• Beesley, J. A., , C. S. Bretherton, , C. Jakob, , E. L. Andreas, , J. M. Intrieri, , and T. A. Uttal, 2000: A comparison of cloud and boundary layer variables in the ECMWF forecast model with observations at Surface Heat Budget of the Arctic Ocean (SHEBA) ice camp. J. Geophys. Res., 105 (D10), 12 33712 349.

    • Search Google Scholar
    • Export Citation
  • Cesana, G., , J. E. Kay, , H. Chepfer, , J. M. English, , and G. de Boer, 2012: Ubiquitous low-level liquid-containing Arctic clouds: New observations and climate model constraints from CALIPSO-GOCCP. Geophys. Res. Lett., 39, L20804, doi:10.1029/2012GL053385.

    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597.

    • Search Google Scholar
    • Export Citation
  • Drobot, S. D., , J. A. Maslanik, , and C. Fowler, 2003: Atmospheric and sea ice conditions during the SHEBA year: Historical and spatial assessment. Polar Geogr., 27, 1537.

    • Search Google Scholar
    • Export Citation
  • ECMWF, 2012: IFS documentation—Cy37r2 operational implementation 18 May 2011: Part IV: Physical processes. European Centre for Medium-Range Weather Forecasts Rep., 174 pp. [Available online at http://www.ecmwf.int/research/ifsdocs/CY37r2/IFSPart4.pdf.]

  • Hazeleger, W., and Coauthors, 2012: EC-Earth V2.2: Description and validation of a new seamless Earth system prediction model. Climate Dyn., 39, 26112629.

    • Search Google Scholar
    • Export Citation
  • Karlsson, J., , and G. Svensson, 2011: The simulation of Arctic clouds and their influence on the winter surface temperature in present-day climate in the CMIP3 multi-model dataset. Climate Dyn., 36, 623635, doi:10.1007/s00382-010-0758-6.

    • Search Google Scholar
    • Export Citation
  • Karlsson, J., , and G. Svensson, 2013: Consequences of poor representation of Arctic sea-ice albedo and cloud-radiation interactions in the CMIP5 model ensemble. Geophys. Res. Lett., 40, 43744379, doi:10.1002/grl.50768.

    • Search Google Scholar
    • Export Citation
  • Kwok, R., , and N. Untersteiner, 2011: The thinning of Arctic sea ice. Phys. Today, 64, 36, doi:10.1063/1.3580491.

  • Meyers, M. P., , P. J. Demott, , and W. R. Cotton, 1992: New primary ice-nucleation parameterizations in an explicit cloud model. J. Appl. Meteor., 31, 708721.

    • Search Google Scholar
    • Export Citation
  • Morrison, H., , G. de Boer, , G. Feingold, , J. Harrington, , M. D. Shupe, , and K. Sulia, 2012: Resilience of persistent arctic mixed-phase clouds. Nat. Geosci., 5, 1117.

    • Search Google Scholar
    • Export Citation
  • Prenni, A. J., and Coauthors, 2007: Can ice-nucleating aerosols affect Arctic seasonal climate? Bull. Amer. Meteor. Soc., 88, 541550.

    • Search Google Scholar
    • Export Citation
  • Serreze, M. C., , M. M. Holland, , and J. Stroeve, 2007: Perspectives on the Arctic's shrinking sea-ice cover. Science, 315, 15331536.

  • Shupe, M. D., 2011: Clouds at Arctic atmospheric observatories. Part II: Thermodynamic phase characteristics. J. Appl. Meteor. Climatol., 50, 645661.

    • Search Google Scholar
    • Export Citation
  • Shupe, M. D., , and J. M. Intrieri, 2004: Cloud radiative forcing of the Arctic surface: The influence of cloud properties, surface albedo, and solar zenith angle. J. Climate, 17, 616628.

    • Search Google Scholar
    • Export Citation
  • Shupe, M. D., , S. Y. Matrosov, , and T. Uttal, 2006: Arctic mixed-phase cloud properties derived from surface-based sensors at SHEBA. J. Atmos. Sci., 63, 697711.

    • Search Google Scholar
    • Export Citation
  • Sorteberg, A., , V. Kattsov, , J. E. Walsh, , and T. Pavlova, 2007: The Arctic surface energy budget as simulated with the IPCC AR4 AOGCMs. Climate Dyn., 29, 131156.

    • Search Google Scholar
    • Export Citation
  • Stramler, K., , A. D. D. Genio, , and W. B. Rossow, 2011: Synoptically driven Arctic winter states. J. Climate,24, 1747–1762.

  • Svensson, G., , and J. Karlsson, 2011: On the Arctic wintertime climate in global climate models. J. Climate,24, 5757–5771.

  • Tjernström, M., , and R. G. Graversen, 2009: The vertical structure of the lower Arctic troposphere analysed from observations and the ERA-40 reanalysis. Quart. J. Roy. Meteor. Soc., 135, 431443.

    • Search Google Scholar
    • Export Citation
  • Uttal, T., and Coauthors, 2002: Surface heat budget of the Arctic Ocean. Bull. Amer. Meteor. Soc., 83, 255276.

  • Verlinde, J., and Coauthors, 2007: The mixed-phase Arctic cloud experiment. Bull. Amer. Meteor. Soc., 88, 205221.

  • Xie, S., , X. Liu, , C. Zhao, , and Y. Zhang, 2013: Sensitivity of CAM5-simulated Arctic clouds and radiation to ice nucleation parameterization. J. Climate, 26, 59815999.

    • Search Google Scholar
    • Export Citation
  • Zhang, T., , K. Stamnes, , and S. A. Bowling, 2001: Impact of the atmospheric thickness on the atmospheric downwelling longwave radiation and snowmelt under clear-sky conditions in the Arctic and Subarctic. J. Climate, 14, 920939.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Frequency distribution of DJFM daily mean LWD values from ERA-Interim between 1990 and 2010. Also shown is the distribution of LWD values from ERA-Interim during the 1997/98 winter (SHEBA winter) and observations from SHEBA covering the same period. (b) Frequency distribution of the DJFM longwave (LW) CRE from ERA-Interim.

  • View in gallery

    (top) SCM column over the SHEBA ice station. Colored shadings represent equivalent potential temperature θe, black contours represent clouds with supercooled liquid water ql > 0.01 g kg−3, and green contours represent clouds with a qi > 0.01 g kg−3. (middle) LWD simulated by the untuned (black dashed) and tuned (black solid) SCM compared with observations (red). (bottom) LWP simulated by the untuned (black dashed) and tuned (black solid) SCM compared with observations (red). Discontinuities in the observations (red lines) are due to the lack of data for the corresponding period.

  • View in gallery

    (a) Frequency distribution of DJFM daily mean LWD values from the tuned EC-Earth SCM forced with ERA-Interim data between 1990 and 2010. Also shown is the distribution of LWD values from the SCM during the 1997/98 winter (SHEBA winter) and observations from SHEBA covering the same period. (b) Frequency distribution of the LW CRE from the EC-Earth SCM during the DJFM period.

  • View in gallery

    Two-dimensional frequency distributions of daily values of LWD and surface pressure during the DJFM period of 1997/98 (SHEBA year) following Morrison et al. (2012). (a) The default version of the SCM. (b) The tuned (l = 0.2) version of the SCM.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 23 23 12
PDF Downloads 21 21 11

The Importance of Representing Mixed-Phase Clouds for Simulating Distinctive Atmospheric States in the Arctic

View More View Less
  • 1 Department of Meteorology, and Bolin Centre for Climate Research, Stockholm University, and Swedish e-Science Research Centre, Stockholm, Sweden
  • 2 Department of Meteorology, and Bolin Centre for Climate Research, Stockholm University, Stockholm, Sweden
  • 3 Department of Meteorology, and Bolin Centre for Climate Research, Stockholm University, and Swedish e-Science Research Centre, Stockholm, Sweden
© Get Permissions
Full access

Abstract

Observations from the Surface Heat Budget of the Arctic Ocean experiment (SHEBA) suggest that the Arctic Basin is characterized by two distinctly different preferred atmospheric states during wintertime. These states appear as two peaks in the frequency distribution of surface downwelling longwave radiation (LWD), representing radiatively clear and opaque conditions. Here, the authors have investigated the occurrence and representation of these states in the widely used ECMWF Interim Re-Analysis (ERA-Interim) dataset. An interannually recurring bimodal distribution of LWD values is not a clearly observable feature in the reanalysis data. However, large differences in the simulated liquid water content of clouds in ERA-Interim compared to observations are identified and these are linked to the lack of a radiatively opaque peak in the reanalysis. Using a single-column model, dynamically controlled by data from ERA-Interim, the authors show that, by tuning the glaciation speed of supercooled liquid clouds, it is possible to reach a very good agreement between the model and observations from the SHEBA campaign in terms of LWD. The results suggest that the presence of two preferred Arctic states, as observed during SHEBA, is a recurring feature of the Arctic climate system during winter [December–March (DJFM)]. The mean increase in LWD during the Arctic winter compared to ERA-Interim is 15 W m−2. This has a substantial bearing on climate model evaluation in the Arctic as it indicates the importance of representing Arctic states in climate models and reanalysis data and that doing so could have a significant impact on winter ice thickness and surface temperatures in the Arctic.

Corresponding author address: Anders Engström, Department of Meteorology, SE-10691 Stockholm University, Stockholm, Sweden. E-mail: anderse@misu.su.se

Supplemental information related to this paper is available at the Journals Online website: http://dx.doi.org/10.1175/JCLI-D-13-00271.s1.

Abstract

Observations from the Surface Heat Budget of the Arctic Ocean experiment (SHEBA) suggest that the Arctic Basin is characterized by two distinctly different preferred atmospheric states during wintertime. These states appear as two peaks in the frequency distribution of surface downwelling longwave radiation (LWD), representing radiatively clear and opaque conditions. Here, the authors have investigated the occurrence and representation of these states in the widely used ECMWF Interim Re-Analysis (ERA-Interim) dataset. An interannually recurring bimodal distribution of LWD values is not a clearly observable feature in the reanalysis data. However, large differences in the simulated liquid water content of clouds in ERA-Interim compared to observations are identified and these are linked to the lack of a radiatively opaque peak in the reanalysis. Using a single-column model, dynamically controlled by data from ERA-Interim, the authors show that, by tuning the glaciation speed of supercooled liquid clouds, it is possible to reach a very good agreement between the model and observations from the SHEBA campaign in terms of LWD. The results suggest that the presence of two preferred Arctic states, as observed during SHEBA, is a recurring feature of the Arctic climate system during winter [December–March (DJFM)]. The mean increase in LWD during the Arctic winter compared to ERA-Interim is 15 W m−2. This has a substantial bearing on climate model evaluation in the Arctic as it indicates the importance of representing Arctic states in climate models and reanalysis data and that doing so could have a significant impact on winter ice thickness and surface temperatures in the Arctic.

Corresponding author address: Anders Engström, Department of Meteorology, SE-10691 Stockholm University, Stockholm, Sweden. E-mail: anderse@misu.su.se

Supplemental information related to this paper is available at the Journals Online website: http://dx.doi.org/10.1175/JCLI-D-13-00271.s1.

1. Introduction

As suggested by Stramler et al. (2011), the Arctic Basin is characterized during wintertime by two distinctly different preferred states that extend from the ocean mixed layer through the troposphere. The basis for their hypothesis is the fact that several observed variables from the Surface Heat Budget of the Arctic Ocean experiment (SHEBA; Uttal et al. 2002), conducted between 1997 and 1998, show bimodal frequency distributions. A combination of synoptic- and local-scale processes has been suggested to be important for the existence and persistence of these states (Morrison et al. 2012). Considering the typically very stably stratified conditions (suppressing turbulence) and absence of solar radiation during the Arctic winter, most of the air–surface energy exchange is mediated through longwave radiation. Therefore, Stramler et al. (2011) separate the two observed Arctic states by the surface net longwave radiation and refer to them as “opaquely cloudy” and “radiatively clear,” respectively. The former is characterized by warmer surface temperatures, lower surface pressure, and a near balance between surface downwelling and emitted longwave radiation, while the latter state is characterized by a negative surface longwave budget, high surface pressure, and cold surface temperatures (Morrison et al. 2012).

A key feature of the opaquely cloudy state is the abundance of mixed-phase clouds (Morrison et al. 2012), which are frequent during the Arctic winter although temperatures are well below 0°C (e.g., Shupe et al. 2006; Shupe 2011). Compared to ice-only clouds, mixed-phase clouds have a stronger radiative signature on the Arctic surface energy budget (Shupe and Intrieri 2004). The persistence of mixed-phase clouds, the opaquely cloudy state, which was up to 10 days during SHEBA winter, is remarkable because the relative humidity below 0°C is higher with respect to ice than water. Thus, ice crystals will grow at the expense of water droplets (the Bergeron–Findeisen process). Therefore, a complex set of local feedback mechanisms has been proposed to explain the persistence of mixed-phase clouds in these thermodynamically unstable conditions (Morrison et al. 2012).

In general, global and regional climate models underestimate the liquid condensate of the Arctic winter (e.g., Prenni et al. 2007; Karlsson and Svensson 2011) and are also associated with too little downwelling longwave radiation (Prenni et al. 2007; Sorteberg et al. 2007; Karlsson and Svensson 2011; Svensson and Karlsson 2011). These are signs that may indicate underestimation of the frequency, or strength, of the opaquely cloudy state. Model evaluation of the Arctic has typically used monthly averages, which rules out assessment of Arctic states. However, using hourly output, Cesana et al. (2012) show that the lack of radiatively opaque states in two climate models can be related to the underestimation of Arctic liquid-containing clouds.

The idea of preferred Arctic states originates from the intense observations during the SHEBA winter. However, the climatological representativeness of the 1997/98 winter has been questioned (Drobot et al. 2003; Uttal et al. 2002), and whether Arctic states is a recurring feature of the Arctic climate system is an open question. Therefore, we investigate the occurrence and representation of key features of winter Arctic states in the European Centre for Medium-Range Weather Forecasts (ECMWF) Interim Re-Analysis (ERA-Interim; Dee et al. 2011). Based on the results from our analysis, we utilize the single-column model (SCM) of the global climate model EC-Earth Consortium (EC-Earth; Hazeleger et al. 2012) to investigate if, given prescribed large-scale conditions, the physics routines of the model are able to reproduce the occurrence of Arctic states during a 20-yr period.

2. Arctic states in ERA-Interim

To determine the representativeness of the SHEBA winter for Arctic wintertime conditions in general, we examine the widely used ERA-Interim dataset. ERA-Interim combines information from a multitude of observations (e.g., surface-based observations, soundings, and satellites) with information from a short forecast started from a previous analysis. This procedure takes into account the uncertainty of both model and observations to provide the best estimate of the historical state of the atmosphere. In ERA-Interim, observations are assimilated as they become available. For example, information from satellites (mainly clear-sky radiances) was included in the early 1990s, and other more or less conventional observational datasets are available for both shorter and longer periods of time (Dee et al. 2011). More specifically and of importance for this study is that temperature and humidity soundings, released twice daily from the SHEBA site during the campaign, are used in ERA-Interim. These soundings were found to reduce the near-surface temperature bias present in the older 40-yr ECMWF Re-Analysis (ERA-40) dataset, and it was also speculated that the timing of specific weather systems was improved (Tjernström and Graversen 2009). Although the two reanalysis datasets use two different assimilation methods and are based on different model versions, it is reasonable to expect that the soundings released during SHEBA has improved the overall accuracy of ERA-Interim during this period as well. It is important to note, however, that the presence and properties of clouds in ERA-Interim is a purely modeled product, which is not affected by the assimilation system, other than indirectly through assimilation of atmospheric state variables.

We study the period December–March (DJFM) for all winters between 1990 and 2010. The ERA-Interim data are retrieved for a point in the Arctic chosen to represent the mean location of the SHEBA station (77°N, 160°W). While Stramler et al. (2011) examined the bimodality of surface net longwave radiation, we study downwelling longwave radiation (LWD). The rationale for this is that the modeled net longwave will include uncertainties related to surface temperature biases in ERA-Interim and hence the description of turbulent surface energy fluxes. By focusing on LWD, we isolate the atmospheric component of the problem. Similar to net longwave radiation, the frequency distribution of LWD from the SHEBA campaign also shows bimodality, which we examine further below. The temporal resolution of ERA-Interim is also lower than the observations from SHEBA and we therefore examine daily mean radiative fluxes constructed from SHEBA hourly data. Although more clearly pronounced in hourly data (e.g., Stramler et al. 2011), a histogram of daily mean LWD values from SHEBA does display a bimodal pattern. In ERA-Interim, radiative fluxes are provided in the forecast fields based on the 0000 and 1200 UTC analysis. The daily mean values of radiation are therefore calculated from the 0000 UTC + 12 h and 1200 UTC + 12 h forecasts.

Figure 1a shows the normalized frequency of daily averaged LWD from SHEBA observations (winter 1997/98) and ERA-Interim (winter 1997/98 and multiyear winter period). The corresponding distribution of each individual year is shown in Fig. S1 (see supplemental material at the Journals Online website: http://dx.doi.org/10.1175/JCLI-D-13-00271.s1). With a few exceptions, the wintertime periods in ERA-Interim generally do not exhibit a clear bimodal distribution of LWD values that corresponds to the observations from SHEBA. The ERA-Interim multiyear median shows a good agreement with SHEBA in the radiatively clear peak. However, ERA-Interim has frequently more values in the range between 150 and 200 W m−2 compared to SHEBA. Within this range, the observations from SHEBA barely fall within the 5th–95th percentile of the multiyear dataset. The opaquely cloudy peak found in observations is not clearly represented in ERA-Interim. The peak, if it exists during a specific DJFM period, is both less frequent and shifted toward lower values compared to SHEBA. ERA-Interim data for the SHEBA winter do show a bimodal distribution of LWD values, but the opaquely cloudy peak is located around 180 W m−2: that is, significantly lower than the observations, which place the peak at 210 W m−2 (Fig. 1a).

Fig. 1.
Fig. 1.

(a) Frequency distribution of DJFM daily mean LWD values from ERA-Interim between 1990 and 2010. Also shown is the distribution of LWD values from ERA-Interim during the 1997/98 winter (SHEBA winter) and observations from SHEBA covering the same period. (b) Frequency distribution of the DJFM longwave (LW) CRE from ERA-Interim.

Citation: Journal of Climate 27, 1; 10.1175/JCLI-D-13-00271.1

Although underestimating the radiatively opaque peak, the simulated time series of LWD in ERA-Interim are in correlative agreement with the observations from SHEBA (r = 0.91), which indicates that ERA-Interim captures the general characteristics of the atmosphere well. The root-mean-square error (RMSE) of LWD for the winter period is 18 W m−2 and there is a negative mean bias of 6 W m−2. One explanation for the lower frequency of LWD values in the radiatively opaque peak could be an underestimate of the simulated water vapor path (WVP). Although well correlated with observations from SHEBA (r = 0.92; RMSE = 0.7 kg m−2), ERA-Interim does have a negative mean bias in WVP of 0.4 kg m−2. At low WVP values, this could introduce a significant error in LWD (Zhang et al. 2001). However, the largest discrepancy in LWD between ERA-Interim and SHEBA is found in the radiatively opaque peak, at higher values of WVP (>5 kg m−2), where the LWD is less sensitive to small errors in WVP (Zhang et al. 2001). The lack of a radiatively opaque peak in ERA-Interim should therefore not be primarily related to an underestimate of clear-sky radiative fluxes. The overall agreement between the simulated WVP and LWD with observations from SHEBA points to the fact that the assimilation of the soundings released during the campaign positively impacted the reanalysis or alternatively that the analysis already captures the large-scale conditions over the site well. However, LWD will not only depend on the temperature and humidity of the atmosphere but also the presence and phase of clouds in the column. Therefore, the less frequent radiatively opaque state in ERA-Interim indicates a possible misrepresentation of the cloudiness or the properties of existing clouds. Comparing the simulated liquid water path (LWP) with the observations from SHEBA shows that ERA-Interim clearly underestimates the observed LWP (Fig. S2 of the supplemental material). In contrast, the reanalysis produces a significantly higher ice water path (IWP) collocated in time with periods of high LWP in SHEBA. However, the longwave radiative effect of ice clouds is weak compared to liquid clouds (Shupe and Intrieri 2004) and clouds with a large longwave cloud radiative effect (CRE) are less frequent than radiatively thinner clouds in ERA-Interim, which may contribute to the lack of a radiatively opaque peak (see Fig. 1b).

The representation of mixed-phase clouds in ERA-Interim is of specific interest during the Arctic winter because the resilience of supercooled liquid clouds will have a significant impact on LWD (Morrison et al. 2012). ERA-Interim is based on cycle 31r2 of the Integrated Forecast System (IFS) at the ECMWF, in which only one prognostic variable for condensed water is used. At temperatures between 250.16 and 273.16 K, condensed water exists in the mixed phase and the partitioning between ice and water is temperature dependent. A previous study, using an earlier version of the ECMWF forecast model that utilizes the same temperature-dependent diagnostic of mixed-phase clouds, indicates that the clouds observed during SHEBA retained liquid at much lower temperatures than predicted by the model (Beesley et al. 2000). According to our analysis, this also appears to be an issue in ERA-Interim. Therefore, drawing conclusions on the representativeness of the bimodality observed in LWD from SHEBA using ERA-Interim alone does not seem appropriate.

3. Single-column model experiments

Given that the large-scale conditions prevailing over the SHEBA ice track appear to be well captured in ERA-Interim, as indicated by the correlative agreement in LWD and the representation of the radiatively clear-sky peak observed during the SHEBA campaign, it becomes increasingly interesting to examine if the model physics are able to represent the bimodal pattern of LWD when these large-scale conditions are prescribed. For this purpose, we use the SCM version of the global climate model EC-Earth, version 3 (Hazeleger et al. 2012). The SCM uses the large-scale conditions of wind, temperature, and specific humidity, as well as surface parameters, from ERA-Interim and is continuously relaxed toward the background state, which is updated every 6 h. This effectively means that we can diagnose the EC-Earth physics from the tendencies, but the tendencies will have little or no impact on the background state of the model. As such, we examine to what extent the model physics are able to produce a bimodality of LWD using the atmospheric state of ERA-Interim. EC-Earth, version 3, is based on the ECMWF IFS cycle 36r4. Compared to ERA-Interim, cycle 36r4 has largely updated physic routines. One main difference between ERA-Interim and EC-Earth is related to the representation of mixed-phase clouds, where cloud ice is now treated as a prognostic variable subject to advection and sedimentation as well as local source and sink processes: for example, convective processes, stratiform condensation, water vapor deposition, melting, evaporation, and autoconversion to snow (ECMWF 2012). Specifically, ice crystal growth by deposition in mixed-phase clouds, at temperatures above −38°C, is described following
e1
where qi (kg kg−1) is the mixing ratio of cloud ice and (s−1) is the glaciation speed, which is a function of the temperature-dependent ice crystal concentration (Meyers et al. 1992) and the supersaturation with respect to ice. To initiate the glaciation process in supercooled liquid clouds, a minimum ice mass mixing ratio is assumed that is dependent on an initial ice particle mass (10−12 kg) and the prescribed temperature-dependent ice crystal concentration. In mixed-phase clouds, supercooled liquid and ice are assumed to be well mixed and distributed uniformly through the cloud (ECMWF 2012).

The formulation enables us to fine tune the speed of the glaciation process by introducing a tuning parameter l in Eq. (1), which will allow for more liquid water to be retained within the clouds. The rationale for using this approach is that it is well known that Arctic wintertime clouds can contain a significant fraction of liquid water (Shupe et al. 2006; Shupe 2011; Morrison et al. 2012) and may not necessarily glaciate as fast as the midlatitude clouds for which these parameterizations are usually assumed valid. Therefore, one can view the applied tuning as a large-scale model approach to increase the resilience of Arctic mixed-phase clouds, although we do not resolve the physical processes that are responsible for upholding the liquid water content. Such unresolved processes are, for example, the sedimentation of ice from a layer of supercooled liquid (on the subgrid scale), cloud-top radiative cooling that forces condensation within the inversion layer, small-scale turbulence within the cloud layer leading to a downward mixing of overlying moist air, or a more accurate representation of aerosols available to act as ice nuclei (Morrison et al. 2012).

By introducing the tuning parameter (i.e., l) and running the model for a range of values between 0 and 1, we attempt to find an optimum value that minimizes the RMSE of LWD compared to the SHEBA winter. This optimum is found at l = 0.2, at which the correlation with LWD values from SHEBA is 0.96 with an RMSE of 7 W m−2. This optimum also represents a minimum in the RMSE of LWP values compared to the SHEBA observations. The main difference between the SCM and ERA-Interim, introduced by the tuning, is the appearance of both liquid and mixed-phase clouds in the column. These clouds now appear for extended periods of time (1–5 days) mainly below 2000 m. They are coinciding with intrusions of warm and moist air in the column and are sometimes located below inversions of equivalent potential temperature and within a relatively warmer and moister air mass. They are also coinciding in time with periods representing the radiatively opaquely peak in the observations from SHEBA (see Fig. 2).

Fig. 2.
Fig. 2.

(top) SCM column over the SHEBA ice station. Colored shadings represent equivalent potential temperature θe, black contours represent clouds with supercooled liquid water ql > 0.01 g kg−3, and green contours represent clouds with a qi > 0.01 g kg−3. (middle) LWD simulated by the untuned (black dashed) and tuned (black solid) SCM compared with observations (red). (bottom) LWP simulated by the untuned (black dashed) and tuned (black solid) SCM compared with observations (red). Discontinuities in the observations (red lines) are due to the lack of data for the corresponding period.

Citation: Journal of Climate 27, 1; 10.1175/JCLI-D-13-00271.1

Using the optimized tuning, the SCM is now run for the full 20-yr period (1990–2010) applying the same relaxation toward ERA-Interim, thereby not letting the physical parameterizations feedback on the atmospheric background state. A frequency histogram for the LWD and the cloud radiative effect, similar to Fig. 1a, is shown in Fig. 3a. With the tuning, a much better agreement between the observations from SHEBA and the tuned SCM is obtained. The SCM displays a clearly defined bimodal distribution of LWD values that corresponds both in frequency and value to SHEBA. A bimodal distribution is also present in the multiyear mean and standard deviation. All DJFM periods now display a bimodal distribution of LWD values (Fig. S3 in the supplemental material). A main contributing factor is the clearly defined peak in the longwave CRE at 65 W m−2, which was previously not present in ERA-Interim (cf. Figs. 1b, 3b). The frequency of clouds with a CRE between 10 and 40 W m−2 has instead decreased, which contributes to the gap in the distribution between the radiatively clear and opaque peak.

Fig. 3.
Fig. 3.

(a) Frequency distribution of DJFM daily mean LWD values from the tuned EC-Earth SCM forced with ERA-Interim data between 1990 and 2010. Also shown is the distribution of LWD values from the SCM during the 1997/98 winter (SHEBA winter) and observations from SHEBA covering the same period. (b) Frequency distribution of the LW CRE from the EC-Earth SCM during the DJFM period.

Citation: Journal of Climate 27, 1; 10.1175/JCLI-D-13-00271.1

4. Discussion

In this study, we have applied a tuning to the bulk glaciation speed of mixed-phase clouds. This approach ameliorates the mismatch between model and observations without targeting any specific component of the parameterization used for ice crystal growth by water vapor deposition. However, other observationally constrained pathways have been proposed to address the generally poor representation of Arctic mixed-phase clouds in models. In Prenni et al. (2007), observations of ice nuclei collected during the Mixed-Phase Arctic Cloud Experiment (M-PACE; Verlinde et al. 2007) was used to adjust the functional form of the Meyers et al. (1992) ice nuclei parameterization. This lead to a better representation of the simulated LWP and net longwave surface fluxes compared to observations. Additionally, in Xie et al. (2013), where ice nuclei concentrations in the Community Atmosphere Model, version 5 (CAM5), were based on the simulated aerosol distribution, an overall increase in LWP, due to a slowdown of the Bergeron–Findeisen process, was found in the Arctic. Another approach could be to include a more detailed, process-based subgrid model of Arctic mixed-phase clouds that would maintain high values of relative humidity despite ongoing glaciation. These approaches, including the bulk approach used within this study, would demand for a regionally dependent parameterization to be implemented in order to obtain an accurate description of the cloud condensate distribution and radiative fluxes in different cloud-type regimes. As an example, the high-level mixed-phase part of tropical deep convective clouds do not likely display the same dynamical structure as Arctic mixed-phase clouds. Also, any approach would most probably have compensating feedbacks because a change in the longwave CRE would impact the surface and atmospheric energy budgets. Therefore, in a coupled model, where atmosphere and ocean are allowed to adjust to the new parameterization globally, the resulting radiative fluxes would not necessarily be in better agreement with observations.

The effect of the tuning is evident in a joint probability density function of LWD and surface pressure (Fig. 4). The magnitude of the shift in LWD due to the applied tuning is clearly dependent on the surface pressure, such that a lower surface pressure gives a larger shift. This seems reasonable given that the appearance of mixed-phase clouds is mainly associated with the intrusion of a warmer and moister air mass, which typically is a sign of cyclone activity. The frequency plot of the tuned version shows excellent agreement with observations from SHEBA [see Fig. 4 in Morrison et al. (2012)]. The fact that the untuned model represents the LWD less well at lower surface pressures (as seen in Fig. 4a) implies that other fluxes of energy, as well as the structure of the boundary layer, are less likely to be representative of the Arctic climate in an opaquely cloudy state. During the SHEBA winter, the radiatively opaque state was associated with a near balance between LWD and surface-emitted longwave radiation (Stramler et al. 2011), which should imply a different setting for turbulent mixing. The loss of energy in a model not representing the radiatively opaque state must therefore be balanced by other fluxes: for example, a latent or sensible heat flux from the atmosphere or freezing of ice.

Compared to observations, global climate models on average underestimate the Arctic longwave CRE during winter by about 10 W m−2 (Karlsson and Svensson 2013) and it is likely that this underestimation is in part related to too little supercooled liquid in models. In terms of the DJFM mean, the tuning applied in this study corresponds to a 15 W m−2 increase in LWD, all else being equal. This difference could have a substantial impact on winter ice thickness and surface temperatures in the Arctic (Kwok and Untersteiner 2011). More specifically, considering the winter ice thickness, a net surface flux of 1 W m−2 over a year would melt 0.1 m of sea ice at its melting point (Serreze et al. 2007). Additionally, with an ongoing Arctic temperature change it is likely that a shift in the frequency of the radiatively clear state in favor of the opaquely cloudy state may occur. Given such a scenario, it is expected that a model underestimating the opaquely cloudy state would be associated with a too small cloud feedback.

Fig. 4.
Fig. 4.

Two-dimensional frequency distributions of daily values of LWD and surface pressure during the DJFM period of 1997/98 (SHEBA year) following Morrison et al. (2012). (a) The default version of the SCM. (b) The tuned (l = 0.2) version of the SCM.

Citation: Journal of Climate 27, 1; 10.1175/JCLI-D-13-00271.1

5. Conclusions

In this study, we have investigated the occurrence and representation of key features of winter Arctic states in ERA-Interim characterized by two distinctive states, a radiatively clear and opaque peak, in downwelling longwave radiation observed during the SHEBA campaign (Stramler et al. 2011; Morrison et al. 2012). In ERA-Interim, an interannually recurring bimodal distribution of LWD values is not a clearly observable feature. Additionally, large differences in the simulated liquid water content of clouds compared to observations were identified. Based on the results from our analysis, we have utilized the single-column model of the global climate model EC-Earth to investigate if, given prescribed large-scale conditions, the model's physics routines are able to reproduce the occurrence of Arctic states during a 20-yr period. We have shown that, by using a simple tuning of the glaciation speed in the single-column version of EC-Earth strongly forced by large-scale conditions from ERA-Interim, it is possible to reach a very good agreement between the model and observations from the SHEBA campaign in terms of LWD. When tuned to agree with observations from SHEBA, the SCM produces recurring Arctic states every winter during December–March between 1990 and 2010. Our tuning can be viewed as a simple way of increasing the resilience of Arctic mixed-phase clouds, which involves a multitude of dynamical and microphysical processes and feedbacks. The results do not preclude the need for improved representation of these processes, which are important for the representation of Arctic mixed-phase clouds, in models. Rather, it highlights that the presence of two preferred Arctic states, as observed during SHEBA, is a recurring feature of the Arctic climate system during winter. The longwave cloud radiative effect associated with the opaquely cloudy peak corresponds to 65 W m−2 during periods of several days. The mean increase in LWD during the Arctic winter (DJFM) compared to ERA-Interim is 15 W m−2. Although the frequency distribution of LWD will be dependent on large-scale synoptic variability, as indicated by the year–year variability found in the simulations, areas characterized by similar boundary conditions as the SHEBA site (i.e., ice-covered ocean during the polar night) are expected to display a similar dependence on the parameterization of mixed-phase clouds. This has a bearing on climate model evaluation in the Arctic as it indicates the importance of representing Arctic states in climate models and reanalysis data and doing so should have an impact on ice thickness and surface temperatures in the Arctic.

Acknowledgments

This study was performed within the Bolin Centre for Climate Research and supported by the Swedish e-Science Research Centre (SERC).

REFERENCES

  • Beesley, J. A., , C. S. Bretherton, , C. Jakob, , E. L. Andreas, , J. M. Intrieri, , and T. A. Uttal, 2000: A comparison of cloud and boundary layer variables in the ECMWF forecast model with observations at Surface Heat Budget of the Arctic Ocean (SHEBA) ice camp. J. Geophys. Res., 105 (D10), 12 33712 349.

    • Search Google Scholar
    • Export Citation
  • Cesana, G., , J. E. Kay, , H. Chepfer, , J. M. English, , and G. de Boer, 2012: Ubiquitous low-level liquid-containing Arctic clouds: New observations and climate model constraints from CALIPSO-GOCCP. Geophys. Res. Lett., 39, L20804, doi:10.1029/2012GL053385.

    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597.

    • Search Google Scholar
    • Export Citation
  • Drobot, S. D., , J. A. Maslanik, , and C. Fowler, 2003: Atmospheric and sea ice conditions during the SHEBA year: Historical and spatial assessment. Polar Geogr., 27, 1537.

    • Search Google Scholar
    • Export Citation
  • ECMWF, 2012: IFS documentation—Cy37r2 operational implementation 18 May 2011: Part IV: Physical processes. European Centre for Medium-Range Weather Forecasts Rep., 174 pp. [Available online at http://www.ecmwf.int/research/ifsdocs/CY37r2/IFSPart4.pdf.]

  • Hazeleger, W., and Coauthors, 2012: EC-Earth V2.2: Description and validation of a new seamless Earth system prediction model. Climate Dyn., 39, 26112629.

    • Search Google Scholar
    • Export Citation
  • Karlsson, J., , and G. Svensson, 2011: The simulation of Arctic clouds and their influence on the winter surface temperature in present-day climate in the CMIP3 multi-model dataset. Climate Dyn., 36, 623635, doi:10.1007/s00382-010-0758-6.

    • Search Google Scholar
    • Export Citation
  • Karlsson, J., , and G. Svensson, 2013: Consequences of poor representation of Arctic sea-ice albedo and cloud-radiation interactions in the CMIP5 model ensemble. Geophys. Res. Lett., 40, 43744379, doi:10.1002/grl.50768.

    • Search Google Scholar
    • Export Citation
  • Kwok, R., , and N. Untersteiner, 2011: The thinning of Arctic sea ice. Phys. Today, 64, 36, doi:10.1063/1.3580491.

  • Meyers, M. P., , P. J. Demott, , and W. R. Cotton, 1992: New primary ice-nucleation parameterizations in an explicit cloud model. J. Appl. Meteor., 31, 708721.

    • Search Google Scholar
    • Export Citation
  • Morrison, H., , G. de Boer, , G. Feingold, , J. Harrington, , M. D. Shupe, , and K. Sulia, 2012: Resilience of persistent arctic mixed-phase clouds. Nat. Geosci., 5, 1117.

    • Search Google Scholar
    • Export Citation
  • Prenni, A. J., and Coauthors, 2007: Can ice-nucleating aerosols affect Arctic seasonal climate? Bull. Amer. Meteor. Soc., 88, 541550.

    • Search Google Scholar
    • Export Citation
  • Serreze, M. C., , M. M. Holland, , and J. Stroeve, 2007: Perspectives on the Arctic's shrinking sea-ice cover. Science, 315, 15331536.

  • Shupe, M. D., 2011: Clouds at Arctic atmospheric observatories. Part II: Thermodynamic phase characteristics. J. Appl. Meteor. Climatol., 50, 645661.

    • Search Google Scholar
    • Export Citation
  • Shupe, M. D., , and J. M. Intrieri, 2004: Cloud radiative forcing of the Arctic surface: The influence of cloud properties, surface albedo, and solar zenith angle. J. Climate, 17, 616628.

    • Search Google Scholar
    • Export Citation
  • Shupe, M. D., , S. Y. Matrosov, , and T. Uttal, 2006: Arctic mixed-phase cloud properties derived from surface-based sensors at SHEBA. J. Atmos. Sci., 63, 697711.

    • Search Google Scholar
    • Export Citation
  • Sorteberg, A., , V. Kattsov, , J. E. Walsh, , and T. Pavlova, 2007: The Arctic surface energy budget as simulated with the IPCC AR4 AOGCMs. Climate Dyn., 29, 131156.

    • Search Google Scholar
    • Export Citation
  • Stramler, K., , A. D. D. Genio, , and W. B. Rossow, 2011: Synoptically driven Arctic winter states. J. Climate,24, 1747–1762.

  • Svensson, G., , and J. Karlsson, 2011: On the Arctic wintertime climate in global climate models. J. Climate,24, 5757–5771.

  • Tjernström, M., , and R. G. Graversen, 2009: The vertical structure of the lower Arctic troposphere analysed from observations and the ERA-40 reanalysis. Quart. J. Roy. Meteor. Soc., 135, 431443.

    • Search Google Scholar
    • Export Citation
  • Uttal, T., and Coauthors, 2002: Surface heat budget of the Arctic Ocean. Bull. Amer. Meteor. Soc., 83, 255276.

  • Verlinde, J., and Coauthors, 2007: The mixed-phase Arctic cloud experiment. Bull. Amer. Meteor. Soc., 88, 205221.

  • Xie, S., , X. Liu, , C. Zhao, , and Y. Zhang, 2013: Sensitivity of CAM5-simulated Arctic clouds and radiation to ice nucleation parameterization. J. Climate, 26, 59815999.

    • Search Google Scholar
    • Export Citation
  • Zhang, T., , K. Stamnes, , and S. A. Bowling, 2001: Impact of the atmospheric thickness on the atmospheric downwelling longwave radiation and snowmelt under clear-sky conditions in the Arctic and Subarctic. J. Climate, 14, 920939.

    • Search Google Scholar
    • Export Citation

Supplementary Materials

Save