Abstract

The eastern Great Basin (GB) in the western United States is strongly affected by droughts that influence water management decisions. Precipitation that falls in the GB, particularly in the Great Salt Lake (GSL) basin encompassed by the GB, provides water for millions of people living along the Wasatch Front Range. Western U.S. precipitation is known to be influenced by El Niño–Southern Oscillation (ENSO) as well as the Pacific decadal oscillation (PDO) in the North Pacific. Historical connectivity between GB precipitation and Pacific Ocean sea surface temperatures (SSTs) on interannual to multidecadal time scales is evaluated for 20 models that participated in phase 5 of the Coupled Model Intercomparison Project (CMIP5). While the majority of the models had realistic ENSO and PDO spatial patterns in the SSTs, the simulated influence of these two modes on GB precipitation tended to be too strong for ENSO and too weak for PDO. Few models captured the connectivity at a quasi-decadal period influenced by the transition phase of the Pacific quasi-decadal oscillation (QDO; a recently identified climate mode that influences GB precipitation). Some of the discrepancies appear to stem from models not capturing the observed tendency for the PDO to modulate the sign of the ENSO–GB precipitation teleconnection. Of all of the models, CCSM4 most consistently captured observed connections between Pacific SST variability and GB precipitation on the examined time scales.

1. Introduction

Precipitation variability in present and future climate scenarios has important impacts on local and regional hydrology, ultimately influencing water availability. Year-to-year fluctuations in precipitation, especially winter snowfall in areas with complex mountainous terrain, create challenges for water managers. The Great Basin (GB) watershed is located in the interior western United States (comprising parts of Utah, Wyoming, Idaho, Oregon, Nevada, and California) and is made up of many smaller snowpack-dominated watersheds, including that of the Great Salt Lake (GSL) in northern Utah. The GSL basin, which encompasses the Wasatch Range, is subjected to substantial interannual and multidecadal precipitation variability (Ropelewski and Halpert 1986; Wang et al. 2010, 2012). Orographic precipitation that occurs in the Wasatch Range is stored as snowpack and then delivered as runoff to the Wasatch Front Range, where more than two million people live and work.

Paleoclimate records indicate that, over most of the past millennium, droughts in the GB were generally more intense and lasted longer than those experienced in the twentieth century, which included the “severe droughts” of the 1930s and 1950s. Studies of tree ring–based drought analysis (Cook et al. 1997; Herweijer et al. 2007) have found that drought frequency has shifted from being centennial and more intense in the early millennium (AD 1000–1400) into being multidecadal and less intense in the late millennium (AD 1800–2000). To what extent such a cyclic feature may change or persist into the future is important information for water management.

The knowledge of precipitation drivers in the GSL basin, which is the largest watershed in the eastern GB, is an important tool for local water managers to guide water resource and supply and infrastructure engineering in preparation for drought or flooding. The GSL basin is situated in the transition boundary of the winter weather pattern that is forced by El Niño–Southern Oscillation (ENSO), known as the ENSO dipole or North American dipole. This positioning means that ENSO has both positive and negative effects on the precipitation received in the basin (Wise 2010; Wang et al. 2010) depending on the phase of the Pacific decadal oscillation (PDO) and possibly also the Atlantic multidecadal oscillation. For much of the twentieth century, El Niño was generally associated with a wet, cool southwestern and a dry, warm northwestern United States, while La Niña was associated with the opposite (Ropelewski and Halpert 1986; Dettinger et al. 1998). It is well known that the PDO, defined as the leading principal component of monthly north-central Pacific SST variability (poleward of 20°N; Mantua et al. 1997), has a significant effect on western U.S. precipitation. The PDO is linked to, and can modulate, the phasing of ENSO, and together they can result in prominent precipitation anomalies in the western United States (Gershunov and Barnett 1998; Gershunov et al. 1999; Mauget 2003). The PDO–ENSO coupling results in a shift of the typical dipole western U.S. precipitation pattern that is related to ENSO (Wise 2010; Brown 2011). Because the GB is situated in the transition of the ENSO dipole, which shifts its wet/dry influences based on the phase of the PDO, it is important to determine how the warming climate will affect the PDO–ENSO teleconnection and the precipitation received in the basin.

In addition to ENSO and the PDO, previous studies have revealed a predominant quasi-decadal oscillation (QDO) associated with precipitation and surface runoff in the GSL basin by examining the GSL surface elevation (Wang et al. 2010). The GSL surface elevation is significantly coherent with the Pacific QDO, defined by anomalous sea surface temperatures (SSTs) in the Niño-4 region (5°S–5°N, 160°E–150°W), which means that precipitation in the GSL basin must be phase-shifted from the Pacific QDO by a quarter phase in order to generate the coherent phase between the Niño-4 and the GSL surface elevation; this creates a lag of about three years between the Niño-4 and the precipitation (and between the precipitation and the GSL surface elevation) (Wang et al. 2010). During this transition phase of the Pacific QDO, a short-wave train pattern forms in the western tropical Pacific and emanates toward North America (Wang et al. 2011). This standing short-wave train is associated with enhanced precipitation in the Intermountain West at a quarter-phase lagged response to the Pacific QDO (i.e., a direct response to the transition-phase QDO anomalies of SSTs; Wang et al. 2011).

There has been an increased interest in the effects that climate change may have on GB hydrology (e.g., Bardsley et al. 2013; Mensing et al. 2013). Climate change assessments rely heavily on climate models. The latest generation of global climate models (GCMs) participating in phase 5 of the Coupled Model Intercomparison Project (CMIP5) contributed to the recently published Intergovernmental Panel on Climate Change (IPCC) Assessment Report 5 (AR5; Stocker et al. 2013). In addition to the atmosphere–ocean global climate models (AOGCMs), this generation of GCMs includes what are known as Earth system models (ESMs), which are AOGCMs coupled with the carbon cycle fluxes between the ocean, atmosphere, and land surface (Taylor et al. 2012). Other advances with CMIP5 include more models, an increased number of experiments and outputs, and enhanced physics packages (Taylor et al. 2012).

On interannual time scales, precipitation variability in CMIP5 models has been shown to be overenergetic compared to observed precipitation variability (Ault et al. 2012). In the western United States, the CMIP5 models tend to underestimate the precipitation variability on decadal to multidecadal time scales (Ault et al. 2012). Because the hydrology in the GB is driven both by the interannual variability (Dettinger et al. 1998) and by quasi-decadal variability (e.g., Wang et al. 2010, 2012), this study presents a CMIP5 model comparison focused on the interannual to multidecadal connections between historical Pacific Ocean SSTs and GB precipitation. Although the limitations of model ranking have been highlighted in recent research (e.g., Mote et al. 2011), identification of models that realistically capture oceanic modulation of GB precipitation over a range of time scales is important for 1) informing an objective weighting of climate projections and 2) selecting models to be used in dynamical downscaling and/or stochastic climate modeling, thus providing climate information usable for collaborators in fields including hydrology, biology, urban planning, and civil engineering. Moreover, because the projected precipitation changes over the GB differ between CMIP5 and the previous generation of models (CMIP3), as was shown in Brekke (2013), there is an urgent need to evaluate the CMIP5 models for the GB hydroclimate.

2. Data and methods

a. Data

The observational precipitation data used in this study are 1° gridded monthly global land surface precipitation data provided by the Global Precipitation Climatology Centre (GPCC), which span from January 1901 to December 2010 (Schneider et al. 2011). We use version 6 of these data in this study. The analysis domain chosen for this study follows Wang et al. (2010) and encompasses the eastern Great Basin in the western United States (37.5–42.5°N, 115°–110°W; Fig. 1). The eastern Great Basin contains the Wasatch Range, a major contributor of water flow in the Great Salt Lake basin. As noted in previous studies (e.g., Wang et al. 2010), even though the analysis domain covers more than just the eastern Great Basin, the low-frequency precipitation variability is consistent over the extent of this domain.

Fig. 1.

The portion of the Great Basin used in the study (37.5°–42.5°N, 115°–110°W). The filled circles indicate the points in the GPCC precipitation dataset that were spatially averaged for analysis. The inset map shows the location of region in the contiguous United States. The color bar indicates elevation in meters above mean sea level.

Fig. 1.

The portion of the Great Basin used in the study (37.5°–42.5°N, 115°–110°W). The filled circles indicate the points in the GPCC precipitation dataset that were spatially averaged for analysis. The inset map shows the location of region in the contiguous United States. The color bar indicates elevation in meters above mean sea level.

The observational sea surface temperature data used in this study (HadISST) are reconstructed 1° gridded monthly data based on in situ and satellite observations. The data span from January 1870 to June 2013 (Rayner et al. 2003) and were obtained from the Met Office Hadley Centre.

For inclusion in the study, CMIP5 models were required to have all-forcing historical precipitation output and sea surface temperature output, both back to at least 1900. Twenty models satisfied this criterion, and Table 1 provides the modeling center, number of ensemble members, and literature references for each. The observational data used in the analysis span 1901–2005 to align with the GPCC precipitation data starting in 1901, and the model data span 1900–2005. The year difference is negligible on account of the filtering described immediately below.

Table 1.

List of CMIP5 models used in this study, the number of ensemble members per model [denoted by No(s).], and their corresponding institutions. (Expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList.)

List of CMIP5 models used in this study, the number of ensemble members per model [denoted by No(s).], and their corresponding institutions. (Expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList.)
List of CMIP5 models used in this study, the number of ensemble members per model [denoted by No(s).], and their corresponding institutions. (Expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList.)

b. Methods

To concentrate on the time scales of interest (i.e., the interannual and quasi-decadal frequencies), we filtered the monthly precipitation and SST data three different ways using a Hamming window (HW) filter (Hamming 1998). We use monthly data by following the method described in Wang et al. (2010). In this way, all seasons are included in the analysis, but the seasonal cycle is filtered out. As discussed in Wang et al. (2010), this windowed filter is most appropriately applied to short-length time series because of its ability to eliminate unwanted frequencies (Iacobucci and Noullez 2005). We applied a 3–7-yr bandpass filter to the data to isolate the high-frequency variability associated with ENSO. We applied a 10–15-yr bandpass filter to obtain the quasi-decadal signal such as that found in Wang et al. (2010).

1) Principal component analysis

We applied a low-pass 7-yr filter on the data to analyze frequencies lower than ENSO, focusing primarily on the PDO. To define the PDO, we computed the leading empirical orthogonal function (EOF; Hannachi et al. 2007) of the detrended, filtered monthly sea surface temperature data. The EOF analysis was performed over the northern Pacific Ocean (20°–66°N, 160°E–110°W) and yielded the well-known east–west dipole characterizing the PDO (e.g., Mantua et al. 1997). We set the sign of the EOF to emulate the positive (warm phase) PDO pattern and regressed the sea surface temperatures onto the associated principal component time series, which gives units of kelvin per standard deviation at each location (as opposed to the arbitrary scaling of the EOF; e.g., Thompson and Wallace 2000).

2) Correlating Great Basin precipitation with Pacific SSTs

In our results, we present maps of the correlation between filtered GB precipitation and SSTs on several time scales for each model. The map domain includes most of the Pacific Ocean (40°S–65°N, 90°E–120°W). For all correlation analyses and displayed maps, the model data were bilinearly interpolated onto the same 1° grid corresponding to the SST observations. To determine the degree to which the model correlation maps matched the observations, we calculated the area-weighted uncentered spatial correlation ():

 
formula

where is an area weighting function, indicates the observational data, and indicates the model data interpolated onto the same grid as the observational data for each location i (, ), where λ is longitude and ϕ is latitude. This statistic was applied by Alexander and Arblaster (2009), and a similar statistic (referred to as congruence) was used by Kiktev et al. (2007). This statistic ranges from −1 to 1, with values closer to 1 indicating that the model pattern is very similar to the observational pattern, values near zero indicating weaker pattern matching, and values close to −1 indicating a strong pattern match but with reversed sign. Where multiple ensemble members were available from a single model, the area-weighted uncentered spatial correlation was calculated for each ensemble member, and the member from each model that most closely matched the observational map is shown for discussion. The statistic does not account for precipitation bias, and we note that all models have a positive bias over the study domain as was found in Mehran et al. (2014). Whether taking the bias as a difference or ratio, the bias and the precipitation variance are positively correlated (, ). These biases are a concern because, if they change in a future climate, they would not simply cancel out in a calculation of climate change.

3) North Pacific streamfunction analysis

To analyze atmospheric teleconnections between Pacific SSTs and GB precipitation as done in Wang et al. (2010, 2012), we computed the geostrophic streamfunction (ψ) over the North Pacific and North America:

 
formula

where is the 200-hPa geopotential height at a given location (, ) and f is the Coriolis parameter . We used 200 hPa because it is the approximate location of the polar jet stream in this sector (Horel and Wallace 1981; Strong and Davis 2008). We obtained NCEP–NCAR Reanalysis 1 monthly mean geopotential height data from 1948 to 2005 (Kalnay et al. 1996) to complete the observational streamfunction analysis. We also calculated the streamfunction using monthly mean geopotential height data over the same time period for each CCSM4 ensemble member. The data were filtered with the HW bandpass 3–7-yr and 10–15-yr filter and then correlated with GB precipitation.

3. Results

a. Interannual (3–7 yr) relationship between Pacific SSTs and GB precipitation

Figure 2 shows the interannual (3–7 yr) contemporaneous correlation spatial patterns between SSTs and GB precipitation for each included CMIP5 model and observations. For observations (Fig. 2u), the contemporaneous correlation map of bandpass 3–7-yr filtered Pacific SSTs and Great Basin precipitation highlights higher-frequency Pacific Ocean variability dominated by ENSO in the tropics with variability also in the extratropics. In the CMIP5 ensemble, we show in Fig. 2 only the member from each model with the map that best matches that of observations based on the area-weighted uncentered spatial correlation coefficients [Eq. (1)]. As discussed in Deser et al. (2014), internal variability within each model differs, and this is evident in each model’s area-weighted uncentered spatial correlation coefficients (crosses, Fig. 3a). Most models, including CCSM4 and CESM1(CAM5.1, FV2), exhibit a stronger than observed correlation between the tropics and GB precipitation compared to observations (cf. all panels to Fig. 2u). Despite being excessively strong in the tropics, the majority of the spatial patterns in the mapped ensemble members do resemble observations [; Fig. 3a; recall Eq. (1)]. However, a few of the models have additional ensemble members that do not capture the observed pattern as well (Fig. 3a; crosses close to zero). In contrast, CCSM4 and CESM1(CAM5.1, FV2) are notable for consistency of among members. This consistency is supported by an analysis of a multicentury preindustrial control run obtained for CCSM4. The values computed between 10 near-century-long blocks of the simulation and observations are shown next to the historical values in Fig. 4. In the bandpass-filtered 3–7-yr data, the preindustrial control values are clustered as they are in the historical runs. The overly strong connection between GB precipitation and ENSO contributes to high variances in GB precipitation for several models with values similar to observations (e.g., MIROC5, CCSM4, GFDL CM3; Fig. 5a), and this is exacerbated in some cases by overly energetic Niño-4 SSTs (Fig. 5d).

Fig. 2.

Contemporaneous correlation maps between bandpass-filtered 3–7-yr SSTs and Great Basin precipitation for models (a)–(t) and observations (u). The model’s ensemble member with the map that most closely matches that of the observations is shown [the match is assessed using the area-weighted uncentered spatial correlation ; Eq. (1)].

Fig. 2.

Contemporaneous correlation maps between bandpass-filtered 3–7-yr SSTs and Great Basin precipitation for models (a)–(t) and observations (u). The model’s ensemble member with the map that most closely matches that of the observations is shown [the match is assessed using the area-weighted uncentered spatial correlation ; Eq. (1)].

Fig. 3.

Bar charts on the left show how well CMIP5 ensemble members capture observed spatial patterns related to Great Basin precipitation (P) and Pacific sea surface temperatures (SSTs), where the statistic used is the area-weighted uncentered spatial correlation [, Eq. (1)]. The compared spatial patterns are (a) correlation between P and SSTs for bandpass of 3–7 yr (i.e., maps in Fig. 2), (b) correlation between P and SSTs for bandpass of 10–15 yr (i.e., maps in Fig. 8), (c) the loading pattern (EOF) defining the Pacific decadal oscillation (maps not shown), and (d) correlation between P and SSTs for bandpass of 10–15 yr with SSTs leading P by 3 yr (i.e., maps in Fig. 9). Bar charts on the right show correlation coefficients for (e) P and Niño-4 SST for bandpass of 3–7 yr, (f) PDO and bandpass 10–15-yr P, (g) PDO and low-pass 7-yr P with PDO leading P by 2 yr, and (h) P and Niño-4 SST for bandpass of 10–15 yr with SST leading P by 3 yr based on the transition-phase teleconnection. Each row orders the models according to the values in the left panel of the row. The ensemble member closest to observations is indicated by the bar, and other available ensemble members are indicated by the cross symbol.

Fig. 3.

Bar charts on the left show how well CMIP5 ensemble members capture observed spatial patterns related to Great Basin precipitation (P) and Pacific sea surface temperatures (SSTs), where the statistic used is the area-weighted uncentered spatial correlation [, Eq. (1)]. The compared spatial patterns are (a) correlation between P and SSTs for bandpass of 3–7 yr (i.e., maps in Fig. 2), (b) correlation between P and SSTs for bandpass of 10–15 yr (i.e., maps in Fig. 8), (c) the loading pattern (EOF) defining the Pacific decadal oscillation (maps not shown), and (d) correlation between P and SSTs for bandpass of 10–15 yr with SSTs leading P by 3 yr (i.e., maps in Fig. 9). Bar charts on the right show correlation coefficients for (e) P and Niño-4 SST for bandpass of 3–7 yr, (f) PDO and bandpass 10–15-yr P, (g) PDO and low-pass 7-yr P with PDO leading P by 2 yr, and (h) P and Niño-4 SST for bandpass of 10–15 yr with SST leading P by 3 yr based on the transition-phase teleconnection. Each row orders the models according to the values in the left panel of the row. The ensemble member closest to observations is indicated by the bar, and other available ensemble members are indicated by the cross symbol.

Fig. 4.

Area weighted uncentered spatial correlations ( values) for all 6 CCSM4 historical runs (black crosses) and one 1050-yr preindustrial control (piControl) run divided into 10 near-century-long blocks (blue crosses) for the (a) bandpass 3–7-yr filtered data and (b) bandpass 10–15-yr filtered contemporaneous data.

Fig. 4.

Area weighted uncentered spatial correlations ( values) for all 6 CCSM4 historical runs (black crosses) and one 1050-yr preindustrial control (piControl) run divided into 10 near-century-long blocks (blue crosses) for the (a) bandpass 3–7-yr filtered data and (b) bandpass 10–15-yr filtered contemporaneous data.

Fig. 5.

Bar charts on the left show the Great Basin precipitation variance for (a) bandpass 3–7-yr filtered data, (b) bandpass 10–15-yr filtered data, and (c) low-pass 7-yr data. Bar charts on the right show the variance of (d) bandpass 3–7-yr Niño-4 SSTs, (e) the PDO index of the ensemble members in (b), and (f) the PDO index of the ensemble members in (c). As in Fig. 3, the ensemble member closest to observations is indicated by the bar, and other available ensemble members are indicated by the cross symbol. The order of the models here follows the first three rows in Fig. 3.

Fig. 5.

Bar charts on the left show the Great Basin precipitation variance for (a) bandpass 3–7-yr filtered data, (b) bandpass 10–15-yr filtered data, and (c) low-pass 7-yr data. Bar charts on the right show the variance of (d) bandpass 3–7-yr Niño-4 SSTs, (e) the PDO index of the ensemble members in (b), and (f) the PDO index of the ensemble members in (c). As in Fig. 3, the ensemble member closest to observations is indicated by the bar, and other available ensemble members are indicated by the cross symbol. The order of the models here follows the first three rows in Fig. 3.

The influences of ENSO on western U.S. precipitation have been studied extensively (e.g., Ropelewski and Halpert 1986, 1989; Halpert and Ropelewski 1992; Hidalgo and Dracup 2003; Mo and Schemm 2008). The Great Basin is situated between the precipitation dipole anomalies associated with ENSO in the western United States. The relatively weak correlations in the observations between SSTs along the equatorial Pacific and GB precipitation (Fig. 2u) support the variable association between ENSO and GB precipitation due to the shifts of the dipole cancelling out the ENSO signal. Nearly all of the models capture the connection between the ENSO region and GB precipitation, but some of them [e.g., CCSM4 and CESM1(CAM5.1, FV2)] have a stronger correlation between ENSO SSTs and GB precipitation than what is observed.

There has been a growing interest in understanding the linkages between SST variability in the North Pacific and western U.S. precipitation (e.g., Latif and Barnett 1994; Kushnir et al. 2002). The midlatitude storm track is known to be affected by sea surface temperatures in the North Pacific (e.g., Namias et al. 1988; Peng and Whitaker 1999; Frankignoul et al. 2011), and we selected CCSM4 to analyze this more closely in part because of its tightly packed set of values across its ensemble members (crosses, Fig. 3a). We calculated the 200-hPa streamfunction, smoothed the data with the HW bandpass filter to focus on the 3–7-yr and 10–15-yr variability, and correlated it with GB precipitation to visualize associated dynamics in the North Pacific (Fig. 6). The bandpass-filtered 3–7-yr observational and averaged CCSM4 data have a negative correlation over the western United States, indicating the overall trough anomalies observed in conjunction with precipitation in the Great Basin (Figs. 6a,b). They are also in agreement about negative streamfunction anomalies over the Bering Sea. However, the individual CCSM4 ensemble members exhibit notably variable streamfunction correlation patterns upstream of the western United States (Figs. 6c–h). The variable streamfunction correlation patterns upstream from the GB in CCSM4 indicate that, even for a model with strong ENSO–GB precipitation correlation, the underlying storm track dynamics can vary markedly.

Fig. 6.

Correlation maps between the (a)–(h) bandpass 3–7-yr filtered and (q)–(x) bandpass 10–15-yr filtered streamfunction and Great Basin precipitation for the observations, the average of all six ensemble members of CCSM4, and each individual ensemble member. (i)–(p) The streamfunction for the bandpass 3–7-yr filtered data for only the years of positive PDO.

Fig. 6.

Correlation maps between the (a)–(h) bandpass 3–7-yr filtered and (q)–(x) bandpass 10–15-yr filtered streamfunction and Great Basin precipitation for the observations, the average of all six ensemble members of CCSM4, and each individual ensemble member. (i)–(p) The streamfunction for the bandpass 3–7-yr filtered data for only the years of positive PDO.

Links between ENSO and the PDO

As noted in the introduction, there is evidence that the PDO modulates ENSO-driven precipitation anomalies over the western United States. To analyze this effect in CMIP5, we computed the 10-yr running correlation between the bandpass-filtered 3–7-yr spatially averaged Niño-4 SSTs and Great Basin precipitation for all models and the observations to depict decadal fluctuations in the ENSO teleconnection. The time series of the running ENSO–GB precipitation correlation () was then compared to the PDO time series (the observational time series are shown in Fig. 7a), and a correlation was calculated between the PDO and to determine the effect the PDO has on the connection between Niño-4 SSTs and GB precipitation. This correlation is shown for observations and each model in Fig. 7b. The observational data show a strong positive running correlation () between Niño-4 SSTs and GB precipitation when the PDO is positive. The correlation between the PDO and is strongly positive because of this tendency (observations bar; Fig. 7b), indicating that El Niño is conducive to precipitation in the Great Basin during a positive (warm phase) PDO. When the PDO is negative (cool phase), La Niña is conducive to precipitation in the Great Basin. In short, the PDO alters the sign of the ENSO–GB precipitation teleconnection.

Fig. 7.

(a) For observations, the blue curve shows the PDO index and the green curve shows the 10-yr running correlation between bandpass 3–7-yr filtered Niño-4 SSTs and Great Basin precipitation (). (b) Correlation between the PDO index and for observations and CMIP5 models. The model order follows that of Figs. 3a,e.

Fig. 7.

(a) For observations, the blue curve shows the PDO index and the green curve shows the 10-yr running correlation between bandpass 3–7-yr filtered Niño-4 SSTs and Great Basin precipitation (). (b) Correlation between the PDO index and for observations and CMIP5 models. The model order follows that of Figs. 3a,e.

To visualize the underlying dynamics, we calculated the correlation coefficients between the 200-hPa streamfunction and GB precipitation for the years of positive PDO (and, consequently, high running correlation coefficients between Niño-4 SSTs and GB precipitation) (Fig. 6i). The trimodal pattern over the Pacific follows that described in Gill (1980) and Alexander et al. (2002) and reflects the canonical atmospheric response to El Niño (note that Fig. 6 only shows the North Pacific portion of the pattern). However, the composite map of all six CCSM4 ensemble members is much weaker than that of the observational map (Fig. 6j), and there is substantial variability among the individual members (Figs. 6k–p).

In general, the models do not capture the observed positive correlation between the PDO and (Fig. 7b). There are some models that have positive correlations between the PDO and , although they are not as strong as observed (Fig. 7b). In CCSM4, is positive for nearly the entire time period (not shown). In this case, positive PDO weakens the ENSO–precipitation correlation but does not render it negative, resulting in the much stronger overall correlation coefficient between Niño-4 SSTs and GB precipitation for CCSM4 (Fig. 3e). Because positive PDO weakens in CCSM4, its correlation coefficient between the PDO index and is negative (Fig. 7b). In BCC_CSM1.1, the PDO has a strong positive correlation with (Fig. 7b). BCC_CSM1.1 also has a correlation coefficient between Niño-4 SSTs and GB precipitation more comparable to that of observations (Fig. 3e). However, its spatial pattern does not match well with observations according to the ru statistic (Fig. 3a). By visual inspection, the BCC_CSM1.1 map in Fig. 2b captures the majority of the observational pattern (Fig. 2u), including the trimodal pattern in the North Pacific. This contrast between the value and visual inspection of the map illustrates limitations in using the area-weighted uncentered spatial correlation as a single summary statistic. A model may display a pattern that is similar to observations, but a spatial shift in the pattern can reduce the value significantly.

b. Quasi-decadal relationship between Pacific SSTs and GB precipitation

The correlation maps between contemporaneous bandpass 10–15-yr filtered SST data and GB precipitation data are generally similar in spatial pattern and strength between the models, although most of them have too-strong correlations in the tropical Pacific compared to observations (Fig. 8). However, a number of the models have relatively high area-weighted uncentered spatial correlations, indicating good overall pattern matches (Fig. 3b). While model agreement with observations seems good at a low frequency, the values decrease more rapidly in order from left to right compared to the higher-frequency mode and even become negative; this suggests a further challenge for models to capture the decadal-scale variability as observed. The various ensemble members of each model also performed variably, with an increased number of them having realizations opposite of that seen in observations (crosses below zero, Fig. 3b). These results suggest that simulated decadal-scale variability is largely internal and stochastic (Guilyardi et al. 2009), and this is further reflected in the large spread of pattern matches at the 10–15-yr time scale in Fig. 4b. This result is also consistent with the finding that SST-related teleconnections may not be reliable predictors of GB precipitation (McAfee 2014; Wise et al. 2015). Finally, the relative lack of agreement in Fig. 3b may reflect the documented challenges associated with simulating Pacific decadal variability (Furtado et al. 2011; Ault et al. 2012).

Fig. 8.

Contemporaneous correlation maps between bandpass 10–15-yr filtered SSTs and Great Basin precipitation for (a)–(t) models and (u) observations. The model’s ensemble member with the map that most closely matches that of the observations is shown.

Fig. 8.

Contemporaneous correlation maps between bandpass 10–15-yr filtered SSTs and Great Basin precipitation for (a)–(t) models and (u) observations. The model’s ensemble member with the map that most closely matches that of the observations is shown.

As with the high-frequency data, we calculated the 200-hPa streamfunction from the bandpass 10–15-yr filtered 200-hPa geopotential height data and correlated it with bandpass-filtered GB precipitation (Fig. 6q). In observations, the negative ψ anomaly over the western United States, which has shifted slightly farther north than the bandpass 3–7-yr map, extends zonally out into the North Pacific. The composite of the CCSM4 ensemble members captures the ψ anomaly centered just off the west coast of North America, but the westward extension of the negative ψ anomaly over the North Pacific is not apparent, reflecting marked variability in the upstream pattern among the ensemble members (Fig. 6r). For example, members such as r2 (Fig. 6t) indicate that the CCSM4 model physics capture the observed zonal upstream pattern given some specific initial conditions, but the diversity of patterns in other members suggests strong internal system variability that can affect the teleconnection pattern and GB precipitation (Deser et al. 2012, 2014).

Connectivity with the PDO

The correlation maps in Fig. 8 between bandpass 10–15-yr filtered SSTs and GB precipitation depict a PDO-like pattern in the North Pacific. To examine further how models simulate the PDO modulation of the ENSO teleconnection, we assess PDO spatial patterns and the correlation between PDO and GB precipitation. We calculated the PDO index for each model and observations following the approach described in section 2a. For each model, we chose the ensemble member that had the highest value; more than half of the models have ensemble members with values at or above 0.8 (Fig. 3c), indicating a good performance in the depiction of PDO. Interestingly, many ensemble members have maps with lower values, with some nearing zero (crosses, Fig. 3c). Contemporaneously, the correlation between the PDO index and bandpass 10–15-yr filtered GB precipitation is highly variable across the ensemble members (Fig. 3f). CCSM4, which has the best spatial pattern match (higher , Fig. 3b), also has the closest correlation coefficient to observations, although the variance of the PDO and precipitation in that ensemble member is much higher than observations (Figs. 5b,e).

The correlation between the PDO index and GB precipitation is maximized at a lag of about two years for observations (Fig. 3g). Only some models capture this lagged correlation (Fig. 3g) even though most can generate realistic PDO spatial patterns (Fig. 3c). MIROC5, HadGEM2-ES, and CCSM4, the models with the highest values in Fig. 3c, have correlations much lower than that of observations at a 2-yr lag. However, some of the other models with values above 0.8 have correlations near that of observations. Some of the models with the lower correlation coefficients in Fig. 3g have PDO index variances that are much higher than the models with the higher correlation coefficients (cf. Figs. 3g and 5f). Visually inspecting the PDO time series shows that the models with the near-zero correlation coefficients and higher variance [i.e., CCSM4, NorESM1-M, and CESM1(CAM5)] have more extreme (high amplitude) PDO fluctuations than found in observations and in models with higher correlation coefficients and lower variance (i.e., MPI-ESM-LR). Additionally, the variance in the low-pass 7-yr filtered GB precipitation (Fig. 5c) is higher in the models than in observations, and it tends to decrease in models with more realistic PDO spatial patterns (, Fig. 3c).

c. Lagged relationship between Pacific QDO and GB precipitation

As noted in the introduction, Wang et al. (2011) described the dynamical connection between the transition phase coupling of the Pacific QDO and GB precipitation. An atmospheric wave train sets up from the western tropical Pacific, where SSTs fluctuate more strongly than in the Niño-4 region, in an arc toward western North America. To show the model representation of this transition-phase teleconnection, we used the bandpass 10–15-yr filtered data to compute correlations with GB precipitation lagging Pacific SSTs by three years, as shown in Fig. 9. The spatial patterns of these correlations are notably variable (recall that we show only the ensemble member best matching the observed pattern). The area-weighted uncentered spatial correlations decrease rapidly across the models and even become negative in some models, indicating only a weak presence of the QDO transition dynamics in CMIP5 (Fig. 3d). One ensemble member of CCSM4 has the highest pattern match in Fig. 3d (shaded bar), yet other members of CCSM4 have maps that are opposite in sign to observed patterns (crosses below zero, Fig. 3d). Generally, as the area-weighted uncentered spatial correlations decrease and become negative, the correlation between Niño-4 SSTs and GB precipitation also decreases and becomes negative (Fig. 3h), suggesting that poor simulations of the Pacific QDO lead to even poorer simulations of its teleconnection impact on GB precipitation. Among all models, CCSM4 stands out as it has some ensemble members that capture both the spatial pattern of this transition-phase teleconnection (Fig. 9c) as well as the contemporaneous relationship between Niño-4 SSTs (where the Pacific QDO manifests itself) and GB precipitation (Fig. 3h).

Fig. 9.

Lagged correlation maps between bandpass 10–15-yr filtered SSTs and Great Basin precipitation with precipitation lagging SSTs by 3 years for (a)–(t) models and (u) observations. The model’s ensemble member with the map that most closely matches that of the observations is shown.

Fig. 9.

Lagged correlation maps between bandpass 10–15-yr filtered SSTs and Great Basin precipitation with precipitation lagging SSTs by 3 years for (a)–(t) models and (u) observations. The model’s ensemble member with the map that most closely matches that of the observations is shown.

4. Summary and discussion

This study focuses on determining the degree to which CMIP5 models are able to simulate observed interannual to multidecadal connections between Pacific SSTs and Great Basin precipitation. All of the models performed reasonably well on the interannual time scale as far as capturing ENSO; however, not all of the models produced the observed connection between Niño-4 SSTs and precipitation in the Great Basin. This discrepancy could be due to the phase of PDO present, as PDO can alter the effect of ENSO on Great Basin precipitation (recall Fig. 7b). Most models can generate a realistic PDO, but the highly variable correlations between the simulated PDO index and GB precipitation are likely due to the too-strong variance in the PDO. For some of the models, the excessively strong correlations between bandpass 3–7-yr SSTs and GB precipitation could be because the PDO did not appear to modulate ENSO at all. In contrast, the quasi-decadal connection (i.e., a dominant mode in the GB) proved to be more difficult for the models to capture. The lagged connection between Niño-4 SSTs and GB precipitation and therefore the transition phase of the Pacific QDO, as discussed in Wang et al. (2010) and Wang et al. (2011), proves to be a challenge for the models.

Of the 20 models evaluated, CCSM4 consistently produced at least one ensemble member that compared well with observations. CCSM4 had high values for the contemporaneous and 3-yr lagged bandpass 10–15-yr filtered data as well as the PDO. The only other model that simulated well all three variabilities more than once was MIROC5, which had an ensemble member with the highest spatial pattern match for the bandpass 3–7-yr filtered contemporaneous correlation and PDO correlation. The high performance of CCSM4 at every time scale [discussed in DeFlorio et al. (2013)] suggests that it might perform better than other models in the simulation of the Pacific SST–GB precipitation variability under future climate.

For the most part, models with more than one ensemble member performed more favorably. The models with only one ensemble member (ACCESS1.3, GFDL-ESM2G, HadGEM2-AO, HadGEM2-CC, and INM-CM4.0) had only one opportunity to capture the connection between Pacific SST variability and GB precipitation while models with more than one ensemble member had more opportunities to capture the connectivity. While this observation echoes other studies (e.g., Kirtman and Shukla 2002; He et al. 2014), the implication from this study is that the connectivity between Pacific SSTs and GB precipitation is highly sensitive to the initial conditions, including the state of the Pacific Ocean. Even slight changes in the initial conditions can result in vastly different outcomes (e.g., Deser et al. 2012, 2014). However, it is important to note that after completing an analysis of a multicentury preindustrial control run from CCSM4, we found that the values of the near-century blocks of the preindustrial control run are similar to the values of the historical runs, suggesting consistency in performance at least in CCSM4.

Finally, models struggled to capture the impact of the PDO on GB precipitation. This may result from the periodicity of PDO being incorrectly simulated or because the teleconnection associated with the PDO transition, which is much weaker in SST anomalies in the tropics, did not achieve the observed pattern. However, using a 2000-yr simulation by GFDL CM2.1, Wang et al. (2012) showed that long control runs are able to depict the transition-phase effect of the PDO on GB precipitation. This aspect echoes the previous argument of models capturing the full spectrum of internal variability, but it nonetheless requires further investigation.

A goal of this study was to narrow down the field of available CMIP5 models in order to determine the “best” model to use for future studies of precipitation in the GB as well as perform dynamical downscaling analyses on a local scale. Ranking of models has been discussed in previous studies (e.g., Mote et al. 2011), and in some cases, choosing one model over another does not lead to significantly different results. A preliminary analysis of future CMIP5 model precipitation output over the western United States yielded quite different outcomes between the top performing models; however, when taking an average of the top models and comparing it to the average of all models, the areas of increased and decreased precipitation did not differ as strongly. In addition, ranking models based on their performance on these two time scales (one that captures ENSO-like variability and one that captures PDO-like variability) is a challenge because we do not fully understand the processes driving these modes of variability. Models that are able to successfully capture the connection between ENSO variability and GB precipitation in the past may not necessarily be the best under climate change (Guilyardi et al. 2009).

This study evaluated 20 of the coupled atmosphere–ocean global climate models that participated in CMIP5 and their ability to capture the historical connections between GB precipitation and the major modes of variability in the Pacific Ocean. By applying a statistical analysis including a correlation map pattern match, we determined which models performed more favorably at the different frequencies. We also evaluated how successful the models are at coupling the frequencies which result in precipitation in the GB. The results presented here may have relevance extending beyond the Great Basin because of the spatial scale of the relevant teleconnections, and the method can also be applied for more specific regional analyses. While considering model ranking and the limitations of ranking, we are able to make decisions on how to best use the models for future analysis of precipitation, including using the outputs as inputs for dynamical downscaling on a small scale.

Acknowledgments

This material is based upon work supported by the National Science Foundation under Grants EPS-1135482, EPS-1135483, and EPS-1208732. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. Support for S. Wang from Grants NNX13AC37G and WaterSMART R13AC80039 and from the Utah Agricultural Experiment Station is appreciated. Additional support for K. Smith was provided by the University of Utah, including a fellowship from the Global Change and Sustainability Center. Provision of computer infrastructure by the Center for High Performance Computing at the University of Utah is gratefully acknowledged.

REFERENCES

REFERENCES
Alexander
,
L. V.
, and
J. M.
Arblaster
,
2009
:
Assessing trends in observed and modelled climate extremes over Australia in relation to future projections
.
Int. J. Climatol.
,
29
,
417
435
, doi:.
Alexander
,
M. A.
,
I.
Bladé
,
M.
Newman
,
J. R.
Lanzante
,
N.-C.
Lau
, and
J. D.
Scott
,
2002
:
The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans
.
J. Climate
,
15
,
2205
2231
, doi:.
Ault
,
T. R.
,
J. E.
Cole
, and
S. St.
George
,
2012
: The amplitude of decadal to multidecadal variability in precipitation simulated by state-of-the-art climate models. Geophys. Res. Lett.,39, L21705, doi:.
Bardsley
,
T.
,
A.
Wood
,
M.
Hobbins
,
T.
Kirkham
,
L.
Briefer
,
J.
Niermeyer
, and
S.
Burian
,
2013
:
Planning for an uncertain future: Climate change sensitivity assessment toward adaptation planning for public water supply
.
Earth Interact.
,
17
, doi:.
Bentsen
,
M.
, and Coauthors
,
2013
:
The Norwegian Earth System Model, NorESM1-M—Part 1: Description and basic evaluation of the physical climate
.
Geosci. Model Dev.
,
6
,
687
720
, doi:.
Bi
,
D.
, and Coauthors
,
2013
:
The ACCESS coupled model: Description, control climate and evaluation
.
Aust. Meteor. Oceanogr. J.
,
63
,
41
64
.
Brekke
,
L.
,
2013
: New CMIP5 downscaled climate projections. 11th Annual Climate Prediction Applications Science Workshop, Logan, UT, NOAA. [Available online at http://gdo-dcp.ucllnl.org/downscaled_cmip_projections/techmemo/downscaled_climate.pdf.]
Brown
,
D. P.
,
2011
:
Winter circulation anomalies in the western United States associated with antecedent and decadal ENSO variability
.
Earth Interact.
,
15
, doi:.
Collins
,
M.
,
S. F. B.
Tett
, and
C.
Cooper
,
2001
:
The internal climate variability of HadCM3, a version of the Hadley Centre coupled model without flux adjustments
.
Climate Dyn.
,
17
,
61
81
, doi:.
Collins
,
W. J.
, and Coauthors
,
2011
:
Development and evaluation of an Earth-system model—HadGEM2
.
Geosci. Model Dev.
,
4
,
1051
1075
, doi:.
Cook
,
E. R.
,
D. M.
Meko
, and
C. W.
Stockton
,
1997
:
A new assessment of possible solar and lunar forcing of the bidecadal drought rhythm in the western United States
.
J. Climate
,
10
,
1343
1356
, doi:.
DeFlorio
,
M. J.
,
D. W.
Pierce
,
D. R.
Cayan
, and
A. J.
Miller
,
2013
:
Western U.S. extreme precipitation events and their relation to ENSO and PDO in CCSM4
.
J. Climate
,
26
,
4231
4243
, doi:.
Deser
,
C.
,
R.
Knutti
,
S.
Solomon
, and
A. S.
Phillips
,
2012
:
Communication of the role of natural variability in future North American climate
.
Nat. Climate Change
,
2
,
775
779
, doi:.
Deser
,
C.
,
A. S.
Phillips
,
M. A.
Alexander
, and
B. V.
Smoliak
,
2014
:
Projecting North American climate over the next 50 years: Uncertainty due to internal variability
.
J. Climate
,
27
,
2271
2296
, doi:.
Dettinger
,
M. D.
,
D. R.
Cayan
,
H. F.
Diaz
, and
D. M.
Meko
,
1998
:
North–south precipitation patterns in western North America on interannual-to-decadal timescales
.
J. Climate
,
11
,
3095
3111
, doi:.
Donner
,
L. J.
, and Coauthors
,
2011
:
The dynamical core, physical parameterizations, and basic simulation characteristics of the atmospheric component AM3 of the GFDL Global Coupled Model CM3
.
J. Climate
,
24
,
3484
3519
, doi:.
Dunne
,
J. P.
, and Coauthors
,
2012
:
GFDL’s ESM2 global coupled climate–carbon Earth system models. Part I: Physical formulation and baseline simulation characteristics
.
J. Climate
,
25
,
6646
6665
, doi:.
Frankignoul
,
C.
,
N.
Sennéchael
,
Y.-O.
Kwon
, and
M. A.
Alexander
,
2011
:
Influence of the meridional shifts of the Kuroshio and the Oyashio Extensions on the atmospheric circulation
.
J. Climate
,
24
,
762
777
, doi:.
Furtado
,
J. C.
,
E.
Di Lorenzo
,
N.
Schneider
, and
N. A.
Bond
,
2011
:
North Pacific decadal variability and climate change in the IPCC AR4 models
.
J. Climate
,
24
,
3049
3067
, doi:.
Gent
,
P. R.
, and Coauthors
,
2011
:
The Community Climate System Model version 4
.
J. Climate
,
24
,
4973
4991
, doi:.
Gershunov
,
A.
, and
T. P.
Barnett
,
1998
:
Interdecadal modulation of ENSO teleconnections
.
Bull. Amer. Meteor. Soc.
,
79
,
2715
2725
, doi:.
Gershunov
,
A.
,
T. P.
Barnett
, and
D. R.
Cayan
,
1999
:
North Pacific interdecadal oscillation seen as factor in ENSO-related North American climate anomalies
.
Eos, Trans. Amer. Geophys. Union
,
80
,
25
30
, doi:.
Gill
,
A. E.
,
1980
:
Some simple solutions for heat-induced tropical circulation
.
Quart. J. Roy. Meteor. Soc.
,
106
,
447
462
, doi:.
Giorgetta
,
M. A.
, and Coauthors
,
2013
: Climate and carbon cycle changes from 1850 to 2100 in MPI-ESM simulations for the Coupled Model Intercomparison Project phase 5. J. Adv. Model. Earth Syst.,5, 572–597, doi:.
Guilyardi
,
E.
,
A.
Wittenberg
,
A.
Fedorov
,
M.
Collins
,
C.
Wang
,
A.
Capotondi
,
G. J.
van Oldenborgh
, and
T.
Stockdale
,
2009
:
Understanding El Niño in ocean–atmosphere general circulation models: Progress and challenges
.
Bull. Amer. Meteor. Soc.
,
90
,
325
340
, doi:.
Halpert
,
M. S.
, and
C. F.
Ropelewski
,
1992
:
Surface temperature patterns associated with the Southern Oscillation
.
J. Climate
,
5
,
577
593
, doi:.
Hamming
,
R.
,
1998
: Digital Filters. 3rd ed. Dover Publications, 296 pp.
Hannachi
,
A.
,
I. T.
Jolliffe
, and
D. B.
Stephenson
,
2007
:
Empirical orthogonal functions and related techniques in atmospheric science: A review
.
Int. J. Climatol.
,
27
,
1119
1152
, doi:.
He
,
J.
,
B. J.
Soden
, and
B.
Kirtman
,
2014
:
The robustness of the atmospheric circulation and precipitation response to future anthropogenic surface warming
.
Geophys. Res. Lett.
,
41
,
2614
2622
, doi:.
Herweijer
,
C.
,
R.
Seager
,
E. R.
Cook
, and
J.
Emile-Geay
,
2007
:
North American droughts of the last millennium from a gridded network of tree-ring data
.
J. Climate
,
20
,
1353
1376
, doi:.
Hidalgo
,
H. G.
, and
J. A.
Dracup
,
2003
:
ENSO and PDO effects on hydroclimatic variations of the upper Colorado River basin
.
J. Hydrometeor.
,
4
,
5
23
, doi:.
Horel
,
J. D.
, and
J. M.
Wallace
,
1981
:
Planetary-scale atmospheric phenomena associated with the Southern Oscillation
.
Mon. Wea. Rev.
,
109
,
813
829
, doi:.
Hurrell
,
J. W.
, and Coauthors
,
2013
:
The Community Earth System Model: A framework for collaborative research
.
Bull. Amer. Meteor. Soc.
, 94, 1339–1360, doi:.
Iacobucci
,
A.
, and
A.
Noullez
,
2005
:
A frequency selective filter for short-length time series
.
Comput. Econ.
,
25
,
75
102
, doi:.
Kalnay
,
E.
, and Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project
.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
, doi:.
Kiktev
,
D.
,
J.
Caesar
,
L. V.
Alexander
,
H.
Shiogama
, and
M.
Collier
,
2007
: Comparison of observed and multimodeled trends in annual extremes of temperature and precipitation. Geophys. Res. Lett.,34, L10702, doi:.
Kirtman
,
B. P.
, and
J.
Shukla
,
2002
: Interactive coupled ensemble: A new coupling strategy for CGCMs. Geophys. Res. Lett.,29 (10), doi: .
Kushnir
,
Y.
,
W. A.
Robinson
,
I.
Bladé
,
N. M. J.
Hall
,
S.
Peng
, and
R.
Sutton
,
2002
:
Atmospheric GCM response to extratropical SST anomalies: Synthesis and evaluation
.
J. Climate
,
15
,
2233
2256
, doi:.
Latif
,
M.
, and
T. P.
Barnett
,
1994
:
Causes of decadal climate variability over the North Pacific and North America
.
Science
,
266
,
634
637
, doi:.
Mantua
,
N. J.
,
S. R.
Hare
,
Y.
Zhang
,
J. M.
Wallace
, and
R. C.
Francis
,
1997
:
A Pacific interdecadal climate oscillation with impacts on salmon production
.
Bull. Amer. Meteor. Soc.
,
78
,
1069
1079
, doi:.
Martin
,
G. M.
, and Coauthors
,
2011
:
The HadGEM2 family of Met Office Unified Model climate configurations
.
Geosci. Model Dev.
,
4
,
723
757
, doi:.
Mauget
,
S. A.
,
2003
:
Intra- to multidecadal climate variability over the continental United States: 1932–99
.
J. Climate
,
16
,
2215
2231
, doi:.
McAfee
,
S. A.
,
2014
:
Consistency and the lack thereof in Pacific decadal oscillation impacts on North American winter climate
.
J. Climate
,
27
,
7410
7431
, doi:.
Mehran
,
A.
,
A.
AghaKouchak
, and
T. J.
Phillips
,
2014
:
Evaluation of CMIP5 continental precipitation simulations relative to satellite-based gauge-adjusted observations
.
J. Geophys. Res.
,
119
,
1695
1707
, doi:.
Mensing
,
S.
, and Coauthors
,
2013
:
A network for observing Great Basin climate change
.
Eos, Trans. Amer. Geophys. Union
,
94
,
105
106
, doi:.
Mignot
,
J.
, and
S.
Bony
,
2013
:
Presentation and analysis of the IPSL and CNRM climate models used in CMIP5
.
Climate Dyn.
,
40
, 2089, doi:.
Mo
,
K. C.
, and
J. E.
Schemm
,
2008
: Relationships between ENSO and drought over the southeastern United States. Geophys. Res. Lett.,35, L15701, doi:.
Mote
,
P.
,
L.
Brekke
,
P. B.
Duffy
, and
E.
Maurer
,
2011
:
Guidelines for constructing climate scenarios
.
Eos, Trans. Amer. Geophys. Union
,
92
,
257
258
, doi:.
Namias
,
J.
,
X.
Yuan
, and
D. R.
Cayan
,
1988
:
Persistence of North Pacific sea surface temperature and atmospheric flow patterns
.
J. Climate
,
1
,
682
703
, doi:.
Peng
,
S.
, and
J. S.
Whitaker
,
1999
:
Mechanisms determining the atmospheric response to midlatitude SST anomalies
.
J. Climate
,
12
,
1393
1408
, doi:.
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:.
Ropelewski
,
C. F.
, and
M. S.
Halpert
,
1986
:
North American precipitation and temperature patterns associated with the El Niño/Southern Oscillation (ENSO)
.
Mon. Wea. Rev.
,
114
,
2352
2362
, doi:.
Ropelewski
,
C. F.
, and
M. S.
Halpert
,
1989
:
Precipitation patterns associated with the high index phase of the Southern Oscillation
.
J. Climate
,
2
,
268
284
, doi:.
Schmidt
,
G. A.
, and Coauthors
,
2014
: Configuration and assessment of the GISS ModelE2 contributions to the CMIP5 archive. J. Adv. Model. Earth Syst.,6, 141–184, doi:.
Schneider
,
U.
,
A.
Becker
,
P.
Finger
,
A.
Meyer-Christoffer
,
B.
Rudolf
, and
M.
Ziese
,
2011
: GPCC full data reanalysis version 6.0 at 1.0°: Monthly land-surface precipitation from rain-gauges built on GTS-based and historic data. Deutscher Wetterdienst, data accessed 13 Feb 2014, doi:.
Stocker
,
T. F.
, and Coauthors
, Eds.,
2013
: Climate Change 2013: The Physical Science Basis. Cambridge University Press, 1535 pp.
Strong
,
C.
, and
R. E.
Davis
,
2008
:
Variability in the position and strength of winter jet stream cores related to Northern Hemisphere teleconnections
.
J. Climate
,
21
,
584
592
, doi:.
Taylor
,
K. E.
,
R. J.
Stouffer
, and
G. A.
Meehl
,
2012
:
An overview of CMIP5 and the experiment design
.
Bull. Amer. Meteor. Soc.
,
93
,
485
498
, doi:.
Thompson
,
D. W. J.
, and
J. M.
Wallace
,
2000
:
Annular modes in the extratropical circulation. Part I: Month-to-month variability
.
J. Climate
,
13
,
1000
1016
, doi:.
Volodin
,
E.
,
N.
Dianskii
, and
A.
Gusev
,
2010
:
Simulating present-day climate with the INMCM4.0 coupled model of the atmospheric and oceanic general circulations
.
Izv. Atmos. Ocean. Phys.
,
46
,
414
431
, doi:.
Wang
,
S.-Y.
,
R. R.
Gillies
,
J.
Jin
, and
L. E.
Hipps
,
2010
:
Coherence between the Great Salt Lake level and the Pacific quasi-decadal oscillation
.
J. Climate
,
23
,
2161
2177
, doi:.
Wang
,
S.-Y.
,
R.
Gillies
,
L.
Hipps
, and
J.
Jin
,
2011
:
A transition-phase teleconnection of the Pacific quasi-decadal oscillation
.
Climate Dyn.
,
36
,
681
693
, doi:.
Wang
,
S.-Y.
,
R. R.
Gillies
, and
T.
Reichler
,
2012
:
Multidecadal drought cycles in the Great Basin recorded by the Great Salt Lake: Modulation from a transition-phase teleconnection
.
J. Climate
,
25
,
1711
1721
, doi:.
Watanabe
,
M.
, and Coauthors
,
2010
:
Improved climate simulation by MIROC5: Mean states, variability, and climate sensitivity
.
J. Climate
,
23
,
6312
6335
, doi:.
Wise
,
E. K.
,
2010
: Spatiotemporal variability of the precipitation dipole transition zone in the western United States. Geophys. Res. Lett.,37, L07706, doi:.
Wise
,
E. K.
,
M. L.
Wrzesien
,
M. P.
Dannenberg
, and
D. L.
McGinnis
,
2015
:
Cool-season precipitation patterns associated with teleconnection interactions in the United States
.
J. Appl. Meteor. Climatol.
,
54
,
494
505
, doi:.
Xin
,
X.-G.
,
T.-W.
Wu
, and
J.
Zhang
,
2013
:
Introduction of CMIP5 experiments carried out with the climate system models of Beijing Climate Center
.
Adv. Climate Change Res.
,
4
,
41
49
, doi:.
Yukimoto
,
S.
, and Coauthors
,
2012
:
A new global climate model of the Meteorological Research Institute: MRI-CGCM3—Model description and basic performance
.
J. Meteor. Soc. Japan
,
90A
,
23
64
, doi:.