mean tropical Pacific precipitation pattern. Part I: The Andes and the eastern Pacific ITCZ. J. Climate, 20 , 3434–3450.
Takahashi, K., and D. S. Battisti, 2007b: Processes controlling the mean tropical Pacific precipitation pattern. Part II: The SPCZ and the southeast Pacific dry zone. J. Climate, 20 , 5696–5706.
Taylor, K. E., 2001: Summarizing multiple aspects of model performance in a single diagram. J. Geophys. Res., 106 , (D7). 7183–7192.
Trenberth, K., 1976: Spatial and temporal variations of the Southern Oscillation. Quart. J. Roy. Meteor. Soc., 102 , 639–653.
van Loon, H., and D. J. Shea, 1985: The Southern Oscillation. Part IV: The precursors south of 15°S to the extremes of the oscillation. Mon. Wea. Rev., 113 , 2063–2074.
van Loon, H., and D. J. Shea, 1987: The Southern Oscillation. Part VI: Anomalies of sea level pressure on the Southern Hemisphere and of Pacific sea surface temperature during the development of a warm event. Mon. Wea. Rev., 115 , 370–379.
Vincent, D. G., 1994: The South Pacific convergence zone (SPCZ): A review. Mon. Wea. Rev., 122 , 1949–1970.
Vincent, E. M., M. Lengaigne, C. E. Menkes, N. C. Jourdain, P. Marchesiello, and G. Madec, 2011: Interannual variability of the South Pacific convergence zone and implications for tropical cyclone genesis. Climate Dyn., in press, doi:10.1007/s00382-009-0716-3.
von Storch, H., H. van Loon, and G. N. Kiladis, 1988: The Southern Oscillation. Part VIII: Model sensitivity to SST anomalies in the tropical and subtropical regions of the South Pacific convergence zone. J. Climate, 1 , 325–331.
Whetton, P., I. Macadam, J. Bathols, and J. O’Grady, 2007: Assessment of the use of current climate patterns to evaluate regional enhanced greenhouse response patterns of climate models. Geophys. Res. Lett., 34 , L14701. doi:10.1029/2007GL030025.
Widlansky, M. J., P. J. Webster, and C. D. Hoyos, 2011: On the location and orientation of the South Pacific convergence zone. Climate Dyn., in press, doi:10.1007/s00382-010-0871-6.
Xie, P., and P. A. Arkin, 1997: Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Amer. Meteor. Soc., 78 , 2539–2558.
Yin, X., A. Gruber, and P. A. Arkin, 2004: Comparison of the GPCP and CMAP merged gauge–satellite monthly precipitation products for the period 1979–2001. J. Hydrometeor., 5 , 1207–1222.
Zhang, X., W. Lin, and M. Zhang, 2007: Toward understanding the double intertropical convergence zone pathology in coupled ocean-atmosphere general circulation models. J. Geophys. Res., 112 , D12102. doi:10.1029/2006JD007878.
DJF seasonal average of (a) CMAP precipitation, (b) NOAA OLR, (c) NCEP–NCAR near-surface wind and divergence, and (d) NCEP–NCAR omega at 500 hPa. Seasonal averages are calculated over the period 1979–99 for precipitation and OLR and 1950–99 for divergence and omega. The SPCZ region is indicated in (a). The SPCZ positions based on linear fit over the longitude range 155°E–140°W to (a) maximum precipitation, (b) minimum OLR, (c) minimum divergence, and (d) minimum omega at 500 hPa are shown as dashed lines.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
DJF seasonal average of (a) CMAP precipitation, (b) NOAA OLR, (c) NCEP–NCAR near-surface wind and divergence, and (d) NCEP–NCAR omega at 500 hPa. Seasonal averages are calculated over the period 1979–99 for precipitation and OLR and 1950–99 for divergence and omega. The SPCZ region is indicated in (a). The SPCZ positions based on linear fit over the longitude range 155°E–140°W to (a) maximum precipitation, (b) minimum OLR, (c) minimum divergence, and (d) minimum omega at 500 hPa are shown as dashed lines.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
DJF seasonal average of (a) CMAP precipitation, (b) NOAA OLR, (c) NCEP–NCAR near-surface wind and divergence, and (d) NCEP–NCAR omega at 500 hPa. Seasonal averages are calculated over the period 1979–99 for precipitation and OLR and 1950–99 for divergence and omega. The SPCZ region is indicated in (a). The SPCZ positions based on linear fit over the longitude range 155°E–140°W to (a) maximum precipitation, (b) minimum OLR, (c) minimum divergence, and (d) minimum omega at 500 hPa are shown as dashed lines.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
DJF seasonal average precipitation from (top) (left) CMAP, (middle) flux-adjusted model mean, (right) nonflux-adjusted model mean; and (remaining panels) CMIP3 models for years 1979–99.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
DJF seasonal average precipitation from (top) (left) CMAP, (middle) flux-adjusted model mean, (right) nonflux-adjusted model mean; and (remaining panels) CMIP3 models for years 1979–99.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
DJF seasonal average precipitation from (top) (left) CMAP, (middle) flux-adjusted model mean, (right) nonflux-adjusted model mean; and (remaining panels) CMIP3 models for years 1979–99.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Zonal average DJF precipitation (155°E–140°W) from CMAP and CMIP3 models for years 1979–99. The flux-adjusted model mean is shown as the solid thick red line, the nonflux adjusted model mean is shown as the solid thick black line, and CMAP is shown as the dashed thick purple line. The equator is indicated with the dashed vertical line.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Zonal average DJF precipitation (155°E–140°W) from CMAP and CMIP3 models for years 1979–99. The flux-adjusted model mean is shown as the solid thick red line, the nonflux adjusted model mean is shown as the solid thick black line, and CMAP is shown as the dashed thick purple line. The equator is indicated with the dashed vertical line.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Zonal average DJF precipitation (155°E–140°W) from CMAP and CMIP3 models for years 1979–99. The flux-adjusted model mean is shown as the solid thick red line, the nonflux adjusted model mean is shown as the solid thick black line, and CMAP is shown as the dashed thick purple line. The equator is indicated with the dashed vertical line.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Seasonal cycle of precipitation averaged over the longitude range 155°E–140°W from CMAP, GPCP, flux-adjusted model mean, nonflux-adjusted model mean, and CMIP3 models for years 1979–99.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Seasonal cycle of precipitation averaged over the longitude range 155°E–140°W from CMAP, GPCP, flux-adjusted model mean, nonflux-adjusted model mean, and CMIP3 models for years 1979–99.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Seasonal cycle of precipitation averaged over the longitude range 155°E–140°W from CMAP, GPCP, flux-adjusted model mean, nonflux-adjusted model mean, and CMIP3 models for years 1979–99.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
As in Fig. 2, but with SPCZ line (red) fitted to points of maximum precipitation in the range 155°E–140°W. CMAP SPCZ is the black dashed line. The slope of the SPCZ line (s) and the mean latitude (lat) are given at upper right of each plot.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
As in Fig. 2, but with SPCZ line (red) fitted to points of maximum precipitation in the range 155°E–140°W. CMAP SPCZ is the black dashed line. The slope of the SPCZ line (s) and the mean latitude (lat) are given at upper right of each plot.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
As in Fig. 2, but with SPCZ line (red) fitted to points of maximum precipitation in the range 155°E–140°W. CMAP SPCZ is the black dashed line. The slope of the SPCZ line (s) and the mean latitude (lat) are given at upper right of each plot.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
The SPCZ mean latitude vs slope for CMIP3 models (A–X), CMAP (filled circle), and GPCP (star). The flux-adjusted model mean is shown as a cross, while the nonflux-adjusted model mean (no symbol) is collocated with model “H.” Model letters are given in Table 1. SPCZ lines were calculated over the range 155°E–140°W from DJF seasonal mean precipitation (1979–99) as shown in Fig. 5.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
The SPCZ mean latitude vs slope for CMIP3 models (A–X), CMAP (filled circle), and GPCP (star). The flux-adjusted model mean is shown as a cross, while the nonflux-adjusted model mean (no symbol) is collocated with model “H.” Model letters are given in Table 1. SPCZ lines were calculated over the range 155°E–140°W from DJF seasonal mean precipitation (1979–99) as shown in Fig. 5.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
The SPCZ mean latitude vs slope for CMIP3 models (A–X), CMAP (filled circle), and GPCP (star). The flux-adjusted model mean is shown as a cross, while the nonflux-adjusted model mean (no symbol) is collocated with model “H.” Model letters are given in Table 1. SPCZ lines were calculated over the range 155°E–140°W from DJF seasonal mean precipitation (1979–99) as shown in Fig. 5.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Taylor diagrams of SPCZ region precipitation for (left) DJF and (right) JJA for years 1979–99 calculated with reference to CMAP precipitation. Model letters are given in Table 1. CMAP is shown as a filled semicircle, GPCP as a star, the flux-adjusted model mean as a cross, and the nonflux-adjusted model mean as a triangle.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Taylor diagrams of SPCZ region precipitation for (left) DJF and (right) JJA for years 1979–99 calculated with reference to CMAP precipitation. Model letters are given in Table 1. CMAP is shown as a filled semicircle, GPCP as a star, the flux-adjusted model mean as a cross, and the nonflux-adjusted model mean as a triangle.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Taylor diagrams of SPCZ region precipitation for (left) DJF and (right) JJA for years 1979–99 calculated with reference to CMAP precipitation. Model letters are given in Table 1. CMAP is shown as a filled semicircle, GPCP as a star, the flux-adjusted model mean as a cross, and the nonflux-adjusted model mean as a triangle.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
The DJF mean latitude of SPCZ line (155°E–140°W) vs Niño-3.4 SST for CMAP (years 1979–99) shown as filled circles and CMIP3 models (years 1950–99) shown as crosses.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
The DJF mean latitude of SPCZ line (155°E–140°W) vs Niño-3.4 SST for CMAP (years 1979–99) shown as filled circles and CMIP3 models (years 1950–99) shown as crosses.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
The DJF mean latitude of SPCZ line (155°E–140°W) vs Niño-3.4 SST for CMAP (years 1979–99) shown as filled circles and CMIP3 models (years 1950–99) shown as crosses.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
As in Fig. 8, but for DJF slope of SPCZ line (155°E–140°W) vs Niño-3.4 SST.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
As in Fig. 8, but for DJF slope of SPCZ line (155°E–140°W) vs Niño-3.4 SST.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
As in Fig. 8, but for DJF slope of SPCZ line (155°E–140°W) vs Niño-3.4 SST.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Precipitation in the SPCZ region in DJF from CMAP, GPCP, and CMIP3 models for El Niño years (Niño-3.4 > 0.75 standard deviations), all years, and La Niña years (Niño-3.4 < −0.75 standard deviations). The results are similar for lower and higher ENSO thresholds. The period 1979–99 is used for CMAP and GPCP, and the period 1950–99 is used for models.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Precipitation in the SPCZ region in DJF from CMAP, GPCP, and CMIP3 models for El Niño years (Niño-3.4 > 0.75 standard deviations), all years, and La Niña years (Niño-3.4 < −0.75 standard deviations). The results are similar for lower and higher ENSO thresholds. The period 1979–99 is used for CMAP and GPCP, and the period 1950–99 is used for models.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Precipitation in the SPCZ region in DJF from CMAP, GPCP, and CMIP3 models for El Niño years (Niño-3.4 > 0.75 standard deviations), all years, and La Niña years (Niño-3.4 < −0.75 standard deviations). The results are similar for lower and higher ENSO thresholds. The period 1979–99 is used for CMAP and GPCP, and the period 1950–99 is used for models.
Citation: Journal of Climate 24, 6; 10.1175/2010JCLI3942.1
Model name (CMIP3 ID), model host institute and country, atmosphere model resolution, and ocean model resolution. Horizontal resolution is expressed as degrees latitude by longitude or spectral truncation (T). Vertical resolution (L) is the number of levels.
Slope and mean latitude of SPCZ line calculated from CMAP, GPCP, and CMIP3 model precipitation; spatial correlation between models and CMAP; and temporal correlation between model SPCZ latitude and Niño-3.4 SST anomaly. All calculations use DJF seasonal mean precipitation for years 1979–99, except model temporal correlations with Niño-3.4, calculated for years 1950–1999. All correlations are significant at the 95% level according to a two-tailed Student’s t test, except those indicated by asterisk (*).
