• Bosart, L. F., , G. J. Hakim, , K. R. Tyle, , M. A. Bedrick, , W. E. Bracken, , J. J. Dickinson, , and D. M. Schultz, 1996: Large-scale antecedent conditions associated with the 12–14 March 1993 cyclone (“Superstorm ‘93”) over eastern North America. Mon. Wea. Rev., 124, 18651891.

    • Search Google Scholar
    • Export Citation
  • Cavallo, S. M., , and G. J. Hakim, 2009: Potential vorticity diagnosis of a tropopause polar cyclone. Mon. Wea. Rev., 137, 13581371.

  • Cavallo, S. M., , and G. J. Hakim, 2010: The composite structure of tropopause polar cyclones from a mesoscale model. Mon. Wea. Rev., 138, 38403857.

    • Search Google Scholar
    • Export Citation
  • Cavallo, S. M., , J. Dudhia, , and C. Snyder, 2011: A multilayer upper-boundary condition for longwave radiative flux to correct temperature biases in a mesoscale model. Mon. Wea. Rev., 139, 19521959.

    • Search Google Scholar
    • Export Citation
  • Chen, C.-C., , G. J. Hakim, , and D. R. Durran, 2007: Transient mountain waves and their interaction with large scales. J. Atmos. Sci., 64, 23782400.

    • Search Google Scholar
    • Export Citation
  • Davis, C. A., , and K. A. Emanuel, 1991: Potential vorticity diagnostics of cyclogenesis. Mon. Wea. Rev., 119, 19291953.

  • Deser, C., , R. Tomas, , M. Alexander, , and D. Lawrence, 2010: The seasonal atmospheric response to projected Arctic sea ice loss in the late twenty-first century. J. Climate, 23, 333351.

    • Search Google Scholar
    • Export Citation
  • Hakim, G. J., 2000: Climatology of coherent structures on the extratropical tropopause. Mon. Wea. Rev., 128, 385406.

  • Hakim, G. J., , and A. K. Canavan, 2005: Observed cyclone–anticyclone tropopause asymmetries. J. Atmos. Sci., 62, 231240.

  • Hakim, G. J., , D. Keyser, , and L. F. Bosart, 1995: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part I: Multiscale case study. Mon. Wea. Rev., 123, 26632692.

    • Search Google Scholar
    • Export Citation
  • Hakim, G. J., , D. Keyser, , and L. F. Bosart, 1996: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part II: Diagnosis using quasigeostrophic potential vorticity inversion. Mon. Wea. Rev., 124, 21762205.

    • Search Google Scholar
    • Export Citation
  • Higgins, M. E., , and J. J. Cassano, 2009: Impacts of reduced sea ice on winter Arctic atmospheric circulation, precipitation, and temperature. J. Geophys. Res., 114, D16107, doi:10.1029/2009JD011884.

    • Search Google Scholar
    • Export Citation
  • Hines, K. M., , and D. H. Bromwich, 2008: Development and testing of polar Weather Research and Forecasting (WRF) model. Part I: Greenland ice sheet meteorology. Mon. Wea. Rev., 136, 19711989.

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

  • Kew, S. F., , M. Sprenger, , and H. C. Davies, 2010: Potential vorticity anomalies of the lowermost stratosphere: A 10-yr winter climatology. Mon. Wea. Rev., 138, 12341249.

    • Search Google Scholar
    • Export Citation
  • McIntyre, M. E., , and T. N. Palmer, 1984: The ‘surf zone’ in the stratosphere. J. Atmos. Terr. Phys., 46, 825849.

  • Morgan, M. C., , and J. W. Nielsen-Gammon, 1998: Using tropopause maps to diagnose midlatitude weather systems. Mon. Wea. Rev., 126, 25552579.

    • Search Google Scholar
    • Export Citation
  • Ogi, M., , and J. M. Wallace, 2007: Summer minimum Arctic sea ice extent and the associated summer atmospheric circulation. Geophys. Res. Lett., 34, L12705, doi:10.1029/2007GL029897.

    • Search Google Scholar
    • Export Citation
  • Pedlosky, J., 1998: Geophysical Fluid Dynamics. 2nd ed. Springer, 728 pp.

  • Peixoto, J. P., , and A. H. Oort, 1992: Physics of Climate. 2nd ed. American Institute of Physics, 520 pp.

  • Rigor, I. G., , J. M. Wallace, , and R. L. Colony, 2002: Response of sea ice to the Arctic Oscillation. J. Climate, 15, 26482663.

  • Schar, C., , and D. R. Durran, 1997: Vortex formation and vortex shedding in continuously stratified flows past isolated topography. J. Atmos. Sci., 54, 534554.

    • Search Google Scholar
    • Export Citation
  • Serreze, M. C., , and A. P. Barrett, 2008: The summer cyclone maximum over the central Arctic Ocean. J. Climate, 21, 10481065.

  • Serreze, M. C., , A. P. Barrett, , A. G. Slater, , M. Steele, , J. Zhang, , and K. E. Trenberth, 2007: The large-scale energy budget of the Arctic. J. Geophys. Res., 112, D11122, doi:10.1029/2006JD008230.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., , J. B. Klemp, , J. Dudhia, , D. O. Gill, , D. M. Barker, , W. Wang, , and J. G. Powers, 2005: A description of the Advanced Research WRF version 2. NCAR Tech. Note NCAR/TN-468+STR, 113 pp. [Available online at http://www.mmm.ucar.edu/wrf/users/docs/arw_v2.pdf.]

  • Solomon, S., , D. Qin, , M. Manning, , M. Marquis, , K. Averyt, , M. M. B. Tignor, , H. L. Miller Jr., , and Z. Chen, Eds., 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

  • Uccellini, L. W., , D. Keyser, , K. F. Brill, , and C. H. Wash, 1985: The Presidents’ Day cyclone of 18–19 February 1979: Influence of upstream trough amplification and associated tropopause folding on rapid cyclogenesis. Mon. Wea. Rev., 113, 962988.

    • Search Google Scholar
    • Export Citation
  • Wilks, D. S., 2006: Statistical Methods in the Atmospheric Sciences. 2nd ed. Academic Press, 627 pp.

  • Woolings, T., , B. Hoskins, , M. Blackburn, , and P. Berrisford, 2008: A new Rossby wave-breaking interpretation of the North Atlantic Oscillation. J. Atmos. Sci., 65, 609626.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Ensemble-mean cyclonic tropopause polar vortex potential temperature amplitude for (a) winter and (b) summer over the period 1990–99. The solid (dashed) contour corresponds to NNRP (WRF). All cyclones survive a minimum of 5 days so that the population is constant for all times displayed.

  • View in gallery

    Cyclonic tropopause polar vortex structural properties of maximum (a),(b) amplitude and (c),(d) average radius for (left) winter and (right) summer over the period 1990–99. The bin interval is 1 K in (a),(b) and 50 km in (c),(d). The solid (dashed) contours correspond to NNRP (WRF).

  • View in gallery

    Cyclonic tropopause polar vortex lifetime distributions for (a) winter and (b) summer over the period 1990–99 from NNRP (“•”) and WRF (“+”). Heavy lines depict linear fits to the exponential distributions from NNRP (black) and WRF (gray). Points excluded in the exponential fits are shaded in light gray. Correlation coefficients for NNRP and WRF respectively are R2 = 0.93, 0.82 (winter) and R2 = 0.91, 0.89 (summer).

  • View in gallery

    Tropopause polar cyclone track density from (a),(b) NNRP; (c),(d) WRF; and (e),(f) WRF minus NNRP for (left) winter and (right) summer over the period 1990–99. Values are equal to the number of unique vortices within a 5° latitude × 15° longitude box divided by the cosine of latitude and normalized by the maximum value. Only vortices lasting at least 2 days and that spent at least 60% of their lifetimes north of 65°N latitude are considered. Contours in (a)–(d) show mean 500-hPa heights from the respective climatology of the entire period with a contour interval of 60 m. The 5250-, 5400-, 5040-, and 5460-m contours are dashed in (a),(b),(c), and (d), respectively. Contours in (e),(f) show 500-hPa height differences (WRF − NNRP) with a contour interval of 20 m and positive (negative) values denoted with the solid (dashed) contours beginning with 20 m near the edge of the domain in (f).

  • View in gallery

    Zonal-mean zonal wind from (a),(b) WRF using full physics and (c),(d) WRF without radiation for (left) winter and (right) summer over the period 1990–99. Contours are shown every 2 m s−1, with zero in gray. Positive (negative) values are denoted by the solid (dashed) contours.

  • View in gallery

    The zonal-mean meridional streamfunction from (a),(b) WRF using full physics and (c),(d) WRF without radiation for (left) winter and (right) summer over the period 1990–99. Contours plotted: (−4, −2, −1.5, −1, −0.6, −0.4, −0.2, 0.2, 0.4, 0.6, 1, 1.5, 2, 4) × 107 kg s−1. The zero contour is plotted in gray. Positive (negative) values are denoted by the solid (dashed) contours.

  • View in gallery

    Ensemble-mean cyclonic tropopause polar vortex properties as a function of time for tropopause potential temperature (a),(b) latitude and (c),(d) amplitude for (left) winter and (right) summer over the period 1990–99. The solid (dashed) contours correspond to the WRF experiment with full physics (no radiation). All cyclones survive a minimum of 5 days so that the population is constant for all times displayed.

  • View in gallery

    Cyclonic tropopause polar vortex structural properties of the maximum (a),(b) amplitude and (c),(d) radius for (left) winter and (right) summer over the period 1990–99. The bin interval is 1 K in (a),(b) and 50 km in (c),(d). The solid (dashed) contours correspond to the WRF experiment with full physics (no radiation).

  • View in gallery

    Cyclonic tropopause polar vortex lifetime distributions for (a) winter and (b) summer over the period 1990–99 from the WRF model experiment with full physics (“•”) and the WRF model experiment with no radiation (“+”). Heavy lines depict linear fits to the exponential distributions from the WRF model experiment with full physics (black) and the WRF model with no radiation (gray). Points excluded in the exponential fits are shaded in light gray. Correlation coefficients for WRF with full physics and WRF with no radiation respectively are R2 = 0.82, 0.84 (winter) and R2 = 0.89, 0.78 (summer).

  • View in gallery

    Tropopause polar cyclone track density from (a),(b) WRF using full physics; (c),(d) WRF without radiation; and (e),(f) WRF without radiation minus WRF using full physics for (left) winter and (right) summer over the period 1990–99. Values are equal to the number of unique vortices within a 5° latitude × 15° longitude box divided by the cosine of latitude and normalized by the maximum value. Only vortices lasting at least 2 days and that spent at least 60% of their lifetimes north of 65°N latitude are considered. Contours in (a)–(d) show mean 500-hPa heights from the respective climatology of the entire period with a contour interval of 60 m. Thick (thin) contours denote the 5220- and 5340- (5160 and 5280) m contours in (a) and (c) and the 5520- and 5640- (5460 and 5580) m contours in (b) and (d). Contours in (e),(f) show 500-hPa height differences (no radiation − full physics) with a contour interval of 20 m, and positive (negative) values denoted with the solid (dashed) contours, greatest with 160 m near the pole in (e).

  • View in gallery

    (left) Winter and (right) summer composite tropopause potential temperature for the experiment with (a),(b) full physics; (c),(d) no radiation; and (e),(f) differences (no radiation − full physics). The contour interval is 5 K in (a)–(d) and 2 K in (e),(f). The first closed contour from the center is 270 K in (a), 295 K in (b), 295 K in (c), and 305 K in (d). Dashed (solid) contours in (e),(f) correspond to negative (positive) differences. Note that only those vortices where the minimum tropopause potential temperature is at least one standard deviation below the domain mean tropopause potential temperature are included in the composite sample shown here.

  • View in gallery

    Composite cross-sectional difference (no radiation − full physics) anomalies in (a),(b) potential temperature, and (c),(d) EPV between the no-radiation and full-physics simulations (no radiation − full physics) for (left) winter and (right) summer. The solid (dashed) black contour is the composite 2-PVU surface for the full-physics (no radiation) experiments. The color interval is 0.5 K in (a),(b) and 0.25 PVU in (c),(d). The zero contour is denoted by the thin, solid contour. Anomalies are computed with respect to the mean difference on a given pressure level.

  • View in gallery

    Composite cross-sectional difference (no radiation − full physics) anomalies in (a),(b) relative humidity; (c),(d) radiative heating rates; and (e),(f) EPV tendency due to radiation between the no-radiation and full-physics simulations (no radiation − full physics) for (left) winter and (right) summer. The solid (dashed) black contour is the composite 2-PVU surface for the full-physics (no radiation) experiments. The color interval is 1% in (a),(b), 0.1 K day−1 in (c),(d), and 0.01 PVU day−1 in (e),(f). The zero contour is denoted by the thin, solid contour. Anomalies are computed with respect to the mean difference on a given pressure level.

  • View in gallery

    Composite 500-hPa heights from the WRF-simulated climatologies with no radiation (dashed), WRF using full physics (solid), and the potential temperature difference (colors; no radiation − full physics) for the (a) winter and (b) summer. The contour interval is 60 m, and the color interval is 1 K. The contours nearest the North Pole are 5160 m in (a), and 5520 and 5460 m in (b), corresponding to the no-radiation and full-physics simulations, respectively. The location of greatest cyclonic TPV density from the no-radiation experiment is shown by the solid, black circle. See text for details.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 96 96 7
PDF Downloads 98 98 5

Radiative Impact on Tropopause Polar Vortices over the Arctic

View More View Less
  • 1 University of Washington, Seattle, Washington
© Get Permissions
Full access

Abstract

Tropopause polar vortices (TPVs) are commonly observed, coherent circulation features of the Arctic with typical radii as large as approximately 800 km. Intensification of cyclonic TPVs has been shown to be dominated by infrared radiation. Here the hypothesis is tested that while radiation alone may not be essential for TPV genesis, radiation has a substantial impact on the long-term population characteristics of cyclonic TPVs.

A numerical model is used to derive two 10-yr climatologies of TPVs for both winter and summer: a control climatology with radiative forcing and an experimental climatology with radiative forcing withheld. Results from the control climatology are first compared to those from the NCEP–NCAR reanalysis project (NNRP), which indicates sensitivity to both horizontal grid resolution and the use of polar filtering in the NNRP. Smaller horizontal grid resolution of 60 km in the current study yields sample-mean cyclonic TPV radii that are smaller by a factor of ~2 compared to NNRP, and vortex track densities in the vicinity of the North Pole are considerably larger compared to NNRP. The experimental climatologies show that winter (summer) vortex maximum amplitude is reduced by 22.3% (38.0%), with a net tendency to weaken without radiation. Moreover, while the number and lifetime of cyclonic TPVs change little in winter without radiation, the number decreases 12% and the mean lifetime decreases 19% during summer without radiation. These results suggest that dynamical processes are primarily responsible for the genesis of the vortices, and that radiation controls their maximum intensity and duration during summer, when the destructive effect of ambient shear is weaker.

Corresponding author address: Steven Cavallo, University of Oklahoma, School of Meteorology, 120 David L. Boren Blvd., Suite 5900, Norman, OK 73072. E-mail: cavallo@ou.edu

Abstract

Tropopause polar vortices (TPVs) are commonly observed, coherent circulation features of the Arctic with typical radii as large as approximately 800 km. Intensification of cyclonic TPVs has been shown to be dominated by infrared radiation. Here the hypothesis is tested that while radiation alone may not be essential for TPV genesis, radiation has a substantial impact on the long-term population characteristics of cyclonic TPVs.

A numerical model is used to derive two 10-yr climatologies of TPVs for both winter and summer: a control climatology with radiative forcing and an experimental climatology with radiative forcing withheld. Results from the control climatology are first compared to those from the NCEP–NCAR reanalysis project (NNRP), which indicates sensitivity to both horizontal grid resolution and the use of polar filtering in the NNRP. Smaller horizontal grid resolution of 60 km in the current study yields sample-mean cyclonic TPV radii that are smaller by a factor of ~2 compared to NNRP, and vortex track densities in the vicinity of the North Pole are considerably larger compared to NNRP. The experimental climatologies show that winter (summer) vortex maximum amplitude is reduced by 22.3% (38.0%), with a net tendency to weaken without radiation. Moreover, while the number and lifetime of cyclonic TPVs change little in winter without radiation, the number decreases 12% and the mean lifetime decreases 19% during summer without radiation. These results suggest that dynamical processes are primarily responsible for the genesis of the vortices, and that radiation controls their maximum intensity and duration during summer, when the destructive effect of ambient shear is weaker.

Corresponding author address: Steven Cavallo, University of Oklahoma, School of Meteorology, 120 David L. Boren Blvd., Suite 5900, Norman, OK 73072. E-mail: cavallo@ou.edu

1. Introduction

Radiation has been shown to play an important role in the intensification and maintenance of cyclonic tropopause polar vortices (TPVs) (Cavallo and Hakim 2010, hereafter CH10). Cyclonic TPVs are frequently observed, high-latitude, cold-core vortices based on the tropopause and play an important role in the formation of surface cyclones. The Arctic is particularly favorable for the maintenance and intensification of cyclonic TPVs (Cavallo and Hakim 2009; CH10), and cyclonic TPVs occasionally exist on time scales of months (Hakim and Canavan 2005). While radiative cooling may directly promote intensification through changes in moisture near TPVs, latent heating in clouds has a weakening effect; however, clouds can also indirectly promote radiative intensification through cloud-top cooling (Cavallo and Hakim 2009). Here, we extend the results of CH10 by examining the impact that radiation has on a simulated climatology of cyclonic TPVs for both winter and summer seasons.

An understanding of TPVs is motivated in part by the fact that surface cyclones develop in association with the dynamics initiated by TPVs (e.g., Hakim et al. 1995; Bosart et al. 1996; Hakim et al. 1996). Furthermore, TPVs are better expressed as vortical rather than wave-like features (e.g., Hakim 2000) and, because of the conservation properties of potential vorticity (PV) and potential temperature, are well identified in these fields and can be tracked for long periods of time on a single PV surface (e.g., Morgan and Nielsen-Gammon 1998). Additional studies have identified cyclonic TPVs as having the following characteristics: a downward intrusion of stratospheric air to ~500 hPa (e.g., Uccellini et al. 1985; Davis and Emanuel 1991; Hakim 2000; Cavallo and Hakim 2009; CH10), radii up to ~1200 km, amplitudes of up to 50 K (e.g., Hakim 2000; Hakim and Canavan 2005; CH10), and nearly saturated atmospheric conditions near the vortex core (Cavallo and Hakim 2009; CH10). Furthermore, the Canadian Archipelago region has been identified as a preferred location for intensification of cyclonic TPVs (Cavallo and Hakim 2009).

The Arctic is a suitable environment for TPVs, partially since they are isolated from the horizontal wind shear associated with the midlatitude jet stream, allowing them to remain in the Arctic for long periods of time. Isolation from the jet stream shields this region from the wave dynamics occurring in the vicinity of the jet stream, and in particular decreases the probability of adverse effects from horizontal shear. Thus, because of isolation of the air in these features over these longer time scales, diabatic processes are able to play an important role in intensity change. As a result of cold temperatures in the Arctic, latent heating rates are considerably lower than in midlatitude or tropical regions, and are lower than radiative cooling rates on average (e.g., Peixoto and Oort 1992, 319–321), providing the potential for radiative processes to dominate PV budgets and control TPV intensification at higher latitudes.

A compositing study of cyclonic TPVs was performed by CH10 using a mesoscale numerical model in a region of frequent vortex intensification. They show that cyclonic TPV intensification is primarily radiatively driven, and while the effects of latent heating are considerable, they are smaller in magnitude. Intensification was shown to result from anomalously large (small) radiative cooling centered below (above) the vortex tropopause—a pattern that is qualitatively similar to the corresponding relative humidity anomalies. Seasonal composites revealed that, despite the large annual variations of solar radiation, vortices intensify because of radiation on average during both summer and winter. Motivated by these results, this study aims to isolate the impact that radiation has on cyclonic TPV population characteristics during both winter and summer. Numerically simulated TPV climatologies are generated with and without radiation under the hypothesis that without radiative feedback, cyclonic TPVs are weaker and have shorter lifetimes. Furthermore, we expect that cyclonic TPVs will be generated in the absence of radiation, implying that while radiative processes may be important for their maintenance, other, dynamical, mechanisms may be more important for their genesis; however, the relative importance of radiation to genesis is unclear at the outset. While processes such as Rossby wave breaking (e.g., McIntyre and Palmer 1984), flow over topography (e.g., Schar and Durran 1997; Chen et al. 2007), and vortex splitting (e.g., Kew et al. 2010) can potentially lead to the generation of vortices on the tropopause, the present study focuses on the intensification processes of existing cyclonic TPVs, where radiation is of primary importance for their maintenance and longevity.

The remainder of the paper is organized as follows. A review of the methods used in Cavallo and Hakim (2009) and CH10 to quantify vortex intensity change is given in section 2. A detailed description of the numerical experiments and procedures is also provided in section 2. Results for the numerically simulated climatologies with and without radiation are presented in section 3, including an examination of the diabatic PV tendencies in relation to changes in the net atmospheric circulation. Conclusions and a summary are given in section 4.

2. Methods

a. Vortex intensity

Ertel potential vorticity (EPV), used here to define vortex intensity, is given by
e1
where ρ is the density, is the absolute vorticity, is the three-dimensional velocity vector, and is the earth’s rotational vector. Potential temperature is defined by
eq1
where T is temperature, p is pressure, po = 105 Pa is a standard constant, R = 287 J K−1 kg−1 is the dry air gas constant, and cp = 1004 J K−1 kg−1 is the specific heat capacity of dry air at constant pressure. Vortex intensity change can then be reasonably well quantified using the EPV tendency equation (e.g., Pedlosky 1998)
e2
when considering diabatic effects alone (Cavallo and Hakim 2009). In (2), the time rate of change following the fluid is given by
eq2
while the gradient operator is defined by
eq3
In a numerical modeling study, CH10 separated (2) into components isolating the radiative , latent heating , and all other1 diabatic processes, such that (2) becomes
e3

Diagnosing (3) from numerical model data, CH10 showed that cyclonic TPV intensity change in the Canadian Arctic was largely due to radiative forcing, which results in a positive EPV tendency within the vortex, on average. Latent heating results in a negative EPV tendency, but with smaller magnitude in comparison to radiation and the net effect is for cyclonic TPVs to intensify. Additionally, no other diabatic forcing term resulted in comparable EPV tendencies near the tropopause in the vortex core. Here we apply this technique by comparing numerically simulated climatologies of the Arctic with and without radiative forcing in order to isolate the impact that radiation has on the population characteristics of cyclonic TPVs. Since radiation is likely the leading mechanism for the intensification of cyclonic TPVs on average, then eliminating this effect is expected to significantly alter the characteristics of cyclonic TPVs.

b. Experimental setup

Sensitivity testing of the TPV climatology to radiation is performed here using the Advanced Research (ARW) version of the Weather Research and Forecasting (WRF) model version 2.2.1 (Skamarock et al. 2005) and the WRF Preprocessing System (WPS) version 2.2. WRF is a mesoscale model capable of resolving a wide range of horizontal scales associated with cyclonic TPVs, which have an average radius of about 360 km in the Canadian Arctic (CH10). Alternatively, a climate model could be used for the present analysis. However, many use a spectral grid, for which there is a singularity at the pole requiring the use of a polar filter. Furthermore, the horizontal resolution of climate models may not be high enough to resolve the wide range of TPV scales. For reference, we compare results for the relatively high-resolution WRF simulation to those for National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis project (NNRP) data (Kalnay et al. 1996), which does employ a polar filter.

Forecasts are initialized using NNRP data, with boundary conditions derived from 6-hourly NCEP–NCAR reanalysis grids. Boundaries are updated using NNRP every 6 h. These forecasts will hereby be referred to simply as WRF. Two sets of numerically simulated forecasts are performed to create a control 1990s climatology ensemble—one representing a winter-simulated climatology [December–February (DJF)] and one representing a summer-simulated climatology [June–August (JJA)]. A vortex census for these simulations will be compared to NNRP, for which winter and summer climatologies will be referred to as NNRP DJF and NNRP JJA, respectively. Repeating this method in the absence of radiative forcing gives what will be referred to as “no-radiation” experiments. In the no-radiation experiments, all radiative transfer calculations are bypassed for both shortwave and longwave radiation. For each experiment, there are a total of 10 model simulations: one for each year during 1990–99. Note that the DJF simulations begin on 1 December of the previous year; for example, the 1990 DJF simulation begins 1 December 1989. The JJA simulations begin 1 June of the respective year. Each simulation is performed for exactly 90 days. Approximate adjustment times for cloud fields to develop are typically ~6–24 h (e.g., Hines and Bromwich 2008). However, since this time period amounts to roughly 2% of all times, and since both the control and no-radiation experiments both experience this effect, we choose to include all times during the simulations for simplicity.

The model configuration of CH10 is employed here, except with a horizontal grid spacing of Δx = Δy = 60 km, 130x × 130y horizontal grid points, and a model time step of Δt = 300 s. Simulations are performed using a polar stereographic map projection, with a center latitude and longitude of 90°N and 100°W. Note that the NNRP data are derived from a spectral model with T62 horizontal resolution—equivalent to a horizontal grid spacing of approximately 210 km, with an output resolution of 2.5°. Although this may lead to a biased comparison with WRF, it serves as a useful benchmark for comparison with previous TPV studies that use NNRP data.

c. Vortex identification

The tropopause is defined by the 2 potential vorticity unit (PVU) surface (1 PVU = 1 × 10−6 K kg−1 s−1 m2), and cyclonic TPVs are identified and tracked using an objective tracking algorithm described in Hakim and Canavan (2005). Cyclonic vortex amplitude is the absolute difference between the last closed contour2 and the local minimum in potential temperature (hereafter the core value). Vortices are filtered, as in Hakim and Canavan (2005) and Cavallo and Hakim (2009), to include only those that spend at least 60% of their lifetimes north of 65°N latitude. Higher horizontal grid resolution can result in a very fine vortex structure leading to the premature identification of the last closed contour. The algorithm is adjusted to account for this possibility by defining the last closed contour by the median radius over all eight radials.

Definitions of the statistics used in the subsequent analysis follow. Vortex genesis is defined to occur by the identification of a local minimum in tropopause potential temperature that is colder than all other locations within a 650-km radius, which is roughly the vortex length scale; results are not sensitive to this choice of radius. Vortex lysis is defined to occur when a local minimum in tropopause potential temperature can no longer be identified with an existing vortex track.

3. Results

a. Mesoscale climatology simulated with full physics

We now compare the control WRF mesoscale vortex-simulated climatology to the NNRP climatology. Table 1 shows the sample sizes used in the subsequent analysis. Note that since the WRF simulations are limited to DJF and JJA of their respective years, the full life cycles of all vortices present at the beginning or end of a season are not identified. For example, it is possible to identify a genesis point but not a corresponding lysis point, and vice versa for each unique vortex track. In general, the number of vortex points categorized in the following mesoscale vortex census is much greater than from NNRP, likely because of higher horizontal grid resolution. Furthermore, while WRF uses a regularly spaced numerical grid across the pole, the latitude–longitude grid and polar filter used by the NNRP likely has an impact on vortices near the pole. As the pole is geographically central to the location of TPVs, such polar filtering of data could have a considerable impact on their statistics in this vicinity.

Table 1.

The sample size of unique vortex tracks, vortex genesis, and vortex lysis events from cyclonic TPV tracks. See text for further details.

Table 1.

A comparison between the NNRP and WRF seasonal composite cyclonic vortex amplitudes over 5-day periods for TPVs in years 1990–99 starting with the vortex origin is shown in Figs. 1a,b. Cyclones in WRF with lifetimes of at least 5 days tend to begin with larger amplitudes, and cyclone growth occurs for longer periods of time. Amplitude differences between WRF and NNRP are greater during the summer than during winter. Cyclone amplitudes are lower during the summer, as shown both by NNRP and WRF, and consistent with the findings of Hakim and Canavan (2005).

Fig. 1.
Fig. 1.

Ensemble-mean cyclonic tropopause polar vortex potential temperature amplitude for (a) winter and (b) summer over the period 1990–99. The solid (dashed) contour corresponds to NNRP (WRF). All cyclones survive a minimum of 5 days so that the population is constant for all times displayed.

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

Probability density functions (PDFs) of the maximum cyclone amplitude of unique cyclones (represented as a percentage) are shown in Figs. 2a,b. WRF maximum cyclone amplitudes are generally larger than in NNRP. Note that here the vortex population is represented by all vortices with lifetimes of at least 2 days, whereas the sample represented in Fig. 1 only includes those with lifetimes of at least 5 days. Maximum cyclone amplitudes average3 23.8 K (18.4 K) in WRF (NNRP) during winter (Fig. 2a) and 20.0 K (13.2 K) in WRF (NNRP) during summer (Fig. 2b). Both NNRP and WRF show a secondary maximum amplitude peak of ~30 and ~35 K, respectively, during winter. No secondary maximum amplitude peak is evident in either dataset during the summer. Given that the same algorithm was used in computing these statistics for both WRF and NNRP, differences appear to be consistent with expectations of smaller grid spacing for the WRF simulation. Cyclone radii for WRF have a mean value of 349 km (375 km) during winter (summer) compared to 734 km (768 km) during winter (summer) in NNRP (Figs. 2c,d). Again, differences in mean cyclone radius are attributed to finer horizontal scales resolved by the WRF simulation.

Fig. 2.
Fig. 2.

Cyclonic tropopause polar vortex structural properties of maximum (a),(b) amplitude and (c),(d) average radius for (left) winter and (right) summer over the period 1990–99. The bin interval is 1 K in (a),(b) and 50 km in (c),(d). The solid (dashed) contours correspond to NNRP (WRF).

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

The lifetime of cyclones can be reasonably approximated by exponentially decaying distributions for both WRF and NNRP (Fig. 3). While the cyclone lifetime pattern in WRF is similar to NNRP during winter (Fig. 3a), WRF produces a greater number of cyclones with longer lifetimes than NNRP during summer (Fig. 3b). The maximum lifetime during the winter is 21.0 days (15.0 days) in WRF (NNRP), with an average lifetime of 5.7 days (3.5 days). During summer, the maximum lifetime increases to 39.3 days in WRF and the average lifetime increases to 8.3 days, while NNRP lifetimes remain largely unchanged with a maximum lifetime of 10.0 days and mean lifetime of 3.6 days (Fig. 3b). With regard to the longer lifetimes during summer in WRF data, we hypothesize that horizontal shear associated with the midlatitude jet stream is weaker, reducing the probability that shear may destroy the vortex, and increasing the probability that radiative processes may intensify the vortex. Moreover, if TPVs are concentrated closer to the pole in WRF compared to NNRP, then TPVs in WRF have a greater opportunity to remain isolated from the jet stream. This topic will be explored in greater detail in the following section.

Fig. 3.
Fig. 3.

Cyclonic tropopause polar vortex lifetime distributions for (a) winter and (b) summer over the period 1990–99 from NNRP (“•”) and WRF (“+”). Heavy lines depict linear fits to the exponential distributions from NNRP (black) and WRF (gray). Points excluded in the exponential fits are shaded in light gray. Correlation coefficients for NNRP and WRF respectively are R2 = 0.93, 0.82 (winter) and R2 = 0.91, 0.89 (summer).

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

The regional variability of cyclones between NNRP and WRF are now compared, beginning with an overview of their spatial track density patterns. Recall that the WRF solutions are constrained only by the NNRP boundary conditions, whereas NNRP assimilates observations. Thus, the following analysis seeks to describe the statistics of the vortices, not how well the WRF forecasts of an individual event verify with analysis estimates.

During winter, NNRP cyclone track densities are highest in the Canadian Archipelago, Baffin Island, Greenland, areas along the northern Siberian coast, and northwestern Alaska (Fig. 4a); a polar minimum in track density is notable. The highest cyclone track densities in WRF are concentrated mainly along the northern and western coast of Greenland and over Baffin Bay, the Beaufort Sea, and the southwestern Barents Sea (Fig. 4c). Track densities are generally greater in WRF over the Arctic Ocean, whereas they are greater in NNRP primarily over Siberia, the Bering Strait, and the Canadian Arctic (Fig. 4e). During the summer, NNRP track densities are greatest near Baffin Island, with additional maxima near Novaya Zemlya, the Greenland Sea, and the East Siberian Sea (Fig. 4b). WRF cyclones track primarily in the western Arctic Ocean, Canadian Arctic, and Greenland–Iceland–Norwegian (GIN) Seas (Fig. 4d). The overall pattern is similar to winter, where cyclone tracks tend to be greater in WRF than in NNRP over the Arctic Ocean (Fig. 4f), lending more evidence that differences are likely due to the polar filter of NNRP discussed earlier, which could act to limit vortex tracks across the pole.

Fig. 4.
Fig. 4.

Tropopause polar cyclone track density from (a),(b) NNRP; (c),(d) WRF; and (e),(f) WRF minus NNRP for (left) winter and (right) summer over the period 1990–99. Values are equal to the number of unique vortices within a 5° latitude × 15° longitude box divided by the cosine of latitude and normalized by the maximum value. Only vortices lasting at least 2 days and that spent at least 60% of their lifetimes north of 65°N latitude are considered. Contours in (a)–(d) show mean 500-hPa heights from the respective climatology of the entire period with a contour interval of 60 m. The 5250-, 5400-, 5040-, and 5460-m contours are dashed in (a),(b),(c), and (d), respectively. Contours in (e),(f) show 500-hPa height differences (WRF − NNRP) with a contour interval of 20 m and positive (negative) values denoted with the solid (dashed) contours beginning with 20 m near the edge of the domain in (f).

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

To understand the differences between NNRP and WRF, we examine the time-mean flow patterns, as represented here by the composite 500-hPa height fields in Fig. 4. During winter, there is a 500-hPa height minimum over Baffin Bay, while there is a relatively strong trough axis over the central part of northern Siberia in NNRP (Fig. 4a). Large cyclone track densities are located along the periphery of the 500-hPa height minimum over Baffin Bay and at the base of the trough axis in Siberia; that is, these climatological troughs are favorable locations for cyclonic TPVs. Note that from this analysis, the physical relationship between the larger-scale climatological troughs and TPVs cannot be distinguished, as at least a portion of larger-scale signal may result from TPVs. For WRF, 500-hPa height minima are centered over northern Canada and over northern Siberia (Fig. 4c). Cyclone track densities are focused primarily along the northern coast of Greenland and downstream of higher topography on Greenland. Genesis density is also greatest along the northern coast of Greenland (not shown); flow over topography is capable of generating potential vorticity by dissipation (e.g., Schar and Durran 1997; Chen et al. 2007), and therefore could be responsible for the higher cyclone density in this region. Locations of greater track densities are also located near regions of relatively high frequencies of cyclonic Rossby wave breaking (e.g., Woolings et al. 2008), as well as vortex splitting and merger events (e.g., Kew et al. 2010), which could be important processes with regard to TPV genesis. DJF difference fields reveal a bias toward higher 500-hPa heights in WRF—greatest near Greenland and over the Arctic Ocean near Severnaya Zemlya (Fig. 4e). During summer, both WRF and NNRP exhibit minimum 500-hPa heights near the North Pole, with the most pronounced trough axis extending equatorward from the 500-hPa height minimum toward Baffin Bay (Figs. 4b,d). A ridge is evident over Greenland, which is surrounded by two trough axes. TPV track density is relatively large upstream of the ridge axis west of Greenland in both NNRP and WRF. Overall, 500-hPa heights are again greater in WRF, with the greatest differences over Baffin Bay, Ellesmere Island, and northern Greenland (Fig. 4f). The positive height bias in WRF is consistent with the tropospheric biases seen with respect to the Global Forecast System (GFS) model in CH10 and Cavallo et al. (2011).

The WRF 1990s winter- and summer-simulated climatologies shown here provide a large, long-term, sample of TPVs with mesoscale spatial resolution. Although the simulations suffer from biases inherent to the model, they provide useful data for performing controlled sensitivity experiments, such as those considered next. This approach is advantageous in that a wider range of vortex scales are resolved than in climate data, particularly NNRP, which has been used in similar studies of TPVs. Furthermore, the mesoscale model used here employs no filtering to data near the North Pole, where TPVs likely exhibit high track density. Although WRF and NNRP differ considerably in the exact locations of cyclonic TPVs, WRF captures life cycles of TPVs in geographically similar regions. In the following section, we compare the WRF control experiment discussed above to an identical simulation, but without explicit radiative forcing.

b. Mesoscale climatology simulated without radiation

We now examine the simulated climatological impact of radiation on cyclonic TPVs during both winter and summer, beginning with the changes to the atmospheric circulation. During winter, the zonal winds are westerly, or cyclonic around the pole in the full-physics simulation (Fig. 5a). The seasonal absence of shortwave radiation leaves only longwave cooling over higher latitudes in both the troposphere and stratosphere, whereas shortwave heating offsets a portion of the longwave cooling in lower latitudes. The result is more radiative cooling at higher latitudes and thermally balanced westerly winds. Without radiation, while the westerly jet stream near the tropopause at 45°N remains, there is an easterly stratospheric jet of ~8 m s−1 near 70°N, which extends close to the surface (Fig. 5c). Note that the effects of radiation are still implicitly included near the lateral boundaries (lower latitudes), resulting in cooling at lower latitudes that is absent near the center of the domain at high latitudes. To maintain thermal wind balance, higher zonal-mean geopotential heights are necessary over higher latitudes, therefore reversing the zonal geopotential height gradient.

Fig. 5.
Fig. 5.

Zonal-mean zonal wind from (a),(b) WRF using full physics and (c),(d) WRF without radiation for (left) winter and (right) summer over the period 1990–99. Contours are shown every 2 m s−1, with zero in gray. Positive (negative) values are denoted by the solid (dashed) contours.

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

During summer, the midlatitude jet stream, indicated by the westerly zonal wind maximum at 45°N and 200 hPa, has less upward extent into the stratosphere than in winter in the full-physics simulation (Fig. 5b). Shortwave heating is relatively greater than longwave cooling over higher latitudes during the summer, especially in the stratosphere since ozone is a strong absorber in shortwave bands. Shortwave heating is strong enough to completely offset longwave cooling in the stratosphere at higher latitudes. At lower latitudes, longwave cooling remains stronger than shortwave heating, resulting in stronger net radiative cooling over lower latitudes than over higher latitudes. Thus, the thermally balanced zonal-mean geopotential height gradient in the stratosphere is of the opposite sign during summer than in winter, with higher zonal-mean geopotential heights at higher latitudes, acting to reduce westerly winds with height above the tropopause. Without radiation, weak easterly zonal winds around 70°N extend downward from the model top to the surface, as the tropospheric zonal-mean geopotential height gradient reverses over higher latitudes as a result of the cooling at lower latitudes near the lateral boundaries as discussed for the winter climatology (Fig. 5d). Note that the easterly winds are considerably weaker during the summer. This is due to weaker net cooling at the lateral boundaries during the summer, as greater amounts of shortwave heating are offsetting the longwave cooling. The result is a weaker meridional temperature gradients and weaker thermally balanced easterlies.

A comparison of the zonal-mean meridional streamfunction with and without radiation is shown in Fig. 6. During both winter and summer, there is a high-latitude positive meridional circulation maximum centered ~70°N around 700 hPa, implying that there is ascent ~60°N and descent poleward of ~80°N. This circulation center weakens without radiation during both seasons, but also shifts slightly poleward (equatorward) during winter (summer). These shifts in the meridional circulation are also accompanied by a poleward (equatorward) shift in the mean latitude of cyclonic TPVs formation during winter (summer) (Figs. 7a,b). Cyclonic TPVs generally track poleward during both winter and summer in both the control and experimental climatologies. This is consistent with the poleward motion expected near the tropopause inferred from the meridional streamfunction at these latitudes (cf. Fig. 6).

Fig. 6.
Fig. 6.

The zonal-mean meridional streamfunction from (a),(b) WRF using full physics and (c),(d) WRF without radiation for (left) winter and (right) summer over the period 1990–99. Contours plotted: (−4, −2, −1.5, −1, −0.6, −0.4, −0.2, 0.2, 0.4, 0.6, 1, 1.5, 2, 4) × 107 kg s−1. The zero contour is plotted in gray. Positive (negative) values are denoted by the solid (dashed) contours.

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

Fig. 7.
Fig. 7.

Ensemble-mean cyclonic tropopause polar vortex properties as a function of time for tropopause potential temperature (a),(b) latitude and (c),(d) amplitude for (left) winter and (right) summer over the period 1990–99. The solid (dashed) contours correspond to the WRF experiment with full physics (no radiation). All cyclones survive a minimum of 5 days so that the population is constant for all times displayed.

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

Regarding cyclone amplitudes, during winter, vortex amplitude increases (decreases) over time in the full-physics (no radiation) simulated climatology (Fig. 7c). Five days after genesis, vortex amplitudes are 25% smaller on average without radiation. During the summer, vortex amplitude increases at nearly the same rate for both climatologies during the first 24 h, but thereafter the amplitude decreases without radiation, and is nearly 30% smaller without radiation after 4 days (Fig. 7d). Thus, on average, cyclones with relatively long lifetimes tend to weaken when radiative forcing is absent.

The distribution of maximum amplitude shows that cyclones have a peak at smaller amplitudes without radiation during both seasons, with averages decreasing from 23.8 (20.0) K to 18.5 (12.4) K during winter (summer) (Figs. 8a,b). A secondary winter peak around 35 K is not present without radiation, which may be an indication that radiation is especially important for stronger cyclones (Fig. 8a). The shift in the amplitude distribution is most pronounced during summer, where few cyclones obtain amplitudes greater than 30 K without radiation (Fig. 8b). The effects of radiation on the number of vortices are modest in comparison, with the number of unique vortex tracks decreasing by 5.2% (12.1%) without radiation during the winter (summer) (recall Table 1). Little change is evident in cyclone radii between the two experiments (Figs. 8c,d). The average radius is 372 (400) km during winter (summer) without radiation, and 349 (375) km during winter (summer) in the full-physics simulated climatology. Therefore, while radiation plays an important role in vortex amplitude, it has a small effect on the horizontal scale of cyclonic TPVs.

Fig. 8.
Fig. 8.

Cyclonic tropopause polar vortex structural properties of the maximum (a),(b) amplitude and (c),(d) radius for (left) winter and (right) summer over the period 1990–99. The bin interval is 1 K in (a),(b) and 50 km in (c),(d). The solid (dashed) contours correspond to the WRF experiment with full physics (no radiation).

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

Cyclone lifetimes follow nearly the same exponential decay pattern during winter with and without radiation, with no statistically significant4 differences in longevity (Fig. 9a). Average lifetimes during winter are 5.7 days, whereas the maximum lifetime is ~21 days in both cases. During summer, cyclones exhibit a statistically significant decrease in lifetime without radiation, as average lifetimes decrease from 8.3 to 7 days while the maximum lifetime decreases from 39.3 to 34 days (Fig. 9b). The relatively greater longevity during summer than during winter, apparent from both the full-physics and no-radiation simulations, may be related to a weaker meridional influence of the summer midlatitude jet stream (cf. Figs. 6a,b). This implies that lateral shear has less of an influence in the regions where TPVs are most prevalent during summer, increasing the likelihood of TPVs remaining isolated in the Arctic. Greater isolation from the jet stream provides more opportunity for the vortices to be maintained or intensified by radiative processes. When eliminating the radiative intensification of TPVs during winter, there is only a minimal impact on lifetime, suggesting that radiation has more of an influence on the vortex population characteristics at longer time scales. That is, during winter, there are more competing weakening mechanisms present that act on shorter time scales, such as dynamics or latent heating. We conclude that, given enough time, radiation will maintain cyclonic TPVs, lengthening their lifetimes.

Fig. 9.
Fig. 9.

Cyclonic tropopause polar vortex lifetime distributions for (a) winter and (b) summer over the period 1990–99 from the WRF model experiment with full physics (“•”) and the WRF model experiment with no radiation (“+”). Heavy lines depict linear fits to the exponential distributions from the WRF model experiment with full physics (black) and the WRF model with no radiation (gray). Points excluded in the exponential fits are shaded in light gray. Correlation coefficients for WRF with full physics and WRF with no radiation respectively are R2 = 0.82, 0.84 (winter) and R2 = 0.89, 0.78 (summer).

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

As discussed earlier, the 500-hPa height pattern is expected to be substantially different between the simulations with full physics and without radiation. Following the discussion of the spatial variations in TPV track density in section 3a, the locations of TPVs are therefore also expected to differ without radiation. While 500-hPa heights are locally lower over the Arctic, over the Canadian Archipelago, and over north-central Siberia in the full-physics simulated climatology during winter (Fig. 10a), 500-hPa heights are locally high along a line from the North Atlantic Ocean across the Arctic Ocean and toward Siberia without radiation (Fig. 10c). As a result, the density of cyclones is reduced over the North Atlantic and central Arctic Ocean, generally where the locally high 500-hPa heights are present without radiation (Fig. 10e). Additionally, most cyclones are focused over the Beaufort Sea and Canadian Archipelago, closer in proximity to locally lower 500-hPa heights and where the 500-hPa flow is weaker. Note that cyclones are filtered to include only those where at least 60% of their lifetimes are poleward of 65°N latitude, and therefore tracks along the southern periphery of the 500-hPa height minima during both seasons without radiation are not included in this analysis. During summer, the lowest 500-hPa heights centered near the North Pole are located farther southward near Baffin Island, Canada, without radiation (Figs. 10b,d). Without radiation, 500-hPa height gradients are generally weaker, which implies the atmospheric flow is weaker at this level. A notable difference is that a ridge of higher 500-hPa heights along the western North American coast is much less pronounced without radiation, which is a region with a considerable increase in cyclone track density without radiation (Figs. 10b,d,f).

Fig. 10.
Fig. 10.

Tropopause polar cyclone track density from (a),(b) WRF using full physics; (c),(d) WRF without radiation; and (e),(f) WRF without radiation minus WRF using full physics for (left) winter and (right) summer over the period 1990–99. Values are equal to the number of unique vortices within a 5° latitude × 15° longitude box divided by the cosine of latitude and normalized by the maximum value. Only vortices lasting at least 2 days and that spent at least 60% of their lifetimes north of 65°N latitude are considered. Contours in (a)–(d) show mean 500-hPa heights from the respective climatology of the entire period with a contour interval of 60 m. Thick (thin) contours denote the 5220- and 5340- (5160 and 5280) m contours in (a) and (c) and the 5520- and 5640- (5460 and 5580) m contours in (b) and (d). Contours in (e),(f) show 500-hPa height differences (no radiation − full physics) with a contour interval of 20 m, and positive (negative) values denoted with the solid (dashed) contours, greatest with 160 m near the pole in (e).

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

We now examine the changes to the composite structure of cyclonic TPVs without radiation. Vortex-relative5 composites of tropopause potential temperature are shown in Fig. 11. The composites are spatially filtered such that only those cyclones for which tropopause potential temperature at the vortex core is at least one standard deviation below the domain mean is included in the composites; fields are averaged relative to the vortex core. The one standard deviation threshold is used as a compromise between sample size and obtaining cyclone statistics on those that differ from the background. In the sample sizes for the WRF full-physics winter and summer cases there are 93 and 57 vortices, respectively, while for the WRF no-radiation winter and summer cases there are 67 and 124 vortices, respectively. In the full-physics winter-simulated climatology, the composite vortex core potential temperature is 269 K, with a vortex amplitude of 30.8 K (Fig. 11a). Vortex core potential temperature is 291 K without radiation, with a vortex amplitude of 21.6 K (Fig. 11c). Note that potential temperature is greater over the whole domain in the no-radiation composite (Fig. 11e) because of the absence of longwave radiative cooling as discussed earlier. However, differences are greatest in the vortex, indicating that the radiative processes have a relatively greater effect on the vortex than on the background environment. In the full-physics summer-simulated climatology, the composite vortex core potential temperature is 291 K with a vortex amplitude of 28.2 K (Fig. 11b). In the no-radiation composite, vortex core tropopause potential temperature is 304 K with an amplitude of 20.7 K (Fig. 11d), with reductions of tropopause potential temperature at the vortex core of ~13 K relative to the full-physics composite (Fig. 11f). As in the winter case, tropopause potential temperature is greater in the no-radiation case, primarily because of domain-wide warmer temperatures from the absence of longwave radiative cooling. The difference is not as pronounced during summer since shortwave heating partially offsets the effect from longwave cooling near the tropopause.

Fig. 11.
Fig. 11.

(left) Winter and (right) summer composite tropopause potential temperature for the experiment with (a),(b) full physics; (c),(d) no radiation; and (e),(f) differences (no radiation − full physics). The contour interval is 5 K in (a)–(d) and 2 K in (e),(f). The first closed contour from the center is 270 K in (a), 295 K in (b), 295 K in (c), and 305 K in (d). Dashed (solid) contours in (e),(f) correspond to negative (positive) differences. Note that only those vortices where the minimum tropopause potential temperature is at least one standard deviation below the domain mean tropopause potential temperature are included in the composite sample shown here.

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

Vertical west–east cross sections through the composite vortex show differences in potential temperature and EPV between the full-physics and no-radiation climatologies (Fig. 12). During winter, potential temperature changes are anomalously negative in the vortex core near the tropopause without radiation, while anomalously positive potential temperature changes are present above and below the negative anomalies, respectively (Fig. 12a). Recalling (1), the anomalously negative (positive) vertical gradients in potential temperature differences around 500 (250) hPa in the vortex result in a reduction (increase) in EPV at the same locations without radiation, thereby reducing the EPV near the vortex tropopause without radiation. Relative humidity near the vortex tropopause is greater without radiation, as the vortices are weaker and there is a higher average tropopause (Fig. 13a). With radiation, relatively strong radiative heating is centered about the vortex tropopause (Fig. 13c), and is located near where the air is relatively dry from the downward intrusion of stratospheric air. This radiative heating pattern is consistent with those found by CH10, and is likely a radiative response to changes in the distribution of water vapor near the vortex; details of this topic are the subject of ongoing research and will be reported in future work.

Fig. 12.
Fig. 12.

Composite cross-sectional difference (no radiation − full physics) anomalies in (a),(b) potential temperature, and (c),(d) EPV between the no-radiation and full-physics simulations (no radiation − full physics) for (left) winter and (right) summer. The solid (dashed) black contour is the composite 2-PVU surface for the full-physics (no radiation) experiments. The color interval is 0.5 K in (a),(b) and 0.25 PVU in (c),(d). The zero contour is denoted by the thin, solid contour. Anomalies are computed with respect to the mean difference on a given pressure level.

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

Fig. 13.
Fig. 13.

Composite cross-sectional difference (no radiation − full physics) anomalies in (a),(b) relative humidity; (c),(d) radiative heating rates; and (e),(f) EPV tendency due to radiation between the no-radiation and full-physics simulations (no radiation − full physics) for (left) winter and (right) summer. The solid (dashed) black contour is the composite 2-PVU surface for the full-physics (no radiation) experiments. The color interval is 1% in (a),(b), 0.1 K day−1 in (c),(d), and 0.01 PVU day−1 in (e),(f). The zero contour is denoted by the thin, solid contour. Anomalies are computed with respect to the mean difference on a given pressure level.

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

Under the vortex tropopause and near the surface, potential temperature is higher, partially from the hydrostatic response of weaker vortices on average, but also because background tropospheric potential temperature is higher near the center of the domain at higher latitudes. Figure 14a shows that given the location of most frequent cyclones (near Cambridge Bay, Canada), a grid-relative horizontal line from west to east would place the higher background potential temperatures near the center of this line. That is, the cross sections have a high probability of sampling vortices across the Arctic, with the higher background potential temperatures near the center of the cross section without radiation (Fig. 12a). Similarly, potential temperatures are higher above the vortex near 125 hPa without radiation because of the absence of longwave cooling that would occur over high latitudes with radiation. Recall that over lower latitudes, some of the longwave cooling is offset by shortwave heating, resulting in a greater potential temperature increase over higher latitudes and near the center of the cross section without radiation. Note that because of the anomalous warming above the vortex near 125 hPa without radiation, there is an anomalous increase in potential temperature with height, resulting in more EPV ~ 250 hPa (Fig. 12c). The net EPV tendency differences reflect the reduced radiative heating and cooling anomalies that are associated with the dry, stratosphere air in the presence of radiative forcings (Fig. 13e).

Fig. 14.
Fig. 14.

Composite 500-hPa heights from the WRF-simulated climatologies with no radiation (dashed), WRF using full physics (solid), and the potential temperature difference (colors; no radiation − full physics) for the (a) winter and (b) summer. The contour interval is 60 m, and the color interval is 1 K. The contours nearest the North Pole are 5160 m in (a), and 5520 and 5460 m in (b), corresponding to the no-radiation and full-physics simulations, respectively. The location of greatest cyclonic TPV density from the no-radiation experiment is shown by the solid, black circle. See text for details.

Citation: Monthly Weather Review 140, 5; 10.1175/MWR-D-11-00182.1

During summer, anomalously positive (negative) differences in potential temperature are present below (above) the vortex tropopause (Fig. 12b), resulting in less EPV in the vortex core without radiation (Fig. 12d). Relative humidity differences are similar to the winter climatology, where weaker vortices lead to relative humidity increases near the tropopause without radiation (Fig. 13b). A notable difference from the winter case is the decreases in tropospheric relative humidity under the vortex core, which reflects the greater amount of tropospheric water vapor during summer. This is also evident in the radiative heating differences where, with radiation, anomalous radiative cooling (heating) is present near areas of anomalous decreases (increases) in relative humidity with height (Fig. 13d), which is consistent with the findings of CH10. The anomalous potential temperature differences follow from the anomalous radiative heating differences, with potential temperature decreases (increases) located above (below) the tropopause without radiation (recall Fig. 12b).

Note that there is an asymmetric potential temperature and EPV change without radiation, with the dipole-like potential temperature pattern extending eastward and upward ~1500 km from the vortex core (Figs. 12b,d). This pattern is also evident in the relative humidity differences (Fig. 13b). In Fig. 14b, it is apparent that higher potential temperatures are more probable east of the composite vortex core in the troposphere, for the same reasons as discussed previously for the winter. The relative humidity asymmetry with higher (lower) values above (below) the tropopause east of the vortex core is primarily a response of the lower (higher) potential temperatures without radiation. The anomalous EPV tendency differences due to radiation are located near the strongest vertical gradients in the anomalous radiative heating differences (Fig. 13f). Therefore, the symmetric, dipolelike differences in EPV above and below the vortex tropopause are primarily a direct radiative response associated with moisture anomalies near the vortex, while the asymmetric changes on the eastern side of the vortex are primarily a response to changes in the background potential temperature, with respect to the locations of the cross section. Although the asymmetric pattern here can be viewed as a rather unrealistic effect of the NNRP-forced boundary conditions, it does emphasize the sensitivity that both temperature and water vapor have on TPVs, which could potentially have a greater impact over the next century as temperatures are projected to increase and sea ice is projected to decrease (Solomon et al. 2007).

4. Summary and conclusions

Radiation has been identified in previous studies as a primary mechanism of cyclonic TPV intensification over a region of frequent cyclone intensification in the Arctic. This study explored the hypothesis that when this intensification mechanism is absent, cyclonic TPVs still form, but they will be weaker and exhibit shorter lifetimes. Here we tested this hypothesis using numerically simulated 10-yr mesoscale climatologies with and without radiation during winter and summer seasons over the Arctic. Changes in vortex characteristics, atmospheric circulation, and dynamics from a potential vorticity perspective were explored.

Results support the hypothesis that radiation plays an important role in the maintenance of cyclonic TPVs. Without radiation, cyclonic TPVs slowly weaken over time on average, and the average amplitude is reduced 22.3% (38.0%) during winter (summer). The larger amplitude reduction during summer corresponds to greater vertical water vapor gradients, as tropospheric water vapor is substantially higher during summer. Vortex lifetimes decrease from an average of 8.3–7 days during summer without radiation; however, no change in lifetimes is evident during winter. This result is consistent with the hypothesis that the vortices are more isolated from the lateral wind shear associated with the midlatitude jet stream during summer, allowing radiation to maintain the vortices for longer periods of time. Additionally, changes in the large-scale atmospheric circulation are observed, particularly with a weakening of the mean meridional streamfunction over higher latitudes without radiative forcings, leading to shifts in the jet stream and corresponding shifts in the locations and movement of TPVs. It is not possible to disentangle the effect of local radiative processes from the large-scale circulation changes on the TPV statistics. However, the fact that the results show stronger and longer-lived vortices, especially in summer, is consistent with previous studies, which showed that this effect was attributable to radiation (e.g., CH10).

The radiative intensification structure seen here is similar to that documented by CH10 for a large sample of cyclonic TPVs over the Canadian Arctic. The CH10 composites showed dipolelike radiative heating anomalies with anomalous radiative cooling (heating) just below (above) the vortex tropopause, which was also qualitatively similar to the relative humidity anomalies. Although the diabatic intensity changes and horizontal scale of vortices studied here are similar to CH10, the horizontal scale is much smaller than those documented in NNRP data both here and in Hakim and Canavan (2005). Given that the same vortex algorithms were used in each study, these differences indicate sensitivity to the horizontal grid spacing used on the native grids. Furthermore, vortex track density is substantially greater over the Arctic Ocean near the North Pole in the WRF-simulated climatologies, likely because of polar filtering NNRP data.

These results emphasize the important role that radiation has on maintaining and intensifying cyclonic TPVs, particularly during summer. With increasing evidence that water vapor is important with regard to the radiative structure of cyclonic TPVs, future work is needed to better quantify this relationship. This may be pursued using an idealized approach, where the complicated interactions between radiation, temperature, water vapor, and latent heating may be more easily isolated in controlled experiments. If water vapor is important to the ultimate evolution of TPVs, then it is also important to understand how changes in the moisture budget associated with warmer Arctic temperatures and decreases in sea ice could affect TPV intensity. This could be a factor, especially during autumn and winter, when the upward-directed surface heat and moisture fluxes are estimated to be largest (Serreze et al. 2007), increasing the potential for diabatic forcing to alter the thermodynamic heat budget of TPVs. Recent studies with the Community Atmospheric Model (CAM) show a strong sensitivity of the lower atmospheric circulation to projected changes in sea ice, alone, during the winter (Deser et al. 2010; Higgins and Cassano 2009), and future studies will examine whether these changes are further communicated above the atmospheric boundary layer to the tropopause. During summer, surface heat fluxes are directed downward from the atmosphere to the surface (e.g., Serreze et al. 2007), for which the low-level atmospheric circulation could have an impact on sea ice movement (e.g., Rigor et al. 2002). Ogi and Wallace (2007) found that summer seasons with anomalously low sea ice concentrations occur when the low-level atmospheric circulation over the Arctic Ocean is anticyclonic, and cyclonic (anticyclonic) sea level pressure patterns have been linked to cyclonic (anticyclonic) 500-hPa flow over the Arctic Ocean (Serreze and Barrett 2008). It is yet to be determined what role TPVs play, if any, on these observations, and this is a fertile topic for future studies. Finally, the results here show that even in the absence of radiation, large numbers of cyclonic TPVs still exist, and for considerable lengths of time. Therefore, while radiation is important for TPV maintenance and intensification, it is apparently not essential for their genesis, which provides another topic for future research.

Acknowledgments

This work represents a portion of the first author’s Ph.D. dissertation at the University of Washington. The authors extend their appreciation to Profs. Dale Durran, Qiang Fu, and Mike Wallace for their thoughtful comments regarding this work. This research was sponsored by the National Science Foundation through Award ATM-0552004 made to the University of Washington.

REFERENCES

  • Bosart, L. F., , G. J. Hakim, , K. R. Tyle, , M. A. Bedrick, , W. E. Bracken, , J. J. Dickinson, , and D. M. Schultz, 1996: Large-scale antecedent conditions associated with the 12–14 March 1993 cyclone (“Superstorm ‘93”) over eastern North America. Mon. Wea. Rev., 124, 18651891.

    • Search Google Scholar
    • Export Citation
  • Cavallo, S. M., , and G. J. Hakim, 2009: Potential vorticity diagnosis of a tropopause polar cyclone. Mon. Wea. Rev., 137, 13581371.

  • Cavallo, S. M., , and G. J. Hakim, 2010: The composite structure of tropopause polar cyclones from a mesoscale model. Mon. Wea. Rev., 138, 38403857.

    • Search Google Scholar
    • Export Citation
  • Cavallo, S. M., , J. Dudhia, , and C. Snyder, 2011: A multilayer upper-boundary condition for longwave radiative flux to correct temperature biases in a mesoscale model. Mon. Wea. Rev., 139, 19521959.

    • Search Google Scholar
    • Export Citation
  • Chen, C.-C., , G. J. Hakim, , and D. R. Durran, 2007: Transient mountain waves and their interaction with large scales. J. Atmos. Sci., 64, 23782400.

    • Search Google Scholar
    • Export Citation
  • Davis, C. A., , and K. A. Emanuel, 1991: Potential vorticity diagnostics of cyclogenesis. Mon. Wea. Rev., 119, 19291953.

  • Deser, C., , R. Tomas, , M. Alexander, , and D. Lawrence, 2010: The seasonal atmospheric response to projected Arctic sea ice loss in the late twenty-first century. J. Climate, 23, 333351.

    • Search Google Scholar
    • Export Citation
  • Hakim, G. J., 2000: Climatology of coherent structures on the extratropical tropopause. Mon. Wea. Rev., 128, 385406.

  • Hakim, G. J., , and A. K. Canavan, 2005: Observed cyclone–anticyclone tropopause asymmetries. J. Atmos. Sci., 62, 231240.

  • Hakim, G. J., , D. Keyser, , and L. F. Bosart, 1995: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part I: Multiscale case study. Mon. Wea. Rev., 123, 26632692.

    • Search Google Scholar
    • Export Citation
  • Hakim, G. J., , D. Keyser, , and L. F. Bosart, 1996: The Ohio Valley wave-merger cyclogenesis event of 25–26 January 1978. Part II: Diagnosis using quasigeostrophic potential vorticity inversion. Mon. Wea. Rev., 124, 21762205.

    • Search Google Scholar
    • Export Citation
  • Higgins, M. E., , and J. J. Cassano, 2009: Impacts of reduced sea ice on winter Arctic atmospheric circulation, precipitation, and temperature. J. Geophys. Res., 114, D16107, doi:10.1029/2009JD011884.

    • Search Google Scholar
    • Export Citation
  • Hines, K. M., , and D. H. Bromwich, 2008: Development and testing of polar Weather Research and Forecasting (WRF) model. Part I: Greenland ice sheet meteorology. Mon. Wea. Rev., 136, 19711989.

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

  • Kew, S. F., , M. Sprenger, , and H. C. Davies, 2010: Potential vorticity anomalies of the lowermost stratosphere: A 10-yr winter climatology. Mon. Wea. Rev., 138, 12341249.

    • Search Google Scholar
    • Export Citation
  • McIntyre, M. E., , and T. N. Palmer, 1984: The ‘surf zone’ in the stratosphere. J. Atmos. Terr. Phys., 46, 825849.

  • Morgan, M. C., , and J. W. Nielsen-Gammon, 1998: Using tropopause maps to diagnose midlatitude weather systems. Mon. Wea. Rev., 126, 25552579.

    • Search Google Scholar
    • Export Citation
  • Ogi, M., , and J. M. Wallace, 2007: Summer minimum Arctic sea ice extent and the associated summer atmospheric circulation. Geophys. Res. Lett., 34, L12705, doi:10.1029/2007GL029897.

    • Search Google Scholar
    • Export Citation
  • Pedlosky, J., 1998: Geophysical Fluid Dynamics. 2nd ed. Springer, 728 pp.

  • Peixoto, J. P., , and A. H. Oort, 1992: Physics of Climate. 2nd ed. American Institute of Physics, 520 pp.

  • Rigor, I. G., , J. M. Wallace, , and R. L. Colony, 2002: Response of sea ice to the Arctic Oscillation. J. Climate, 15, 26482663.

  • Schar, C., , and D. R. Durran, 1997: Vortex formation and vortex shedding in continuously stratified flows past isolated topography. J. Atmos. Sci., 54, 534554.

    • Search Google Scholar
    • Export Citation
  • Serreze, M. C., , and A. P. Barrett, 2008: The summer cyclone maximum over the central Arctic Ocean. J. Climate, 21, 10481065.

  • Serreze, M. C., , A. P. Barrett, , A. G. Slater, , M. Steele, , J. Zhang, , and K. E. Trenberth, 2007: The large-scale energy budget of the Arctic. J. Geophys. Res., 112, D11122, doi:10.1029/2006JD008230.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., , J. B. Klemp, , J. Dudhia, , D. O. Gill, , D. M. Barker, , W. Wang, , and J. G. Powers, 2005: A description of the Advanced Research WRF version 2. NCAR Tech. Note NCAR/TN-468+STR, 113 pp. [Available online at http://www.mmm.ucar.edu/wrf/users/docs/arw_v2.pdf.]

  • Solomon, S., , D. Qin, , M. Manning, , M. Marquis, , K. Averyt, , M. M. B. Tignor, , H. L. Miller Jr., , and Z. Chen, Eds., 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

  • Uccellini, L. W., , D. Keyser, , K. F. Brill, , and C. H. Wash, 1985: The Presidents’ Day cyclone of 18–19 February 1979: Influence of upstream trough amplification and associated tropopause folding on rapid cyclogenesis. Mon. Wea. Rev., 113, 962988.

    • Search Google Scholar
    • Export Citation
  • Wilks, D. S., 2006: Statistical Methods in the Atmospheric Sciences. 2nd ed. Academic Press, 627 pp.

  • Woolings, T., , B. Hoskins, , M. Blackburn, , and P. Berrisford, 2008: A new Rossby wave-breaking interpretation of the North Atlantic Oscillation. J. Atmos. Sci., 65, 609626.

    • Search Google Scholar
    • Export Citation
1

Includes diabatic processes resulting from the convective and planetary boundary layer physics schemes, as well as all other explicit mixing.

2

To define the last closed contour, eight equally spaced radials are defined from the location of minimum tropopause potential temperature. The last closed contour is the minimum potential temperature value along the eight radials at the location of the sign reversal.

3

All differences in average amplitude and radius values are statistically significant at the 95% significance level from a two-sample difference of a mean standard Gaussian z test given by Wilks [2006, his Eq. (5.8)].

4

Computed by testing the statistical significance of the two-sample best-fit rates of exponential decay using the Student’s t test and a 95% significance level.

5

In computing vortex-relative composites, the grid is shifted such that the center grid point is located at the grid points corresponding to minima in tropopause potential temperature. Fields are subsequently averaged on this shifted (nonrotated) grid.

Save