1. Introduction
Extratropical cyclones represent the dominant mode of extreme weather to impact the midlatitudes during the winter months. Damaging winds and heavy freezing or frozen precipitation are among the threats these systems pose to local economies and infrastructure. Owing to a high population density, the megalopolis in the northeastern United States, from Washington, D.C., to Boston, is particularly susceptible to threats posed by extratropical cyclones. A one-day shutdown of business and infrastructure in this region typically costs state economies on the order of hundreds of millions of dollars, with the state of New York topping out at just over $700 million (American Highway Users Alliance and IHS Global Insight 2010). Given their vast socioeconomic impacts, it is important to assess how extratropical cyclones may be impacted by anthropogenic climate change.
Previous studies concerning extratropical cyclones and climate change have investigated a wide range of topics including shifts in the midlatitude storm tracks and changes in the frequency and intensity of extratropical cyclones. Many of these studies, using general circulation models (GCMs), project the total number of extratropical cyclone events both globally and in the Northern Hemisphere to decrease with warming (e.g., Geng and Sugi 2003; Lambert and Fyfe 2006; Bengtsson et al. 2009; Champion et al. 2011; Catto et al. 2011; Mizuta et al. 2011; Zappa et al. 2013). Changes in the frequency of intense extratropical cyclones within the Northern Hemisphere are still unclear, with some studies projecting decreases (Catto et al. 2011; Zappa et al. 2013) while others project increases (Geng and Sugi 2003; Lambert and Fyfe 2006; Mizuta et al. 2011). Precipitation associated with extratropical cyclone events is also projected to increase with warming (Watterson 2006; Bengtsson et al. 2009; Champion et al. 2011; Zappa et al. 2013). This is consistent with the prevailing idea that increases in temperature should result in atmospheric water vapor pressure increases that scale with the Clausius–Clapeyron relation (Frei et al. 1998; Allen and Ingram 2002; Soden et al. 2005; Held and Soden 2006; Wentz et al. 2007). Recent work by Colle et al. (2013), using data from phase 5 of the Coupled Model Intercomparison Project (CMIP5) GCM simulations, found that the number of intense cyclone events over the eastern United States increases with warming where minimum sea level pressure is used as the metric for intensity. Little change is found in the local Eady growth rate, upper-tropospheric jet strength, or low-level temperature gradient with warming. However, large precipitation increases are seen in the future projections, leading Colle et al. (2013) to hypothesize that the intensity increases may be largely attributable to enhanced latent heat release.
The importance of latent heat release to extratropical cyclogenesis is well documented in the synoptic-dynamic literature, and has been quantified using potential vorticity (PV) inversion (e.g., Davis and Emanuel 1991; Davis 1992; Davis et al. 1993; Balasubramanian and Yau 1994; Stoelinga 1996; Huo et al. 1999). Latent heat release is crucial to type-C cyclogenesis in which the intensification of the surface cyclone is driven by latent heat release and the formation of an intense low-level diabatic PV anomaly (Plant et al. 2003; Ahmadi-Givi et al. 2004). Latent heating has also been shown to be vital to the development of storm-scale features such as the low-level jet (Lackmann and Gyakum 1999; Lackmann 2002; Brennan et al. 2008). Although increases in precipitation associated with extratropical cyclones are projected with warming in many studies (e.g., Karl and Knight 1998; Trenberth 1999; Trenberth et al. 2003; Watterson 2006; Bengtsson et al. 2009; Champion et al. 2011; Colle et al. 2013), fewer studies have attempted to quantify the dynamical role of the enhanced latent heat release on cyclone structure and intensity. Simulations investigating the sensitivity of an idealized baroclinic wave to varying levels of moisture show that increases on the order of those projected by GCMs only result in small intensity increases (Booth et al. 2013). It is unclear whether this holds true when considering projected changes in both moisture and temperature.
Some of the environmental changes projected to occur with warming in the Northern Hemisphere include increased low-latitude tropospheric static stability, increased upper-tropospheric baroclinicity, and decreased lower-tropospheric baroclinicity (Yin 2005). Each of these changes may impact future extratropical cyclone track/intensity and are well resolved by GCMs. However, GCMS typically lack the horizontal resolution necessary to adequately resolve moist diabatic processes within extratropical cyclones (Willison et al. 2013). Consequently, the structure and amount of condensational heating within extratropical cyclones, and thus their intensity, may be underrepresented by GCMs. Similarly, storm-scale features sensitive to condensational heating, such as the low-level jet, may not be adequately represented. To investigate whether the feedback between latent heating and the low-level jet may amplify in the future, Lackmann (2013) employed a form of regional downscaling to simulate an extreme flooding event and its future replication; this technique is known as “pseudo-global-warming” (PGW; e.g., Schär et al. 1996; Frei et al. 1998; Sato et al. 2007; Hara et al. 2008; Rasmussen et al. 2011). The same approach has also been used to investigate the impacts of climate change on Atlantic hurricanes (Mallard et al. 2013).
Using this PGW approach, the focus of the current paper is to address the hypothesis posed by Colle et al. (2013) that changes in latent heating are responsible for increases in extratropical cyclone intensity along the U.S. East Coast. Likewise, this study seeks to identify changes in storm-scale cyclone dynamics using the PV framework. Simulations of present-day cyclone events and observed events with projected thermodynamic conditions are performed using a high-resolution numerical weather prediction model. Changes in the track, structure, and intensity of these events due to warming are then analyzed using cyclone compositing. This will not only identify how these events may change in the future, but also allows an assessment of why those changes are occurring from a dynamical standpoint.
2. Data and methods
a. Selection of cases
The current study focuses on “Miller-A” extratropical cyclone events, which track continuously along the U.S. East Coast (Miller 1946; Kocin and Uccellini 2004, section 2a) during the climatological winter months (December–February). The North American Regional Reanalysis (NARR; Mesinger et al. 2006) is used to manually identify and categorize events that occurred between 1979 and 2011 for use in this study. An extratropical cyclone is categorized as a Miller-A event if it meets the following criteria during its evolution. First, the cyclone must originally develop over or within 500 km of the Gulf of Mexico. For the purposes of this study, a cyclone is said to have developed once a closed surface low is observed using a sea level pressure contour interval of 2 hPa. Second, the cyclone must track no farther than 400 km east of the coastline until it reaches the latitude of Cape May, New Jersey. This criterion is invoked in order to isolate events that track in close proximity to the coastline. Additionally, the cyclone must attain a minimum sea level pressure below 995 hPa within 36 h of its centering time (T0); T0 is defined as the time when the cyclone tracks closest to Cape Hatteras, North Carolina. A threshold of minimum central pressure below 995 hPa ensures the sampling of events that are of at least moderate intensity. These track and intensity criteria for Miller-A cyclones are chosen to focus on what the authors deem to be strongly developing systems with high socioeconomic impacts.
Finally, if an event cannot unmistakably be defined as a Miller-A (i.e., if it displays characteristics of a Miller-B coastal redevelopment event; see Miller 1946) then it is discarded to avoid ambiguity in the dataset. The same selection criteria applied to the NARR are also applied to our present-day simulations of individual cyclone events to be described in section 2b. This is done to maximize homogeneity within the dataset. Ultimately we are left with a dataset consisting of the 10 events that are listed in Table 1. Since strongly developing Miller-A events are the focus of this study, it should be kept in mind that any results presented here are not necessarily applicable to all cyclone events. Marciano (2014) has performed additional analysis for Miller-B events, however.
Starting and ending times for simulations of each of the 10 events.
b. Numerical model, configuration, and present-day simulations
The Advanced Research Weather Research and Forecasting (WRF-ARW) Model version 3.2.1 (Skamarock et al. 2008) is used to perform all simulations in this study. Each simulation is run for 84 h and configured with 36-km horizontal grid spacing and 41 vertical levels. Simulations were also performed with 12-km and 4-km horizontal grid spacing in order to assess the sensitivity of results to horizontal resolution. Decreasing the horizontal grid spacing results in larger but qualitatively similar changes with warming (Marciano 2014, section 4.4). Therefore, in an effort to limit computational expense, 36-km was deemed sufficient to represent the diabatic processes in which we are interested. Each simulation is initialized 36 h prior to T0 and output is generated every three hours. The Betts–Miller–Janjic (BMJ) cumulus parameterization is used to represent subgrid-scale precipitation while the planetary boundary layer is represented using the Mellor–Yamada–Janjic (MYJ) scheme. Microphysics are handled using the Weather Research and Forecasting (WRF) single-moment 5-class (WSM5) scheme. Longwave and shortwave radiation are represented using the Rapid Radiative Transfer Model (RRTM) and Dudhia schemes, respectively. The sensitivity of results to the above physics parameterization choices was tested by performing an ensemble of simulations using different configurations. The main results did not exhibit strong sensitivity to model physics (Marciano 2014, section 4.1). This lack of physics sensitivity is likely a result of our choice of relatively coarse horizontal grid spacing. The NARR, which has 32-km horizontal grid spacing and 45 vertical levels, is used as the initial and lateral boundary conditions for each present-day simulation. Fields are available in the NARR only up to the 100-hPa level; therefore, all simulations performed in this study have a model top of 100 hPa. This upper boundary is not constrained to follow the NARR. Both the NARR domain and the domain used for each simulation are presented in Fig. 1.
Domains for the North American Regional Reanalysis (NARR) and Advanced Research Weather Research and Forecasting (ARW) model simulations are shown in purple and pink, respectively.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
c. Future simulations
To perform “future” simulations of present-day events, thermodynamic changes are applied to the initial and lateral boundary conditions in a manner consistent with the aforementioned PGW approach (Sato et al. 2007; Hara et al. 2008). The thermodynamic changes applied here are derived from a five-member ensemble of GCMs using the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) A2 emissions scenario. The five GCMs that make up the ensemble used here are the Bjerknes Centre for Climate Research Bergen Climate Model version 2 (BCCR BCM2), Centre National de Recherches Météorologiques Coupled Global Climate Model version 3 (CNRM-CM3), Institute of Numerical Mathematics Coupled Model version 3 (INM-CM3.0), Max Planck Institute (MPI) ECHAM5, and Hadley Centre Coupled Model, version 3 (HadCM3). These GCMs were chosen because they had more complete pressure-level data relative to other Coupled Model Intercomparison Project phase 3 (CMIP3) GCMs.
Monthly averages of temperature at the surface (soil and sea surface), 2 m, and at each isobaric level are computed from the GCM ensemble. These monthly averages are used to compute decadal averages for each month where the periods 1990–99 and 2090–99 represent the present and future climates, respectively. The decadal average temperature change is then applied to the NARR initial and lateral boundary conditions for each simulation. An event that occurs in the month of December would have the December monthly decadal average change at each grid point applied to it, with analogous logic used for other months. Consistent with previous studies, relative humidity is held constant with the applied temperature increases such that atmospheric water vapor pressure varies with the Clausius–Clapeyron relation (Frei et al. 1998; Allen and Ingram 2002; Soden et al. 2005; Held and Soden 2006; Wentz et al. 2007).
The composite temperature and moisture changes at hour 00 are presented in Fig. 2. In the lower troposphere (Fig. 2a) the strongest warming occurs at high latitudes while in the upper troposphere (Fig. 2b) the strongest warming is found in the tropics. The change in specific humidity at the 850-hPa level (Fig. 2c) shows that the largest lower-tropospheric moisture increases are found at low latitudes. In addition to the upper- and lower-tropospheric temperature changes mentioned above, a north–south cross section (Fig. 2d) indicates stratospheric cooling at high latitudes. It is also evident that the applied changes make the lower and middle tropospheric lapse rate more stable at low-to-mid latitudes and less stable at high latitudes. As these are the composite changes, the temperature changes applied to individual cases vary slightly month by month but are qualitatively similar to those described above (not shown). After thermodynamic changes have been applied, the updated initial and lateral boundary conditions drive future simulations of extratropical cyclone events using the same model configurations described in the previous section. Prior to model initialization, the WRF preprocessing software (WPS) recalculates the geopotential height field from the virtual temperature such that it is in hydrostatic balance. This introduces some imbalance between the mass and wind fields; however, the applied thermodynamic changes are sufficiently spatially smooth so as to avoid introducing strong gravity wave adjustment in the model. The applied thermodynamic changes in this study are projected over a period of 100 years; therefore, the future simulations can be thought of as how the same synoptic-scale extratropical cyclone events might evolve if they were to take place 100 years later. Rather than using the GCM ensemble, simulations were also performed using the thermodynamic changes from each of the individual GCMs, and this was not found to significantly alter the overall results (Marciano 2014, section 4.2).
(a) Hour 00 composite temperature change (°K) at 925 hPa. (b)–(d) As in (a), but for composite temperature change (°K) at 250 hPa, composite specific humidity change (10−4 kgwv kgair−1) at 850 hPa, and north–south cross section of composite temperature change, respectively. Cross section is taken along the 75°W latitudinal band from 18° to 57°N. These endpoints represent the southern and northern limits of the domain at this longitude.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
d. Cyclone compositing
Cyclone compositing is used to assess the intensity, track, and an assortment of storm-scale changes due to warming. Unless explicitly stated, all analyses are presented using storm-relative coordinates. Given their similar tracks, the need to rotate storms relative to their travel path is not deemed necessary. Storm-relative analyses allow a focus on the present-to-future changes common among the events and eliminate artificial smoothing due to variations in storm track and propagation speed. Storm-relative composites have previously been used to analyze impacts of climate change on downscaled GCM simulations of autumn extratropical cyclone events in the northwest Atlantic (Jiang and Perrie 2007; Perrie et al. 2010). The storm-relative composites in the current study utilize a grid measuring 1116 km × 1116 km centered on the sea level pressure minimum. This 1116 km × 1116 km grid represents ±15 grid points from the cyclone center. Prior to recording the minimum sea level pressure and its location at each output time, a 20-point Gaussian weighted smoothing function is applied to the sea level pressure field in order to remove any mesoscale lows (due to convection along fronts, etc.) that may generate discontinuities in the track. While the filter also reduces the magnitudes of the pressure minima used to compute the mean, the overall results are not sensitive to this change. Composite analysis begins at simulation hour 18, because this is the first time that a continuous track exists for all present and future cyclones. Because the WRF domain is limited on its northern and eastern peripheries by the lateral boundaries of the NARR domain (Fig. 1) many of the simulated events approach the lateral boundaries before the conclusion of the simulations. To limit possible contamination from the lateral boundaries, composite analysis concludes at hour 63. This issue also dictated the size of the storm-relative grid where 1116 km × 1116 km is the maximum grid size possible without lateral boundary contamination.
e. Potential vorticity
3. Assessment of present-day simulations
Before analyzing any present to future changes, it is important to determine how well the present-day cyclone simulations match the reanalysis data. The root-mean-square (RMS) error is computed for minimum sea level pressure and track for each WRF simulation versus the NARR (Table 2). Track errors are computed as the average distance between the NARR and WRF tracks for all points along the cyclone tracks. Eight of the 10 simulations exhibit RMS intensity errors less than 5 hPa and seven of the 10 simulations exhibit RMS track errors less than 300 km. Although some of the individual simulations have larger errors than would be preferred, the average simulation does a satisfactory job overall, and with a focus on the changes from present to future, these errors are deemed acceptable. The composite storm tracks (Fig. 3a) reveal systematic biases that are primarily in the along-track direction (i.e., differences in propagation speed along the track). The composite tracks superimposed on the individual tracks are also presented for the NARR (Fig. 3b) and WRF simulations (Fig. 3c). Composite analyses of the minimum sea level pressure evolutions (Fig. 3d) reveal that the average minimum sea level pressure is too high in the WRF simulations after hour 30. The average case has an RMS intensity error of ~3.3 hPa and RMS track error of ~238 km. We note that the lack of observations over the oceans available for the reanalysis may contribute to errors in the NARR central pressures and cyclone tracks.
Root-mean-square (RMS) error for each simulation in terms of minimum sea level pressure and cyclone track expressed in units of hectopascals (hPa) and kilometers (km), respectively.
(a) Composite storm track with time points using the NARR (solid black line, red crosses) and present-day WRF model simulations (solid green line, pink crosses) for hours 18–63. (b) Composite (mean) storm track (thick solid black line) and individual storm tracks (thin lines, see shared legend for colors) using the NARR for hours 18 to 63. (c) As in (b), but for the present-day WRF model simulations. (d) Composite minimum sea level pressure evolution using the NARR (solid black line) and present-day WRF simulations (solid green line) for hours 18–63.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
4. Analysis of present-to-future changes
a. Cyclone track
The mean cyclone track is determined by averaging the latitudes and longitudes of the 10 individual cyclone tracks. The mean present-day cyclone moves northeast out of the coastal Gulf of Mexico region and up along the U.S. East Coast as expected (Fig. 4). A similar track is seen for the mean future cyclone, but the track is shifted eastward, attaining a displacement of ~190 km by simulation hour 63. An eastward track shift is evident in each of the 10 future simulations (not shown). The mean future track lies well within the one standard deviation (1-sigma) cone of its present counterpart, meaning that the variability in cyclone track among the individual simulations is greater than the changes that are projected to occur due to warming.
Tracks of the mean present-day (solid blue) and future (solid red) surface cyclones for simulation hours 18–63. Time points are labeled every 9 h and indicated with black crosses. The one standard deviation (e.g., 1-sigma) cones for the average present-day and average future track are shown by blue dashed and red dashed lines, respectively.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
The speed at which the upper-tropospheric disturbance propagates can dictate the development and general motion of a surface cyclone. Composite pressure and wind on the dynamic tropopause, here defined as the 1.5-PVU (PV unit) surface, are computed in an Earth-relative sense in order to assess whether the future upper disturbance propagates faster. At simulation hour 48 (Fig. 5) the highest pressures on the dynamic tropopause are located over the southeastern U.S. coastline in both the present-day and future composites. The leading edge of the dynamic tropopause extending to or below the 300-hPa level is situated farther east in the future (Fig. 5b) relative to the present (Fig. 5a). This is seen more clearly in the present-to-future pressure change on the dynamic tropopause (Fig. 6). The couplet of pressure increases to the east and decreases to the west indicates a clear eastward shift of the upper trough. The pressure increases to the east are also greater in magnitude than the decreases to the west, meaning that the 1.5-PVU surface extends closer to the surface in the future, such that the UPV anomaly is stronger. The eastward shift of the upper-level trough is consistent with the observed shift of the surface cyclone track.
Pressure (hPa; contour) and wind (kt; fill) on the dynamic tropopause (1.5-PVU surface) at simulation hour 48 for (a) present-day and (b) future Earth-relative composites.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
Present-to-future change in Earth-relative composite pressure (hPa) on the dynamic tropopause (1.5-PVU surface) at simulation hour 48.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
Differences in Earth-relative composite zonal wind are computed at simulation hour 06 to assess the change in the upper tropospheric westerly flow due to the imposed thermodynamic changes. In the future simulations, the wind field has had sufficient time to adjust to the applied thermodynamic changes by hour 06 and is thus deemed representative of the imposed zonal wind change. Analysis of this zonal wind speed change at the 350-hPa level (Fig. 7) reveals modest increases in the upper-tropospheric westerly flow with an area average increase of ~3% in the region shown. Since the imposed upper-tropospheric changes in temperature (Fig. 2b) and moisture (not shown) strengthen the meridional virtual temperature gradient aloft, the resulting increase in westerly wind shear is to be expected, consistent with thermal wind balance. This increase in zonal wind speed is consistent with an increase in the eastward propagation speed of the upper-tropospheric trough and thus explains the eastward surface cyclone track shift seen in future simulations.
Present-to-future change in Earth-relative composite 350-hPa zonal wind speed (m s−1) at simulation hour 06.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
b. Precipitation
Consistent with previous studies, storm-relative composite analysis of total precipitation from simulation hours 18 to 63 reveals that the present-day cyclones (Fig. 8a) produce less precipitation than their future replications (Fig. 8b). While the present-day composite exhibits a very limited area of total precipitation exceeding 100 mm, the future composite clearly shows an extensive region of total precipitation exceeding this threshold. The present-to-future storm-relative change in total precipitation (Fig. 8c) reveals large increases in most areas. Several areas of precipitation increase are found to be statistically significant at the 95% confidence level using a Wilcoxon signed-rank test. The area-average increase in total precipitation is found to be ~33%. Given an approximate 3.5-K increase in area-average 850-hPa temperature, the expected increase in atmospheric water vapor content projected by the Clausius–Clapeyron relation is ~25%. The future precipitation increases at a rate which exceeds the Clausius–Clapeyron moisture increase for these simulations of significant East Coast cyclones. However, estimating the Clausius–Clapeyron-driven change in precipitation from the mean temperature change does not account for changes in temperature within the cyclones due to shifts in cyclone track.
Total precipitation (mm) for simulation hours 18–63 in the (a) present-day and (b) future storm-relative composites. (c) Present-to-future change in total precipitation (mm). Regions of statistically significant change at the 95% confidence level are contoured in black. The position of the composite storm center is represented by an asterisk. Distances from the storm center (km) in each direction are shown on the ordinate and abscissa.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
To isolate the change due to the geographic shift in track with warming, the present sea surface temperatures (SSTs) are output along the individual tracks for times when both the present and future cyclones are over water. The mean SST along all present and future tracks are then computed and used to determine the temperature change due solely to the track shift. This shift in track adds an additional ~1.5 K of warming resulting in a net area-average temperature increase of ~5 K. Taking the warming due to the track shift into consideration, the expected water vapor increase is ~35%, meaning that the area-averaged precipitation increase is slightly less than the moisture increase corresponding to the Clausius–Clapeyron relation. Nevertheless, there are localized super–Clausius–Clapeyron increases in precipitation that are consistent with previous work investigating moisture transport in midlatitude cyclones. Performing a series of idealized baroclinic wave experiments, Boutle et al. (2011) found slightly super–Clausius–Clapeyron precipitation increases when surface temperature was increased with relative humidity held constant. However, unlike the current study, they did not change the middle- and upper-tropospheric stability in their simulations. The large increase in precipitation over the same duration of time also indicates increased future rain rates. This idea is supported by statistically significant increases in simulated radar reflectivity (not shown). Area-averaged 3-hourly storm-relative precipitation increases are found to range from ~16% at hour 18 to over 50% by hour 63. This indicates an overall upward trend in precipitation rate that increases over the lifetime of the future composite event.
Storm-relative liquid equivalent total snowfall is primarily confined to the northwest quadrant of the composite cyclone for present-day simulations (Fig. 9a), producing more snowfall than the future composite (Fig. 9b). The present-day composite exhibits a large swath of liquid equivalent snowfall values exceeding 40 mm, while values exceeding this threshold are generally limited in the future composite. This same swath shows up as an area of substantial present-to-future decreases in total snowfall (Fig. 9c) with an area-average decrease of ~33%. A large portion of the liquid equivalent snowfall decreases are found to be statistically significant at the 95% confidence level. Overall, these results suggest a trend toward more rain and less snow in future East Coast cyclone events.
Liquid equivalent total snowfall (mm) for simulation hours 18–63 in the (a) present-day and (b) future storm-relative composites. (c) Present-to-future change in liquid equivalent total snowfall (mm). Regions of statistically significant change at the 95% confidence level are contoured in black. The position of the composite storm center is represented by an asterisk. Distances from the storm center (km) in each direction are shown on the ordinate and abscissa.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
c. Intensity and storm-relative diagnostics
The future storm-relative composite cyclone attains a minimum sea level pressure ~3.5 hPa deeper than that found in the present composite by simulation hour 63 (Fig. 10a). Simulation hour 63 is also the time of deepest composite minimum sea level pressure. The future composite also exhibits a higher minimum sea level pressure at hour 18 than the present-day composite, indicating enhanced deepening rates in the cyclone evolution for the future composite. The maximum 12-h deepening rate of the future composite is found to be ~15% greater than that of the present composite. Enhanced deepening occurs after enhanced rain rates are seen in the future composite, implying that diabatic heating is the mechanism driving the greater intensification. The difference in minimum sea level pressure between the present-day and future composites (Fig. 10b) shows this trend more clearly. Nine of the 10 events attain future minimum sea level pressures less than that of their present-day counterpart. Consistent with this trend, the decrease in average minimum sea level pressure by simulation hour 63 is found to be statistically significant at the 95% confidence level using a Wilcoxon signed-rank test. The largest decrease in minimum sea level pressure among the individual cases is found to be ~8 hPa (Fig. 10b; case 4). This value is much smaller than that found in previous work where comparable increases in SST were applied during the simulation of an extratropical cyclone event along the U.S. East Coast (Booth et al. 2012). That study found a minimum sea level pressure decrease of ~15 hPa; however, the applied changes were limited to SSTs whereas the experimental design here also introduces changes in vertical stability and moisture. It is likely this difference in largest minimum sea level pressure decrease stems from the competing effects introduced by the temperature, stability, and moisture changes in the current study.
(a) Average minimum sea level pressure (hPa) evolution for the present-day (solid blue line) and future (solid red line) composites. Error bars denote the one standard deviation range. (b) Average minimum sea level pressure difference (hPa) between the present-day and future composites (thick solid red line). The minimum sea level pressure differences between the present-day and future simulations of each event (see legend for line colors) are also presented.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
In addition to minimum sea level pressure, changes in storm-relative 10-m wind speed and low-level jet intensity are also used to evaluate changes in surface cyclone intensity; these parameters hold implications for sensible weather impacts. Analysis of storm-relative sea level pressure and 10-m wind speed for the present-day (Fig. 11a) and future (Fig. 11b) composites indicate increases in intensity. The sea level pressure field in the future composite clearly exhibits a lower central pressure as well as a tightening of the sea level pressure contours surrounding the cyclone center. Consistent with this increase in the magnitude of the sea level pressure gradient, stronger winds are seen at the 10-m level. With the increase in precipitation associated with the future cyclone composite, one would also expect to see an enhancement of the low-level jet. This is indeed the case as the future composite shows expanded and strengthened 850-hPa wind speeds along the eastern flank of the cyclone (Figs. 11c,d). Likewise, the 850-hPa geopotential heights are clearly lower with a reduction in spacing of the height contours evident as well.
(left) Present-day and (right) future storm-relative composite analysis of (a),(b) 10-m wind speed (fill; m s−1) and sea level pressure (contour; hPa) and (c),(d) 850-hPa wind speed (fill; m s−1) and geopotential height (contour; m) at simulation hour 63. To focus on the low-level jet signal, only 850-hPa wind speeds greater than 25 m s−1 are shown. The position of the composite storm center is represented by an asterisk. Distances from the storm center (km) in each direction are shown on the ordinate and abscissa.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
The present-to-future changes in 10-m (Fig. 12a) and 850-hPa wind speed (Fig. 12b) indicate strengthening of the cyclonic circulation. Large areas of statistically significant increases in 10-m wind speed at the 95% confidence level are evident on the western, northern, and eastern edges of the cyclone. It should be noted that the increases on the western side of the storm may be partly because the future systems track farther away from the coastline so there is less frictional dissipation of near-surface winds there. Along the eastern flank of the composite cyclone there are widespread increases in 850-hPa wind speed, and the changes over much of the region are statistically significant at the 95% confidence level. This provides evidence to support the hypothesis that enhanced latent heat release results in a strengthened low-level jet within these systems.
Present-to-future change in storm-relative composite (a) 10-m wind speed (m s−1) and (b) 850-hPa wind speed (m s−1) at simulation hour 63. Regions of statistically significant change at the 95% confidence level are contoured in black. The position of the composite storm center is represented by an asterisk. Distances from the storm center (km) in each direction are shown on the ordinate and abscissa.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
To determine whether or not enhanced latent heat release is responsible for the increase in future surface cyclone intensity, additional storm-relative diagnostics are computed using the PV framework. Time-averaged fields for hours 51–63 are shown here to assess the PV diagnostics during the mature stages of the composite cyclones. As noted in section 2e, the UPV field is assessed by computing pressure and wind on the dynamic tropopause (1.5-PVU surface). Comparing the present-day and future time-averaged UPV fields for simulation hours 51–63 reveals slight differences in the vertical extent and positioning of the UPV (Figs. 13a,b). The UPV is situated west of the surface cyclone center in both the present-day and future composites. Wind speeds are also found to be larger on the dynamic tropopause in the future composite. We speculate this increase is in part a manifestation of strengthened diabatic outflow. The present-day and future DPV fields, time-averaged for the same period, exhibit somewhat similar structure but different intensity (Figs. 13c,d). The DPV maximum develops along the western flank of the cyclone center in each composite with this feature stronger in the future. The maximum DPV values seen in the present-day and future composites both exceed 1.25 PVU, but the area in which these values occur is visibly larger in the future composite. Small but noticeable differences are seen between the present-day and future time-averaged 
Time-averaged (left) present-day and (right) future storm-relative composite analysis of (a),(b) pressure (hPa; contour) and wind (m s−1; fill) on the dynamic tropopause (1.5-PVU surface), (c),(d) 900–750-hPa layer average Ertel PV (PVU; contour/fill), and (e),(f) sea level pressure (hPa; solid contour) and 2-m potential temperature anomaly (K; dashed contour/fill). Time-averaged quantities are computed for simulation hours 51–63. The position of the composite storm center is represented by an asterisk. Distances from the storm center (km) in each direction are shown on the ordinate and abscissa.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
One possible explanation for the weakening of the future 
The composite difference field of pressure on the dynamic tropopause, time-averaged for simulation hours 51–63, reveals some slight structural changes in terms of the relative positioning and extent of the UPV (Fig. 14a). South and west of the surface cyclone center the UPV extends closer to the surface (i.e., higher tropopause pressures in future) indicating a slight strengthening of the feature. However, statistically significant changes at the 95% confidence level are limited when analyzing individual output times (not shown). Present-to-future change in composite DPV, time-averaged for the same period, shows that the DPV within the cyclone core increases in strength with warming (Fig. 14b). Analysis of individual output times during the later stages of the cyclone evolution shows that the increase in future DPV strength is statistically significant at the 95% confidence level (not shown). This is consistent with the hypothesis proposed by Colle et al. (2013) and provides additional evidence that the increase in extratropical cyclone intensity along the U.S. East Coast is linked to stronger latent heating in the future. Change in the time-averaged 
Time-averaged present-to-future change in storm-relative composite (a) pressure (hPa) on the dynamic tropopause (1.5-PVU surface), (b) 900–750-hPa layer average Ertel PV, and (c) 2-m potential temperature anomaly (K). Time-averaged quantities are computed for simulation hours 51–63. The position of the composite storm center is represented by an asterisk. Distances from the storm center (km) in each direction are shown on the ordinate and abscissa.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
Overall, the weakening of the 
5. Discussion
Likewise, it is conceivable that the increases in precipitation found here are due in significant part to a storm track that is more conducive to heavy precipitation (i.e., increased overwater fetch) rather than the large-scale thermodynamic changes that have been imposed. To investigate this possibility, we compute the tracks for the five rainiest and five driest present simulations defined using area-average total precipitation (Fig. 15a). There does not appear to be any systematic differences between the tracks of the rainiest and driest events, although it is worth noting that two of the driest events track well inland initially. To investigate this further, we also compute the storm-relative composite total precipitation for the five present simulations that track farthest east (Fig. 15b) and farthest west (Fig. 15c). Here we define the farthest east/west by computing the time-average longitude of each cyclone along its track. The composite precipitation totals for the five events that track farthest east are visibly less than those of the five events that track farthest west. To the extent that future increases in cyclone intensity result from enhanced diabatic heating in the future, this suggests that the future increases in intensity are not a result of the future eastward shift in cyclone track.
(a) Individual cyclone tracks for five driest (brown lines) and five rainiest (green lines) present-day WRF simulations. (b) Storm-relative composite total precipitation (mm) for the five present-day WRF simulations that tracked farthest east. (c) As in (b), but for the five present-day WRF simulations that tracked farthest west.
Citation: Journal of Climate 28, 2; 10.1175/JCLI-D-14-00418.1
6. Conclusions
Previous investigations of climate warming impacts on extratropical cyclones suggest a decrease in the total number of cyclones, while indicating an increase in the most intense cyclones. Past studies have primarily made use of GCMs to diagnose hemispheric and regional changes in cyclone intensity. Most of these changes have been attributed to changes in environmental baroclinicity with latent heating also hypothesized to play some role (Geng and Sugi 2003; Lambert and Fyfe 2006; Watterson 2006; Teng et al. 2008; Bengtsson et al. 2009; Catto et al. 2011; Mizuta et al. 2011). Given their fairly coarse horizontal grid spacing, typically ~100 km or more, GCMs provide a limited means of diagnosing changes at storm scale. To our knowledge, few studies have investigated the impacts of warming on the storm-scale dynamics of extratropical cyclones.
In the present study we use the WRF model to perform simulations of wintertime extratropical cyclone events along the U.S. East Coast in present and future thermodynamic environments. The horizontal grid spacing used in these simulations allows for the diagnosis of changes in moist diabatic processes that may be unresolved or underresolved by GCMs. Storm-relative cyclone composites are generated from these simulations in order to assess the storm-scale impacts of warming on coastal extratropical cyclone events. The potential vorticity framework is also used in this work to test the hypothesis posed by Colle et al. (2013) that enhanced latent heating is responsible for increases in surface cyclone intensity with warming.
Earth-relative composite analysis of the present-day and future cyclone simulations indicate that an enhancement of upper-trough phase speeds result in a trend toward extratropical cyclone events that propagate eastward more quickly in the future. Storm-relative total precipitation is found to significantly increase in the future with area-average composite precipitation increasing at a rate slightly less than the moisture increase dictated by the Clausius–Clapeyron relation. Increased precipitation over the same time frame suggests enhanced precipitation rates associated with future extratropical cyclone events along the U.S. East Coast. Snowfall exhibits a statistically significant decrease in the future storm-relative composite. The strength of these high-impact cyclone events is found to increase using several different metrics for intensity. The storm-relative minimum sea level pressure of the future composite is ~3.5 hPa less than that of the present composite. Statistically significant increases in 10-m wind speed result from the deepening of the future cyclone and subsequent enhancement of the sea level pressure gradient. Enhanced latent heat release in the future cyclones is shown to strengthen the low-level jet through an apparent augmentation of the positive feedback between condensational heating and low-level moisture transport (Lackmann 2002). Storm-relative PV diagnostics indicate a strengthening of the low-level DPV and weakening of 
The goal of the current study was to determine how climate change might influence high-impact extratropical cyclones along the U.S. East Coast. Overall, our results suggest that future Miller-A events will be more intense in the future although the magnitude of this change is somewhat modest in terms of sensible impacts. A limitation to our approach is that it is focused on a specific category of high-impact cyclone event. We have only simulated cyclone events that developed strongly in the present climate, but it is possible that some events that were weak in the present climate may become quite intense in the future. While the composite cyclone is shown to increase in intensity with warming, this is not true for all individual events. Investigating why some events strengthen and others weaken with warming is a crucial next step in understanding how climate change may influence cyclone dynamics.
Acknowledgments
This research was supported by NSF Grant AGS-1007606, awarded to North Carolina State University. WRF is made available by the National Center for Atmospheric Research, funded by the National Science Foundation. The Program for Climate Model Diagnosis and Intercomparison (PCMDI) is acknowledged for making the GCM data used in this study available. We thank three anonymous reviewers who provided valuable feedback on earlier versions of this article. We also thank Dr. Anantha Aiyyer, who provided valuable suggestions on the M.S. thesis of C. Marciano, upon which this article is based.
REFERENCES
Ahmadi-Givi, F., G. C. Graig, and R. S. Plant, 2004: The dynamics of a midlatitude cyclone with very strong latent-heat release. Quart. J. Roy. Meteor. Soc., 130, 295–323, doi:10.1256/qj.02.226.
Allen, M. R., and W. J. Ingram, 2002: Constraints on future changes in climate and the hydrologic cycle. Nature, 419, 224–232, doi:10.1038/nature01092.
American Highway Users Alliance and IHS Global Insight, 2010: The economic costs of disruption from a snowstorm. 9 pp. [Available online at http://www.highways.org/2014/02/economic-impact-of-snowstorms/.]
Balasubramanian, G., and M. K. Yau, 1994: The effects of convection on a simulated marine cyclone. J. Atmos. Sci., 51, 2397–2417, doi:10.1175/1520-0469(1994)051<2397:TEOCOA>2.0.CO;2.
Bengtsson, L., K. I. Hodges, and N. Keenlyside, 2009: Will extratropical storms intensify in a warmer climate? J. Climate, 22, 2276–2301, doi:10.1175/2008JCLI2678.1.
Booth, J. F., L. Thompson, J. Patoux, and K. A. Kelly, 2012: Sensitivity of midlatitude storm intensification to perturbations in the sea surface temperature near the Gulf Stream. Mon. Wea. Rev., 140, 1241–1256, doi:10.1175/MWR-D-11-00195.1.
Booth, J. F., S. Wang, and L. Polvani, 2013: Midlatitude storms in a moister world: Lessons from idealized baroclinic life cycle experiments. Climate Dyn., 41, 787–802, doi:10.1007/s00382-012-1472-3.
Boutle, I. A., S. E. Belcher, and R. S. Plant, 2011: Moisture transport in midlatitude cyclones. Quart. J. Roy. Meteor. Soc., 137, 360–373, doi:10.1002/qj.783.
Brennan, M. J., G. M. Lackmann, and K. M. Mahoney, 2008: Potential vorticity (PV) thinking in operations: The utility of nonconservation. Wea. Forecasting, 23, 168–182, doi:10.1175/2007WAF2006044.1.
Bretherton, F. P., 1966: Baroclinic instability and the short wavelength cut-off in terms of potential vorticity. Quart. J. Roy. Meteor. Soc., 92, 335–345, doi:10.1002/qj.49709239303.
Catto, J. L., L. C. Shaffrey, and K. I. Hodges, 2011: Northern Hemisphere extratropical cyclones in a warming climate in the HiGEM high-resolution climate model. J. Climate, 24, 5336–5352, doi:10.1175/2011JCLI4181.1.
Champion, A. J., K. I. Hodges, L. O. Bengtsson, N. S. Keenlyside, and M. Esch, 2011: Impact of increasing resolution and a warmer climate on extreme weather from Northern Hemisphere extratropical cyclones. Tellus, 63A, 893–906, doi:10.1111/j.1600-0870.2011.00538.x.
Colle, B. A., Z. Zhang, K. A. Lombardo, E. Chang, P. Liu, and M. Zhang, 2013: Historical evaluation and future prediction of eastern North American and western Atlantic extratropical cyclones in the CMIP5 models during the cool season. J. Climate, 26, 6882–6903, doi:10.1175/JCLI-D-12-00498.1.
Davis, C. A., 1992: A potential-vorticity diagnosis of the importance of initial structure and condensational heating in observed extratropical cyclogenesis. Mon. Wea. Rev., 120, 2409–2428, doi:10.1175/1520-0493(1992)120<2409:APVDOT>2.0.CO;2.
Davis, C. A., and K. A. Emanuel, 1991: Potential vorticity diagnostics of cyclogenesis. Mon. Wea. Rev., 119, 1929–1953, doi:10.1175/1520-0493(1991)119<1929:PVDOC>2.0.CO;2.
Davis, C. A., M. T. Stoelinga, and Y.-H. Kuo, 1993: The integrated effect of condensation in numerical simulations of extratropical cyclogenesis. Mon. Wea. Rev., 121, 2309–2330, doi:10.1175/1520-0493(1993)121<2309:TIEOCI>2.0.CO;2.
Eady, E. T., 1949: Long waves and cyclone waves. Tellus, 1, 33–52, doi:10.1111/j.2153-3490.1949.tb01265.x.
Frei, C., C. Schär, D. Lüthi, and H. C. Davies, 1998: Heavy precipitation processes in a warmer climate. Geophys. Res. Lett., 25, 1431–1434, doi:10.1029/98GL51099.
Geng, Q., and M. Sugi, 2003: Possible change of extratropical cyclone activity due to enhanced greenhouse gases and sulfate aerosols—Study with a high-resolution AGCM. J. Climate, 16, 2262–2274, doi:10.1175/1520-0442(2003)16<2262:PCOECA>2.0.CO;2.
Hara, M., T. Yoshikane, H. Kawase, and F. Kimura, 2008: Estimation of the impact of global warming on snow depth in Japan by the pseudo-global-warming method. Hydrol. Res. Lett., 2, 61–64, doi:10.3178/hrl.2.61.
Held, I. M., and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. J. Climate, 19, 5686–5699, doi:10.1175/JCLI3990.1.
Huo, Z., D.-L. Zhang, and J. R. Gyakum, 1999: Interaction of potential vorticity anomalies in extratropical cyclogenesis. Part I: Static piecewise inversion. Mon. Wea. Rev., 127, 2546–2562, doi:10.1175/1520-0493(1999)127<2546:IOPVAI>2.0.CO;2.
Jiang, J., and W. Perrie, 2007: The impacts of climate change on autumn North Atlantic midlatitude cyclones. J. Climate, 20, 1174–1187, doi:10.1175/JCLI4058.1.
Karl, T. R., and R. W. Knight, 1998: Secular trends of precipitation amount, frequency, and intensity in the USA. Bull. Amer. Meteor. Soc., 79, 231–241, doi:10.1175/1520-0477(1998)079<0231:STOPAF>2.0.CO;2.
Kocin, P. J., and L. W. Uccellini, 2004: Overview. Vol. 1, Northeast Snowstorms, Amer. Meteor. Soc., 296 pp.
Lackmann, G. M., 2002: Potential vorticity redistribution, the low-level jet, and moisture transport in extratropical cyclones. Mon. Wea. Rev., 130, 59–74, doi:10.1175/1520-0493(2002)130<0059:CFPVMT>2.0.CO;2.
Lackmann, G. M., 2013: The south-central U.S. flood of May 2010: Present and future. J. Climate, 26, 4688–4709, doi:10.1175/JCLI-D-12-00392.1.
Lackmann, G. M., and J. R. Gyakum, 1999: Heavy cold-season precipitation in the northwestern United States: Synoptic climatology and an analysis of the flood of 17–18 January 1986. Wea. Forecasting, 14, 687–700, doi:10.1175/1520-0434(1999)014<0687:HCSPIT>2.0.CO;2.
Lambert, S. J., and J. C. Fyfe, 2006: Changes in winter cyclone frequencies and strengths simulated in enhanced greenhouse warming experiments: Results from the models participating in the IPCC diagnostic exercise. Climate Dyn., 26, 713–728, doi:10.1007/s00382-006-0110-3.
Mallard, M. S., G. M. Lackmann, and A. Aiyyer, 2013: Atlantic hurricanes and climate change. Part I: Experimental design and isolation of thermodynamic effects. J. Climate, 26, 4876–4893, doi:10.1175/JCLI-D-12-00182.1.
Marciano, C. G., 2014: Climate change and wintertime East Coast cyclones. M.S. thesis, Dept. of Marine, Earth and Atmospheric Sciences, North Carolina State University, 167 pp. [Available online at http://repository.lib.ncsu.edu/ir/bitstream/1840.16/9243/1/etd.pdf.]
Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343–360, doi:10.1175/BAMS-87-3-343.
Miller, J. E., 1946: Cyclogenesis in the Atlantic coastal region of the United States. J. Meteor., 3, 31–44, doi:10.1175/1520-0469(1946)003<0031:CITACR>2.0.CO;2.
Mizuta, R., M. Matsueda, H. Endo, and S. Yukimoto, 2011: Future change in extratropical cyclones associated with change in the upper troposphere. J. Climate, 24, 6456–6470, doi:10.1175/2011JCLI3969.1.
Morgan, M. C., and J. W. Nielsen-Gammon, 1998: Using tropopause maps to diagnose midlatitude weather systems. Mon. Wea. Rev., 126, 2555–2579, doi:10.1175/1520-0493(1998)126<2555:UTMTDM>2.0.CO;2.
Perrie, W., Y. Yao, and W. Zhang, 2010: On the impacts of climate change and the upper ocean on midlatitude northwest Atlantic landfalling cyclones. J. Geophys. Res., 115, D23110, doi:10.1029/2009JD013535.
Plant, R. S., G. C. Craig, and S. L. Gray, 2003: On a threefold classification of extratropical cyclogenesis. Quart. J. Roy. Meteor. Soc., 129, 2989–3012, doi:10.1256/qj.02.174.
Rasmussen, R., and Coauthors, 2011: High-resolution coupled climate runoff simulations of seasonal snowfall over Colorado: A process study of current and warmer climate. J. Climate, 24, 3015–3048, doi:10.1175/2010JCLI3985.1.
Reed, R. J., G. Grell, and Y.-H. Kuo, 1993: The ERICA IOP 5 storm. Part II: Sensitivity tests and further diagnosis based on model output. Mon. Wea. Rev., 121, 1595–1612, doi:10.1175/1520-0493(1993)121<1595:TEISPI>2.0.CO;2.
Sato, T., F. Kimura, and A. Kitoh, 2007: Projection of global warming onto regional precipitation over Mongolia using a regional climate model. J. Hydrol., 333, 144–154, doi:10.1016/j.jhydrol.2006.07.023.
Schär, C., C. Frei, D. Lüthi, and H. C. Davies, 1996: Surrogate climate-change scenarios for regional climate models. Geophys. Res. Lett., 23, 669–672, doi:10.1029/96GL00265.
Skamarock, W. C., and Coauthors, 2008: A description of the Advanced Research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, 113 pp.
Soden, B. J., D. L. Jackson, V. Ramaswamy, M. D. Schwarzkopf, and X. Huang, 2005: The radiative signature of upper tropospheric moistening. Science, 310, 841–844, doi:10.1126/science.1115602.
Stoelinga, M. T., 1996: A potential vorticity–based study of the role of diabatic heating and friction in a numerically simulated baroclinic cyclone. Mon. Wea. Rev., 124, 849–874, doi:10.1175/1520-0493(1996)124<0849:APVBSO>2.0.CO;2.
Teng, H., W. M. Washington, and G. A. Meehl, 2008: Interannual variations and future change of wintertime extratropical cyclone activity over North America in CCSM3. Climate Dyn., 30, 673–686, doi:10.1007/s00382-007-0314-1.
Trenberth, K. E., 1999: Conceptual framework for changes of extremes of the hydrological cycle with climate change. Climatic Change, 42, 327–339, doi:10.1023/A:1005488920935.
Trenberth, K. E., A. Dai, R. M. Rasmussen, and D. B. Parsons, 2003: The changing character of precipitation. Bull. Amer. Meteor. Soc., 84, 1205–1217, doi:10.1175/BAMS-84-9-1205.
Watterson, I. G., 2006: The intensity of precipitation during extratropical cyclones in global warming simulations: A link to cyclone intensity? Tellus, 58A, 82–97, doi:10.1111/j.1600-0870.2006.00147.x.
Wentz, F. J., L. Ricciardulli, K. Hilburn, and C. Mears, 2007: How much more rain will global warming bring? Science, 317, 233–235, doi:10.1126/science.1140746.
Willison, J., W. A. Robinson, and G. M. Lackmann, 2013: The importance of resolving mesoscale latent heating in the North Atlantic storm track. J. Atmos. Sci., 70, 2234–2250, doi:10.1175/JAS-D-12-0226.1.
Yin, J. H., 2005: A consistent poleward shift of the storm tracks in simulations of 21st century climate. Geophys. Res. Lett., 32, L18701, doi:10.1029/2005GL023684.
Zappa, G., L. C. Shaffrey, K. I. Hodges, P. G. Sansom, and D. B. Stephenson, 2013: A multimodel assessment of future projections of North Atlantic and European extratropical cyclones in the CMIP5 climate models. J. Climate, 26, 5846–5862, doi:10.1175/JCLI-D-12-00573.1.