1. Introduction
A distinctive feature of mid-Holocene climate is the dominant influence of precessional insolation forcing (Fig. 1a) and its effect on strengthening the Northern Hemisphere summer monsoon climates. The prediction of precessional influence on monsoons by Kutzbach (1981) and the subsequent confirmation of the mid-Holocene monsoon increase from various different paleoproxy studies (e.g., Rossignol-Strick 1985; Pokras and Mix 1987) stand out as a significant achievement of climate modeling. The purpose of this note is to make a case for another pronounced feature of the mid-Holocene climate: the possible weakening of the transient eddy activity over the North Pacific midlatitude jet region and associated North Pacific basinwide climate conditions during boreal winter.
In a previous study, we found significant reduction in the transient eddy activity in a coupled model—a full atmospheric general circulation model coupled to a reduced gravity ocean—in a set of runs used to investigate mid-Holocene El Niño–Southern Oscillation (ENSO) behavior (Chiang et al. 2009). In that paper, we suggested that reduced ENSO activity during that time was associated with the reduced stochastic forcing of ENSO resulting from the reduced transient eddy activity. The study, however, did not address the origins of the reduced eddy activity; we now address it here.
We show that the mid-Holocene simulation resembles two observed (and related) features of today’s climate: (i) interannual variations in wintertime North Pacific storminess and (ii) the seasonal phenomenon of midwinter suppression whereby North Pacific transient eddy activity in today’s climate is reduced in midwinter, a phenomenon first identified by Nakamura (1992). The large-scale background climate anomalies associated with both these observed analogs—in particular, the strengthening of the North Pacific subtropical jet—are readily generated by the mid-Holocene orbital forcing. We argue that it is the generation of these background conditions by the mid-Holocene orbital forcing that leads to reduced North Pacific storminess.
2. Model and simulations
The model we use, as in Chiang et al. (2009), is the Community Climate Model version 3.6 (CCM3) running at T42 resolution coupled to a 1.5-layer reduced gravity ocean (RGO) with a resolution of 2° longitude × 1° latitude (hereafter referred to as the CCM3-RGO). The CCM3 is a widely used atmospheric model consisting of a spectral dynamical core and full physics package including radiation, convection, boundary layer, and a diagnostic treatment of clouds, as well as a prescribed land surface and sea ice (Kiehl et al. 1998). The 1.5-layer reduced gravity ocean model had been extensively used for simulating coupled tropical ocean processes; the version used in this model is same as the one documented in Chang (1994) except that ours uses a time-fixed but spatially varying mixed layer depth estimated from Levitus data, and the variation of entrained subsurface temperature is parameterized in terms of the variation of thermocline depth using a multivariate linear relationship (Fang 2005). Furthermore, the reduced gravity ocean has been extended to simulate the entire global ocean (and not just the tropics); poleward of 30°N and 30°S, the parameterized influence of thermocline depth variations on the temperature of entrained upwelling water is switched off, since that mechanism is not important for determining extratropical SST anomalies.
The imposition of a fixed mixed layer depth is somewhat problematic for simulating extratropical SSTs, since seasonal variations thereof can be quite large. To partly compensate for the missing surface ocean model physics, we impose an annual-mean flux correction to the surface layer of the ocean model, allowing the simulated SST annual mean climatology to reasonably match the observed SST. Seasonal variations in the climate are generally reasonably well simulated compared to observations, although one deficiency of note is that the strength of the Aleutian low in the model wintertime is ∼25% weaker than in observations. We refer to Fang (2005) and Chiang et al. (2009) for details of the ocean model and coupling, as well as the fidelity of the model simulations.
The main simulations we use in our analyses are (i) a control simulation with present day (A.D. 1980) orbital conditions, land surface, and CO2 set at 355 ppm and (ii) a “mid-Holocene” (also called “6KBP”) simulation with orbital values of eccentricity = 0.018 67, obliquity = 24.103°, and angular precession = 1.37°, with the orbital parameters computed by the model using a formula from Berger (1978) and the year set to 4020 B.C. The subsequent change to the incoming insolation at the top of the atmosphere of the mid-Holocene simulation is shown in Fig. 1a. All other parameters and boundary conditions in the mid-Holocene simulation were kept the same as the control. In addition, we also ran “3KBP” and “9KBP” simulations in which the model year was set to 1020 and 7020 B.C., respectively.
For each of the above cases, we did an ensemble of three simulations each lasting 30 years, starting from different initial conditions extracted from a long-term present-day control simulation. We discard the first 10 years of each simulation as startup—since the ocean model does not simulate the deep ocean circulation, it equilibrates fairly rapidly—and use the subsequent 20 years for computing monthly climatological mean fields. The three individual climatologies are then combined to produce an ensemble mean. We also saved daily averaged geopotential height in the latter 20 years of each simulation for computation of transient eddy statistics.
For all other simulations used to test model sensitivity, we only ran a single 30-yr simulation; otherwise the data processing is similar to the above.
3. Boreal wintertime changes in the North Pacific
a. Background mid-Holocene Pacific climate changes
Figure 2 shows January–March (JFM) averaged changes to transient eddy activity over the North Pacific in the mid-Holocene simulation relative to the control. The measure of eddy activity used here is the variance of high-passed daily 500-mb geopotential height. The high pass is done with a simple 24-h difference filter (Wallace et al. 1988), where the data point for the previous day is subtracted from the present day, at every grid point. The reduction is apparent across the entire midlatitude basin, reduced by up to 30% of the climatological mean.
Figure 3 shows the seasonal evolution of the CCM3-RGO transient eddy activity averaged over the western North Pacific for both the control and mid-Holocene simulations (for comparison, the observed transient eddy activity is also shown). The mid-Holocene reduction primarily occurs during December–April, with the most pronounced period in January–March (Fig. 3d). Prior to November, the transient eddy activity is comparable and even greater for the mid-Holocene as for the control. The reduced transient eddy activity is corroborated by a comparable decrease in the transient eddy heat transport over the western midlatitude North Pacific, as measured by 866 mb
We note here that while in the CCM-RGO mid-Holocene ENSO activity is significantly reduced (as documented in Chiang et al. 2009), this is not the cause of the reduced North Pacific transient eddy activity. We did a complementary set of simulations —present-day and 6000 years BP—in which the ENSO in the model was essentially switched off by setting the thermocline depth anomalies in the reduced gravity ocean to zero. This effectively shuts off the thermocline–SST coupling that is integral to the ENSO dynamics, and ENSO in these simulations is essentially removed. In these simulations, the mid-Holocene transient eddy activity was reduced in a manner closely resembling Fig. 3.
The wintertime mid-Holocene northern and tropical Pacific is marked by pronounced changes to the background climate (Fig. 4). Sea level pressure (SLP) in the midlatitude North Pacific is reduced, deepening the Aleutian low by up to 6 mb (Fig. 4a). The midlatitude westerlies are significantly weakened (by up to 8 m s−1 at 200 mb; Fig. 4b) and the subtropical jet is considerably strengthened (by up to 10 m s−1). The reduced midlatitude westerlies are coincident with the reduced transient eddy activity (cf. Figs. 2 and 4b), consistent with the midlatitude westerlies being an eddy-driven feature.
Sea surface temperatures show cooling over 30°–40°N extending from the east coast of China in a slightly northwestward direction, extending into the central midlatitude North Pacific (Fig. 4c). The cooling is up to 2°C at the coastline. North of this feature are warmer anomalies straddling the high-latitude North Pacific from around 50°N to past 80°N. This combined warming to the north and cooling to the south weakens the North Pacific subarctic front; this weakening may be related to the decrease to the transient eddy activity (Nakamura et al. 2004). The warming portion also connects with a larger feature that extends into the North American continent and then, southward, parallels the west coast of North America. The 200-mb wind structure appears consistent, in a thermal wind sense, with the increased meridional temperature gradient thus presented between 20° and 30°N and the reduced meridional gradients between 40° and 70°N.
The tropical rainfall structure (Fig. 4d) shows a pronounced southward shift of the intertropical convergence zone (ITCZ) over the tropical Pacific, and indeed over all tropical basins (not shown), with southward anomalous cross-equatorial flow and ITCZ rainfall increasing in the Southern Hemisphere and decreasing in the Northern Hemisphere. This response indicates a stronger winter Hadley cell extending from the southern tropical summer latitudes into the winter northern tropical latitudes, consistent with Hadley cell driving of a stronger subtropical jet through increased angular momentum transport by the mean meridional circulation.
b. Mid-Holocene anomalies across the PMIP2 model simulations
Similar North Pacific climate changes arise in coupled model simulations of the mid-Holocene as part of the Paleoclimate Model Intercomparison Project 2 (PMIP2) archive (Braconnot et al. 2007). The pattern of boreal wintertime climate anomalies seen in our CCM3-RGO mid-Holocene simulations (Fig. 4) are also the most representative anomalies across the seven PMIP2 models we considered (Table 1). We objectively extract this common response by applying a multivariate empirical orthogonal function (EOF) analysis (e.g., Wilks 2006, 477–479) on four JFM mid-Holocene climate anomalies—precipitation, 200-mb zonal wind U, surface temperature, and SLP—across the PMIP2 models of Table 1.1 In other words, we replace the usual “time” dimension of the EOF with the various model representations. If there is a dominant common response to the mid-Holocene orbital forcing in all the PMIP2 model simulations, the EOF method is capable of objectively extracting it.
The EOF was taken over the northern and tropical Pacific from 40°S to 70°N and 100°E to 80°W. Prior to taking the multivariate EOF, we (i) interpolated each model’s output to the spatial grid resolution of the Fast Ocean–Atmosphere Model (FOAM) (the FOAM model being the coarsest resolution of the PMIP2 models we used), (ii) normalized the data by the square root of the cosine of latitude to account for the latitudinally varying surface area represented by the grid point, and (iii) normalized each field by dividing by the standard deviation of all of that field’s data grid points. Note that we do not remove the composite mean (i.e., the average anomaly across all the models) from each model’s anomaly when preparing the data for the EOF. We computed the EOF by applying a singular value decomposition on the resulting data matrix.
Results of the multivariate EOF analysis are shown in Figs. 5 and 6. EOF1 dominates the variance (∼50% of the total), and the corresponding principal component (PC) loadings are all of the same sign and comparable magnitude; in other words, all PMIP models exhibit behavior as represented by EOF1, albeit with varying strength. It is well separated from EOF2, which explains just over 20% of the total variance. The EOF1 pattern (Fig. 6) shows a striking resemblance to the JFM mid-Holocene anomalies of the CCM3-RGO (Fig. 4)—namely, a deeper North Pacific midlatitude low, an increased northern subtropical jet and reduced midlatitude westerlies, cooling of SST in the Kuroshio extension region, and a southward-shifted ITCZ precipitation. In the PMIP2 case, the southward shift in rainfall is represented largely by a reduction to the equatorial rainfall and an increase in the rainfall along the South Pacific convergence zone (SPCZ). Harrison et al. (2003) noted similar wintertime mid-Holocene climate anomalies in the North Pacific in two coupled model simulations [the Fast Ocean–Atmosphere Model and the National Center for Atmospheric Research (NCAR) Climate System Model].
On the basis of the good match between the PMIP2 EOF1 (the most representative response across all models) and the CCM3-RGO mid-Holocene anomalies, we argue that the climate response we obtained in our CCM3-RGO simulations is a meaningful climate response to mid-Holocene orbital forcing. The spatial resemblance between the PMIP2 EOF 1 results with the CCM3-RGO mid-Holocene simulation is, however, not perfect. Most prominently, the PMIP2 EOF1 shows a pronounced response in the eastern equatorial Pacific reminiscent of La Niña–like conditions; and the rainfall pattern, while showing a more pronounced SPCZ structure, also has some resemblance to the precipitation response to La Niña–type conditions. The relative magnitude of the tropical rainfall response (as compared to the deepening Aleutian low) is also weaker in the PMIP2 EOF1 as compared to the CCM3-RGO response. We speculate on the origins of the tropical Pacific differences between the PMIP2 and CCM3-RGO in section 6.
For the PMIP2 mid-Holocene simulations for which daily 500-mb geopotential height output was available, we find some tendency toward a decrease in mid-Holocene transient eddy activity, but it is not consistent across all models (Fig. 7). Both ECHAM5 and the Model for Interdisciplinary Research on Climate (MIROC) showed behavior most consistent with our CCM3-RGO simulations, with substantial decreases in mid-Holocene eddy activity during the midwinter. The Meteorological Research Institute (MRI) model also exhibited decreased transient eddy activity, but the change was relatively small. FOAM showed an increase in the eddy activity between October and December, but with a sharp decrease thereafter; we note, however, that the mean transient eddy activity in this model is located too far poleward. Mid-Holocene transient eddies in the third climate configuration of the Met Office Unified Model (HadCM3) increased for December–January and decreased thereafter.
4. Present-day climate analogs
a. Interannual variations in North Pacific wintertime storminess
We now make an argument for the close resemblance between the mid-Holocene climate anomalies and climate anomalies associated with observed interannual variations in North Pacific wintertime storminess. Interannual variations in North Pacific midwinter storminess were investigated by Zhang (1997) in the context of atmospheric general circulation model simulations and were also observed by Chang (2001). In both cases, they found that decreases in the North Pacific transient eddy activity are associated with increases in the North Pacific subtropical jet strength. Nakamura et al. (2002) further documented the interannual and decadal modulations of Pacific storm-track activity and found that decreases in Pacific storminess were associated not only with a strengthening of the subtropical westerlies but also with a deepening of the Aleutian low and cooling of the northwestern Pacific.
The resemblance between the climate changes found in Nakamura et al. (2002) and our model mid-Holocene changes motivated us to extend this comparison. To create an interannual index of North Pacific storminess, we first computed the January–February mean variance of daily high-pass filtered 500-mb geopotential height from National Centers for Environmental Prediction (NCEP) reanalysis (Kalnay et al. 1996), spanning 1980 to 2008, and then averaged this field over a box from 40° to 60°N and 150°E to 130°W to create the interannual index. We then regressed the normalized index against various climate fields over the same time period (Fig. 8). The regressed fields indeed closely resemble the CCM3-RGO January–March mid-Holocene climate anomalies (Fig. 4), exhibiting a deeper Aleutian low, increased North Pacific subtropical jet, cooler SST in the Kuroshio Extension region, and a reduced North Pacific ITCZ and enhanced SPCZ.
In response to a suggestion by a reviewer, we tested the PMIP2 models to examine whether the relative lack of storminess response in the PMIP2 mid-Holocene simulations might be tied to how well the models simulate the observed interannual relationship between wintertime North Pacific jet strength and storminess. We used a method similar to Chang (2001) to assess the correlation between year-to-year values of January–February 200-mb U with the corresponding storm-track activity; the results are shown in Table 2. In short, all models appear to possess the appropriate negative correlation (exceeding 95% confidence) between the subtropical jet strength with midlatitude storm-track activity, with r values in the range of what was found in Chang (2001). Hence (apart from FOAM, which clearly does not simulate storm-track activity well), we cannot say which of the PMIP2 model results in Fig. 7 should be more reliable.
b. Midwinter suppression
There is also a pronounced seasonal reduction in North Pacific storminess during midwinter, a phenomenon known as “midwinter suppression.” First documented by Nakamura (1992), it refers to a significant reduction to observed baroclinic wave activity in midwinter in the North Pacific compared to late autumn and early spring, despite increased baroclinicity in the large-scale atmospheric conditions during that time (Fig. 3a, contours). Nakamura associated the reduced activity with the climatological strength of the North Pacific westerlies: when the subtropical jet exceeds 45 m s−1, baroclinic eddy activity diminishes (Fig. 3a, shading). The interannual variations in North Pacific storminess described in section 4a can be viewed as a modulation of midwinter suppression—for example, midwinter suppression was pronounced in the early–mid-1980s, only to almost disappear in the late 1980s and early 1990s (Nakamura et al. 2002).
Could the mid-Holocene climate changes be associated with an enhancement of this phenomenon? Indeed, the model mid-Holocene northern and tropical Pacific climate appears, by various measures, to be more winterlike. The CCM3-RGO mid-Holocene subtropical jet is strengthened in midwinter compared to the control and is in sync with the timing of the reduced mid-Holocene transient eddy activity (cf. Figs. 3b,c). Figures 9a–d show the same fields as Figs. 4a–d, but for the deviation of the January–March CCM3-RGO control climatology from the annual mean. The resemblance between the seasonal deviation and the mid-Holocene anomalies is striking. Both exhibit reduced SLP in the midlatitude North Pacific (Figs. 4a and 9a), an increased subtropical jet and reduced midlatitude westerlies (Figs. 4b and 9b), enhanced cooling in the Kuroshio Extension region (Figs. 4c and 9c), and southward-shifted tropical rainfall anomalies (Figs. 4d and 9d).
Given that our mid-Holocene climate changes are “more winterlike,” it suggests that the reduced model transient eddy activity in the mid-Holocene simulation can be viewed simply as an enhancement of this phenomenon. A more winterlike mid-Holocene Northern Hemisphere is not unexpected, given that the same precessional forcing that gives rise to more boreal summertime insolation also gives rise to less boreal wintertime insolation (Fig. 1). Specifically, for the fall insolation leading up to the winter months [September–November (SON)], the mid-Holocene insolation anomalies resemble the seasonal deviation in insolation leading up to the wintertime, with decreases in the Northern Hemisphere and increases to the Southern Hemisphere (cf. the September–November anomalies in Fig. 1a and the seasonal variation in Fig. 1b). Since the ocean responds to the insolation with a delay of a few months, it may be these insolation anomalies that determine the Pacific Ocean surface conditions during boreal winter. The orbital climate forcings are thus consistent with the mid-Holocene climate anomalies in giving a more winterlike northern and tropical Pacific climate. We expand on this interpretation in section 5.
5. Origins of mid-Holocene North Pacific climate anomalies
The question remains as to how the large-scale North Pacific mid-Holocene climate changes come about from the orbital forcing. We suggested in section 3b that it is the more winterlike insolation anomalies during the boreal fall and early winter that produces the more winterlike mid-Holocene. This interpretation is supported with an idealized simulation in which the mid-Holocene insolation was applied to a CCM3-RGO model simulation for six months of the year only, from the beginning of September through the end of February. The wintertime North Pacific climate anomalies resulting from them bear close resemblance to that for the standard mid-Holocene simulation (not shown; they are similar to Fig. 4 for the standard mid-Holocene simulation). The complementary simulation, applying insolation anomalies from the beginning of March through the end of August, did not produce appreciable wintertime midlatitude North Pacific climate anomalies (not shown).
Simulations of North Pacific transient eddy activity across various time slices during the Holocene provide confirmation of our interpretation. Whereas present-day perihelion occurs in boreal winter, perihelion at the start of the Holocene epoch occurred in boreal summer. Thus, as we go further back in time from the present day to the early Holocene, precessional changes led to insolation changes with an increasingly winterlike climate. If our interpretation is correct, wintertime North Pacific transient eddy activity should reduce as we go back in time from the present; indeed, this is what the CCM-RGO simulates. Table 3 shows a measure of North Pacific transient eddy activity for several time slices during the Holocene (present day, 3KBP, 6KBP, and 9KBP). In accordance with our interpretation, North Pacific wintertime transient eddy activity decreases monotonically as one goes back in time to 9KBP.
The mid-Holocene North Pacific climate anomalies appear to be a consequence of global insolation changes and resulting global climate adjustment, not simply a response to northern extratropical insolation changes. To show this, we perform two additional idealized simulations, with the mid-Holocene insolation anomalies applied only north of 27°N and then only south of 27°N. Both simulations show climate anomalies resembling the full mid-Holocene simulation (cf. Fig. 4)—deeper Aleutian low, colder north Pacific SST, and stronger subtropical jet—but each possesses roughly half the magnitude (not shown). The exception is the tropical rainfall response, which is more pronounced in the simulation applying the insolation anomalies south of 27°N. Adding the anomalies of these two idealized simulations returns approximately the anomalies of the standard mid-Holocene simulation.
How do insolation changes in the tropics and Southern Hemisphere lead to North Pacific climate anomalies? We believe it comes about through the tropical rainfall response: Yin (2002) had previously suggested a link between rainfall changes over the tropical western Pacific and strengthening of the North Pacific subtropical jet. The insolation anomalies leading up to winter have an “interhemispheric-like” character with less insolation in the north and more in the south, and the tropical rainfall responds to the resulting interhemispheric thermal gradient with a southward shift (Broccoli et al. 2006; Chiang and Bitz 2005). The southward-shifted rainfall should lead to a stronger winter Hadley cell (Lindzen and Hou 1988) that in turn increases the angular momentum transport that drives the North Pacific subtropical jet. Thus, the origins of the increased subtropical jet in the wintertime mid-Holocene North Pacific—which most directly relates to the midwinter suppression, according to Nakamura (1992)—is a product of both the tropical climate changes and changes in the mid- and high-latitude North Pacific.
6. Discussion
Our claimed link between mid-Holocene storminess and midwinter suppression assumes that the large-scale conditions associated with midwinter suppression are in fact what drive the reduced transient eddy activity. The current dynamical literature on midwinter suppression largely supports this assumption. Midwinter suppression has been most closely linked to the strength of the subtropical jet and its effect on eddy growth rates, starting from Nakamura’s (1992) seminal analysis. Nakamura suggested that the faster advection of eddies by the increased jet may be the reason for the suppression, although Harnik and Chang (2004) pointed out that the decrease in the growth rate with a narrower jet may also play a role. There is evidence from diagnostic studies of interannual/decadal variations in North Pacific wintertime storminess for a decreased efficiency in the conversion of mean flow available potential energy to eddy kinetic energy as the westerlies strengthen (Chang 2001; Nakamura et al. 2002). More recently, Lee and Kim (2003) argued that a sufficiently strong subtropical jet encourages baroclinic eddy growth in the subtropical latitudes at the expense of midlatitude eddies, thus reducing eddy activity in the middle and higher latitudes and weakening the subpolar jet. Probably the most convincing evidence for North Pacific basic-state control of transient eddy activity comes from Zhang and Held (1999): they showed that a linear stochastic model used to simulate midlatitude storm activity is able to capture the midwinter suppression quantitatively, suggesting that the origins thereof lie in the basic-state climate.
There is an alternate hypothesis that reduced seeding of the Pacific storm track from upstream sources may be the reason for midwinter suppression (Orlanski 2005; Penny et al. 2010), rather than changes to the downstream development. This, however, may not be inconsistent with the basic-state control of midwinter suppression, since the basic state may control the rate of seeding as well. Indeed, building on previous suggestions by Son et al. (2009) that topography modifies downstream storm-track intensity, Park et al. (2010) showed in a CCM3 simulation with the central Asian mountains removed that North Pacific midwinter suppression no longer appears, implying that it is the presence of those mountains that leads to midwinter suppression. Park et al. attribute part of the midwinter suppression to reduced baroclinic conversion over the central Asian mountains, consistent with Penny et al. (2010). The interesting aspect of this, however, is that Son et al. (2009) would argue that topography suppresses downstream transient eddy activity when the basic state flow is in a strong single-jet (as opposed to a weak double-jet) state—and such a single-jet state usually occurs in midwinter.
It is not the purpose of this note to elucidate the mechanisms of transient eddy suppression in our model. Given the multitude of existing hypotheses for midwinter suppression, it is unlikely that we would be able come up with a definitive mechanism for our own simulations. Instead, our purpose is simply to make the connection between reduced transient eddy activity in our mid-Holocene simulation, with similar observed occurrences in today’s climate, and in so doing to argue for the plausibility of a “less stormy” North Pacific wintertime climate.
The major difference in the structure of the PMIP2 mid-Holocene response with the CCM3-RGO (as discussed in section 3b) occurred in the tropical Pacific: the PMIP2 gave a more La Niña-like structure, and the relative magnitudes of the precipitation anomalies are smaller. We speculate that this arises from the difference between using a reduced-gravity tropical ocean (in the CCM3-RGO) and a full dynamical ocean. In particular, the reduced-gravity ocean is designed to be responsive to the equatorial thermocline depth–SST relationship in the tropics; in fact, the CCM3-RGO was designed to give a good representation of tropical ocean–atmosphere dynamics, specifically ENSO. On the other hand, many fully coupled models have difficulty producing a reasonable equatorial mean state and annual cycle (Davey et al. 2002). The relative lack of response in the PMIP2 simulations may have to do with the fact that the tropical thermocline tends to be too diffuse in fully coupled models because of large vertical diffusivity; this dampens the equatorial thermocline depth–SST relationship, and as a result ENSO variations in these models tend to be weak (Meehl et al. 2001). The tropical sensitivity directly affects our question since a shift in the Hadley cell influences the strength of the subtropical jet (as discussed in section 5) and hence the transient eddy activity. In this regard, it is telling that the subtropical jet response in PMIP2 EOF1 relative to the deeper Aleutian low response is smaller than that for the CCM3-RGO; this may be a reason why we do not see a convincing signal of reduced transient eddy activity in the PMIP2 mid-Holocene simulations.
The lack of a convincing signal of reduced transient eddy activity in the PMIP2 mid-Holocene simulations is the weakest link in our argument. However, given the complexity of the midlatitude dynamics underlying midwinter suppression, it may not be surprising that we cannot yet obtain a consistent response in transient eddy activity across the various model simulations. We note that of the five PMIP2 model simulations in Fig. 7 that had daily 500-mb geopotential height output accessible, only three (ECHAM5, MRI, and HadCM3) appear to simulate midwinter suppression adequately. Related to this, we note that Nishii et al. (2009) found that only eight of the 19 coupled models that participate in the World Climate Research Programme’s Coupled Model Intercomparison Project Phase 3 (CMIP3) twentieth-century climate simulations adequately reproduce the midwinter minimum in North Pacific storm-track activity; they used these eight models only in assessing future changes to the midwinter suppression. Since we only consider eight PMIP2 model simulations, we unfortunately did not have the luxury to discard models that could not adequately simulate midwinter suppression.
Is there paleoproxy evidence for reduced storminess in the mid-Holocene? Paleoproxies directly influenced by midlatitude storminess are unlikely, so any evidence is likely to come from changes to the North Pacific basic state climate. One such basic-state climate variable is the position of the subtropical westerlies. A stronger winter subtropical jet may imply that midlatitude rainfall may have been consistently shifted southward; over the west coast of North America, this may imply a wetter Californian and drier Pacific Northwest climate. This rainfall shift can be clearly seen in the PMIP2 model results in Fig. 6d; for the CCM3-RGO, the mid-Holocene January–March rainfall in the grid points around Northern California increases by ∼20% compared to the present-day simulation. As far as paleoproxy observations go, Kirby et al. (2007) documents wetter early–mid-Holocene conditions relative to today in Southern California from a lake sediment core and attributes this change to precessional forcing changing the location of the wintertime storm tracks. Thompson et al. (1993) come to a similar conclusion. On the other hand, oxygen isotope records from a lake sediment core in Alaska indicate drier conditions in the early and mid-Holocene, progressively becoming wetter into the late Holocene (Anderson et al. 2001). Thus, western North American paleoprecipitation records appear consistent with a southward-shifted subtropical jet during the early and mid-Holocene. It is beyond the scope of our analysis for a detailed assessment of Pacific paleoclimate records during the mid-Holocene, but such an undertaking would certainly be of value.
Acknowledgments
The coupling between the CCM3 and RGO was initially developed by one of us (YF) under the direction of Dr. Ping Chang; we also thank Dr. Chang for useful discussions. The authors acknowledge financial support from NSF ATM-0438201 and OCE-0902774, and the Comer Science and Education Foundation. The PMIP2 output was provided by several international modeling groups and collected and archived by the Laboratoire des Sciences du Climat et de l’Environnement. The PMIP2/MOTIF Data Archive is supported by CEA, CNRS, the EU project MOTIF (EVK2-CT-2002-00153), and the Programme National d’Etude de la Dynamique du Climat (PNEDC). More information is available online at http://pmip2.lsce.ipsl.fr/ and http://motif.lsce.ipsl.fr/.
REFERENCES
Anderson, L., M. B. Abbott, and B. P. Finney, 2001: Holocene climate inferred from oxygen isotope ratios in lake sediments, Central Brooks Range, Alaska. Quat. Res., 55 , 313–321.
Berger, A., 1978: Long-term variations of daily insolation and Quaternary climate changes. J. Atmos. Sci., 35 , 2362–2367.
Braconnot, P., and Coauthors, 2007: Results of PMIP2 coupled simulations of the mid-Holocene and Last Glacial Maximum. Part I: Experiments and large-scale features. Climate Past, 3 , 261–277.
Broccoli, A. J., K. A. Dahl, and R. J. Stouffer, 2006: Response of the ITCZ to Northern Hemisphere cooling. Geophys. Res. Lett., 33 , L01702. doi:10.1029/2005GL024546.
Chang, E. K. M., 2001: GCM and observational diagnoses of the seasonal and interannual variations of the Pacific storm track during the cool season. J. Atmos. Sci., 58 , 1784–1800.
Chang, P., 1994: A study of the seasonal cycle of sea surface temperature in the tropical Pacific Ocean using reduced gravity models. J. Geophys. Res., 99 , (C4). 7725–7741.
Chiang, J. C. H., and C. M. Bitz, 2005: Influence of high latitude ice cover on the marine intertropical convergence zone. Climate Dyn., 25 , 477–496.
Chiang, J. C. H., Y. Fang, and P. Chang, 2009: Pacific climate change and ENSO activity in the mid-Holocene. J. Climate, 22 , 923–939.
Davey, M. K., and Coauthors, 2002: STOIC: A study of coupled model climatology and variability in tropical ocean regions. Climate Dyn., 18 , 403–420.
Fang, Y., 2005: A coupled model study of the remote influence of ENSO on tropical Atlantic SST variability. Ph.D. thesis, Texas A&M University, 93 pp.
Harnik, N., and E. K. M. Chang, 2004: The effects of variations in jet width on the growth of baroclinic waves: Implications for midwinter Pacific storm track variability. J. Atmos. Sci., 61 , 23–40.
Harrison, S. P., J. E. Kutzbach, Z. Liu, P. J. Bartlein, B. Otto-Bliesner, D. Muhs, I. C. Prentice, and R. S. Thompson, 2003: Mid-Holocene climates of the Americas: A dynamical response to changed seasonality. Climate Dyn., 20 , 663–688.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–471.
Kiehl, J. T., J. J. Hack, G. B. Bonan, B. A. Boville, D. L. Williamson, and P. J. Rasch, 1998: The National Center for Atmospheric Research Community Climate Model: CCM3. J. Climate, 11 , 1131–1149.
Kirby, M., S. Lund, M. Anderson, and B. Bird, 2007: Insolation forcing of Holocene climate change in Southern California: A sediment study from Lake Elsinore. J. Paleolimnol., 38 , 395–417.
Kutzbach, J. E., 1981: Monsoon climate of the early Holocene: Climate experiment with the earth’s orbital parameters for 9000 years ago. Science, 214 , 59–61.
Lee, S., and H. Kim, 2003: The dynamical relationship between subtropical and eddy-driven jets. J. Atmos. Sci., 60 , 1490–1503.
Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Prof. Paper 13, 173 pp. and 17 microfiche.
Lindzen, R. S., and A. Y. Hou, 1988: Hadley circulations for zonally averaged heating centered off the equator. J. Atmos. Sci., 45 , 2416–2427.
Meehl, G. A., P. R. Gent, J. M. Arblaster, B. L. Otto-Bliesner, E. C. Brady, and A. Craig, 2001: Factors that affect the amplitude of El Niño in global coupled climate models. Climate Dyn., 17 , 515–526.
Nakamura, H., 1992: Midwinter suppression of baroclinic wave activity in the Pacific. J. Atmos. Sci., 49 , 1629–1641.
Nakamura, H., T. Izumi, and T. Sampe, 2002: Interannual and decadal modulations recently observed in the Pacific storm track activity and East Asian winter monsoon. J. Climate, 15 , 1855–1874.
Nakamura, H., T. Sampe, Y. Tanimoto, and A. Shimpo, 2004: Observed associations among storm tracks, jet streams, and midlatitude oceanic fronts. Earth’s Climate: The Ocean–Atmosphere Interaction, Geophys. Monogr., Vol. 147, Amer. Geophys. Union, 329–345.
Nishii, K., T. Miyasaka, Y. Kosaka, and H. Nakamura, 2009: Reproducibility and future projection of the midwinter storm-track activity over the Far East in the CMIP3 climate models in relation to “Haru-Ichiban” over Japan. J. Meteor. Soc. Japan, 87 , 581–588.
Orlanski, I., 2005: A new look at the Pacific storm track variability: Sensitivity to tropical SSTs and to upstream seeding. J. Atmos. Sci., 62 , 1367–1390.
Park, H. S., J. C. H. Chiang, and S. W. Son, 2010: The role of the central Asian mountains on the midwinter suppression of North Pacific storminess. J. Atmos. Sci., in press.
Penny, S., G. H. Roe, and D. S. Battisti, 2010: The source of the midwinter suppression in storminess over the North Pacific. J. Climate, 23 , 634–648.
Pokras, E. M., and A. C. Mix, 1987: Earth’s precession cycle and Quaternary climatic change in tropical Africa. Nature, 326 , 486–487.
Rossignol-Strick, M., 1985: Mediterranean Quaternary sapropels, an immediate response of the African monsoon to variation of insolation. Palaeogeogr. Palaeoclimatol. Palaeoecol., 49 , 237–263.
Son, S-W., M. Ting, and L. M. Polvani, 2009: The effect of topography on storm-track intensity in a relatively simple general circulation model. J. Atmos. Sci., 66 , 393–411.
Thompson, R. S., C. H. Whitlock, P. J. Bartlein, S. P. Harrison, and W. G. Spaulding, 1993: Climatic changes in the western United States since 18,000 yr BP. Global Climates since the Last Glacial Maximum, H. E. Wright Jr., et al., Eds., University of Minnesota Press, 468–513.
Wallace, J. M., G. H. Lim, and M. L. Blackmon, 1988: Relationship between cyclone tracks, anticyclone tracks and baroclinic waveguides. J. Atmos. Sci., 45 , 439–462.
Wilks, D. S., 2006: Statistical Methods in the Atmospheric Sciences. 2nd ed. International Geophysics Series, Vol. 91, Academic Press, 627 pp.
Xie, P. 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, J., 2002: The peculiar behavior of baroclinic waves during the midwinter suppression of the Pacific storm track. Ph.D. thesis, University of Washington, 121 pp.
Zhang, Y. Q., 1997: On the mechanisms of the mid-winter suppression of the Pacific storm track. Ph.D. thesis, Princeton University, 152 pp.
Zhang, Y. Q., and I. M. Held, 1999: A linear stochastic model of a GCM’s midlatitude storm tracks. J. Atmos. Sci., 56 , 3416–3455.
(a) Top-of-atmosphere (TOA) changes to insolation during the mid-Holocene (6000 years BP), as applied to our mid-Holocene model simulations. (b) The TOA insolation as applied to the present-day control run, but with the annual mean removed. For both (a) and (b), the contour units are W m−2, and positive (unshaded) values are directed toward the earth. This plot emphasizes the seasonal variation to TOA insolation. Comparison of (a) and (b) shows that the months leading up to NH midwinter (Oct–Dec) are more winterlike during the mid-Holocene.
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
Mid-Holocene changes to the high-passed geopotential height variance at 500 mb, averaged over JFM. The contour interval (CI) is 200 m2; negative values are shaded. Note that the zero contour is not plotted.
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
(a)–(c) Monthly-mean high-passed 500-mb geopotential height variance (contours, contour units × 1000 m2) and 200-mb zonal wind (lighter shading >40 m s−1, darker shading >60 m s−1), both zonally averaged over the western North Pacific between 130°E and 180°, for (a) NCEP reanalysis, (b) the CCM3-RGO control simulation, and (c) the mid-Holocene simulation. (d) The difference in the monthly-mean high-passed 500-mb geopotential height variance between the mid-Holocene and control simulations; shaded regions are for negative values, contour units × 1000 m2.
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
Difference in JFM climatology between the mid-Holocene and control simulations: (a) sea level pressure (CI 1 mb), (b) 200-mb zonal wind (CI 2 m s−1), (c) SST (CI 0.2 K), and (d) precipitation (CI 1 mm day−1). Shaded regions are negative; the zero contour is not plotted.
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
(a) Normalized eigenvalues of a multivariate (SLP, 200-mb U, surface temperature, and precipitation) EOF analysis of JFM mid-Holocene anomalies across all PMIP2 models. The domain of the EOF is the northern and tropical Pacific, 40°S–70°N and 100°E–80°W. (b) PC loadings corresponding to model 1. The models used in the EOF are listed in Table 1; the model number on the x axis corresponds to the number in the table.
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
EOF 1 of the PMIP2 model fields plotted in the manner of Fig. 4. Note that the contour intervals are as in Fig. 4, except that the precipitation contour interval is 0.5 mm day−1. The scaling of the EOF is such that the associated principal component (PC) matrix has unit norm.
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
Monthly-mean high-passed 500-mb geopotential height variance, zonally averaged over the western North Pacific between 130°E and 180°, for various PMIP2 simulations as indicated: (left) present-day simulations (values above 6000 m2 are shaded) and (right) differences between the present-day and mid-Holocene simulations (the latter minus the former; negative values are shaded). Contour units are ×1000 m2.
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
Regression of a normalized interannual index of Jan–Feb North Pacific transient eddy activity on various Jan–Feb climate fields, from 1980 to 2008. See text for the definition of the interannual index. (a) SLP (CI 1 mb), (b) 200-mb zonal wind (CI 2 m s−1), (c) SST (CI 0.2 K), and (d) precipitation (CI 1 mm day−1). Regression coefficients have been multiplied by −2, so the regression fields represent climate anomalies associated with weak storminess. Shaded regions are negative; the zero contour is not shown. The precipitation fields come from Xie and Arkin (1997) and the other climate fields from NCEP reanalysis (Kalnay et al. 1996). These fields should be compared to the CCM3-RGO mid-Holocene anomalies in Fig. 4.
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
As in Fig. 4, but for the deviation of the JFM climatology from the annual mean climatology for the control simulation. Contour intervals are as in Fig. 4 (except for SST, which is 0.4 K).
Citation: Journal of Climate 23, 14; 10.1175/2010JCLI3510.1
PMIP2 model simulations used in this study. See Braconnot et al. (2007) for a general reference and the PMIP2 Web site (http://pmip2.lsce.ipsl.fr/) for details of the models and experiments. Note that all simulations used here are ones that do not include mid-Holocene land vegetation changes. Since our analysis was performed, mid-Holocene simulations using the Australian Commonwealth Scientific and Industrial Research Organisation Mark version 3.0 (CSIRO Mk3) coupled model were included in the database; we did not use the CSIRO model output in our analysis.
Correlation between (a) year-to-year Jan–Feb 200-mb zonal wind in a 20° lon × 13° lat box centered at 150°E at the latitude of the maximum jet, with (b) the corresponding Jan–Feb storm-track activity in a 20° lon × 13° lat box centered at 180° at the latitude of the maximum storminess. Our storminess measure was the variance of high-passed 500-mb daily geopotential height averaged for Jan to Feb. All models show a significant negative correlation (exceeding 95% confidence, assuming the number of degrees of freedom corresponds to the number of years of the simulation) between the subtropical jet strength with midlatitude storm-track activity, in accordance with the observed relationship.
Wintertime North Pacific transient eddy activity at various time slices in the Holocene. Transient eddy activity here is measured by the January–March 500-mb high-passed geopotential height variance (as shown in Fig. 2), but averaged over the basinwide North Pacific from 40° to 60°N and 150°E to 130°W. Values shown are relative to the present-day control simulation. The results support the interpretation that the model North Pacific transient eddy activity decreases under more winterlike conditions.
