Structure of a Warm Front: Helsinki Testbed Observations and Model Simulation

Mirja L. Kemppi Department of Physics, University of Helsinki, Helsinki, Finland

Search for other papers by Mirja L. Kemppi in
Current site
Google Scholar
PubMed
Close
and
Victoria A. Sinclair Department of Physics, University of Helsinki, Helsinki, Finland

Search for other papers by Victoria A. Sinclair in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The purpose of this study is to document the structure of a warm front in northeast Europe, identify the effects that the Finnish coastline has on the evolution of the front, and investigate factors that influence the speed that the warm front moves at within, and above, the boundary layer. The warm front formed over Estonia, traveled northward across the Gulf of Finland, and then crossed the southern coastline of Finland. Surface-based measurements from the Helsinki Testbed are analyzed together with output from a high-resolution numerical weather prediction model, Application of Research to Operations at Mesoscale (AROME). During the early stages of development, the warm front interacted with a stationary baroclinic zone and, consequently, evolved into an S shape. As the front approached the southern coast of Finland, the temperature gradient at 1000 hPa increased, as it merged with a diabatically generated temperature gradient. At 1000 hPa, the front stalled at the coastline due to friction-enhanced convergence, while the front’s speed at 860 hPa was almost uniform and unaffected by the coastline. At both 860 and 1000 hPa, the front moved slower than the wind speed. Hence, the front’s movement had a propagating component that was directed in the opposite direction to that of the front’s movement. The distribution of the ageostrophic winds showed that the front’s propagation component was produced by the front’s secondary circulation and surface friction. These results highlight the importance of surface sensible heat fluxes and friction on the evolution and movement of warm fronts.

Corresponding author address: Mirja L. Kemppi, University of Helsinki, P.O. Box 64, Gustaf Hällströmin katu 2 FI-00014, Helsinki, Finland. E-mail: mirja.kemppi@alumni.helsinki.fi

Abstract

The purpose of this study is to document the structure of a warm front in northeast Europe, identify the effects that the Finnish coastline has on the evolution of the front, and investigate factors that influence the speed that the warm front moves at within, and above, the boundary layer. The warm front formed over Estonia, traveled northward across the Gulf of Finland, and then crossed the southern coastline of Finland. Surface-based measurements from the Helsinki Testbed are analyzed together with output from a high-resolution numerical weather prediction model, Application of Research to Operations at Mesoscale (AROME). During the early stages of development, the warm front interacted with a stationary baroclinic zone and, consequently, evolved into an S shape. As the front approached the southern coast of Finland, the temperature gradient at 1000 hPa increased, as it merged with a diabatically generated temperature gradient. At 1000 hPa, the front stalled at the coastline due to friction-enhanced convergence, while the front’s speed at 860 hPa was almost uniform and unaffected by the coastline. At both 860 and 1000 hPa, the front moved slower than the wind speed. Hence, the front’s movement had a propagating component that was directed in the opposite direction to that of the front’s movement. The distribution of the ageostrophic winds showed that the front’s propagation component was produced by the front’s secondary circulation and surface friction. These results highlight the importance of surface sensible heat fluxes and friction on the evolution and movement of warm fronts.

Corresponding author address: Mirja L. Kemppi, University of Helsinki, P.O. Box 64, Gustaf Hällströmin katu 2 FI-00014, Helsinki, Finland. E-mail: mirja.kemppi@alumni.helsinki.fi

1. Introduction

Conceptual models of warm fronts have changed little from what was first conceived by the Norwegian school (Bjerknes 1919; Bjerknes and Solberg 1922). Today’s textbooks (e.g., Bluestein 1993; Ahrens 1994; Wallace and Hobbs 2006) present warm fronts as sloping zones that tilt forward over the prefrontal cold air and extend from the surface to the midtroposphere. In such conceptual models (e.g., Ahrens 1994, chapter 12) the passage of the surface warm front is indicated by a coincident temperature increase, a cyclonic wind shift, and a minimum in surface pressure. Conceptual models depict warm-front-related precipitation as being continuous, light to moderate in intensity, and located on the cold side (ahead) of the surface front (e.g., Wallace and Hobbs 2006, p. 339, Fig. 8.34).

The few improvements over the last 90 years of the conceptual model of warm fronts may be related to the limited research devoted to warm fronts. Warm fronts are generally weaker, more difficult to observe (especially at the surface), and have less dramatic weather associated with them than cold fronts. This has led to many papers documenting the structure of cold fronts in detail (e.g., Sanders 1955; Browning and Harrold 1970; Shapiro 1984), but notably fewer cases reporting on the structure of warm fronts. This relative lack of studies on warm fronts has acted as the main source of motivation for this study.

Previous warm-front-related research has often concentrated on the precipitation patterns associated with warm fronts (Browning et al. 1973; Houze et al. 1981). Warm fronts are often accompanied by rainbands embedded within the cloud mass on the cold side of the surface front (e.g., Austin and Houze 1972; Roach and Hardman 1975; Heymsfield 1979). For example, Heymsfield (1979) analyzed the wind and precipitation structure of a warm front crossing the Chicago area using Doppler radar and observed two types of organized precipitation bands that were ahead of, and oriented diagonally to, the surface warm front.

Locatelli and Hobbs (1987) used various mesoscale measurements (radar, aircraft, and rawinsondes), as well as synoptic and satellite data, to document a warm front approaching the Washington coast. Some of the features of the analyzed warm front were consistent with textbook conceptual models: the temperature rise and the veering of the winds were concentrated in a forward-sloping frontal zone, and the clouds and the precipitation were produced by the ascent of the warm air along the frontal surface. Some observed features were quite unlike the conceptual model: the warm front was not a uniformly sloping zone, but a staircaselike zone of ascent with nearly horizontal parts and steeply sloping parts. Neiman et al. (1993) observed similar structures within a warm front and referred to them as the elevator–escalator model. The “escalator” referred to regions where ascent is slantwise along gently sloped sections of the frontal surface, and the “elevator” represented mesoconvective updrafts embedded within the frontal zone.

More recently, Wakimoto and Bosart (2001) analyzed a warm front of a mature cyclone over the ocean using airborne Doppler radar data and dropsondes. Close to the cyclone center, the surface warm front was well defined, though not a discontinuity. Approximately 130 km away from the cyclone center, the front was observed aloft but not at the surface. Wakimoto and Bosart (2001) suggested that the different structures of the surface front at the two locations were due to an evolving large-scale flow. At both locations, the front had a larger horizontal temperature gradient aloft (~4 km) than at the surface, which Wakimoto and Bosart (2001) hypothesized may be due to turbulent fluxes of heat and momentum from the ocean surface acting to diffuse the surface front. This observation differs from the analytical model of frontogenesis presented by Hoskins and Bretherton (1972), which predicts that the strongest horizontal temperature gradients will develop near a rigid boundary, such as the earth’s surface, and not in the free atmosphere. This discrepancy demonstrates that boundary layer processes, such as surface fluxes, can have a large impact on the structure of fronts within the boundary layer.

Doyle and Bond (2001) documented the structure of a warm front that made landfall along the Washington coast, an area with significant coastal orography. While offshore, the front had a classical structure; a gradual wind shift and temperature increase were observed over a 20–30-km zone and the front sloped gently forward. However, the movement of the front was significantly impeded by the presence of the coast. The speed of the front decreased from 10 m s−1 when offshore to 3 m s−1 as it approached the coast.

The warm front examined in the present paper was chosen for three reasons: its intensity, its orientation, and that it occurred in autumn. In Helsinki, the temperature increased by 6°C in 3 h, rain rates of up to 5 mm h−1 were measured, and over 14 mm of rain fell in total. Given that Helsinki is located far from any major cyclogenesis regions, and is situated at the end of the climatological storm track (Petterssen 1956; Wernli and Schwierz 2006; Dacre and Gray 2009), fronts that are observed in Helsinki tend to be mature, and therefore the temperature differences between the cold and warm air masses are often small. Hence, this warm front can be considered to be an intense front. The warm front was oriented almost parallel to the southern coast of Finland, which offered a good opportunity to document the effects of the coastline on the front.

The warm front also occurred during the night at the end of October. In autumn, the difference between the surface temperature of the land and the sea is large, especially at night. The mean air temperature in Helsinki’s city center is 6.2°C (1.4°C) in October (November) (Drebs et al. 2002), while the sea surface temperature in the Gulf of Finland is 9.0°C (5.0°C) in October (November) (Lehmann and Tschersich 2006). This temperature difference should ensure that the effects of the horizontal gradient in surface type on the boundary layer structure are maximized. In spring, large temperature contrasts also exist between the sea and land. The mean air temperature in March is −1.5°C and the sea is still frozen, and in April the mean air temperature is 3.3°C and the mean sea surface temperature is 1.0°C. However, the temperature gradient is in the opposite direction compared to that in autumn and, potentially, would act to weaken northward-moving warm fronts, whereas in autumn the land–sea temperature gradient potentially acts to intensify northward-moving warm fronts.

The aim of this paper is to describe the structure and evolution of a warm front that crossed over southern Finland during the night of 30/31 October 2008 using observations from the Helsinki Testbed mesoscale network and high-resolution numerical output from the Application of Research to Operations at Mesoscale (AROME) model. Within this aim, the paper attempts to answer the following questions. What are the main characteristics of this warm front? How does the southern coastline of Finland affect the evolution of the warm front’s structure? What determines the speed of the warm front?

In section 2, the data sources and the methods are introduced. Section 3 presents the synoptic situation, and a mesoscale analysis based on synop observations is presented in section 4. Observations of the frontal passage at three locations are presented in section 5 and are compared to AROME output to verify the forecast. In section 6, the movement of the front is described and the three-dimensional evolution of the front is presented. In section 7, the structure of the warm front is compared to previous studies, and hypotheses for the front’s movement are discussed. The main conclusions are presented in section 8.

2. Data sources and methods

a. Observations

There are four sources of surface-based observations: Helsinki Testbed stations, Finnish synop stations, Estonian synop stations, and the Kivenlahti instrumented tower in southern Finland.

The Helsinki Testbed (Koskinen et al. 2011) is a mesoscale observing network with automatic weather stations that cover most of southern Finland (boxed area in Fig. 1b). Current observations are available online (http://testbed.fmi.fi). The Helsinki Testbed network consists of over 60 measuring stations, of which data from 43 were analyzed in this study. All stations are equipped with either Vaisala’s Weather Transmitter WXT510 or the updated version, WXT520 (Vaisala 2011). The transmitters measure air temperature, air pressure, relative humidity, precipitation, and wind speed and direction. Testbed stations have between one and four measurement levels and, at the majority of the stations, the transmitter is placed on a preexisting mast. In this study, two measurement levels are used; the lowest level, which is between 3 and 6 m, and the uppermost level, which is usually located between 65 and 120 m. The output files contain parameter values averaged over the previous 5 min except for wind data values, which are averaged over the previous 10 min. The data are not subject to any routine quality control and, hence, the raw data files occasionally contain missing values and some unrealistic observations. Therefore, before we performed any analysis with these data, manual quality control (removing unrealistic values and systematic errors) was essential.

Fig. 1.
Fig. 1.

Map of the study region. (a) The large box, marked by boldfaced lines, shows the domain over which AROME was run. The small box shows the location of the Helsinki Testbed. (b) Expanded view of the area within the large box in (a). Numbers relate to testbed stations: 1, Hyrsylä; 2, Liljendal; 3, Orrengrund; 4, Roihupelto; and 5, Strömsö. Here, K refers to the Kivenlahti tower and A and B to the Estonian synop stations Roomassaare and Heltermaa, respectively; J and V refer to the sounding stations Jokioinen and Visby, respectively; and GF refers to the Gulf of Finland. Shading shows the surface elevation (m).

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

The Kivenlahti tower (marked as K in Fig. 1) is a radio mast located about 15 km west of the center of Helsinki (Karppinen et al. 2002). The base of the mast is 44 m above sea level and is approximately 8 km from the coastline. The mast has nine measuring levels between 5 and 327 m above ground level. Temperature is measured at all levels except the highest, but we do not include temperature data from the 266-m level in our analysis as this sensor was found to have a warm bias. Wind speed and direction are measured at four levels (26, 93, 218, and 327 m). The 10-min-averaged values are recorded, along with maximum and minimum values.

b. Mesoscale model description

A numerical weather prediction (NWP) model, AROME (Seity et al. 2010), is used to complement the available observations and gain additional insights into the three-dimensional structure and evolution of the warm front. AROME is currently run operationally by the Finnish Meteorological Institute (FMI) over Finland with a horizontal grid spacing of 2.5 km and 40 nonuniformly distributed levels. The lowest model level is located ~30 m above the ground and there are nine model levels below 1 km, allowing boundary layer structures to be well resolved. We used the standard operational configuration, but the forecast was rerun over a smaller domain (shown in Fig. 1a) to allow additional diagnostics to be calculated. The model was initialized at 1200 UTC 30 October 2008 and run for 24 h. Initial and lateral boundary conditions are taken from the synoptic-scale NWP model, High-Resolution Limited-Area Model (HIRLAM), which FMI currently runs operationally at a horizontal grid spacing of 7.5 km over a domain covering all of northern Europe. The boundary conditions of AROME are updated every hour and history files of the model output are written every 1 h.

AROME is a nonhydrostatic, semi-Lagrangian, spectral model that includes a full range of modern physical parameterization schemes. The dynamical core of AROME (Bénard et al. 2010) is taken from the Météo-France NWP model, Aire Limitée Adaption Dynamique Initialisation-Non Hydrostatic (ALADIN-NH). Microphysics and precipitation are parameterized by the three-class ice parameterization scheme (ICE3; Pinty and Jabouille 1998; Lascaux et al. 2006), which includes six prognostic 3D variables (cloud water, cloud fraction, precipitating rain, ice, snow, and graupel). Turbulent mixing is parameterized by the Cuxart 1D turbulence scheme (Cuxart et al. 2000), which solves a prognostic equation for turbulent kinetic energy. Shallow convection is represented by an eddy diffusivity mass-flux (EDMF) scheme, which combines both eddy diffusivity and mass flux approaches (Siebesma et al. 2007). The eddy diffusion term is computed using the Cuxart turbulence scheme and the mass flux term used in the Soares et al. (2004) updraft model. Radiation is parameterized following Morcrette (1991) and surface processes are represented by the Externalized Surface (SURFEX) scheme (Le Moigne 2009).

c. Locating the front

To calculate the speed of the front at different stages of its evolution, an automatic method for locating the front was used. Objective identification of fronts was first presented by Renard and Clarke (1965), who defined the thermal frontal parameter (TFP) as
e1
where τ is either the potential temperature, equivalent potential temperature, or wet-bulb potential temperature. In their research Renard and Clarke (1965) used potential temperature, as reliable humidity data were not available. The TFP represents “the gradient of the magnitude of the gradient of a thermodynamic scalar quantity, resolved into the direction of the gradient of that quantity” (Hewson 1998). Subsequent authors have applied and modified the method of Renard and Clarke (1965) [see Table 1 of Hewson (1998) for a summary]. Hewson (1998) presented a fully automated method of locating fronts, which is based on the TFP. Zero contours of the TFP are drawn and then segments that do not represent fronts are eliminated by applying two masking variables. A closely related method is used in this study.

Whereas Hewson (1998) used wet-bulb potential temperature as the thermal variable (τ), we use potential temperature when analyzing AROME output and temperature when analyzing the observations. Recently, the advantages of using potential temperature (and to a lesser extent temperature) to locate surface fronts have been highlighted by Sanders and Doswell (1995) and Sanders (1999). Additionally, by using variables that do not include moisture, we ensure that only fronts that are thermal gradients, and not moisture gradients, are identified.

1) Model output

Hewson’s (1998) methods were developed for models with a resolution considerably lower than that of AROME. The high resolution of AROME results in noisy model output, especially in thermal fields near the surface, making it difficult to apply the exact methods of Hewson (1998). Recently, the dilemma of using Hewson’s (1998) methods for high-resolution NWP output has been acknowledged by Jenkner et al. (2010) in their study of fronts in the European Alps, where they used an NWP model with a horizontal grid spacing of 7 km. Jenkner et al. (2010) used the TFP with masking variables to locate fronts but found it necessary to first apply a simple diffusive smoothing filter with a weighted moving average for a Cartesian grid [Jenkner et al. (2010), their Eq. (1)] to the output from the high-resolution model.

To locate the warm front from AROME output, we also apply a smoothing filter to the thermal fields using the same method as Jenkner et al. (2010) but adapted for a non-Cartesian grid. The effects of the smoothing filter are shown in Fig. 3 of Jenkner et al. (2010). Once the model output has been smoothed, the frontal zones are located by calculating the TFP [Eq. (1)] with τ equal to potential temperature and plotting the zero contour of this field. To remove localized thermal gradients that are not synoptic-scale fronts, the two-dimensional frontogenesis function F,
e2
is used as a masking term. This is the Petterssen (1936) form of frontogenesis, and is also the sum of terms 2, 3, 6, and 7 in Eq. (2.3.21) in Bluestein (1993, p. 253). Two thresholds of the Petterssen frontogenesis function were selected, 2.16 K (100 km)−1 h−1 at 1000 hPa and 1.0 K (100 km)−1 h−1 at 860 hPa, and only the segments of the fronts that exceeded these threshold were retained. The threshold values were selected to produce the best match with a manual frontal analysis.

2) Observational data

Temperature time series were used to identify the time of the surface front passage. First, the temperature data (consisting of 5-min averages) from the testbed stations were smoothed by taking a 30-min running mean. This smoothing was necessary as the 5-min averages contain turbulent fluctuations, as well as mesoscale variations due to the passage of the front. A 30-min running average was selected, as this time period is the averaging time scale recommended (e.g., Stull 1988, p. 39) to remove turbulent variations. The second derivative of temperature with respect to time was then calculated, and the front’s location in time was taken from where the second derivative was a minimum.

3. Synoptic overview

Sea level pressure and 1000–500-hPa thickness from the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS) are shown in Fig. 2. At 1200 UTC 29 October (not shown), the cyclone of interest was beginning to develop in the Gulf of Genoa. By 1200 UTC 30 October, the cyclone had traveled northeastward and was located over central Europe with a central pressure of less than 988 hPa (Fig. 2a). A pronounced trough in the surface pressure stretched eastward across the Baltic countries to western Russia. Warm-air advection was occurring over large areas of eastern Europe, and strong warm-air advection was evident over the Baltic Sea and southwest Finland. A large trough in the 500-hPa geopotential height was centered over western Europe, and a strong ridge was located over Russia (Fig. 2b). Embedded within the large-scale trough, over central Europe, was a short-wave trough that was oriented northwest to southeast. An east–west-oriented jet streak at 300 hPa was present over central Finland at 1200 UTC 30 October (Fig. 2b). The warm front was evident in the routine radiosonde sounding taken from Visby, Sweden, at 1200 UTC 30 October (Fig. 3). A temperature inversion and a pronounced wind shift (from northeasterly to southerly) with height were both located at ~750 hPa.

Fig. 2.
Fig. 2.

Sea level pressure (black contours, every 4 hPa) and 1000–500-hPa thickness (gray contours every 6 dam) at (a) 1200 UTC 30 Oct and (c) 0000 UTC 31 Oct 2008, and 300-hPa wind speeds (shading according to color bar) and 500-hPa height at (b) 1200 UTC 30 Oct and (d) 0000 UTC 31 Oct 2008. The H indicates the location of Helsinki. This figure was generated from the GFS analysis on a 1.0° × 1.0° latitude–longitude grid. Local time in Finland is UTC + 2.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

Fig. 3.
Fig. 3.

Soundings from (top) Visby (57.65°N, 18.35°E), Gotland, Sweden, at 1200 UTC 30 Oct 2008 and (bottom) Jokioinen (60.81°N, 23.50°E), Finland, at 0000 UTC 31 Oct 2008. (Figures provided by the University of Wyoming, information online at http://weather.uwyo.edu/upperair/sounding.html.)

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

Twelve hours later (0000 UTC 31 October), when the surface front passed over Helsinki, the cyclone’s center was located to the east of southern Sweden and the minimum pressure remained less than 988 hPa (Fig. 2c). The 500-hPa large-scale trough over western Europe had extended farther south, and a weak ridge was evident over southern Finland (Fig. 2d). The short-wave trough had moved northeastward, intensifying slightly, and strong warm-air advection was occurring over southern and central Finland. The sounding taken from Jokioinen, 90 km north–west of Helsinki, at 0000 UTC 31 October (Fig. 3), showed the warm front to be at ~950 hPa; a temperature inversion and a wind shift were collocated at this level.

4. Mesoscale analysis

To document the evolution and movement of the front as it approached and passed over Helsinki, standard synop observations from Finland and Estonia are analyzed. At 1800 UTC 30 October (Fig. 4a), the warm front was located across southern Estonia, with temperatures of 11°C behind the front, while typical temperatures in southern Finland were 4°–5°C. Winds were southerly in southern Estonia and northeasterly in southern Finland. Three hours later at 2100 UTC (Fig. 4b), the warm front had passed over all of Estonia and was located in the Gulf of Finland. The eastern stations in the Gulf of Finland were recording warmer temperatures (8.1° and 9.7°C) than those farther west (5.8° and 6.0°C) and a similar distance from the coast. Between 2100 and 0000 UTC (Fig. 4c), the warm front moved across the Gulf of Finland, and by 0000 UTC had only begun moving onshore. (Note that the Helsinki synop station, located at the airport approximately 20 km inland, recorded a temperature of 4.6°C at 2100 UTC and only increased to 5.6°C by 0000 UTC.) At 0300 UTC 31 October (Fig. 4d), inland stations in southern central Finland had recorded small increases in temperature, whereas stations located in the sea to the southwest of Finland had recorded larger increases in temperature, as had stations in southeastern Finland. These observations demonstrate that the front advanced farther north more rapidly over the Baltic Sea to the west of Finland, and over eastern Finland, than it did in the Helsinki area.

Fig. 4.
Fig. 4.

Synop temperature and wind observations from Finland and Estonia at (a) 1800 UTC 30 Oct, (b) 2100 UTC 30 Oct, (c) 0000 UTC 31 Oct, and (d) 0300 UTC 31 Oct 2008. Wind speeds are marked as half barb, 2.5 m s−1; full barb, 5 m s−1; and pennant, 25 m s−1.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

5. Station observations and forecast validation

Before a more detailed analysis of the front is presented (section 6), the AROME forecast of the warm frontal passage is compared to observations from three different weather stations. The stations, two in Estonia and one in Finland, are located in areas where the front’s mesoscale structure evolved considerably over a short time period. First, the observed structure of the front at these three stations is described.

a. Station observations

Both observed and forecast time series of temperature, wind speed and direction, and mean sea level pressure at two Estonian synop stations are presented in Fig. 5. At the time of the frontal passage, no rapid increase in temperature was observed at Roomassaare (marked as A in Fig. 1b). Instead, the temperature increased slowly, indicating warm-air advection rather than a tight cross-frontal temperature gradient (Fig. 5a). A clockwise shift in wind direction at ~1800 UTC 30 October (Fig. 5b) marked the passage of the warm front at Roomassaare. The 10-m wind speed increased slightly (by ~2.5 m s−1) ahead of the surface front, but decreased after the front passed (0000 UTC). As the warm front approached the station, the pressure decreased from 999 to 993 hPa, and behind the front, in the warm sector, it remained constant.

Fig. 5.
Fig. 5.

The 23-h time series of 2-m temperature (°C), wind speed and direction, and mean sea level pressure at (a),(b) Roomassaare and (c),(d) at Heltermaa. Solid lines show the observations, and dashed lines the model forecast. Wind speeds are marked as half barb, 2.5 m s−1; full barb, 5 m s−1; and pennant, 25 m s−1. Shaded gray areas indicate frontal zones.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

At the second Estonian synop station, Heltermaa (marked as B in Fig. 1b), a rapid increase in temperature of 4.5°C in 4 h was observed (Fig. 5c). The surface front was observed at 2100 UTC, and had a coincident clockwise wind shift. At the same time, the mean sea level pressure decreased until the passage of the front, reaching a minimum value of 993 hPa. The 10-m wind speed at Heltermaa decreased from 7.5 m s−1 ahead of the front to 2.5 m s−1 in the warm sector (Fig. 5d).

The rate of the temperature increase differed considerably between Heltermaa and Roomassaare, despite the small distance (70 km) between the two stations. Satellite images (not shown) indicate that both stations had low and extensive cloud cover during the frontal passage, and therefore the difference in the rate of temperature increase cannot be attributed to differences in solar insolation. Evaporative cooling due to precipitation may also explain why the temperature increased more rapidly at Heltermaa than at Roomassaare. Widespread precipitation was observed at both Estonian stations at 1500 UTC (Fig. 6a), but was heavier at Heltermaa than at Roomassaare. By 1800 UTC, rain was no longer falling at Roomassaare and a few scattered showers were located north of Heltermaa (Fig. 6b). Therefore, it is unlikely that evaporative cooling is the cause of the different temperature increases. Given that the frontal zone was narrower at Heltermaa, it appears that kinematic processes, rather than diabatic processes, have intensified the frontal zone. This will be revisited in section 6a.

Fig. 6.
Fig. 6.

Pseudo-CAPPI radar composite at 500 m over Finland and northern Estonia at (a) 1500 UTC 30 Oct, (b) 1800 UTC 30 Oct, and (c) 0000 UTC 31 Oct 2008. Light precipitation is shown in blue and heavy precipitation in red. The scale for the radar reflectivity factor is described in Saltikoff et al. (2010). Model-simulated radar reflectivity at 940 hPa at (d) 1500 UTC 30 Oct, (e) 1800 UTC 30 Oct, and (f) 0000 UTC 31 Oct 2008. The thick black line shows the location of the front at 1000 hPa as diagnosed using the method described in section 2c. Note that color scales are not comparable.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

The characteristics of this warm front in southern Finland are analyzed using the observations from the Kivenlahti tower (K in Fig. 1b). Figure 7 shows time series of temperature, mean sea level pressure, precipitation rates, and wind speed and direction from both observations and the AROME forecast. Before the frontal passage, the temperature decreased by ~0.5°C at the five lowest measurement levels (Fig. 7a). This decrease was caused by evaporative cooling due to heavy precipitation, which was observed by radar and nearby rain gauges (Figs. 6b and 7b), at around 1800 UTC. A small increase in the dewpoint temperature at Kivenlahti (not shown) at 1800 UTC supports this claim. The warm front was first observed at 0000 UTC 31 October at the highest measurement level of 296 m and, finally, 4 h later, at the lowest level of 5 m (Fig. 7a). Due to the front, the temperature at Kivenlahti increased by 6°C in 3 h, and a clockwise wind shift of 110° (from northeast to southeast) coincident with the temperature increase occurred. The mean sea level pressure at Juhanila (pressure observations are not available at Kivenlahti, so the closest station equipped with a pressure sensor—6 km east of the Kivenlahti tower—is used) decreased from 1005 hPa ahead of the front to a minimum of 996 hPa in the warm sector.

Fig. 7.
Fig. 7.

Time series of temperature (°C) at Kivenlahti and mean sea level pressure (hPa) at Juhanila (gray line) from (a) observations and (c) AROME forecast. Wind barbs at Kivenlahti and hourly precipitation (bars) at Juhanila from (b) observations and (d) AROME forecast. Wind speeds are marked as half barb, 2.5 m s−1; full barb, 5 m s−1; and pennant, 25 m s−1.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

Most of the warm-front precipitation at Kivenlahti was observed between 1500 and 1800 UTC (Figs. 6a, 6b, and 7b). During these 3 h, almost continuous rain was observed and 14 mm of rain was measured at Juhanila. By 0000 UTC 31 October, when the front was first observed at Kivenlahti, only a few light showers were located over southern Finland (Fig. 6c). Therefore, the frontal precipitation was first observed at Kivenlahti 9 h before the passage of the surface front.

b. AROME forecast validation

A summary of the AROME forecasts at Roomassaare, Heltermaa, and Kivenlahti is presented in Table 1. The AROME forecast for the nearest grid point to the Estonian synop station, Roomassaare, agrees well with the observations. The timing of the temperature increase is correct, although the temperature in the warm sector (between 1800 and 0300 UTC) is overestimated by ~1°C (Fig. 5a). AROME also correctly captures the timing of the wind shift at Roomassaare, but ahead of the front (1200–1800 UTC) the forecast winds are too northerly and too weak (Fig. 5b).

Table 1.

Overview of the AROME forecasts at Roomassaare, Heltermaa, and Kivenlahti.

Table 1.

At Heltermaa, the AROME forecast 2-m temperature is 1°C warmer than the observations both ahead of the front (1200–1800 UTC) and in the warm sector (2100–0600 UTC). The timing of the surface front (at ~2100 UTC) in the AROME forecast matches reasonably well with the observations (Fig. 5b); however, the forecast frontal warming starts ~1 h later, and is more rapid, than the observed warming. AROME correctly forecasts the timing of the wind shift (2100 UTC) at Heltermaa but overestimates the wind speed throughout the whole simulation by 2.5–5 m s−1 (Fig. 5d). At both stations, AROME forecasts the decrease in mean sea level pressure well, although at both stations the observed surface pressure is slightly lower than the forecast values.

The measurements from the Kivenlahti tower compare well to the AROME forecast. The timing of the frontal passage is forecast to be at 0130 UTC at upper levels and 0230 UTC at lower levels, which roughly corresponds to the observed passage of the surface front at Kivenlahti. The temperature increase is forecast to be slightly smaller (5.5°C; Fig. 7c) than that observed (6°C; Fig. 7a). The AROME forecast of wind direction agrees well with the observations, but the forecast wind speed at lower levels is overestimated by 2.5 m s−1 (Fig. 7d). In the AROME forecast, no sign of the evaporative cooling at 1800 UTC is visible, and hence the forecast temperature at low levels is overestimated by ~1°C. The time difference between the forecast frontal passage at the highest and lowest level is ~1 h, so the slope of the front at the Kivenlahti tower is forecast to be steeper near the surface than the slope calculated from the observations. The pressure decrease is forecast correctly between 1200 and 0200 UTC, but, after this, the observed pressure continued to decrease by 1 hPa, whereas the forecast pressure begins to rise.

To validate the precipitation forecast, AROME-simulated radar reflectivities are compared to composite radar images, and AROME rainfall totals are compared to rain gauge observations from Helsinki Testbed stations. At 1500 UTC, AROME correctly predicts the location of the leading edge of the frontal precipitation (Fig. 6d), but does not capture the lighter band of precipitation to the northwest of the main rainband. The double-banded structure is better represented by AROME at 1800 UTC (Fig. 6e), but AROME predicts the main precipitation band to be too wide; the leading edge is correctly located but the trailing edge is too far south. Additionally, AROME produces rain on the warm side of the surface front in the Gulf of Finland at 1800 UTC. At 0000 UTC (Fig. 6f), the leading edge of the precipitation is correctly located, the showery nature of the precipitation is captured, and the beginnings of the cold front are also correctly modeled west of Estonia. However, the region of precipitation remains too extensive. Rain gauge data confirm that the AROME forecast of precipitation in southern Finland is better between 1500 and 1900 UTC, when continuous rain was observed, than after 1900 UTC, when the precipitation consisted of showers.

Overall, at all three stations AROME produced a forecast that can be used to investigate the structure of this warm front. Variables that are dominated by the synoptic-scale flow (surface pressure) or by the mesoscale flow (the timing and intensity of the front) are well forecast. Variables such as wind speed and absolute temperature are not as well forecast, as these are more dependent on boundary layer processes and hence the boundary layer parameterization scheme. Wind speed is the most poorly predicted variable, which is likely due to the coastal locations of all stations. This is especially true for Heltermaa, which is located immediately next to the sea.

6. Movement and evolution

We now present an analysis of the front’s structure and evolution over the whole AROME domain. We investigate the early evolution of the front, the effects that the Finnish coastline had on the structure of the front, the factors that controlled the speed of the front, the vertical tilt of the front, and the evolution of the front once it advanced inland across the coast.

a. Early evolution of the front

At 1200 UTC 30 October, two distinct baroclinic zones are evident at 990 hPa (Fig. 8a). The first extends from the eastern Gulf of Finland into western Russia (referred to as the “eastern section”) and the second is located in central Estonia (referred to as the “western section”) and is oriented southwest to northeast. Both baroclinic zones have moderate potential temperature gradients; the eastern section has a potential temperature gradient of 4.5 K (100 km)−1 and the western section a gradient of 2.75 K (100 km)−1. The eastern section has a wind shift and region of low-level convergence (Fig. 9a) collocated with the potential temperature gradient, but the western section has a less pronounced wind shift associated with it, and no localized convergence. By 1500 UTC (not shown), both sections have intensified slightly but the eastern section has remained almost stationary whereas the western section has moved northward. In contrast, at 860 hPa only one diffuse baroclinic zone is present (Figs. 8b and 9b).

Fig. 8.
Fig. 8.

Potential temperature (K, filled contours) and mean sea level pressure (hPa, thick black contours) at (left) 990 and (right) 860 hPa at (a),(b) 1200 UTC 30 Oct, (c),(d) 1800 UTC 30 Oct, and (e),(f) 0000 UTC 31 Oct 2008. Thick black line, labeled A–B, in (c) and (e) marks the location of the cross section shown in Figs. 10 and 15.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

Fig. 9.
Fig. 9.

Convergence (according to shading) and wind vectors at (left) 990 and (right) 860 hPa at (a),(b) 1200 UTC 30 Oct, (c),(d) 1800 UTC 31 Oct, and (e),(f) 0000 UTC 31 Oct 2008.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

The almost stationary eastern section is not part of the main warm front (the western baroclinic zone) at 1200 UTC 30 October, but is a separate zone of baroclinicity. A strong, stationary high pressure area is located over central Russia and zonal flow dominates over central Finland and northwest Russia. This pattern results in prolonged low-level warm-air advection over western Russia, which, once maintained for several days, leads to the formation of a temperature gradient in the broad zone of convergence between the southerly winds on the western edge of the high pressure region and the westerly winds in northern Russia. The front remains relatively stationary as it is embedded within a large-scale confluent region associated with the entrance to the jet streak (Fig. 2c). Satellite images (not shown) support this hypothesis, revealing two separate cloud bands of which the band related to the eastern section remains stationary.

At 1800 UTC 30 October, the horizontal potential temperature gradient across the western section of the front has increased as the cyclone’s center and the front have moved northward. Hence, at 1800 UTC, a strong warm front is situated over the islands west of Estonia (Fig. 8c). The increase in the horizontal potential temperature gradient at 990 hPa is believed to result, at least partially, from the change in surface friction. The increased surface roughness over the islands compared to over the sea decreases the wind speeds and deflects the winds away from their geostrophic direction. Therefore, on the southern coastline of both islands, wind speeds decrease, leading to localized regions of convergence, which amplifies the preexisting horizontal temperature gradient. Strong convergence occurs at 990 hPa on the southern coast of the southernmost island, as well as along the front over western Estonia (Fig. 9c). However, additional evidence to support the hypothesis that frictional convergence is partially responsible for the front’s intensification is found by comparing the pattern of evolution of the potential temperature and the wind field at 860 hPa to the evolution at 990 hPa between 1200 (Figs. 8a, 8b, 9a, and 9b) and 1800 UTC (Figs. 8c, 8d, 9c, and 9d). The horizontal potential temperature gradient intensifies less at 860 hPa than at 990 hPa during these 6 h. The 860-hPa wind direction at 1200 UTC shows a moderate wind shift associated with the western section of the front, but, unlike at 990 hPa, the wind shift does not become well defined by 1800 UTC. At 1800 UTC, the wind shift consists of a gradual veering collocated with the axis of the pressure trough and a second, sharper zone farther north. Additionally, at 0000 UTC, once the front is much farther north (Fig. 8e), there is still a region of convergence located on the southern coast of the Estonian island (Fig. 9e).

At 1800 UTC, the warm front (the western section) begins to merge with the stationary baroclinic zone (eastern section) forming a distinct S-shaped frontal zone across Estonia and the Gulf of Finland (Fig. 8c). As the front merges with the stationary baroclinic zone, the temperature gradient across the baroclinic zone increases (Fig. 8c). In the vertical, the front extends from the surface to 650 hPa and active kinematic (Petterssen) frontogenesis is occurring (Fig. 10a). The region of kinematic frontogenesis is almost vertical between the surface and 900 hPa (over the Gulf of Finland) and slopes forward with a constant slope above 850 hPa.

Fig. 10.
Fig. 10.

Cross sections of potential temperature (black contours, contour interval = 2 K), Pettersson frontogenesis calculated using Eq. (2) [values exceeding 1.8 K (100 km)−1 h−1 are shaded gray], and wind barbs at (a) 1800 UTC 30 Oct and (b) 0000 UTC 31 Oct 2008. Location of the cross section is shown as a thick black line in Figs. 8c and 8e. Wind speeds are marked as half barb, 2.5 m s−1; full barb, 5 m s−1; and pennant, 25 m s−1. Land is marked as black lines at the bottom of the figures.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

A third zone of baroclinicity is evident ~75 km ahead of the western section of the warm front at 1800 UTC. Associated with this third baroclinic zone is a weak, localized temperature gradient (Fig. 8c), a change in wind direction, and a region of convergence at 990 hPa (Fig. 9c). The origin of this prefrontal wind shift is unclear, and it appears to be a temporary feature because, by 2100 UTC (not shown), there is no evidence of multiple baroclinic zones, which suggests that this prefrontal feature has merged with the frontal zone. Previous studies have shown that cyclones can have multiple warm-front-like baroclinic zones (Metz et al. 2004) and that cold fronts can have prefrontal wind shifts [see Schultz (2005) for a review]. Therefore, this transient prefrontal wind shift is not believed to be unusual.

b. Effect of the Finnish coastline on the intensity of the front

Previously, Hines and Mechoso (1993) showed that large values of surface drag act to significantly weaken warm fronts. In their idealized study, they found intense, shallow, bent-back warm fronts developed when surface drag values representative of an ocean surface were used, but that when continental surface drag values were used, distinct warm fronts failed to develop. Also in an idealized modeling study, but of cold fronts, Muir and Reeder (2010) showed that including a coastline led to increased convergence, which intensified the front within the boundary layer and increased the speed that the front moved at once onshore. Therefore, we investigate if the step change in surface roughness across the Finnish coastline affects the intensity of the warm front within the boundary layer, and its movement.

At 1200 UTC, the wind direction at 990 hPa does not change across the coastline, and there is a slight decrease in wind speeds onshore compared to over the sea (Fig. 9a). Consequently, there is no coastal-induced convergence (Fig. 9a), and since the wind is blowing offshore, there is weak divergence along the coastline at 1200 UTC. By 1800 UTC (Fig. 9c), the wind speeds over the sea, but on the cold side of the front, have increased, but the coastline does not induce a change in wind direction or any localized convergence. At 0600 UTC 31 October (not shown), once the front is well inland, there are two bands of convergence of equal magnitude; one due to the front and the second located along the coastline. The coastline induces convergence at 0600 UTC as the winds are southerly, and decrease in speed inland.

At 1800 UTC (Fig. 11a) strong kinematic frontogenesis [30 K (100 km)−1 h−1] is occurring along both the eastern and western sections of the front. At 0000 UTC (Fig. 11b), stronger kinematic frontogenesis [20 K (100 km)−1 h−1] is found along the eastern part of the front, which is oriented along the coastline, than along the western section of the front [8 K (100 km)−1 h−1], which is located over the sea.

Fig. 11.
Fig. 11.

Pettersson frontogenesis [shading according to color bar, K (100 km)−1 (h)−1] at (a) 1800 UTC 30 Oct and (b) 0000 UTC 31 Oct 2008 and frontogenesis due to surface sensible heat fluxes at (c) 1800 UTC 30 Oct and (d) 0000 UTC 31 Oct 2008. Surface pressure is shown by black lines, every 2 hPa.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

Surface heat fluxes have also previously been shown to modify the structure of fronts within the boundary layer. At 1800 UTC (Fig. 12a) there are two horizontal gradients in the surface sensible heat flux. The first is collocated with the surface front and the second is associated with the coastline. Frontogenesis due to the surface sensible heat fluxes (Fig. 11c) shows that at 1800 UTC surface sensible heat fluxes are frontolytical across the front but frontogenetical across the coastline. The surface sensible heat fluxes act to create a localized “front” across the coastline, which is confined to within the boundary layer. When the front arrives at the coastline at 0000 UTC, there is no longer a gradient in surface sensible heat fluxes across the coast, and consequently, no frontogenesis due to the surface sensible heat fluxes. However, the front merges with the low-level coastal-induced “front,” which acts to increase the temperature gradient across the front within the boundary layer.

Fig. 12.
Fig. 12.

Surface sensible heat fluxes (shading according to color bar, W m−2) and surface momentum fluxes (vectors) at (a) 1800 and (b) 0000 UTC. Positive heat fluxes warm the boundary layer and negative heat fluxes cool the boundary layer.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

c. Factors controlling the speed of the front

The location of the front in the AROME forecast, as defined by the method given in section 2c, is presented in Fig. 13. The front is located every 2 h between 1600 UTC 30 October and 0400 UTC 31 October at two pressure levels: 860 and 1000 hPa. Identifying the location of the front near the surface (1000 hPa) and at a level above the boundary layer (860 hPa) allows us to identify the impacts that the Finnish coastline has on the movement of the front. To quantify the movement of the front, the speed of the front has been calculated by measuring the distance between the frontal locations in Fig. 13 and then dividing by the time difference. This method assumes that the frontal speed is constant over the 2-h period.

Fig. 13.
Fig. 13.

Location of the front at (a) 860 and (b) 1000 hPa based on the AROME simulations at 1600 (black), 1800 (blue), 2000 (green), 2200 (yellow), 0000 (orange), 0200 (red), and 0400 UTC (brown).

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

At 860 hPa, the front’s movement over the Gulf of Finland and Finland was almost uniform and showed little spatial variability (Fig. 13a). Between 1800 and 2000 UTC, the front passed over the Finnish coastline. The front’s speed was approximately constant at 7 m s−1 between 1900 and 2300 UTC and decreased to 5.5 m s−1 by 0300 UTC (Fig. 14a). Because the front had a similar speed when it crossed the coastline to its speed once inland, we can conclude that the coastline had no influence at this level (which was expected) but also that the synoptic-scale forcing determining the frontal movement remained approximately constant.

Fig. 14.
Fig. 14.

The speed of the front (solid lines) and the wind speed behind the front estimated from AROME output (dashed lines) at (a) 860 hPa approximately along the vertical cross section shown in Figs. 10 and 15, (b) 1000 hPa in the western part of the AROME domain (over the sea), (c) 1000 hPa in the same location as in (a), and (d) at 1000 hPa in the eastern part of the AROME domain.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

In contrast, at 1000 hPa, the movement of the front varies in both time and space, with the western part of the front located over the sea moving faster than the eastern part, which encounters the Finnish coastline (Fig. 13b). At 2300 UTC, the western part of the front moves at 12 m s−1 (Fig. 14b), while the central and eastern sections move at 2 m s−1 (Figs. 14c and 14d). This suggests that the movement of the front at 1000 hPa is impeded by the Finnish coastline.

Smith and Reeder (1988) reviewed theories and observations of cold front movement with the aim of identifying factors that determine the speed that fronts move at, but found no clear consensus. Frontal movement can be considered as the sum of advection plus propagation (Markowski and Richardson 2010, chapter 5), and if the frontal speed differs from the wind speeds behind the front, then the front is said to be propagating. The propagating component of the frontal movement can either be in the same direction as the advection component so that the front will move faster than the wind speed, or in the opposite direction to the advection component, causing the front to move slower than the wind speed. Figure 14 shows that, at both 860 and 1000 hPa, and at all times, the warm front moves slower than the wind speeds, demonstrating that this front has a propagating component in the opposite direction to the frontal movement. In the central part of the Gulf of Finland, the front at 860 hPa moved at 60% of the wind speed as it approached the coastline and 45% of the wind speed when over land (Fig. 14a). In the same location, at 1000 hPa, the front moved at 15% of the wind speed as it approached the coastline and 30% of the wind speed once inland (Fig. 14c).

Warm fronts can have a propagation component to their movement that is in the opposite direction to the front’s movement due to the thermally direct ageostrophic circulation. Ascent in the warm air (behind the surface warm front) and descent in the cold air (ahead of the surface front) mean that at low levels the horizontal ageostrophic wind component is directed from the cold air to the warm air and, hence, in the opposite direction to the front’s movement. At upper levels, the ageostrophic flow is from warm to cold air and, thus, in the same direction as the front’s movement. Therefore, it is plausible that the ageostrophic circulation decreases the speed of warm fronts at low levels and increases it at upper levels. Figure 15 shows the horizontal component of the ageostrophic wind along the cross section marked in Fig. 8c. However, the ageostrophic wind does not result solely from the cross-front circulation; additional processes, such as surface friction, will contribute to the ageostrophic component of the wind.

Fig. 15.
Fig. 15.

Cross sections of potential temperature (black contours, contour interval = 2 K) and horizontal ageostrophic wind component parallel to the cross section at (a) 1800 UTC 30 Oct and (b) 0000 UTC 31 Oct 2008. Location of the cross section is shown as a thick black line in Figs. 8c and 8e. Wind speeds are marked as half barb, 2.5 m s−1; full barb, 5 m s−1; and pennant, 25 m s−1. Land is marked with black lines at the bottoms of the figures. Gray shading shows areas where the ageostrophic wind is directed in the opposite direction to the movement of the front.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

At 1800 UTC, when the front is located in the Gulf of Finland, the ageostrophic wind is directed in the opposite direction of the front’s movement on the cold side of the front, from the surface up to 700 hPa, and in the same direction as the front’s movement on the warm side of the front, but only above 950 hPa. Within the boundary layer (approximately below 950 hPa), the ageostrophic wind is stronger than in the free troposphere. Additionally, the ageostrophic winds are stronger in the boundary layer over land than in the marine boundary layer, indicating that surface friction, as well as the front’s secondary circulation, are contributing to the ageostrophic wind component within the boundary layer. This is in agreement with Keyser and Anthes (1982), who compared the component of the ageostrophic wind driven by the acceleration term in the momentum equations to the component driven by the friction term, and found that in the boundary layer the friction term was much larger.

At 0000 UTC (Fig. 15b), the ageostrophic wind remains directed in the opposite direction of the front’s movement at low levels and in the same direction at upper levels. At both 860 and 1000 hPa at 1800 and 0000 UTC, the ageostrophic wind is acting in the opposite direction of the front’s movement, but the ageostrophic wind is much stronger at 1000 hPa (typically 15–20 m s−1) than at 860 hPa (typically 5–10 m s−1). Therefore, the warm front likely moves faster at 860 than at 1000 hPa in part due to the weaker component of the ageostrophic wind directed in the opposite direction to the front’s movement.

It is also plausible that the stability of the prefrontal boundary layer affected the movement of the warm front, especially as the front advanced onshore. Idealized modeling by Muir and Reeder (2010) showed that a cold front was impeded from advancing onshore when the onshore boundary layer was convective and well mixed. Two reasons were given for this: first, strong turbulence reduced the low-level wind speeds, and, second, the heating onshore due to surface heat fluxes was greater than the cooling due to cold-air advection behind the cold front. When the onshore boundary layer was stably stratified, the front advanced onshore without delay as low-level winds were stronger and surface fluxes no longer heated the onshore boundary layer. In the case of this warm front, at 1800 UTC, the onshore boundary layer is shallow (400 m) and stably stratified (Fig. 16a), and the low-level wind speeds are moderate (7.5 m s−1). The temperature increase due to warm-air advection was estimated to be greater (+2.3 K h−1) than the temperature decrease due to the onshore surface heat fluxes (−0.2 K h−1), and hence, the stable stratification is unlikely to prevent the warm front from advancing onshore.

Fig. 16.
Fig. 16.

Vertical profiles of potential temperature taken from the AROME forecast at 25°E at evenly spaced latitudes between 59° and 61°N. Solid black lines show profiles over Estonia, dashed black lines are over the Gulf of Finland, and solid gray lines are over Finland at 1800 UTC 30 Oct and 0000 UTC 31 Oct 2008.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

d. Vertical tilt of the front

Local lapse rates are examined to quantify the stratification of the prefrontal boundary layer and the frontal zone, which can be related to the vertical tilt of the front. At Kivenlahti, observations show that the air was conditionally unstable between 1200 and 1500 UTC (local lapse rates are between 6° and 9°C km−1; Fig. 17). As the front approached, the lapse rate decreased, indicating a stable prefrontal boundary layer, which is in agreement with the negative heat fluxes forecast by AROME at 1800 UTC (Fig. 12a). Within the frontal zone (2100 UTC 30 October–0400 UTC 31 October), the air was very stable (minimum lapse rate ~−12°C km−1 at 0000 UTC), yet ascent and hence precipitation still developed due to adequate isentropic uplift. AROME overestimates the lapse rates at all times, but especially within the frontal zone; the minimum lapse rate in AROME is ~−4°C km−1.

Fig. 17.
Fig. 17.

Local lapse rate (−dT/dz) between the highest and the lowest levels at the Kivenlahti tower from observations (solid line) and AROME simulation (dashed line). Units are K km−1. Horizontal gray dashed lines show the dry-adiabatic lapse rate (9.8 K km−1), the approximate moist-adiabatic lapse rate (6.0 K km−1), and the isothermal lapse rate (0 K km−1).

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

The slope of the front can be estimated from the Kivenlahti tower observations and from AROME output. Using the average speed of the front in southern Finland calculated from the testbed observations (~3.2 m s−1), the slope of the front, as observed by the Kivenlahti tower, is estimated to be ~1:180. AROME forecasts the front to be too vertical in the lowest 300 m; the time between the front arriving at the highest and lowest levels is 1 h compared to 4 h in the observations (Fig. 7). The slope of the front in the AROME output is also estimated by calculating the distance between the location of the front at 1000 and 860 hPa, which allows the slope of the front above the surface layer to be quantified. Figures 10a, 10b, 13a, and 13b, show that the slope of the warm front over southern Finland decreases with time, from ~1:56 at 2000 UTC, to ~1:153 at 0200 UTC on 31 October. The magnitude of the front’s slope at 0200 UTC between 1000 and 860 hPa (approximately 1.2 km) agrees with the ratio calculated from the Kivenlahti tower observations. This suggests that AROME misdiagnosed the slope of the front more significantly in the surface layer than above the surface layer. Excessive turbulent mixing in the surface layer may explain why the front is too vertical in the AROME forecast, since mixing results in vertical isentropes [see Fig. 4 of Peng et al. (2001) for a schematic illustration of this process]. The onshore boundary layer at 0000 UTC 31 October is well mixed up to 960 hPa (Fig. 16b) and thus supports this hypothesis.

The slope of this warm front is shallow when compared to previous studies. Wakimoto and Bosart (2001) studied a warm front that had a slope of 1:60, and Heymsfield (1979) observed a warm front that had a slope of 1:100. For clouds, and hence precipitation, to form, air parcels must be lifted slantwise along the sloping frontal surface until they reach their lifting condensation level (LCL). As this warm front has a shallow slope (1:180), the air parcels will travel considerable horizontal distances before they reach their LCL. Hence, if precipitation is caused only by the frontal circulation, and not by synoptic-scale processes such as thermal advection, this warm front will have its precipitation located farther ahead of the surface front in the cold air (recall that at Kivenlahti rain began 9 h before the surface front and there was a 3-h dry period before the surface front was observed) than steeper warm fronts.

e. Evolution of the front once inland

Once the surface front crossed the Finnish coastline, its structure and movement could be examined in detail using observations from Helsinki Testbed stations. The surface front was first observed at the easternmost marine station, Orrengrund (number 3 in Fig. 1b). At 2300 UTC, there was a 4°C temperature increase (Fig. 18a), which coincided with a rapid clockwise wind shift (Fig. 18b). The temperature increase was not constant across the frontal zone (between 1800 and 2300 UTC), but became more rapid toward the warm side of the front. At stations close to the coast, Roihupelto and Strömsö (station numbers 4 and 5 in Fig. 1b), the temperature increased by 6°C, and again there was a coincident clockwise change in wind direction. The temperature increase was 2°C greater at these coastal stations compared to the marine station Orrengrund, but the observed temperature behind the front was similar at all three stations.

Fig. 18.
Fig. 18.

Time series of (a) potential temperature and (b) wind speed and direction and precipitation [mm (3 h)−1] at five Helsinki Testbed stations. Numbers refer to those in Fig. 1: 1, western inland station of Hyrsylä (magenta); 2, eastern inland station of Liljendal (gray); 3, marine station of Orrengrund in eastern part of the testbed (blue); 4, central station of Roihupelto near Helsinki (red); and 5, coastal station of Strömsö in western part of the testbed (green). Wind speeds are marked as half barb, 2.5 m s−1; full barb, 5 m s−1; and pennant, 25 m s−1.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

As the front moved farther inland, alongfront variations began to develop. At Liljendal (station 2 in Fig. 1b), a rapid temperature increase of 5°C in ~3 h was observed, whereas, farther to the west at Hyrsylä (station number 1 in Fig. 1b), the temperature increased much more gradually (5.5°C in ~6 h). The timing of the front differed considerably at these two stations; the surface front was observed at 0200 UTC at Liljendal compared to at 0600 UTC at Hyrsylä. The location of the front (Fig. 19) shows that the front moved farther inland over eastern Finland than in the Helsinki region, which is in general agreement with the location of the fronts diagnosed from the AROME output at 1000 hPa (Fig. 13a).

Fig. 19.
Fig. 19.

Location of the front based on the Helsinki Testbed observations. The front’s location is marked every 30 min from 0000 UTC 31 Oct to 0500 UTC 31 Oct 2008. The front is not located in the northern part of the plotted domain as there are no testbed stations there.

Citation: Monthly Weather Review 139, 9; 10.1175/MWR-D-10-05003.1

The speed of the front, calculated from the testbed observations, varied as it moved across the coastline and inland throughout the testbed region. Near Helsinki, the movement of the front decreased from ~3.5 to ~1.5 m s−1 as the front approached the coast. This is in general agreement with the AROME output, which shows the speed of the eastern part of the front to increase between 2300 and 0300 UTC to ~5 m s−1, but the speed of the central part of the front to increase slightly between 2300 and 0100 UTC and then to decrease between 0100 and 0300 UTC. The different frontal movements are thought to be related to the topography. In southwestern Finland, the topography is notably (>50 m) higher than in southeastern Finland where there are numerous large lakes (Fig. 1b).

7. Discussion

We have analyzed the evolution of a warm front that crossed the southern Finnish coastline during the night of 30/31 October 2008. Both dynamical and boundary layer processes were found to affect the front’s structure and speed.

a. Frontal structure

In its early stages, the front’s structure was affected by synoptic-scale, dynamical processes. The front merged with a stationary baroclinic zone, forming an S-shaped front that was approximately 650 km long. Warm fronts at the end of the climatological storm tracks have previously been shown to be short or “stubby” features due to diffluence (Schultz et al. 1998 and references therein). Diffluence at the end of zonally oriented jet streams occurs in concert with stretching deformation, which is located in regions where the axes of dilation are meridionally oriented [see Fig. 1b of Black and Dole (2000) for a graphic illustration of this]. However, despite being located at the end of the climatological storm track, this front was not a short feature, most likely due to the anomalous synoptic pattern. The warm front merged with a stationary baroclinic zone that existed due to an area of confluence that was collocated with an entrance region of a jet streak.

As the front moved across the Finnish coastline, the structure of the front in the boundary layer was modified. The temperature gradient across the front increased as the front merged with an along-coast temperature gradient that was generated by differential surface sensible heat fluxes. The surface sensible heat fluxes were positive over the warmer sea and negative over the cooler land. Previous idealized modeling studies (Reeder 1986; Physick 1988) have shown landfalling cold fronts to intensify when surface sensible heat fluxes are positive over land and smaller over the sea. Therefore, despite the opposite gradient of sensible heat fluxes, our results agree with these earlier studies. When the front attempted to move inland, the winds became directed onshore. Due to the change in surface roughness, and surface friction, this increased convergence along the coastline. The convergence acted on the preexisting temperature gradient, decreasing the width of the frontal zone.

b. Frontal speed

At 860 hPa, the front moved slower than the wind speeds behind the front. At this level, above the boundary layer, the front’s speed was not affected by surface friction or by the stratification of the boundary layer. Therefore, we conclude that the front is not simply advected by wind. This conclusion is very likely applicable to all warm fronts. Bjerknes and Solberg (1921) analyzed a warm front that also moved slower than the wind speed and state that the difference “represents the movement of the warm air upwards along the cold wedge.” Hence, it is possible that this warm front moved slower than the mean horizontal wind as a proportion of the horizontal wind was directed upward along the slope of the front. The vertical velocity was estimated to be ~8 cm s−1 by considering the slope of the front (1:60) and the difference between the horizontal wind speed and the front’s speed (5 m s−1). This estimate of the vertical velocity compares well with the vertical velocities in the AROME simulation (not shown) at this level. Alternatively, the front’s secondary circulation, as inferred from the ageostrophic wind, which produced winds in the opposite direction to that of the front’s movement at both 1000 and 860 hPa, may explain why the front moved slower than the wind speeds behind the front. It appears very likely, especially within the free troposphere, that the front’s secondary circulation is a source of the front’s propagating component.

The difference between the warm front’s speed and the horizontal wind speeds may also be interpreted as air parcels are able to move through the frontal zone and that the front is not a material surface. Entrainment through cold fronts within the boundary layer has previously been investigated. Well-mixed boundary layers are commonly found behind cold fronts (e.g., Bond and Fleagle 1988; Nuss 1989; Sinclair et al. 2010). Two explanations exist for this: cold-air advection over warmer surfaces leads to positive surface heat fluxes that destabilize the boundary layer, or warm air is entrained through the frontal zone into the cold air. In an observational study of an intense cold front, Sanders (1955) hypothesized that entrainment of warm air through the frontal zone was occurring. Recently, Schultz and Roebber (2008) reanalyzed the same cold front using a high-resolution NWP model and found that, in a model simulation with no surface fluxes, a well-mixed postfrontal boundary layer still developed, indicating that entrainment through the frontal zone was occurring. Although these previous studies have concentrated on cold fronts, it is likely that air can be horizontally entrained through warm fronts.

The speed of the front at 1000 hPa differed from the speed of the front at 860 hPa, indicating that within the boundary layer additional processes were affecting the front’s movement. As the front approached the Finnish coast, the speed of the front near the surface decreased, and the front appeared to stall along the coastline. Previous studies have investigated the effects that coastlines have on frontal movement and structure. However, they have tended to be either idealized studies (e.g., Reeder 1986; Physick 1988), about cold fronts (e.g., Garratt 1988), or about fronts making landfall in areas of coastal mountains (e.g., Colle et al. 1999; Neiman et al. 2004). Doyle and Bond (2001) showed a warm front to decelerate from 10 to 3 m s−1 as it approached a coastline. The flow ahead of this warm front was blocked by steep orography, so it is not easily comparable with the warm front that was studied here. In this study, the warm front decelerated when it encountered the relatively flat terrain of southern Finland, suggesting that coastlines without complex topography inland can also alter the motion of warm fronts.

Two possible explanations exist for why the warm front decelerated as it approached the Finnish coastline: the change in surface roughness altered the surface friction, generating a zone of convergence along the coast, effectively tying the front to the coast, or the stable stratification of the onshore boundary layer impeded the movement of the warm front onshore.

Evidence that supports the hypothesis that the change in surface friction, and low-level winds, caused the decrease in the front’s speed includes that the surface momentum fluxes from AROME changed in magnitude and direction at the coastline. Before the front reached the coastline, there was no coastal convergence as the winds were offshore. However, once the front attempted to move inland, the winds became onshore, creating additional convergence. Additionally, the ageostrophic winds were larger in the boundary layer than in the free troposphere, indicating that surface friction, as well as the front’s secondary circulation, were contributing to the ageostrophic wind component. Within the boundary layer, the ageostrophic winds were directed in the opposite direction to the front’s movement, impeding the front’s progress.

Evidence that suggests that the stability hypothesis is false includes that the surface sensible heat fluxes from the AROME forecast changed little across the coast at 0000 UTC (Fig. 12b) and that the prefrontal stable layer is very shallow (Fig. 16a) and only weakly stable (the temperature decreased by 1°C over the depth of the Kivenlahti tower). If the prefrontal boundary layer had been more stably stratified, it is possible that the front would have been lifted over the stable layer and not observed at the surface. Since warm fronts slope gently forward, it is plausible that near the surface warm fronts move due to vertical turbulent mixing in the cold air ahead of the surface front. Warm air from aloft, behind the warm front, may be transported down to the surface by turbulence. If the boundary layer had been very strongly stratified, this process may not have been possible, or may have acted on a much slower time scale. Xu and Gu (2002) highlighted this method of maintaining a temperature front at the surface. In an idealized modeling study, they showed that when no-slip boundary conditions were used, surface friction reduced the thermal advection near the surface and that the maximum temperature gradient was generated about the surface layer. The temperature gradient at the surface was maintained by vertical diffusion.

Once inland, the eastern part of the warm front moved faster than the western part and we hypothesize that this is due to the large lakes in eastern Finland. Gallus and Segal (1999) showed a cold front that accelerated as it moved over Lake Michigan due to the decreased surface roughness and increased thermal stratification behind the front. Both effects caused a decrease in postfrontal turbulence, which allowed the wind speed behind the front to increase. It is likely that similar processes are occurring in the warm front presented here.

8. Conclusions

This study is, to our knowledge, the first detailed description of an intense warm front in northern Europe at the end of the climatological storm track. The goals of the analysis presented here were to document the structure of the warm front, to identify the role that the Finnish coastline had in modifying the structure of the warm front, and to investigate what controls the speed of the warm front.

The warm front developed in a region of the cyclone favorable for kinematic frontogenesis. Initially, two baroclinic zones were evident, a stationary, intense zone over Russia, and a weaker, more mobile zone over Estonia that was south of the western baroclinic zone. Within 6 h, the two baroclinic zones merged forming an S-shaped front along the Gulf of Finland. As the front came onshore, the potential temperature gradient near the surface intensified, as it merged with a diabatically generated temperature gradient that was oriented along the coast. Coastal-induced convergence then delayed the movement of the front inland. Specific conclusions arising from the analysis include the following:

  • Many aspects of the warm front’s structure agree with conceptual models. The temperature increase and the veering of the winds were collocated within a forward-sloping frontal zone, and the precipitation was continuous and preceded the surface front.

  • As the front approached the southern coast of Finland, the temperature gradient intensified and the front’s speed decreased.

  • The strongest potential temperature gradient across the front was found near, but not at, the surface. This agrees with the idealized, analytical model of Hoskins and Bretherton (1972), but disagrees with the observations of Wakimoto and Bosart (2001). The strong temperature gradient at low levels is due to limited turbulent mixing in the boundary layer.

  • The front had a constant slope in the AROME vertical cross sections and did not exhibit a staircaselike structure as had been previously found by Locatelli and Hobbs (1987) and Neiman et al. (1993), indicating that deep embedded convection was not present in this warm front.

  • Model output showed the front’s movement to be spatially uniform at 860 hPa but not at 1000 hPa. At both pressure levels the front’s movement had a propagating component, directed in the opposite direction to that of the front’s movement. At 860 hPa, the front’s speed was between 45% and 60% of the wind speed behind the front, and at 1000 hPa, the speed of the front varied from 15% to 30% of the postfront wind speeds.

  • The ageostrophic wind, calculated from AROME output, was directed in the opposite direction from the front’s movement at both 860 and 1000 hPa, but was larger within the boundary layer than in the free troposphere. This suggests that the front’s propagation component was produced by the front’s secondary circulation and by surface friction.

  • The topography and lakes of southern Finland caused the front to move nonuniformly inland. The front moved faster over the smooth lakes in the east than over the higher topography and rougher surface (forest) in the west.

This case study has highlighted the importance of boundary layer processes, especially surface friction, on the near-surface structure of warm fronts and has attempted to identify the processes that control the speed of warm fronts. Although only one case study was presented, the results of which cannot be generalized to all warm fronts, this study addressed the long-unanswered question of what causes fronts to move. Previously, authors have investigated what affects the movement of cold fronts, but still no firm conclusions have been reached on this topic. Considerably less research has been conducted on warm fronts than cold fronts, and consequently the literature is currently devoid of explanations as to why warm fronts move. To address this situation, a more systematic approach needs to be considered. This will include analyzing observations, real-data modeling studies, and abstracting the problem to an idealized setting.

Acknowledgments

We thank the Finnish Meteorological Institute for making the numerical model, AROME, available for our use; Miina Krabbi from the Estonian Meteorological and Hydrological Institute for the observational data from Estonia; Jani Poutiainen from the Finnish Meteorological Institute for providing the Helsinki Testbed and Kivenlahti tower data; and Pauli Jokinen from the Finnish Meteorological Institute for providing the Finnish synop data. We also thank David Schultz, Hannu Savijärvi, Tom Galarneau, and an anonymous reviewer who provided valuable comments on earlier versions of this manuscript. Both authors are funded by Grant 126853 from the Academy of Finland.

REFERENCES

  • Ahrens, C. D., 1994: Meteorology Today: An Introduction to Weather, Climate, and the Environment. 5th ed. West Publishing, 528 pp.

  • Austin, P. M., and R. A. Houze Jr., 1972: Analysis of the structure of precipitation patterns in New England. J. Appl. Meteor., 11, 926935.

    • Search Google Scholar
    • Export Citation
  • Bénard, P., J. Vivoda, J. Mašek, P. Smolíková, K. Yessad, C. Smith, R. Brožková, and J.-F. Geleyn, 2010: Dynamical kernel of the Aladin-NH spectral limited-area model: Revised formulation and sensitivity experiments. Quart. J. Roy. Meteor. Soc., 136, 155169.

    • Search Google Scholar
    • Export Citation
  • Bjerknes, J., 1919: On the structure of moving cyclones. Geofys. Publ., 1 (2), 18.

  • Bjerknes, J., and H. Solberg, 1921: Meteorological conditions for the formation of rain. Geofys. Publ., 2 (3), 361.

  • Bjerknes, J., and H. Solberg, 1922: Life cycle of cyclones and the polar front theory of atmospheric circulation. Geofys. Publ., 3 (1), 138.

    • Search Google Scholar
    • Export Citation
  • Black, R. X., and R. M. Dole, 2000: Storm tracks and barotropic deformation in climate models. J. Climate, 13, 27122728.

  • Bluestein, H., 1993: Observations and Theory of Weather Systems. Vol. II, Synoptic–Dynamic Meteorology in Midlatitudes, Oxford University Press, 594 pp.

    • Search Google Scholar
    • Export Citation
  • Bond, N. A., and R. G. Fleagle, 1988: Prefrontal and postfrontal boundary layer processes over the ocean. Mon. Wea. Rev., 116, 12571273.

    • Search Google Scholar
    • Export Citation
  • Browning, K., and T. Harrold, 1970: Air motion and precipitation growth at a cold front. Quart. J. Roy. Meteor. Soc., 96, 369389.

  • Browning, K., M. Hardman, T. Harrold, and C. Pardoe, 1973: The structure of rainbands within a mid-latitude depression. Quart. J. Roy. Meteor. Soc., 99, 215231.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., C. F. Mass, and B. F. Smull, 1999: An observational and numerical study of a cold front interacting with the Olympic Mountains during COAST IOP5. Mon. Wea. Rev., 127, 13101334.

    • Search Google Scholar
    • Export Citation
  • Cuxart, J., P. Bougeault, and J. L. Redelsperger, 2000: A turbulence scheme allowing for mesoscale and large-eddy simulations. Quart. J. Roy. Meteor. Soc., 126, 130.

    • Search Google Scholar
    • Export Citation
  • Dacre, H. F., and S. L. Gray, 2009: The spatial distribution and evolution characteristics of North Atlantic cyclones. Mon. Wea. Rev., 137, 99115.

    • Search Google Scholar
    • Export Citation
  • Doyle, J. D., and N. A. Bond, 2001: Research aircraft observations and numerical simulations of a warm front approaching Vancouver Island. Mon. Wea. Rev., 129, 978998.

    • Search Google Scholar
    • Export Citation
  • Drebs, A., A. Nordlund, P. Karlsson, J. Helminen, and P. Rissanen, 2002: Tilastoja suomen ilmastosta 1971–2000. Climatological Statistics of Finland 1971–2000, Finnish Meteorological Institute, 99 pp.

    • Search Google Scholar
    • Export Citation
  • Gallus, W. A., Jr., and M. Segal, 1999: Cold front acceleration over Lake Michigan. Wea. Forecasting, 14, 771781.

  • Garratt, J. R., 1988: Summertime cold fronts in southeast Australia—Behavior and low-level structure of main frontal types. Mon. Wea. Rev., 116, 636649.

    • Search Google Scholar
    • Export Citation
  • Hewson, T., 1998: Objective fronts. Meteor. Appl., 5, 3765.

  • Heymsfield, G., 1979: Doppler radar study of a warm frontal region. J. Atmos. Sci., 36, 20932107.

  • Hines, K. M., and C. R. Mechoso, 1993: Influence of surface drag on the evolution of fronts. Mon. Wea. Rev., 121, 11521175.

  • Hoskins, B. J., and F. P. Bretherton, 1972: Atmospheric frontogenesis models: Mathematical formulation and solution. J. Atmos. Sci., 29, 1137.

    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., S. A. Rutledge, T. J. Matejka, and P. V. Hobbs, 1981: The mesoscale and microscale structure and organization of clouds and precipitation in midlatitude cyclones. III: Air motions and precipitation growth in a warm-frontal rainband. J. Atmos. Sci., 38, 639649.

    • Search Google Scholar
    • Export Citation
  • Jenkner, J., M. Sprenger, I. Schwenk, C. Schwierz, S. Dierer, and D. Leuenberger, 2010: Detection and climatology of fronts in a high-resolution model reanalysis over the Alps. Meteor. Appl., 17, 118.

    • Search Google Scholar
    • Export Citation
  • Karppinen, A., S. Joffre, and J. Kukkonen, 2002: Evaluation of meteorological data measured at a radio tower in the Helsinki Metropolitan Area. Proc. COST 715 Expert Meeting, Toulouse, France, European Commission EUR 20451, 89–98.

    • Search Google Scholar
    • Export Citation
  • Keyser, D., and R. Anthes, 1982: The influence of planetary boundary layer physics on frontal structure in the Hoskins–Bretherton horizontal shear model. J. Atmos. Sci., 39, 17831802.

    • Search Google Scholar
    • Export Citation
  • Koskinen, J. T., and Coauthors, 2011: The Helsinki Testbed: A mesoscale measurement, research, and service platform. Bull. Amer. Meteor. Soc., 92, 325342.

    • Search Google Scholar
    • Export Citation
  • Lascaux, F., E. Richard, and J. Pinty, 2006: Numerical simulations of three different MAP IOPs and the associated microphysical processes. Quart. J. Roy. Meteor. Soc., 132, 19071926.

    • Search Google Scholar
    • Export Citation
  • Lehmann, A., and G. Tschersich, 2006: Trends in sea surface temperature of the Baltic Sea since 1990. BALTEX Newsletter, No. 9, International BALTEX Secretariat, Geesthacht, Germany, 7–9.

    • Search Google Scholar
    • Export Citation
  • Le Moigne, P., 2009: SURFEX scientific documentation. CNRM Tech. Rep., Météo-France/CNRS, Toulouse, France, 211 pp.

  • Locatelli, J., and P. Hobbs, 1987: The mesoscale and microscale structure and organization of clouds and precipitation in midlatitude cyclones. XIII: Structure of a warm front. J. Atmos. Sci., 44, 22902309.

    • Search Google Scholar
    • Export Citation
  • Markowski, P., and Y. Richardson, 2010: Mesoscale Meteorology in Midlatitudes. Wiley-Blackwell, 407 pp.

  • Metz, N. D., D. M. Schultz, and R. H. Johns, 2004: Extratropical cyclones with multiple warm-front-like baroclinic zones and their relationship to severe convective storms. Wea. Forecasting, 19, 907916.

    • Search Google Scholar
    • Export Citation
  • Morcrette, J.-J., 1991: Radiation and cloud radiative properties in the ECMWF operational weather forecast model. J. Geophys. Res., 96D, 91219132.

    • Search Google Scholar
    • Export Citation
  • Muir, L. C., and M. J. Reeder, 2010: Idealized modelling of landfalling cold fronts. Quart. J. Roy. Meteor. Soc., 136, 21472161.

  • Neiman, P. J., M. A. Shapiro, and L. S. Fedor, 1993: The life cycle of an extratropical marine cyclone. Part II: Mesoscale structure and diagnostics. Mon. Wea. Rev., 121, 21772199.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., F. M. Ralph, P. O. G. Persson, A. B. White, D. P. Jorgensen, and D. E. Kingsmill, 2004: Modification of fronts and precipitation by coastal blocking during an intense landfalling winter storm in southern California: Observations during CALJET. Mon. Wea. Rev., 132, 242273.

    • Search Google Scholar
    • Export Citation
  • Nuss, W., 1989: Air–sea interaction influences on the structure and intensification of an idealized marine cyclone. Mon. Wea. Rev., 117, 351369.

    • Search Google Scholar
    • Export Citation
  • Peng, M. S., J. H. Powell, R. T. Williams, and B. F. Jeng, 2001: Boundary layer effects on fronts over topography. J. Atmos. Sci., 58, 22222239.

    • Search Google Scholar
    • Export Citation
  • Petterssen, S., 1936: Contribution to the theory of frontogenesis. Geofys. Publ., 11, 727.

  • Petterssen, S., 1956: Weather and Weather Systems. Vol. II, Weather Analysis and Forecasting, McGraw-Hill, 266 pp.

  • Physick, W., 1988: Mesoscale modeling of a cold front and its interaction with a diurnally heated land mass. J. Atmos. Sci., 45, 31693187.

    • Search Google Scholar
    • Export Citation
  • Pinty, J.-P., and P. Jabouille, 1998: A mixed-phase cloud parameterization for use in mesoscale non-hydrostatic model: Simulations of a squall line and of orographic precipitations. Preprints, Conf. of Cloud Physics, Everett, WA, Amer. Meteor. Soc., 217–220.

    • Search Google Scholar
    • Export Citation
  • Reeder, M. J., 1986: The interaction of a surface cold front with a prefrontal thermodynamically well-mixed boundary layer. Aust. Meteor. Mag., 34, 137148.

    • Search Google Scholar
    • Export Citation
  • Renard, R. J., and L. C. Clarke, 1965: Experiments in numerical objective frontal analysis. Mon. Wea. Rev., 93, 547556.

  • Roach, W., and M. Hardman, 1975: Mesoscale air motions derived from wind finding dropsonde data: The warm front and rainbands of 18 January 1971. Quart. J. Roy. Meteor. Soc., 101, 437462.

    • Search Google Scholar
    • Export Citation
  • Saltikoff, E., H. Huuskonen, J. Koistinen, and H. Järvinen, 2010: Quality assurance in the FMI Doppler radar network. Boreal Environ. Res., 15, 579594.

    • Search Google Scholar
    • Export Citation
  • Sanders, F., 1955: An investigation of the structure and dynamics of an intense surface frontal zone. J. Meteor., 12, 542552.

  • Sanders, F., 1999: A proposed method of surface map analysis. Mon. Wea. Rev., 127, 945955.

  • Sanders, F., and C. A. Doswell III, 1995: A case for detailed surface analysis. Bull. Amer. Meteor. Soc., 76, 505521.

  • Schultz, D. M., 2005: A review of cold fronts with prefrontal troughs and wind shifts. Mon. Wea. Rev., 133, 24492472.

  • Schultz, D. M., and P. J. Roebber, 2008: The fiftieth anniversary of Sanders (1955): A Mesoscale model simulation of the cold front of 17–18 April 1953. Synoptic–Dynamic Meteorology and Weather Analysis and Forecasting: A Tribute to Fred Sanders, Meteor. Monogr., No. 55, Amer. Meteor. Soc., 126–143.

    • Search Google Scholar
    • Export Citation
  • Schultz, D. M., D. Keyser, and L. F. Bosart, 1998: The effect of large-scale flow on low-level frontal structure and evolution in midlatitude cyclones. Mon. Wea. Rev., 126, 17671791.

    • Search Google Scholar
    • Export Citation
  • Seity, Y., P. Brousseau, S. Malardel, G. Hello, P. Bénard, F. Bouttier, C. Lac, and V. Masson, 2010: The AROME-France convective-scale operational model. Mon. Wea. Rev., 139, 976991.

    • Search Google Scholar
    • Export Citation
  • Shapiro, M., 1984: Meteorological tower measurements of a surface cold front. Mon. Wea. Rev., 112, 16341639.

  • Siebesma, A. P., P. M. M. Soares, and J. Teixeira, 2007: A combined eddy-diffusivity mass-flux approach for the convective boundary layer. J. Atmos. Sci., 64, 12301248.

    • Search Google Scholar
    • Export Citation
  • Sinclair, V. A., S. E. Belcher, and S. L. Gray, 2010: Synoptic controls on boundary-layer characteristics. Bound.-Layer Meteor., 134, 387409.

    • Search Google Scholar
    • Export Citation
  • Smith, R. K., and M. J. Reeder, 1988: On the movement and low-level structure of cold fronts. Mon. Wea. Rev., 116, 19271944.

  • Soares, P. M. M., P. M. A. Miranda, A. P. Siebesma, and J. Teixeira, 2004: An eddy-diffusivity/mass-flux parametrization for dry and shallow cumulus convection. Quart. J. Roy. Meteor. Soc., 130, 33653384.

    • Search Google Scholar
    • Export Citation
  • Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. 1st ed. Kluwer Academic, 685 pp.

  • Vaisala, cited 2011: WXT520 datasheet. [Available online at http://www.vaisala.com/en/products/multiweathersensors/Pages/WXT520.aspx.]

  • Wakimoto, R., and B. Bosart, 2001: Airborne radar observations of a warm front during FASTEX. Mon. Wea. Rev., 129, 254274.

  • Wallace, J. M., and P. V. Hobbs, 2006: Atmospheric Science: An Introductory Survey. 2nd ed. Academic Press, 483 pp.

  • Wernli, H., and C. Schwierz, 2006: Surface cyclones in the ERA-40 dataset (1958–2001). Part I: Novel identification method and global climatology. J. Atmos. Sci., 63, 24862507.

    • Search Google Scholar
    • Export Citation
  • Xu, Q., and W. Gu, 2002: Semigeostrophic frontal boundary layer. Bound.-Layer Meteor., 104, 99110.

Save
  • Ahrens, C. D., 1994: Meteorology Today: An Introduction to Weather, Climate, and the Environment. 5th ed. West Publishing, 528 pp.

  • Austin, P. M., and R. A. Houze Jr., 1972: Analysis of the structure of precipitation patterns in New England. J. Appl. Meteor., 11, 926935.

    • Search Google Scholar
    • Export Citation
  • Bénard, P., J. Vivoda, J. Mašek, P. Smolíková, K. Yessad, C. Smith, R. Brožková, and J.-F. Geleyn, 2010: Dynamical kernel of the Aladin-NH spectral limited-area model: Revised formulation and sensitivity experiments. Quart. J. Roy. Meteor. Soc., 136, 155169.

    • Search Google Scholar
    • Export Citation
  • Bjerknes, J., 1919: On the structure of moving cyclones. Geofys. Publ., 1 (2), 18.

  • Bjerknes, J., and H. Solberg, 1921: Meteorological conditions for the formation of rain. Geofys. Publ., 2 (3), 361.

  • Bjerknes, J., and H. Solberg, 1922: Life cycle of cyclones and the polar front theory of atmospheric circulation. Geofys. Publ., 3 (1), 138.

    • Search Google Scholar
    • Export Citation
  • Black, R. X., and R. M. Dole, 2000: Storm tracks and barotropic deformation in climate models. J. Climate, 13, 27122728.

  • Bluestein, H., 1993: Observations and Theory of Weather Systems. Vol. II, Synoptic–Dynamic Meteorology in Midlatitudes, Oxford University Press, 594 pp.

    • Search Google Scholar
    • Export Citation
  • Bond, N. A., and R. G. Fleagle, 1988: Prefrontal and postfrontal boundary layer processes over the ocean. Mon. Wea. Rev., 116, 12571273.

    • Search Google Scholar
    • Export Citation
  • Browning, K., and T. Harrold, 1970: Air motion and precipitation growth at a cold front. Quart. J. Roy. Meteor. Soc., 96, 369389.

  • Browning, K., M. Hardman, T. Harrold, and C. Pardoe, 1973: The structure of rainbands within a mid-latitude depression. Quart. J. Roy. Meteor. Soc., 99, 215231.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., C. F. Mass, and B. F. Smull, 1999: An observational and numerical study of a cold front interacting with the Olympic Mountains during COAST IOP5. Mon. Wea. Rev., 127, 13101334.

    • Search Google Scholar
    • Export Citation
  • Cuxart, J., P. Bougeault, and J. L. Redelsperger, 2000: A turbulence scheme allowing for mesoscale and large-eddy simulations. Quart. J. Roy. Meteor. Soc., 126, 130.

    • Search Google Scholar
    • Export Citation
  • Dacre, H. F., and S. L. Gray, 2009: The spatial distribution and evolution characteristics of North Atlantic cyclones. Mon. Wea. Rev., 137, 99115.

    • Search Google Scholar
    • Export Citation
  • Doyle, J. D., and N. A. Bond, 2001: Research aircraft observations and numerical simulations of a warm front approaching Vancouver Island. Mon. Wea. Rev., 129, 978998.

    • Search Google Scholar
    • Export Citation
  • Drebs, A., A. Nordlund, P. Karlsson, J. Helminen, and P. Rissanen, 2002: Tilastoja suomen ilmastosta 1971–2000. Climatological Statistics of Finland 1971–2000, Finnish Meteorological Institute, 99 pp.

    • Search Google Scholar
    • Export Citation
  • Gallus, W. A., Jr., and M. Segal, 1999: Cold front acceleration over Lake Michigan. Wea. Forecasting, 14, 771781.

  • Garratt, J. R., 1988: Summertime cold fronts in southeast Australia—Behavior and low-level structure of main frontal types. Mon. Wea. Rev., 116, 636649.

    • Search Google Scholar
    • Export Citation
  • Hewson, T., 1998: Objective fronts. Meteor. Appl., 5, 3765.

  • Heymsfield, G., 1979: Doppler radar study of a warm frontal region. J. Atmos. Sci., 36, 20932107.

  • Hines, K. M., and C. R. Mechoso, 1993: Influence of surface drag on the evolution of fronts. Mon. Wea. Rev., 121, 11521175.

  • Hoskins, B. J., and F. P. Bretherton, 1972: Atmospheric frontogenesis models: Mathematical formulation and solution. J. Atmos. Sci., 29, 1137.

    • Search Google Scholar
    • Export Citation
  • Houze, R. A., Jr., S. A. Rutledge, T. J. Matejka, and P. V. Hobbs, 1981: The mesoscale and microscale structure and organization of clouds and precipitation in midlatitude cyclones. III: Air motions and precipitation growth in a warm-frontal rainband. J. Atmos. Sci., 38, 639649.

    • Search Google Scholar
    • Export Citation
  • Jenkner, J., M. Sprenger, I. Schwenk, C. Schwierz, S. Dierer, and D. Leuenberger, 2010: Detection and climatology of fronts in a high-resolution model reanalysis over the Alps. Meteor. Appl., 17, 118.

    • Search Google Scholar
    • Export Citation
  • Karppinen, A., S. Joffre, and J. Kukkonen, 2002: Evaluation of meteorological data measured at a radio tower in the Helsinki Metropolitan Area. Proc. COST 715 Expert Meeting, Toulouse, France, European Commission EUR 20451, 89–98.

    • Search Google Scholar
    • Export Citation
  • Keyser, D., and R. Anthes, 1982: The influence of planetary boundary layer physics on frontal structure in the Hoskins–Bretherton horizontal shear model. J. Atmos. Sci., 39, 17831802.

    • Search Google Scholar
    • Export Citation
  • Koskinen, J. T., and Coauthors, 2011: The Helsinki Testbed: A mesoscale measurement, research, and service platform. Bull. Amer. Meteor. Soc., 92, 325342.

    • Search Google Scholar
    • Export Citation
  • Lascaux, F., E. Richard, and J. Pinty, 2006: Numerical simulations of three different MAP IOPs and the associated microphysical processes. Quart. J. Roy. Meteor. Soc., 132, 19071926.

    • Search Google Scholar
    • Export Citation
  • Lehmann, A., and G. Tschersich, 2006: Trends in sea surface temperature of the Baltic Sea since 1990. BALTEX Newsletter, No. 9, International BALTEX Secretariat, Geesthacht, Germany, 7–9.

    • Search Google Scholar
    • Export Citation
  • Le Moigne, P., 2009: SURFEX scientific documentation. CNRM Tech. Rep., Météo-France/CNRS, Toulouse, France, 211 pp.

  • Locatelli, J., and P. Hobbs, 1987: The mesoscale and microscale structure and organization of clouds and precipitation in midlatitude cyclones. XIII: Structure of a warm front. J. Atmos. Sci., 44, 22902309.

    • Search Google Scholar
    • Export Citation
  • Markowski, P., and Y. Richardson, 2010: Mesoscale Meteorology in Midlatitudes. Wiley-Blackwell, 407 pp.

  • Metz, N. D., D. M. Schultz, and R. H. Johns, 2004: Extratropical cyclones with multiple warm-front-like baroclinic zones and their relationship to severe convective storms. Wea. Forecasting, 19, 907916.

    • Search Google Scholar
    • Export Citation
  • Morcrette, J.-J., 1991: Radiation and cloud radiative properties in the ECMWF operational weather forecast model. J. Geophys. Res., 96D, 91219132.

    • Search Google Scholar
    • Export Citation
  • Muir, L. C., and M. J. Reeder, 2010: Idealized modelling of landfalling cold fronts. Quart. J. Roy. Meteor. Soc., 136, 21472161.

  • Neiman, P. J., M. A. Shapiro, and L. S. Fedor, 1993: The life cycle of an extratropical marine cyclone. Part II: Mesoscale structure and diagnostics. Mon. Wea. Rev., 121, 21772199.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., F. M. Ralph, P. O. G. Persson, A. B. White, D. P. Jorgensen, and D. E. Kingsmill, 2004: Modification of fronts and precipitation by coastal blocking during an intense landfalling winter storm in southern California: Observations during CALJET. Mon. Wea. Rev., 132, 242273.

    • Search Google Scholar
    • Export Citation
  • Nuss, W., 1989: Air–sea interaction influences on the structure and intensification of an idealized marine cyclone. Mon. Wea. Rev., 117, 351369.

    • Search Google Scholar
    • Export Citation
  • Peng, M. S., J. H. Powell, R. T. Williams, and B. F. Jeng, 2001: Boundary layer effects on fronts over topography. J. Atmos. Sci., 58, 22222239.

    • Search Google Scholar
    • Export Citation
  • Petterssen, S., 1936: Contribution to the theory of frontogenesis. Geofys. Publ., 11, 727.

  • Petterssen, S., 1956: Weather and Weather Systems. Vol. II, Weather Analysis and Forecasting, McGraw-Hill, 266 pp.

  • Physick, W., 1988: Mesoscale modeling of a cold front and its interaction with a diurnally heated land mass. J. Atmos. Sci., 45, 31693187.

    • Search Google Scholar
    • Export Citation
  • Pinty, J.-P., and P. Jabouille, 1998: A mixed-phase cloud parameterization for use in mesoscale non-hydrostatic model: Simulations of a squall line and of orographic precipitations. Preprints, Conf. of Cloud Physics, Everett, WA, Amer. Meteor. Soc., 217–220.

    • Search Google Scholar
    • Export Citation
  • Reeder, M. J., 1986: The interaction of a surface cold front with a prefrontal thermodynamically well-mixed boundary layer. Aust. Meteor. Mag., 34, 137148.

    • Search Google Scholar
    • Export Citation
  • Renard, R. J., and L. C. Clarke, 1965: Experiments in numerical objective frontal analysis. Mon. Wea. Rev., 93, 547556.

  • Roach, W., and M. Hardman, 1975: Mesoscale air motions derived from wind finding dropsonde data: The warm front and rainbands of 18 January 1971. Quart. J. Roy. Meteor. Soc., 101, 437462.

    • Search Google Scholar
    • Export Citation
  • Saltikoff, E., H. Huuskonen, J. Koistinen, and H. Järvinen, 2010: Quality assurance in the FMI Doppler radar network. Boreal Environ. Res., 15, 579594.

    • Search Google Scholar
    • Export Citation