• Barnston, A. G., , and R. E. Livezey, 1987: Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev., 115 , 10831126.

    • Search Google Scholar
    • Export Citation
  • Black, R. X., , and R. M. Dole, 1993: The dynamics of large-scale cyclogenesis over the North Pacific Ocean. J. Atmos. Sci., 50 , 421442.

    • Search Google Scholar
    • Export Citation
  • Cash, B. A., , and S. Lee, 2001: Observed nonmodal growth of the Pacific–North American teleconnection pattern. J. Climate, 14 , 10171028.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., , S. Lee, , and K. L. Swanson, 2002: Storm track dynamics. J. Climate, 15 , 21632183.

  • Chavez, F. P., , J. Ryan, , S. E. Lluch-Cota, , and M. Niquen, 2003: From anchovies to sardines and back: Multidecadal change in the Pacific Ocean. Science, 299 , 217221.

    • Search Google Scholar
    • Export Citation
  • Deser, C., , A. Phillips, , and J. Hurrell, 2004: Pacific interdecadal climate variability: Linkages between the Tropics and the North Pacific during boreal winter since 1900. J. Climate, 17 , 31093124.

    • Search Google Scholar
    • Export Citation
  • Dole, R., 1986a: Persistent anomalies of the extratropical Northern Hemisphere wintertime circulation: Structure. Mon. Wea. Rev., 114 , 178207.

    • Search Google Scholar
    • Export Citation
  • Dole, R., 1986b: The life cycles of persistent anomalies and blocking over the North Pacific. Advances in Geophysics, Vol. 29, Academic Press, 31–69.

  • Dole, R., 1989: Life cycles of persistent anomalies. Part I: Evolution of 500 mb height fields. Mon. Wea. Rev., 117 , 177211.

  • Dole, R., , and N. D. Gordon, 1983: Persistent anomalies of the extratropical Northern Hemisphere wintertime circulation: Geographical distribution and regional persistence characteristics. Mon. Wea. Rev., 111 , 15671586.

    • Search Google Scholar
    • Export Citation
  • Dole, R., , and R. X. Black, 1990: Life cycles of persistent anomalies. Part II: The development of persistent negative height anomalies over the North Pacific Ocean. Mon. Wea. Rev., 118 , 824846.

    • Search Google Scholar
    • Export Citation
  • Duchon, C. E., 1979: Lanczos filtering in one and two dimensions. J. Appl. Meteor., 18 , 10161022.

  • Feldstein, S. B., 2000: The timescale, power spectra, and climate noise properties of teleconnection patterns. J. Climate, 13 , 44304440.

    • Search Google Scholar
    • Export Citation
  • Feldstein, S. B., 2002: Fundamental mechanisms of the growth and decay of the PNA teleconnection pattern. Quart. J. Roy. Meteor. Soc., 128 , 775796.

    • Search Google Scholar
    • Export Citation
  • Hare, S. R., , and N. J. Mantua, 2000: Empirical evidence for North Pacific regime shifts in 1977 and 1989. Progress in Oceanography, Vol. 47, Pergamon Press, 103–145.

  • Honda, M., , H. Nakamura, , J. Ukita, , I. Kousaka, , and K. Takeuchi, 2001: Interannual seesaw between the Aleutian and Icelandic lows. Part I: Seasonal dependence and life cycle. J. Climate, 14 , 10291042.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kleist, D. T., , and M. C. Morgan, 2005: Interpretation of the structure and evolution of adjoint-derived forecast sensitivity gradients. Mon. Wea. Rev., 133 , 467485.

    • Search Google Scholar
    • Export Citation
  • Lu, J., , R. J. Greatbatch, , and K. A. Peterson, 2004: Trend in Northern Hemisphere winter atmospheric circulation during the last half of the twentieth century. J. Climate, 17 , 37453760.

    • Search Google Scholar
    • Export Citation
  • 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 , 10691079.

    • Search Google Scholar
    • Export Citation
  • Molteni, F., , and T. N. Palmer, 1993: Predictability and finite-time instability of the northern winter circulation. Quart. J. Roy. Meteor. Soc., 119 , 269298.

    • Search Google Scholar
    • Export Citation
  • Namias, J., 1978: Multiple causes of the North American abnormal winter 1976-77. Mon. Wea. Rev., 106 , 279295.

  • Overland, J. E., , and P. J. Stabeno, 2004: Is the climate of the Bering Sea warming and affecting the ecosystem? Eos, Trans. Amer. Geophys. Union, 85 , 309316.

    • Search Google Scholar
    • Export Citation
  • Overland, J. E., , J. M. Adams, , and N. A. Bond, 1999: Decadal variability of the Aleutian low and its relation to high-latitude circulation. J. Climate, 12 , 15421548.

    • Search Google Scholar
    • Export Citation
  • Palmer, T. N., 1988: Medium and extended range predictability, and the stability of the PNA mode. Quart. J. Roy. Meteor. Soc., 114 , 691713.

    • Search Google Scholar
    • Export Citation
  • Quadrelli, R., , and J. M. Wallace, 2004: Varied expressions of the hemispheric circulation observed in association with contrasting polarities of prescribed patterns of variability. J. Climate, 17 , 42454253.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K., 1990: Recent observed interdecadal climate changes in the Northern Hemisphere. Bull. Amer. Meteor. Soc., 71 , 988993.

  • Trenberth, K., , and J. W. Hurrell, 1994: Decadal atmosphere-ocean variations in the Pacific. Climate Dyn., 9 , 303319.

  • Wallace, J. M., , and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109 , 784812.

    • Search Google Scholar
    • Export Citation
  • Wu, Q., , and D. M. Straus, 2004: AO, COWL, and observed climate trends. J. Climate, 17 , 21392156.

  • Yasunaka, S., , and K. Hanawa, 2002: Regime shifts found in the Northern Hemisphere SST field. J. Meteor. Soc. Japan, 80 , 119135.

  • Zou, X. F., , F. Vandenberghe, , M. Pondeca, , and Y-H. Kuo, 1997: Introduction to adjoint techniques and the MM5 adjoint modeling system. NCAR Tech. Note NCAR/TN-435+STR, 110 pp.

  • View in gallery

    (a) Line graph of 1500 6-h time periods for the fall and winter months of 2000–01 depicting 500-hPa height anomalies at KP1 (thin), and filtered 500-hPa height anomalies (boldface). (b) A close-up of (a), depicting how an event was declared by identifying a period of at least 10 days where filtered 500-hPa height anomalies were below −100 m. Arrows indicate the duration of one particular event from 1200 UTC 16 Dec 2000 to 0000 UTC 7 Jan 2001.

  • View in gallery

    Histogram of the total number of LSC events vs the length of events at KP1.

  • View in gallery

    The total number of events from November to March and the locations for KP1 and KP2 for (a) 1977–2003, (b) 1977–88, and (c) 1988–2003.

  • View in gallery

    Composite 500-hPa height anomalies for all KP1 events from 1977 to 2002 (shaded; interval 60 m, beginning at ±30 m) and 500-hPa height from 1977 to 2003 (contoured every 6 dam) at (a) 5 days prior to onset, (c) onset (X indicates the position of KP1), and (e) 6 days after onset. Composite sea level pressure for all KP1 events from 1977 to 2003 (contoured every 4 hPa), and 1000–500-hPa thickness (dashed, interval 6 dam) at (b) 5 days prior to onset, (d) onset, and (f) 6 days after onset. The 540-dam thickness is shown as a boldface dotted line.

  • View in gallery

    Same as in Fig. 4 but for KP2.

  • View in gallery

    Difference (KP2 − KP1) at 7 days after onset of (a) 500-hPa geopotential height anomalies (contours; interval 30 m) and (b) mean sea level pressure (contours; interval 2 hPa). Negative contours are dashed. Regions with a statistical significance at or above the 95% confidence level are shaded.

  • View in gallery

    KP1 onset +6 composite 500-hPa height anomalies (shaded; interval 60 m, beginning at ±30 m) and 500-hPa height (contoured every 6 dam) for (a) 1977–88 events and (c) 1988–2003 events. KP1 onset +6 composite sea level pressure (contoured every 4 hPa), and 1000–500-hPa thickness (dotted; interval 6 dam) for (b) 1977–88 events and (d) 1988–2003 events. The 540-dam thickness is shown as a boldface dotted line.

  • View in gallery

    Same as in Fig. 7 but for KP2.

  • View in gallery

    Difference (pre-1988 minus post-1988) 500-hPa composite geopotential height anomalies (contours; interval 30 m) at 8.5 days after onset for (a) KP1 and (b) KP2. Regions with a statistical significance at or above the 95% confidence level are shaded.

  • View in gallery

    Composite RMS of unfiltered 500-hPa height anomalies for all events (shaded; interval 20 m) and a schematic for primary cyclone tracks (arrows) for (a) KP1 events, (b) KP2 events, and (c) sequence of times when LSC events are not occurring (nonevents).

  • View in gallery

    Composite 500-hPa height anomalies for all KP1 events from 1977 to 2002 for day 5 after onset (shaded; same interval as Fig. 4), and a schematic of the PNA teleconnection pattern index areas. PNA pattern index areas are denoted with a box. Positions of KP1 and KP2 are denoted with an X.

  • View in gallery

    (a) Sea level pressure (contours; interval 4 hPa), 1000–500-hPa thickness (dotted; interval 6 dam), and relative vorticity (shaded; interval 4 × 105 s−1) at 0000 UTC 5 Mar 1998. (b) The 310-K potential vorticity (contours; interval 1 PVU) and 310-K winds (m s−1) at 0000 UTC 5 Mar 1998. (c) Same as in (a) but for 0000 UTC 6 Mar 1998. (d) Same as in (b) but for 0000 UTC 6 Mar 1998. (e) Same as in (a) but for 0000 UTC 7 Mar 1998. (f) Same as in (b) but for 0000 UTC 7 Mar 1998.

  • View in gallery

    The 48-h (a) sea level pressure (red contours; interval 4 hPa), σ = 0.775 relative vorticity (black contours; interval 6 × 10−5 s−1), 1000–500-hPa thickness (dotted; interval 6 dam), and the sensitivity of R1 to relative vorticity on the σ = 0.775 surface (color fill; interval 6 × 106 m2); (b) isotachs on the σ = 0.375 surface (light contours; interval 10 m s−1), sensitivity of R2 to relative vorticity on the σ = 0.375 surface (color fill; 6 × 106 m2 s−1); (c) cross section [orientation indicated in (a)] of relative vorticity (contours; interval 4 × 10−5 s−1) and sensitivity of R1 to relative vorticity (color fill; interval 18 × 106 m2); (d) cross section [orientation indicated in (b)] of isotachs (contours, interval 5 m s−1) and sensitivity of R2 to relative vorticity (color fill; interval 6 × 106 m2 s−1). In (b) the heavier contours indicate the region over which R2 was defined.

  • View in gallery

    (a) Sensitivity of R1 to vorticity at 500 hPa (color fill; interval 6 × 108 m2), 500-hPa height (blue contours; interval 6 dam), and 500-hPa relative vorticity (black contours; interval 4 × 10−5 s−1) at 0000 UTC 5 Mar 1998; (b) cross section from 54.7°N, 119.0°E to 36.4°N, 164.8°E depicting the sensitivity of R1 to vorticity at 500 hPa (color fill; interval 6 × 108 m2), and 500-hPa relative vorticity (black contours; interval 2 × 10−5 s−1) at 0000 UTC 5 Mar 1998. (c) Same as in (a) but for 0000 UTC 6 March 1998. (d) Same as in (b) but for 0000 UTC 6 Mar 1998, and from 39.6°N, 145.4°E to 40.3°N, 169.8°W. (e) Same as in (a) but for 1800 UTC 6 Mar 1998. (f) Same as in (d) but for 1800 UTC 6 Mar 1998. The gray line in (a),(c),(e) depicts the cross section shown at corresponding times in (b),(d),(f), respectively.

  • View in gallery

    (a) The 550–950-hPa layer averaged sensitivity of R2 to relative vorticity (color fill; interval 3 × 105 m2 s−1), 500-hPa height (blue contours; interval 6 dam), and 500-hPa relative vorticity (black contours; interval 4 × 10−5 s−1) at 0000 UTC 5 Mar 1998. (b) Sensitivity of R2 to temperature at 750 hPa (color fill; interval 15 m2 s−2 K−1), 750-hPa temperature (black contours; interval 3°C), and sea level pressure (orange contours; interval 4 hPa) at 0000 UTC 5 Mar 1998. (c) Same as in (a) but for 0000 UTC 6 Mar 1998. (d) Same as in (b) but for 0000 UTC 6 Mar 1998. (e) Same as in (a) but for 1800 UTC 6 Mar 1998. (f) Same as in (b) but for 1800 UTC 6 Mar 1998.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 31 31 6
PDF Downloads 10 10 1

Synoptic–Dynamic Climatology of Large-Scale Cyclones in the North Pacific

View More View Less
  • 1 Department of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, Madison, Wisconsin
© Get Permissions
Full access

Abstract

A climatology of large-scale, persistent cyclonic flow anomalies over the North Pacific was constructed using the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) global reanalysis data for the cold season (November–March) for 1977–2003. These large-scale cyclone (LSC) events were identified as those periods for which the filtered geopotential height anomaly at a given analysis point was at least 100 m below its average for the date for at least 10 days. This study identifies a region of maximum frequency of LSC events at 45°N, 160°W [key point 1 (KP1)] for the entire period. This point is somewhat to the east of regions of maximum height variability noted in previous studies. A second key point (37.5°N, 162.5°W) was defined as the maximum in LSC frequency for the period after November 1988. The authors show that the difference in location of maximum LSC frequency is linked to a climate regime shift at about that time. LSC events occur with a maximum frequency in the period from November through January.

A composite 500-hPa synoptic evolution, constructed relative to the event onset, suggests that the upper-tropospheric precursor for LSC events emerges from a quasi-stationary long-wave trough positioned off the east coast of Asia. In the middle and lower troposphere, the events are accompanied by cold thickness advection from a thermal trough over northeastern Asia. The composite mean sea level evolution reveals a cyclone that deepens while moving from the coast of Asia into the central Pacific. As the cyclone amplifies, it slows down in the central Pacific and becomes nearly stationary within a day of onset. Following onset, at 500 hPa, a stationary wave pattern, resembling the Pacific–North American teleconnection pattern, emerges with a ridge immediately downstream (over western North America) and a trough farther downstream (from the southeast coast of the United States into the western North Atlantic). The implications for the resulting sensible weather and predictability of the flow are discussed. An adjoint-derived sensitivity study was conducted for one of the KP1 cases identified in the climatology. The results provide dynamical confirmation of the LSC precursor identification for the events. The upper-tropospheric precursor is seen to play a key role not only in the onset of the lower-tropospheric height falls and concomitant circulation increases, but also in the eastward extension of the polar jet across the Pacific. The evolution of the forecast sensitivities suggest that LSC events are not a manifestation of a modal instability of the time mean flow, but rather the growth of a favorably configured perturbation on the flow.

Corresponding author address: Linda M. Keller, Department of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, 1225 W. Dayton Street, Madison, WI 53706. Email: lmkeller@wisc.edu

Abstract

A climatology of large-scale, persistent cyclonic flow anomalies over the North Pacific was constructed using the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) global reanalysis data for the cold season (November–March) for 1977–2003. These large-scale cyclone (LSC) events were identified as those periods for which the filtered geopotential height anomaly at a given analysis point was at least 100 m below its average for the date for at least 10 days. This study identifies a region of maximum frequency of LSC events at 45°N, 160°W [key point 1 (KP1)] for the entire period. This point is somewhat to the east of regions of maximum height variability noted in previous studies. A second key point (37.5°N, 162.5°W) was defined as the maximum in LSC frequency for the period after November 1988. The authors show that the difference in location of maximum LSC frequency is linked to a climate regime shift at about that time. LSC events occur with a maximum frequency in the period from November through January.

A composite 500-hPa synoptic evolution, constructed relative to the event onset, suggests that the upper-tropospheric precursor for LSC events emerges from a quasi-stationary long-wave trough positioned off the east coast of Asia. In the middle and lower troposphere, the events are accompanied by cold thickness advection from a thermal trough over northeastern Asia. The composite mean sea level evolution reveals a cyclone that deepens while moving from the coast of Asia into the central Pacific. As the cyclone amplifies, it slows down in the central Pacific and becomes nearly stationary within a day of onset. Following onset, at 500 hPa, a stationary wave pattern, resembling the Pacific–North American teleconnection pattern, emerges with a ridge immediately downstream (over western North America) and a trough farther downstream (from the southeast coast of the United States into the western North Atlantic). The implications for the resulting sensible weather and predictability of the flow are discussed. An adjoint-derived sensitivity study was conducted for one of the KP1 cases identified in the climatology. The results provide dynamical confirmation of the LSC precursor identification for the events. The upper-tropospheric precursor is seen to play a key role not only in the onset of the lower-tropospheric height falls and concomitant circulation increases, but also in the eastward extension of the polar jet across the Pacific. The evolution of the forecast sensitivities suggest that LSC events are not a manifestation of a modal instability of the time mean flow, but rather the growth of a favorably configured perturbation on the flow.

Corresponding author address: Linda M. Keller, Department of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, 1225 W. Dayton Street, Madison, WI 53706. Email: lmkeller@wisc.edu

1. Introduction

Changes in circulation patterns over the North Pacific are manifest in a number of variables, including the sea level pressure and 500-hPa geopotential height fields, and the intensity and position of the midlatitude storm tracks (Wu and Straus 2004; Lu et al. 2004; Chang et al. 2002). There are certain anomalous weather patterns that persist on time scales longer than typical synoptic-scale variability. These events, which include blocking, teleconnection patterns, and other realizations of persistent flow anomalies (e.g., Dole 1986a, b, 1989; Dole and Black 1990; Wallace and Gutzler 1981; Barnston and Livezey 1987) are associated with significant and important changes in the extratropical general circulation and related weather sequences. One of the most widely recognized patterns in the North Pacific is the Pacific–North American (PNA) teleconnection pattern (Wallace and Gutzler 1981). Several observational and modeling studies indicate that the PNA, a stationary wave pattern, completes its cycle of growth and decay in about two weeks (Feldstein 2000; Cash and Lee 2001; Feldstein 2002).

It is widely believed that predictability in the medium range (7–14 days) is diminished during the development of these flow patterns, and it has been observed that once the flows are established, predictability is enhanced. Palmer (1988) and Molteni and Palmer (1993) have demonstrated the dependence of predictability on the phase of the PNA teleconnection pattern. The differences in predictability could be related to changes in the characteristics of the most rapidly growing perturbations of the observed flow.

Episodic large-scale cyclogenesis events (LSCs) over the North Pacific (Dole and Gordon 1983, hereafter DG; Dole 1986a, b, 1989; Dole and Black 1990; Black and Dole 1993) are a particularly interesting class of a medium-range weather phenomenon that potentially has a significant impact on predictability over North America. The articles cited above examined the life cycle and dynamics of LSCs that occurred during the period 1963–87 and determined that the mechanism for development of these persistent anomalies was either a large-scale instability of, or an initial value development upon, the time-mean flow. Additionally, the potential vorticity (PV)-based analyses of Black and Dole (1993) revealed temporal variations in the structure of the developing cyclones that suggested the role of nonmodal growth of the LSCs.

In this study, the results of an updated and extended climatology of LSC events are presented for the period 1977–2003. In section 2, the data used and the methodology are described. The possibility of a regime change, which is defined as a change in the large-scale atmospheric circulation patterns as well as the sea surface temperature patterns, during the period is explored. Results of the climatology of the events are presented in section 3 and the composite evolution of the events is found in section 4. The initial results of an adjoint-based forecast sensitivity study, designed to identify the characteristics of the precursors to LSC events, are presented in section 5. Future research directions are briefly described in section 6.

2. Data and methodology

The dataset used in the initial phase of this work is the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) global reanalysis data (Kalnay et al. 1996) of 500-hPa geopotential height and sea level pressure. The period covered in this study is the cold season (November–March) for 1977/78 through 2002/03. The reanalysis data are on a 2.5° × 2.5° grid available every 6 h. For the purposes of this study, only the Northern Hemisphere will be examined.

Twenty-five years of the study period (1977/78–2001/02) were used to calculate the climatology. At each grid point, the 25 values for each 6-h analysis time were averaged to form the climatology. Deviations of the geopotential height from climatology were calculated by subtracting the climatology from the analyses for each time period. The time series of these deviations at each grid point were low-pass filtered with a 151-point Lanczos filter (Duchon 1979) with a cutoff frequency of 10 days. This filtering allows for an examination of the background or lower-frequency flow. An example of the application of the filter for the period July 2000–June 2001 is shown in Fig. 1a. A closer examination of an event in December 2000 is shown in Fig. 1b. The time series used all months of the year and was started in July, so any end effects of the filtering are well removed from the time interval being studied. The height anomalies were normalized by the sine of the latitude (DG) so that the geopotential height gradient is proportional to the geostrophic wind. These normalized, filtered geopotential height deviations will hereafter be referred to as height anomalies.

Following DG, the filtered time series at each grid point was examined for each cold season for 500-hPa height anomalies, which lasted for at least 10 days and were below −100 m. An examination of our climatology showed that the number of events decreased rapidly as the duration increased (Fig. 2). The decision to pick 10 days for the duration of the events was based on the desire to have events that were longer than typical synoptic-scale variations and to have a sufficient number of events to work with. In this study, an additional requirement was that events had to be separated by at least 10 days so that precursor events could be isolated. While this study follows the general method found in DG, there are some specific differences in the anomaly dataset. The anomalies in DG were calculated as the observed geopotential heights minus the long-term seasonal trend, which was defined by the least squares quadratic fit of the time series of geopotential height for each calendar day at each grid point. Even though this study used a somewhat different methodology than DG, many of the LSC events in the overlapping years of both studies were identified, although the starting times of the events differed. Thus, these events are quite robust, not being dependent on the choice of dataset or exact replication of method.

3. Climatology

Examination of the frequency of occurrence of large-scale cyclogenesis events in the North Pacific for the entire 25-yr period shows several areas of maximum activity (Fig. 3a). The grid point at 45°N, 160°W was chosen to represent the area of maximum activity and will be referred to as key point 1 (KP1). We note that this region of maximum large-scale cyclone activity is to the east and southeast of the DG analysis, which was around 45°N, 170°W. We attribute the differences in location of the maximum activity to the differences in and preparation of the data noted above.

Variations in the North Pacific sea level pressure exhibit quasi-decadal shifts in magnitude and position of the Aleutian low. The shift in 1976–77 has been well documented (Trenberth 1990; Trenberth and Hurrell 1994; Honda et al. 2001; Deser et al. 2004). The possibility of another “regime shift” in 1988–89 has been explored by Yasunaka and Hanawa (2002). In addition, many reports of changes in marine ecosystems during 1988–89 indicate a climate shift may have occurred (Hare and Mantua 2000; Mantua et al. 1997; Chavez et al. 2003; Overland and Stabeno 2004). The Arctic Oscillation also shows a shift in 1988 (Overland et al. 1999) with a return to more neutral conditions after 1995 (NOAA/Climate Prediction Center; see information online at http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/month.ao.gif).

To examine the effect of this possible change in background state on the climatology of the LSCs, the events were partitioned into those occurring before November 1988 and those occurring after this time. The events before 1988 (hereafter referred to as events from pre-1988) are located mainly north of 40°N and extend from 180° to 150°W in an arc across the central Pacific (Fig. 3b). The maximum area of occurrence for events for the period from 1988 on (post-1988) also lies between 180° and 150°W but is found primarily south of 40°N (Fig. 3c). The grid point at 37.5°N, 162.5°W was chosen to represent this area and will be referred to as key point 2 (KP2). For the 26-yr period of the study, 33 events were identified at KP1 and 29 events were identified at KP2 (Table 1). The events at KP1 and KP2 are not necessarily independent in time because of their large spatial scale and duration.

For KP1 most of the events occur from November to January with events spread evenly before and after 1988. January has the maximum number of events. For KP2, most events occurred in November–February with the maximum still occurring in January. In the pre-1988 dataset, events are more likely to occur from January to March, while in the post-1988 dataset the events occur more frequently from November to January.

4. Anomaly structure and evolution

Composited hemispheric 500-hPa height and sea level pressure analyses were produced for events identified for the two key points (animations for the composite evolution and individual events may be found online at http://aurora.aos.wisc.edu/lazear/LSC_climatology.html). The onset time for each event was defined as the time when the 500-hPa height anomaly at the representative key point dropped below −100 m. The composites cover a time span of 15 days (60 time periods) starting 5 days prior to onset to elucidate the flow pattern evolution of an event.

a. Composites for 1977–2002

The composite evolution of the 500-hPa anomalies, full geopotential height fields, sea level pressure, and thickness fields for KP1 for all 25 yr are shown in Fig. 4 for days −5, 0, and +6 relative to onset. Five days prior to onset of the event (Fig. 4a), the composite 500-hPa height field reveals weak positive geopotential height anomalies near KP1. A trough in the 500-hPa height field is anchored over the east coast of Asia. At the surface, a large cyclonic circulation center is situated to the southeast of the Kamchatka Peninsula with an extension into the Gulf of Alaska (Fig. 4b). This feature is maintained by a succession of lows coming from Kamchatka and south of Japan. Downstream of a large Mongolian anticyclone, a 1000–500-hPa thickness trough extends from Siberia to eastern Asia and Japan reflecting the cold air associated with this high. Farther east, thermal ridges have formed ahead of the surface low in the Pacific and south of Greenland.

By onset (Fig. 4c), the positive height anomalies in the central Pacific are replaced by negative height anomalies, and the 500-hPa height trough in the Pacific extends eastward to south of the Aleutians. The surface low moves to a position near the western Gulf of Alaska, the cold thermal trough extends out over the central Pacific (Fig. 4d), and the thermal ridge continues to build off the west coast of North America.

By onset +6 (days), the composite negative height anomalies deepen below −150 m (Fig. 4e). Downstream, positive height anomalies are located over western North America associated with the amplifying ridge, with the largest anomalies over Alberta and Saskatchewan, Canada. As the anomalies continue to amplify, a negative height anomaly develops off the southeastern United States. This hemispheric pattern is similar to the second principal component in Quadrelli and Wallace (2004), which they refer to as “PNA-like.” The surface low has continued to deepen and covers the entire North Pacific basin while the thermal ridge remains over the west coast of North America (Fig. 4f). A surface anticyclone covers most of the United States while the strong thermal trough continues over the East Coast. The anticyclone that develops over the United States advects air of mostly Pacific origin into southwestern Canada as the strength and size of the cyclone in the Pacific effectively cuts off flow from Siberia to Canada to the central plains of the United States.

For KP2 at onset −5, the 500-hPa height trough is anchored in its climatological position over the east coast of Asia, with weak, positive height anomalies in the central Pacific (Fig. 5a). The surface low has a broad center from south of Kamchatka into the northern Gulf of Alaska. The cold thermal trough extends farther into the Pacific than was seen for KP1 at this time (Fig. 5b).

At the onset of the large-scale cyclogenesis event, negative geopotential height anomalies develop over Japan (Fig. 5c), and negative height anomalies appear in the Pacific. The 500-hPa height trough is not as extended as for KP1, and the anomalies are farther south. A broad area of low pressure at the surface is also farther south and less intense at this stage than for KP1 (Fig. 5d). The extension of the anticyclone from Siberia across the Arctic into North America allows very cold air to funnel into the central United States (“the Siberian express”).

At onset +6 the central Pacific anomalies are deeper than −150 m, and the trough in the 500-hPa height field extends across the Pacific (Fig. 5e). A positive height anomaly develops over the west coast of the United States, and an area of negative height anomalies begins to appear over Nova Scotia. This negative area eventually will cover most of the east coast of the United States. At the surface, the anticyclone over Mongolia moves eastward and shuts down the source region for cyclones forming south of Japan (Fig. 5f).

To determine if the composite anomaly evolution of the KP1 events are significantly different than the anomaly evolution of KP2 events, a difference-of-means test was performed. Using a two-sided t test, the null hypothesis of no significant difference can be rejected at the 5% significance level. Starting 1.5 days before onset of an LSC event, areas of significant differences over the central Pacific begin to appear at 500 hPa. As the LSC events mature, these areas spread into southern Asia and the western Pacific as well as the Aleutians (e.g., onset +7, Fig. 6a). At the surface, the areas of significant differences begin in the central Pacific. Areas in the western Pacific and over eastern Asia develop in the storm-track regions as the event matures (Fig. 6b).

b. Comparisons between composites of 1977–88 and 1988–2002

To examine the effect of the regime shift described earlier, composites of 500-hPa geopotential heights and anomalies were made for the periods 1977–88 and 1988–2002 for KP1 and KP2. The anomaly patterns for KP1 for the two periods present somewhat different scenarios. For the pre-1988 period, as the event develops in the Pacific an area of negative height anomalies develops on the east coast of North America, and positive anomalies extend all the way from the Rockies to Greenland (e.g., onset +6; Fig. 7a). At sea level, a large anticyclone sits over the western and central United States, but the coldest air stays in Siberia (Fig. 7b). A strong thermal ridge amplifies in the eastern Atlantic as the event develops. The post-1988 period (1988–2002) has both positive and negative 500-hPa geopotential height anomalies that are more intense than those formed during the pre-1988 period (Fig. 7c). A large anticyclone in Mongolia broadens to the southeast and eliminates the area south of Japan as a source region for lows moving into the developing event. As the LSC develops in the Pacific and rolls into the Gulf of Alaska, the Siberian air is pinched off from the west as in the pre-1988 period (Fig. 7d). A cyclone that develops west of Greenland and extends westward effectively closes the path for the Siberian air. As the event matures, a classic PNA pattern of anomalies develops in both time periods.

As the KP2 events develop in the pre-1988 period, negative 500-hPa height anomalies are found over eastern Asia, the central Pacific, the east coast of North America and western Europe (e.g., onset +6; Fig. 8a). Positive height anomalies develop over the Rockies several days after onset. Surface cyclones remain south of the Aleutians and move into the Gulf of Alaska (Fig. 8b). Upstream, the large anticyclone resides over Mongolia and moves southeastward. The Pacific cyclones intensify earlier in the pre-1988 period with a tighter gradient associated with the thermal trough off the east coast of Asia. Because the KP2 LSCs in the Pacific stay south of the Aleutians, a path is open for Siberian air to plummet into the central United States. Strong cyclones develop over the Maritime Provinces of Canada while a thermal ridge develops in the 1000–500-hPa thickness field over the Atlantic. For the post-1988 period, the positive height anomalies develop over the Rockies shortly after onset (Fig. 8c), but the negative height anomalies off the East Coast develop quite late in the event and then are positioned farther north and east than for the pre-1988 period. The Pacific cyclone moves closer to the Alaskan coast, and the cyclones in the Maritime Provinces move east of Greenland, allowing periodic injections of cold air into the central United States (Fig. 8d).

Difference of means tests were performed on both KP1 and KP2 events to see if the pre- and post-1988 events for each had significant geographical or temporal differences. An examination of the anomaly patterns at day +8.5 shows the locations of the main differences between the pre- and post-1988 periods (Fig. 9). For KP1, the null hypothesis of no difference can be rejected at the 5% level for a two-sided t test for the main anomaly areas in the central Pacific and for the Gulf of Alaska and the Rockies (Fig. 9a). For KP2, the anomaly area in the central Pacific is significant at the 95% confidence level. Other significant anomaly areas are over Asia and the Middle East (Fig. 9b). For the area of interest in the Pacific, the areas of significant differences for KP1 are more widespread, whereas for KP2, the areas are more compact.

c. Nonevent composites

Two sets of composite evolutions of 15-day sequences of 500-hPa height anomalies were constructed for periods during which LSC events were not occurring in the Pacific. We refer to these sequences as nonevents. The composites were prepared to correspond to the temporal distribution of the large-scale events for KP1 and KP2 (e.g., the same number of nonevents in November, December, etc., as there were events, the same frequency distribution before and after 1988, etc.). The most notable characteristic of the anomaly patterns is the much smaller anomaly amplitude (not shown). Furthermore, the classic PNA pattern does not emerge in the composites.

d. Cyclone tracks

Root-mean-square (RMS) fields for 500-hPa unfiltered height anomalies were calculated to examine the tracks of the cyclones associated with the large-scale cyclone events (Fig. 10). RMS fields for each event were averaged together to produce mean storm tracks for KP1 and KP2. Both key points feature tracks across the Pacific, starting near Japan. KP1 events move into the eastern Pacific around 45°N and curve toward Alaska (Fig. 10a). KP2 events also move into the eastern Pacific but stay farther south, closer to 35°N. The cyclones have a much more pronounced curve to the north and northwest into the Bering Sea (Fig. 10b). In addition, KP2 RMS fields indicate a second track extending from Kamchatka into the large-scale cyclones. The nonevent composite for KP1 shows a more zonal track extending across the Pacific and into California (Fig. 10c) along with less variability.

e. Relation to PNA patterns

Because PNA episodes have characteristics such as an extended Asian jet in common with LSC events, an examination of how the LSC events are related to the PNA patterns was performed. The PNA index for normalized 500-hPa height anomalies at each observation time (i.e., every 6 h) was calculated using the four areas defined by the Climate Diagnostics Center (15°–25°N, 180°–140°W; 40°–50°N, 180°–140°W; 45°–60°N, 125°–105°W; 25°–35°N, 90°–70°W; Fig. 11). We defined a PNA episode as whenever the PNA index was 1.5 standard deviations above its long-term mean. Episodes that were 2 days or shorter were removed. Examining our set of LSC events for KP1, 27 of the 31 cases occurred when a PNA episode was identified by our definition. The PNA episode started at onset or later in 25 of those 27 events. Because KP1 is included in one of the areas used to compute the PNA index, the large number of coincident PNA episodes with LSC events is not surprising.

During the period of study, there were 55 PNA episodes. Decreasing the duration or using less intense anomalies for an LSC event added a few more LSC events to those occurring during a PNA episode, but having a PNA episode in place did not guarantee that an LSC event would occur.

KP2 events are characterized by an extended Asian jet that lies farther south than the jet in KP1 events. This places KP2 south of the second area of the PNA index. Because of the way we define the PNA index, fewer KP2 events are associated with PNA episodes. Only 16 out of 27 LSC events are coincident with a PNA episode. In addition, the PNA episodes start at onset or later for only 11 out of those 16 events. A PNA-like pattern does develop, but with the centers offset from the classic PNA locations.

5. Adjoint case study of event onset

An examination of individual KP1 LSC events reveals both the emergence of an upper-tropospheric, short-wave trough from the east Asian coast prior to Pacific large-scale cyclogenesis and the subsequent development of a large-scale, mid-Pacific, lower-tropospheric cyclonic circulation and concomitant extended Pacific jet that are characteristic of LSC events. While the synoptic evolution of the 500-hPa heights and mean sea level pressure for both the individual events and composite event suggest that the upper-tropospheric short wave is a precursor to the surface cyclone and associated extended jet, we seek dynamical evidence for this link and wish to determine how changes in the structure of the presumed precursor trough might affect the subsequent LSC development. To this end, we employ the adjoint of a forecast model as a tool to derive the sensitivities of cyclone and jet intensity at LSC onset to the initial conditions of a simulation of an LSC event. If the adjoint-derived forecast sensitivities confirm the identification of an East Asian short-wave upper trough as the precursor to the LSC event, this information could be used to design observational targeting strategies to better improve forecasts of LSC onset and enhance predictability of the flow over the Pacific and North America. Below, we present the results of such a sensitivity study for a single KP1 LSC event. While acknowledging the dangers of generalizing the results of a single case, the results presented below are typical of several other cases we have also examined. We intend to construct a composite of the sensitivity fields in a subsequent paper.

a. Case description

The case chosen in this paper was a KP1 LSC event that began 0000 UTC 7 March 1998 and ended 1200 UTC 20 March 1998. Of the 31 cases at KP1, this case had the distinction of having the smallest RMS differences between the 500-hPa height anomalies for the case and the composite anomalies for the 5-day period leading up to and including the onset of the event.

Two days prior to onset of the event, 0000 UTC 5 March 1998, two separate cyclogenesis events occurred in the western and central Pacific. The cyclogenesis event directly associated with the decrease of 500-hPa geopotential heights to more than 100 m below the climatological average for that date at KP1 was associated with a surface cyclone (mean sea level pressure 1003 hPa) located to the southeast of the Kamchatka Peninsula (Fig. 12a). This cyclone (hereafter referred to as “cyclone A”) was located in an environment of enhanced baroclinicity along the periphery of an upper trough extending from the Kamchatka Peninsula to the east of Hokkaido. A 500-hPa short-wave trough was located upshear (in a thermal wind sense) of the surface of the cyclone. The configuration of the 500-hPa trough and attendant relative vorticity were favorable for modest baroclinic development over the next 24 h. By 0000 UTC 6 March, the cyclone had deepened to 996 hPa (Fig. 12c). Intensification of cyclone A over the subsequent 18 h leading to onset was weaker as the upper trough had nearly overtaken the surface cyclone (Fig. 12e).

The second, more robust surface, baroclinic development (hereafter referred to as “cyclone B”) occurred in a region of enhanced baroclinicity farther south along and west of cyclone A. Cyclone B was triggered by an upstream, upper trough located over Kyushu on 0000 UTC 5 March (Fig. 12a). Cyclone B then deepened substantially from 1007 to 990 hPa by 0000 UTC 6 May (Fig. 12c). Particularly notable about this cyclogenesis event was the amplification of the upper-tropospheric ridge (characterized by low potential vorticity) downstream of cyclone B (Figs. 12b,d,f). The evolution of the PV on the 310-K isentropic surface suggested the ridge amplification was attributable, in part, to nonconservative processes, particularly mid- and lower-tropospheric latent heating.

b. Forecast and adjoint models

For this study, version one of the fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model (MM5) Adjoint Modeling System (Zou et al. 1997) is utilized. This modeling system includes the nonlinear model, its tangent linear model, and corresponding adjoint. The MM5 is a nonhydrostatic, limited-area, primitive equation model that uses a terrain-following sigma vertical coordinate.

The model domain used for this study, whose lateral extent is shown in Fig. 12, is at a 90-km horizontal resolution with 20 evenly spaced sigma levels up to a model top at 100 hPa. The simulation is initialized using reanalysis data interpolated to the MM5 grid, and the lateral boundaries are updated using the reanalysis data. Physical parameterizations used in the simulation include the Grell convective scheme, a bulk aerodynamic formulation of the planetary boundary layer, horizontal and vertical diffusion, dry convective adjustment, and explicit treatment of cloud water, rain, snow, and ice. Further description of the modifications made to the adjoint code are described in Kleist and Morgan (2005). The simulation was 48 h in length, verifying at 0000 UTC 7 March 1998. The evolution and final forecasted positions and intensities of cyclones A and B (Fig. 13a) agreed well with analyses (Fig. 12e).

c. Definition of response functions

For this case two response functions (defined at a forecast time of 48 h) are chosen that characterize the cyclogenesis event (associated with cyclone A) in the central Pacific and concomitant onset of the large-scale cyclone event. The first response function R1 is defined as the circulation about a box (shaded green in Fig. 13a) encircling the lower-tropospheric vorticity maximum associated with the surface cyclone:
i1520-0493-134-12-3567-eq1
where V is the horizontal wind vector, and dl is a differential element along the path enclosing that box. This response function measures the intensity of the lower-tropospheric cyclonic circulation. The sensitivity of R1 to the initial vorticity (∂R1/∂ζ) on the σ = 0.775 surface at 48 h (initial condition for adjoint model) is shown in Fig. 13a.1 An approximate north–south cross section of this sensitivity and the 48-h forecast of relative vorticity are shown in Fig. 13c. It is noted that the vorticity is nearly barotropic at this time with two maxima—one in the lower troposphere associated with the surface cyclone, the other in the upper troposphere with the upper trough.
The second response function R2 is defined as twice the volume-integrated kinetic energy per unit mass within the 30 m s−1 isotach maximum associated with the leading edge of the jet extension across the Pacific,
i1520-0493-134-12-3567-eq2
where ℜ defines the region for which the magnitude of the 48-h forecast wind is at least 30 m s−1. The sensitivity of R2 to vorticity (∂R2/∂ζ) on the σ = 0.375 surface at 48 h along with the isotachs of wind speeds above 30 m s−1 are shown in Fig. 13b. An approximate north–south cross section taken through the jet shows the distribution of the sensitivities of R2 to vorticity (Fig. 13d). Note that these sensitivities are characterized by a dipole—implying that increasing (decreasing) the vorticity to the north (south) leads to a more intense zonal jet at 48 h. That the distribution of these sensitivities for R2 extends outside 30 m s−1 is consistent with the fact that the wind attributed to a point vortex extends over a much broader region than the vortex itself. As a consequence, placing a point vortex in the sensitive region outside of 30 m s−1 can still have an effect on R2.

d. Sensitivity results

1) Surface cyclone intensity and precursor identification

Inspection of the 500-hPa distributions of relative vorticity and the sensitivities of R1 with respect to the initial relative vorticity distribution (∂R1/∂ζ; Fig. 14a) reveals maximum positive sensitivities in the vicinity of the upper-tropospheric short-wave trough (blue shaded), which triggered the surface cyclone. This region of positive sensitivities extends over much of Japan to the south-southwest of the upper trough. The configuration of sensitivities relative to the initial vorticity distribution implies that a more intense upper trough at this location at the initial forecast time would lead to a larger cyclonic circulation at later times. The region of positive sensitivities extends eastward into the central Pacific. That the positive sensitivities extend over the downstream shortwave ridge suggests that a weakened ridge would support a more vigorous lower-tropospheric cyclone 48 h later. The sensitivities display a baroclinic upshear tilt at the initial time (Fig. 14b). Within 24 h, the positive (negative) sensitivity patterns become more localized to the upper trough (downstream ridge) over the Pacific (Fig. 14c), while the positive sensitivities to the upstream ridge have decreased in magnitude. By this time, the baroclinic tilt of the sensitivities has decreased as has the tilt between the vorticity maximum associated with cyclone A and its upstream, upper-tropospheric precursor (Fig. 14d). By 42 h, the maximum sensitivities at 500 hPa have decreased in magnitude and are localized to the vicinity of the upper-tropospheric short wave (Figs. 14e,f).

2) Jet streak intensity

The sensitivities of the magnitude of the upper-tropospheric jet streak to the relative vorticity at various times along the forecast trajectory show large, positive sensitivities to vorticity with the upper-tropospheric precursor short wave (Fig. 15a). Unlike the sensitivities for the lower-tropospheric circulation, these sensitivities are geographically more sequestered to the upper trough, though there are some relatively large, positive sensitivities over the downstream ridge. It is of interest to note that the upper trough associated with the more robust cyclone southwest of the lead cyclone lies in regions of negative (Fig. 15a), relatively neutral (Fig. 15c), and then positive sensitivities (Fig. 15e) over the 48-h forecast period. This suggests that early on, the final time jet streak would be more intense with a weaker upstream trough during the first 18 h of the simulation, while during the next 18 h or so, the final time jet streak would be stronger if the upstream vortex were more intense. The region of negative sensitivities amplifies downstream while propagating faster than the short-wave trough. At 24 h into the simulation (Fig. 15c), the region of negative sensitivity to vorticity encompasses much of the ridge downstream of short-wave B. By construction, the negative sensitivities end up to the south of the upper trough at 42 h (Fig. 15e). The evolution of this region of negative sensitivities to vorticity from 24 to 42 h into the simulation suggest that mechanisms that would favor upper-tropospheric 500-hPa ridge development (including enhanced lower-tropospheric warm advection and latent heating) would lead to a more intense upper jet maximum.

Sensitivities of R2 with respect to lower-tropospheric temperature (∂R2/∂T) at the initial time (Fig. 15b) reveal that the largest sensitivities lie in the regions of greatest baroclinicity close to the upper-tropospheric short waves associated with cyclones A and B. The largest negative sensitivities to lower-tropospheric temperature are found upstream of the surface low. These large, negative sensitivities remain nearly collocated with the thermal trough to west of the surface cyclone A through 48 h. Positive temperature sensitivities initially to the west of the cyclone B are found in that cyclone’s warm sector in 24 h (Fig. 15d) and in and to the east of the thermal ridge axis 18 h later (Fig. 15f). Based upon the distribution of sensitivities to temperature at this time, the implied temperature perturbations required to increase R2 is consistent with thermal wind balance. Increasing the magnitude of the thermal gradient in the lower troposphere would be associated with increased wind shear and concomitant large jet speeds. While the results of the sensitivity of R1 to vorticity indicate the importance of the cyclone A precursor short wave and associated cold air, sensitivities to R2 highlight the importance of the thermal structure, possibly amplified by latent heating, associated with cyclone B.

6. Summary and directions for future work

a. Summary

Among the unresolved issues associated with LSCs are the dynamics of their onset (i.e., are they a realization of a modal instability of the large-scale flow, as opposed to an amplification of a favorably configured perturbation to the larger-scale flow?). Cash and Lee (2001), using a linear-stochastic model of the Northern Hemisphere wintertime circulation, demonstrate that the PNA arises through the optimal nonmodal growth associated with a constructive interference of decaying empirical normal modes of the unforced operator of their model. To the extent that the dynamics of the PNA and LSCs are closely related, the Cash and Lee results suggest that LSC development might be attributed to the amplification of an optimally configured perturbation of the large-scale flow over the North Pacific. Closely related to an improved understanding of LSC development are the predictability of LSC onset and the predictability of the resultant downstream flow. Because LSCs are associated with strong realizations of one phase of the PNA pattern, they are of particular interest as they are associated with significant weather anomalies over much of North America (Namias 1978). Over the eastern half of North America, cold-air outbreaks have been associated with the PNA teleconnection pattern. As a consequence, understanding the development of LSCs is likely important for medium- and extended-range forecasts during the cold season over North America.

The results of a North Pacific LSC climatology for the period November 1977–March 2002 identified at KP1 a maximum frequency of occurrence of 500-hPa, large-scale, persistent, cyclonic flow anomalies. At KP1, 31 events were observed over the 25 yr of the climatology. A second key point (KP2), where 27 events occurred, was defined as the maximum in LSC frequency for the period after November 1988. The shift in location of maximum event frequency is associated with a regime shift that occurred at about that time. LSC events are primarily a cold season phenomenon (November–March) with maximum frequency occurring in the period from November to January. While there are no significant changes in event frequency across the regime shift at KP1, at KP2 there are more events in November and January after 1988. The shift in the location of maximum event frequency is associated with a southwestward shift in the 500-hPa storm track as defined by the 500-hPa RMS geopotential height anomaly differences. These differences provide further evidence of a signal for a Pacific-wide weather regime change. Furthermore, the nonevent composite RMS fields reveal lower variability through the central Pacific when compared with LSC events. This suggests that fewer or perhaps much less vigorous eddies cross the Pacific during nonevent time intervals.

Important differences exist between events at KP1 and those at KP2. During the development of events at KP1, the storm track is directed eastward and then northward into the Gulf of Alaska, while for KP2, the storm track is farther to the south with a more exaggerated curve to the north and northwest into the Gulf of Alaska. The differences between the storm tracks during the onset of these events have important implications for weather over the continent of North America—cold-air outbreaks appear to be favored during KP2 LSC events, while advection of air of Pacific origin into southwest Canada during KP1 events mitigates the effects of cold-air outbreaks. Development at both key points generally occurs during a PNA-like episode, although the centers for the PNA episodes during KP2 events are rotated counterclockwise from the classic positions.

An adjoint-derived sensitivity study was conducted for KP1 using response functions separately measuring the kinetic energy of the jet streak and the lower-tropospheric vorticity at LSC onset near KP1. The forecast sensitivities calculated for this case highlight the importance of the precursor upper-tropospheric trough associated with the cyclone arriving at KP1 at onset (cyclone A). Amplification of this synoptic feature 48 h prior to the onset of the case would lead to a more robust lower-tropospheric cyclonic circulation and stronger isotach maximum in the upper troposphere at the case onset. The sensitivity calculations also suggest the possible importance of nonconservative processes, associated with the upstream cyclone (cyclone B), in increasing and extending the upper-tropospheric jet farther east at onset. That the adjoint-derived forecast sensitivities are confined and maximized in regions of strong baroclinic and barotropic shear, and relatively remote from the Tropics and subtropics suggests that the dynamics of the onset, for this case, may largely be driven by midlatitude baroclinic and barotropic growth of a favorably configured perturbation of the North Pacific flow, perhaps amplified by latent heat release in the ascent region of cyclone B. This result is consistent with that of Black and Dole (1993) who suggested that LSCs were not forced by anomalous tropical convection. Furthermore, given the evidence linking the occurrence of the PNA with LSC onset at KP1, this result is also consistent with the results of Cash and Lee (2001) who suggest that PNA initiation is dependent on the location and configuration of an optimal perturbation relative to the stationary wave pattern.

b. Future work

The results of this study suggest several directions for further work. The authors have begun constructing a climatology of adjoint-derived forecast sensitivity patterns for the response functions presented in section 5 to better understand the distribution and variability of the sensitivity to initial conditions for LSC events. The results of this work would provide information useful in assessing where additional observations might be taken to improve forecasts of LSC events. Additionally a climatology of singular vector (SV) amplification factors and initial structure for energy norm SVs localized over the north-central Pacific and North America optimized for 48–96-h forecasts verifying at LSC onset time (onset +5) is also ongoing. The results of these studies over the Pacific coupled with the results presented here should reveal the relationship between the disturbances antecedent to LSC development and the flow within which these disturbances develop. The results for the studies at onset +5 for verifying regions in the North Pacific (North America) will provide important information on the susceptibility of the larger-scale flow to small perturbations that could subsequently disrupt the developed pattern of low-frequency variability over the central Pacific.

While this study focuses on cold season LSC events, LSC events do occur during the warm season. During the April–October periods from 1977 to 2002, 23 warm season events were identified at KP1. The authors are planning to examine the synoptic climatology of these warm season events to gain a greater understanding of the minimum conditions necessary for the large-scale events to form.

Acknowledgments

The authors thank Professors Robert X. Black of the Georgia Institute for Technology and Daniel J. Vimont of the University of Wisconsin—Madison for their helpful comments at various stages of this work. The authors also thank an anonymous reviewer for several helpful suggestions that improved the paper. The authors acknowledge use of the NCAR Data Support Services archive to access the NCEP–NCAR Global Reanalysis dataset (ds090.0). This project was supported by National Science Foundation Grant ATM-0121186.

REFERENCES

  • Barnston, A. G., , and R. E. Livezey, 1987: Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev., 115 , 10831126.

    • Search Google Scholar
    • Export Citation
  • Black, R. X., , and R. M. Dole, 1993: The dynamics of large-scale cyclogenesis over the North Pacific Ocean. J. Atmos. Sci., 50 , 421442.

    • Search Google Scholar
    • Export Citation
  • Cash, B. A., , and S. Lee, 2001: Observed nonmodal growth of the Pacific–North American teleconnection pattern. J. Climate, 14 , 10171028.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., , S. Lee, , and K. L. Swanson, 2002: Storm track dynamics. J. Climate, 15 , 21632183.

  • Chavez, F. P., , J. Ryan, , S. E. Lluch-Cota, , and M. Niquen, 2003: From anchovies to sardines and back: Multidecadal change in the Pacific Ocean. Science, 299 , 217221.

    • Search Google Scholar
    • Export Citation
  • Deser, C., , A. Phillips, , and J. Hurrell, 2004: Pacific interdecadal climate variability: Linkages between the Tropics and the North Pacific during boreal winter since 1900. J. Climate, 17 , 31093124.

    • Search Google Scholar
    • Export Citation
  • Dole, R., 1986a: Persistent anomalies of the extratropical Northern Hemisphere wintertime circulation: Structure. Mon. Wea. Rev., 114 , 178207.

    • Search Google Scholar
    • Export Citation
  • Dole, R., 1986b: The life cycles of persistent anomalies and blocking over the North Pacific. Advances in Geophysics, Vol. 29, Academic Press, 31–69.

  • Dole, R., 1989: Life cycles of persistent anomalies. Part I: Evolution of 500 mb height fields. Mon. Wea. Rev., 117 , 177211.

  • Dole, R., , and N. D. Gordon, 1983: Persistent anomalies of the extratropical Northern Hemisphere wintertime circulation: Geographical distribution and regional persistence characteristics. Mon. Wea. Rev., 111 , 15671586.

    • Search Google Scholar
    • Export Citation
  • Dole, R., , and R. X. Black, 1990: Life cycles of persistent anomalies. Part II: The development of persistent negative height anomalies over the North Pacific Ocean. Mon. Wea. Rev., 118 , 824846.

    • Search Google Scholar
    • Export Citation
  • Duchon, C. E., 1979: Lanczos filtering in one and two dimensions. J. Appl. Meteor., 18 , 10161022.

  • Feldstein, S. B., 2000: The timescale, power spectra, and climate noise properties of teleconnection patterns. J. Climate, 13 , 44304440.

    • Search Google Scholar
    • Export Citation
  • Feldstein, S. B., 2002: Fundamental mechanisms of the growth and decay of the PNA teleconnection pattern. Quart. J. Roy. Meteor. Soc., 128 , 775796.

    • Search Google Scholar
    • Export Citation
  • Hare, S. R., , and N. J. Mantua, 2000: Empirical evidence for North Pacific regime shifts in 1977 and 1989. Progress in Oceanography, Vol. 47, Pergamon Press, 103–145.

  • Honda, M., , H. Nakamura, , J. Ukita, , I. Kousaka, , and K. Takeuchi, 2001: Interannual seesaw between the Aleutian and Icelandic lows. Part I: Seasonal dependence and life cycle. J. Climate, 14 , 10291042.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kleist, D. T., , and M. C. Morgan, 2005: Interpretation of the structure and evolution of adjoint-derived forecast sensitivity gradients. Mon. Wea. Rev., 133 , 467485.

    • Search Google Scholar
    • Export Citation
  • Lu, J., , R. J. Greatbatch, , and K. A. Peterson, 2004: Trend in Northern Hemisphere winter atmospheric circulation during the last half of the twentieth century. J. Climate, 17 , 37453760.

    • Search Google Scholar
    • Export Citation
  • 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 , 10691079.

    • Search Google Scholar
    • Export Citation
  • Molteni, F., , and T. N. Palmer, 1993: Predictability and finite-time instability of the northern winter circulation. Quart. J. Roy. Meteor. Soc., 119 , 269298.

    • Search Google Scholar
    • Export Citation
  • Namias, J., 1978: Multiple causes of the North American abnormal winter 1976-77. Mon. Wea. Rev., 106 , 279295.

  • Overland, J. E., , and P. J. Stabeno, 2004: Is the climate of the Bering Sea warming and affecting the ecosystem? Eos, Trans. Amer. Geophys. Union, 85 , 309316.

    • Search Google Scholar
    • Export Citation
  • Overland, J. E., , J. M. Adams, , and N. A. Bond, 1999: Decadal variability of the Aleutian low and its relation to high-latitude circulation. J. Climate, 12 , 15421548.

    • Search Google Scholar
    • Export Citation
  • Palmer, T. N., 1988: Medium and extended range predictability, and the stability of the PNA mode. Quart. J. Roy. Meteor. Soc., 114 , 691713.

    • Search Google Scholar
    • Export Citation
  • Quadrelli, R., , and J. M. Wallace, 2004: Varied expressions of the hemispheric circulation observed in association with contrasting polarities of prescribed patterns of variability. J. Climate, 17 , 42454253.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K., 1990: Recent observed interdecadal climate changes in the Northern Hemisphere. Bull. Amer. Meteor. Soc., 71 , 988993.

  • Trenberth, K., , and J. W. Hurrell, 1994: Decadal atmosphere-ocean variations in the Pacific. Climate Dyn., 9 , 303319.

  • Wallace, J. M., , and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109 , 784812.

    • Search Google Scholar
    • Export Citation
  • Wu, Q., , and D. M. Straus, 2004: AO, COWL, and observed climate trends. J. Climate, 17 , 21392156.

  • Yasunaka, S., , and K. Hanawa, 2002: Regime shifts found in the Northern Hemisphere SST field. J. Meteor. Soc. Japan, 80 , 119135.

  • Zou, X. F., , F. Vandenberghe, , M. Pondeca, , and Y-H. Kuo, 1997: Introduction to adjoint techniques and the MM5 adjoint modeling system. NCAR Tech. Note NCAR/TN-435+STR, 110 pp.

Fig. 1.
Fig. 1.

(a) Line graph of 1500 6-h time periods for the fall and winter months of 2000–01 depicting 500-hPa height anomalies at KP1 (thin), and filtered 500-hPa height anomalies (boldface). (b) A close-up of (a), depicting how an event was declared by identifying a period of at least 10 days where filtered 500-hPa height anomalies were below −100 m. Arrows indicate the duration of one particular event from 1200 UTC 16 Dec 2000 to 0000 UTC 7 Jan 2001.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 2.
Fig. 2.

Histogram of the total number of LSC events vs the length of events at KP1.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 3.
Fig. 3.

The total number of events from November to March and the locations for KP1 and KP2 for (a) 1977–2003, (b) 1977–88, and (c) 1988–2003.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 4.
Fig. 4.

Composite 500-hPa height anomalies for all KP1 events from 1977 to 2002 (shaded; interval 60 m, beginning at ±30 m) and 500-hPa height from 1977 to 2003 (contoured every 6 dam) at (a) 5 days prior to onset, (c) onset (X indicates the position of KP1), and (e) 6 days after onset. Composite sea level pressure for all KP1 events from 1977 to 2003 (contoured every 4 hPa), and 1000–500-hPa thickness (dashed, interval 6 dam) at (b) 5 days prior to onset, (d) onset, and (f) 6 days after onset. The 540-dam thickness is shown as a boldface dotted line.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 5.
Fig. 5.

Same as in Fig. 4 but for KP2.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 6.
Fig. 6.

Difference (KP2 − KP1) at 7 days after onset of (a) 500-hPa geopotential height anomalies (contours; interval 30 m) and (b) mean sea level pressure (contours; interval 2 hPa). Negative contours are dashed. Regions with a statistical significance at or above the 95% confidence level are shaded.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 7.
Fig. 7.

KP1 onset +6 composite 500-hPa height anomalies (shaded; interval 60 m, beginning at ±30 m) and 500-hPa height (contoured every 6 dam) for (a) 1977–88 events and (c) 1988–2003 events. KP1 onset +6 composite sea level pressure (contoured every 4 hPa), and 1000–500-hPa thickness (dotted; interval 6 dam) for (b) 1977–88 events and (d) 1988–2003 events. The 540-dam thickness is shown as a boldface dotted line.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 8.
Fig. 8.

Same as in Fig. 7 but for KP2.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 9.
Fig. 9.

Difference (pre-1988 minus post-1988) 500-hPa composite geopotential height anomalies (contours; interval 30 m) at 8.5 days after onset for (a) KP1 and (b) KP2. Regions with a statistical significance at or above the 95% confidence level are shaded.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 10.
Fig. 10.

Composite RMS of unfiltered 500-hPa height anomalies for all events (shaded; interval 20 m) and a schematic for primary cyclone tracks (arrows) for (a) KP1 events, (b) KP2 events, and (c) sequence of times when LSC events are not occurring (nonevents).

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 11.
Fig. 11.

Composite 500-hPa height anomalies for all KP1 events from 1977 to 2002 for day 5 after onset (shaded; same interval as Fig. 4), and a schematic of the PNA teleconnection pattern index areas. PNA pattern index areas are denoted with a box. Positions of KP1 and KP2 are denoted with an X.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 12.
Fig. 12.

(a) Sea level pressure (contours; interval 4 hPa), 1000–500-hPa thickness (dotted; interval 6 dam), and relative vorticity (shaded; interval 4 × 105 s−1) at 0000 UTC 5 Mar 1998. (b) The 310-K potential vorticity (contours; interval 1 PVU) and 310-K winds (m s−1) at 0000 UTC 5 Mar 1998. (c) Same as in (a) but for 0000 UTC 6 Mar 1998. (d) Same as in (b) but for 0000 UTC 6 Mar 1998. (e) Same as in (a) but for 0000 UTC 7 Mar 1998. (f) Same as in (b) but for 0000 UTC 7 Mar 1998.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 13.
Fig. 13.

The 48-h (a) sea level pressure (red contours; interval 4 hPa), σ = 0.775 relative vorticity (black contours; interval 6 × 10−5 s−1), 1000–500-hPa thickness (dotted; interval 6 dam), and the sensitivity of R1 to relative vorticity on the σ = 0.775 surface (color fill; interval 6 × 106 m2); (b) isotachs on the σ = 0.375 surface (light contours; interval 10 m s−1), sensitivity of R2 to relative vorticity on the σ = 0.375 surface (color fill; 6 × 106 m2 s−1); (c) cross section [orientation indicated in (a)] of relative vorticity (contours; interval 4 × 10−5 s−1) and sensitivity of R1 to relative vorticity (color fill; interval 18 × 106 m2); (d) cross section [orientation indicated in (b)] of isotachs (contours, interval 5 m s−1) and sensitivity of R2 to relative vorticity (color fill; interval 6 × 106 m2 s−1). In (b) the heavier contours indicate the region over which R2 was defined.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 14.
Fig. 14.

(a) Sensitivity of R1 to vorticity at 500 hPa (color fill; interval 6 × 108 m2), 500-hPa height (blue contours; interval 6 dam), and 500-hPa relative vorticity (black contours; interval 4 × 10−5 s−1) at 0000 UTC 5 Mar 1998; (b) cross section from 54.7°N, 119.0°E to 36.4°N, 164.8°E depicting the sensitivity of R1 to vorticity at 500 hPa (color fill; interval 6 × 108 m2), and 500-hPa relative vorticity (black contours; interval 2 × 10−5 s−1) at 0000 UTC 5 Mar 1998. (c) Same as in (a) but for 0000 UTC 6 March 1998. (d) Same as in (b) but for 0000 UTC 6 Mar 1998, and from 39.6°N, 145.4°E to 40.3°N, 169.8°W. (e) Same as in (a) but for 1800 UTC 6 Mar 1998. (f) Same as in (d) but for 1800 UTC 6 Mar 1998. The gray line in (a),(c),(e) depicts the cross section shown at corresponding times in (b),(d),(f), respectively.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Fig. 15.
Fig. 15.

(a) The 550–950-hPa layer averaged sensitivity of R2 to relative vorticity (color fill; interval 3 × 105 m2 s−1), 500-hPa height (blue contours; interval 6 dam), and 500-hPa relative vorticity (black contours; interval 4 × 10−5 s−1) at 0000 UTC 5 Mar 1998. (b) Sensitivity of R2 to temperature at 750 hPa (color fill; interval 15 m2 s−2 K−1), 750-hPa temperature (black contours; interval 3°C), and sea level pressure (orange contours; interval 4 hPa) at 0000 UTC 5 Mar 1998. (c) Same as in (a) but for 0000 UTC 6 Mar 1998. (d) Same as in (b) but for 0000 UTC 6 Mar 1998. (e) Same as in (a) but for 1800 UTC 6 Mar 1998. (f) Same as in (b) but for 1800 UTC 6 Mar 1998.

Citation: Monthly Weather Review 134, 12; 10.1175/MWR3260.1

Table 1.

Start and end dates and duration in days of LSC events for KP1 and KP2.

Table 1.

1

Refer to Kleist and Morgan (2005) for the details of this calculation.

Save