1. Introduction
Extratropical transition (ET) is a process during which tropical cyclones (TCs) interact with extratropical flow patterns, undergoing a transformation from symmetric, warm-core tropical systems to asymmetric, cold-core extratropical systems (Jones et al. 2003; Evans et al. 2017). Transitioning TCs often exhibit an increase in the radius of gale-force winds and the development of frontal structures. During transition, changes in the TC environment include an increase in vertical wind shear, Coriolis parameter, meridional temperature and moisture gradient magnitudes, and a decrease in sea surface temperature (SST; e.g., Harr and Elsberry 2000; Klein et al. 2000; Atallah and Bosart 2003; Jones et al. 2003; Evans and Hart 2008; Evans et al. 2017). Through these environmental and subsequent structural changes, a subset of the transitioning TCs can develop into intense cyclonic systems that bring TC-like conditions (e.g., torrential precipitation, large ocean waves, storm surge, and hurricane-force winds) to areas far removed from the original TC in late summer and early autumn. For example, Hurricane Irene (2011) brought heavy rainfall and catastrophic flooding to the U.S. East Coast and New England during ET (e.g., Avila and Cangialosi 2011; Liu and Smith 2016). Beyond impacts in their immediate vicinity, transitioning TCs can also influence midlatitude weather through their outflow, which can initiate or modify midlatitude baroclinic wave packets (Wirth et al. 2018). Such wave trains can be associated with high-impact weather in regions far downstream of the transitioning TCs (e.g., Harr and Dea 2009; Keller et al. 2014; Archambault et al. 2015; Quinting and Jones 2016; Keller 2017; Pohorsky et al. 2019; Keller et al. 2019).
Several recent studies have investigated the potential response of ET to climate change (e.g., Liu et al. 2017; Jung and Lackmann 2019; Michaelis and Lackmann 2019; Liu et al. 2020; Bieli et al. 2020). Michaelis and Lackmann (2019) conducted an analysis of multiseasonal general circulation model (GCM) simulations with “TC-allowing” resolution (a 15-km grid in the Northern Hemisphere). They found that the future simulations exhibited a more TC-favorable background environment in parts of the North Atlantic, along with a shift toward stronger lower-tropospheric TC warm-core structures, and poleward shifts in the latitude of TC genesis and peak intensity. In these model simulations, the percentage and overall frequency of ET increased in the North Atlantic for the late twenty-first century. In contrast, Bieli et al. (2020) found little overall change in ET frequency, and an increase in the likelihood of ET in the western North Pacific basin.
In addition to Michaelis and Lackmann (2019), several other recent studies suggest that the frequency of TCs undergoing ET will increase with climate change in the North Atlantic due to TC-favorable environmental changes (e.g., warmer SSTs and reduced vertical wind shear) by the late twenty-first century, including Haarsma et al. (2013), Baatsen et al. (2015), and Liu et al. (2017). Poleward shifts in TC genesis, in conjunction with the poleward extension of TC-favorable conditions (e.g., Sharmila and Walsh 2018), are also consistent with studies that find an increase in the latitude of maximum TC intensity (e.g., Kossin et al. 2014). Using Hurricane Irene (2011) as a case study, Jung and Lackmann (2019) and Liu et al. (2020) determined that Irene exhibited an increased ability to extend TC-like conditions poleward in the future simulations, consistent with the aforementioned results.
In summary, studies of ET and climate change to date suggest that the number of TCs experiencing extratropical transition is likely to increase in some regions, implying that locations such as the northeastern United States and western Europe could experience increased exposure to threats associated with ET. However, there is not yet a consensus around the question of how overall ET frequency will change in a warming climate. Furthermore, there are critical questions left unanswered regarding the potential influence of climate change on ET, including changes in storm-scale characteristics and hazards as well as potential changes in downstream impacts. Before a consensus can develop, consistent results must be obtained from a variety of experimental methods, models, and datasets. Some limitations in the methods of previous studies include coarse model resolution, requiring use of convective parameterization, and examination of a limited number of ET cases. For coarse resolution TC simulations, the use of convective parameterization to simulate subgrid-scale precipitation precludes explicit representation of a portion of the TC secondary circulation, which affects simulated TC intensity and hinders representation of realistic TC structure. Gentry and Lackmann (2010) demonstrated that a grid spacing of 4 km or less adequately captured the TC structure and intensity of Hurricane Ivan (2004). Knutson and Tuleya (2004) also showed the use of convective parameterization can have a considerable impact on the rainband and inner-core features associated with the simulated TCs, demonstrating the importance of explicit convection to the representation of the responses of TC precipitation to greenhouse-induced warming. Without capturing the full strength of tropical cyclones in numerical simulations, it is difficult to assess changes in the ET process and associated impacts in response to climate change (Davis 2018).
Many previous studies have used the cyclone phase space (CPS) method of Hart (2003) to quantify and diagnose ET. However, this method was developed at a time when coarser grid spacing was the rule for analyses and simulations, and our experience with high-resolution simulations suggests that transition is delayed in the CPS framework due to the stronger and more persistent presence of a warm-core vortex in such simulations. Thus, we introduce an alternate, experimental method for diagnosis of ET.
As mentioned above, in addition to direct threats, transitioning systems can potentially impact remote regions via downstream dispersion of Rossby wave energy. With increased water vapor content in a warming atmosphere, an associated increase in latent heating and diabatic outflow could result, enhancing downstream energy dispersion associated with ET (e.g., Steinfeld and Pfahl 2019). To the best of the authors’ knowledge, this aspect has not been investigated to date for ET events; we will present an initial analysis here, and a recent paper by Michaelis and Lackmann (2021) present preliminary analysis of this aspect using global model simulations.
Studies of ET and climate change based on single case studies, such as Jung and Lackmann (2019) and Liu et al. (2020), raise the question of representativeness. To address this limitation, we introduce a quasi-idealized model experimental design using composites of North Atlantic ET cases over 40-yr period from 1979 to 2018. Using composite-derived initial and lateral boundary conditions, in conjunction with high-resolution, convection-allowing resolution, our goal is to provide a representative and detailed analysis of storm-scale changes in the ET process in a changing climate. Our simulations use an “aquaplanet” domain to eliminate complications associated with land interactions, as discussed by Jung and Lackmann (2019). The quasi-idealized approach allows us to analyze how climate change will affect a generic North Atlantic recurving TC undergoing ET. The basic method is similar to the pseudo–global warming (PGW) approach used in many previous studies (e.g., Schär et al. 1996; Hara et al. 2008; Mallard et al. 2013; Lackmann 2015; Patricola and Wehner 2018; Jung and Lackmann 2019; Reed et al. 2020), providing downscaled present-day and end-of-century counterpart simulations in highly similar synoptic environments. The unique contributions of this paper toward understanding how ET responds to climate change are (i) a novel experimental design that is quasi-idealized, allowing more general results than individual case studies, (ii) convection-allowing model resolution, providing full-strength TC representation, and (iii) a new method for diagnosis of ET that is effective for high-resolution model simulations.
In section 2, we describe our method of case classification, present a new, experimental method of defining ET completion, provide information about the storm-relative compositing techniques, and detail the model experimental design. Section 3 provides an overview of synoptic and storm-scale evolutions in the reanalysis-based composite as well as in present-day, composite-based simulations. In section 4, we present projected future changes in North Atlantic ET; general findings are discussed and the paper is concluded in section 5.
2. Data and methods
a. Data and classification of ET cases
Ritchie and Elsberry (2003, 2007) and Hart et al. (2006) demonstrated that the evolution and intensity changes during ET are highly sensitive to the temporal and spatial scales of interaction between a mobile upper-tropospheric trough and the transitioning TC. While many factors are involved in the scales of interaction between the upper- and lower-level systems, TC tracks and associated synoptic patterns play an important role in dictating the timing and phasing of the trough–TC interaction, resulting in a variety of evolutionary pathways in the ET process. North Atlantic TC tracks are determined in part by large-scale steering flows such as the North Atlantic subtropical high, which can be affected by various climate features including El Niño–Southern Oscillation (ENSO) and the North Atlantic Oscillation (NAO; e.g., Bove et al. 1998; Elsner et al. 2000; Xie et al. 2005; Kossin et al. 2010; Colbert and Soden 2012). Thus, employing a track-based TC classification method brings several benefits: It allows stratification of ET events by synoptic pattern and can help to identify potential impacts from climate variability (e.g., ENSO, Atlantic multidecadal oscillation).
Here, North Atlantic TCs that underwent ET between 1979 and 2018 are examined and classified into one of three categories according to track. Our categorization is based on the classification scheme presented by Colbert and Soden (2012): Straight moving (SM) TCs remain south of 25°N until they cross 80°W, eventually threatening the U.S. Gulf Coast and the western Caribbean Sea. Recurving landfall (RCL) TCs cross the 70°W meridian north of 25°N or cross the 65°W meridian north of 40°N and threaten the U.S. East Coast. Recurving ocean (RCO) TCs do not cross either threat boundary, but travel to the north of 25°N latitude and recurve into the open ocean (Fig. 1). This study focuses on TCs that formed in the Atlantic main development region, defined as the area between 7° and 22.5°N and between 17.5° and 82°W, and are classified as RCO cases. This selection process is designed to minimize uncertainties from TCs experiencing different degrees of land interaction and synoptic steering flows.
Track boundaries for classifying ET events. TCs categorized as straight-moving (SM) cross the green boundary line to threaten the Gulf Coast and western Caribbean. Recurving landfall (RCL) TCs cross the red boundary and threaten the U.S. East Coast. Recurving oceanic TCs, shown as blue tracks, did not directly threaten the United States. The selected 21 RCO TC tracks for use in compositing over the 40-yr reanalysis period are shown in light blue contours, with dots indicating TC location at 6-hourly intervals. Tracks obtained from HURDAT2 dataset.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
To identify ET events and for TC track information, we used the NHC “best track” Atlantic Hurricane Database (HURDAT2; Landsea and Franklin 2013). For atmospheric fields, we used the European Centre for Medium-Range Weather Forecasts reanalysis product, ERA5 (ECMWF 2017; Hersbach et al. 2020), available with an approximate grid length of 0.25°. The TC track-based classification procedure, with the specific latitude and longitude track thresholds described above, yields 49 RCO ET cases over the 40-yr period (Fig. 1). While there is considerable variability in the time taken to complete transition, Evans and Hart (2003) and Hart et al. (2006) found that ~70% of North Atlantic TCs that underwent ET took less than 36 h for this process to complete. To reduce spread within the composite sample, we applied this threshold to the 49 RCO ET cases. Because HURDAT2 only provides the time of ET completion and does not include temporal information such as ET duration, the cyclone phase space method (CPS; Hart 2003) is adopted to determine the length of ET for the selected 49 cases. We found that 21 RCO ET cases complete CPS-defined ET in 36 h or less, out of the 49 RCO ET cases (Fig. 1). We note that ~10 of the RCO ET cases (out of 49) either merged with a midlatitude trough or decayed before completion of ET, based on the CPS analysis. ET duration of the rest of the unselected ET cases ranges from 42 to 72 h. We present more information regarding determination of ET onset and ET completion in section 2c.
b. Numerical model simulations
Comparisons of numerical model perturbation experiments that change initial conditions, physics parameterizations, or boundary conditions can yield spurious results due to the rapid propagation of numerical noise within a model (Ancell et al. 2018). Thus, we use a small ensemble with varying initial and lateral boundary conditions to provide robust results and minimize the influence of spurious numerical noise. The ensembles are composed of six members with varying initial and lateral boundary conditions using a novel quasi-idealized experimental design.
The simulations are initialized (and use lateral boundary conditions) from composites of the 21 North Atlantic RCO ET events discussed in section 2a. To create realistic initial and lateral boundary conditions for simulations of North Atlantic ET events, we utilized storm-relative coordinates in a fixed domain. This is a four-step procedure: First, the latitude and longitude position of all 21 storm centers is determined, and the average of the latitude and longitude is set as the 21-case average storm center position. Next, we recenter the 21 ET cases’ coordinates relative to the storm center at each simulation time increment (i.e., 22 time increments for 5.25 days with a 6-h ERA5 time interval; from 60 h prior to ET onset to 66 h after ET onset). After this step, the x and y axes of each case domain are measured in degrees relative to the storm center. In the third step, we fix the domain based on the 21-case average storm track (this domain is used as the WRF domain shown in Figs. 2a–d) and the storm-center-relative x and y axes are reassigned based on the 21-case average location of the storm center at each time. In other words, we put 21 storms in storm-relative coordinates into the same location based on the 21-case average storm center position in fixed latitude and longitude coordinates. Finally, the 21 RCO ET cases with varying locations are centered to the same location at a given time, with the same domain size, allowing us to generate composite environmental fields. We are then able to use these as initial and lateral boundary conditions for a six-member ensemble of high-resolution simulations, which are able to attain full TC and ET intensity relative to the coarser reanalysis; this also allows us to perform experimental simulations.
Ensemble-mean sea surface temperature (K) in WRF simulation domain for the (a) present-day and (b) future simulations at the initial time in the quasi-idealized, all-oceanic model domain. (c) Imposed ensemble-mean sea surface temperature change (K) in the future simulations at the initial time. (d) Changes in summer and fall sea surface temperature climatology between 2080–99 and 1980–99, derived from the 20-model ensemble of CMIP5 RCP8.5 projections. (e) Vertical cross section of imposed, zonally averaged (between 80° and 25°W) ensemble-mean temperature change for the future simulation at the initial time. Coastlines and U.S. state boundaries are shown for reference in (a)–(d).
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
Varying initial and lateral boundary conditions for the 6-member ensemble are derived via random selection of 15 of the 21 RCO ET cases, resulting in differing composites of the storm-relative environmental fields. Randomized 15-case composite initial conditions are generated 60 h prior to ET onset, determined by the CPS computation from ERA5 grids. The simulations are run for 126 h to capture the entire ET and post-ET process. We use the Weather Research and Forecasting (WRF) Model version 4.0 (Skamarock et al. 2019).
We conducted convection-allowing simulations by adding a two-way, TC-following nest with 4-km grid length within an outer domain featuring 12-km grid length. The dimension of the mobile inner nest is 1636 km × 1636 km, sufficient to capture the circulation of the transitioning system. In the vertical, 50 sigma levels are used with higher resolution in the lower and upper troposphere. The model top is set at 50 hPa. The 4-km domain allows us to simulate realistic TC structures, including the TC secondary circulation. Because the tracks of the simulated storms vary among the 6 ensemble members, we apply a similar procedure to that used in generating the initial and lateral boundary conditions to the ensemble output in order to obtain composite synoptic environment fields relative to the simulated storms. For fields such as precipitation and wind speed, we use a 24° × 24° grid box centered on the grid cell of minimum sea level pressure (SLP), allowing us to identify common characteristics between ET events and to detect changes in these characteristics between future and present-day storms.
Model physics include the Kain-Fritsch (KF) cumulus parameterization (Kain 2004) for the outer domain only, the WSM6 single 6-class microphysics scheme, Yonsei University (YSU) planetary boundary layer, and the Rapid Radiative Transfer Model-Global (RRTMG) scheme for longwave and shortwave radiation. These physics schemes are chosen based on their performance reproducing TC tracks, rainband structures, and intensities in the Hurricane Irene simulations of Jung and Lackmann (2019). All ensemble members are run with the physics choices described above. We ran the WRF digital filter initialization (DFI) technique in all simulations to balance the momentum and mass fields and populate hydrometeors fields at the initial time. This reduces the spinup time and ensures a more balanced initial state in both present-day and future simulations (Lynch and Huang 1992, 1994; Lackmann 2015). This is particularly important since we are using composite initial conditions, which are not necessarily in a state of dynamical balance.
For the projected “future” simulations, we used a pseudo–global warming (PGW) approach similar to that of Jung and Lackmann (2019). Using a 20-GCM ensemble of Coupled Model Intercomparison Project (CMIP5) RCP 8.5 projections for the years 2080–99, and the years 1980–99 in historical runs, we compute thermodynamic differences in July–September climatological averages (Fig. 2d; See Jung and Lackmann (2019) for a full list of 20 CMIP5 GCMs adopted in this study, and other details of the PGW approach used here). The GCM-average thermodynamic “delta” values (Figs. 2c,e), including 2-m air temperature, SST, and atmospheric temperature at each isobaric level, are regridded to the composite domain and added to the present-day composite initial and lateral boundary conditions.
Additionally, we set carbon dioxide (CO2) concentrations to 936 ppm in the future simulations, consistent with Meinshausen et al. (2011). Relative humidity is held constant in the modification of initial conditions for the future simulations, in accordance with previous studies (e.g., Allen and Ingram 2002; Soden et al. 2005; Hill and Lackmann 2011; Lackmann 2015), resulting in relatively larger increases in water vapor mixing ratio in locally warm regions. This choice avoids unrealistic RH conditions that would arise if GCM-derived moisture changes were applied. The imposed thermodynamic changes on the future environmental fields could potentially cause imbalances between the wind and mass fields, but use of the DFI technique as discussed above alleviates this issue.
c. Determination of the onset and completion of ET
Despite its widespread use, several previous studies (e.g., Studholme et al. 2015; Liu et al. 2017; Zarzycki et al. 2017; Bieli et al. 2020) show that the CPS definitions brought about false alarm ET events or did not adequately reflect the ET process due to the reasons discussed by Evans et al. (2017, their section 2). The CPS method may work best for datasets that do not fully resolve the intense inner warm-core TC structure; the CPS results are thus sensitive to the resolution of the data used. Similarly, we encountered issues where the completion of ET was indeterminate by the CPS method in some of our high-resolution simulations because a strong remnant warm-core vortex persisted; both the simulated ensemble future and present-day storms retained their warm core in the lower troposphere even after they developed thermally asymmetric structures. In those cases, a parameter describing the existence of a lower-tropospheric cold core defined by CPS does not appear to capture the time of ET completion such as deduced from subjective interpretation of simulated satellite imagery (Fig. 3).
Analysis of WRF-simulated ET ensembles. (a),(e) Cyclone phase space (CPS) diagrams showing
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
Results using this method were subjectively compared to simulated satellite imagery, and CPS results (Fig. 3). The experimental method provided results that are qualitatively consistent with storm characteristics in simulated satellite imagery; at the time of diagnosed ET, imagery depicted highly asymmetric systems attended by warm- and cold-frontal structures (Figs. 3c,d,g,h). As the model output is only available every 3 h, we use the 3-h interval for determining the completion of ET (i.e., 65 h for the present-day, 75 h for the future; Figs. 3d,f). For our high-resolution simulations, we replace the procedure described in section 2a with this method to determine ET completion. Although our new method reflected the ET process relatively well, it has important limitations: the diabatic generation of Ke, partly created by surface latent heat flux, projects onto the baroclinic conversion. Further, the method was developed and assessed using a limited number of ET cases. Thus, more work is necessary to assess the reliability of this method and to establish threshold ET values, using datasets of varying resolution and quality with a large number of ET cases.
d. Eddy kinetic energy budget analysis
3. Storm evolution in the composite and simulation
To show the realism of the composite ET event evolution, which we use as initial and lateral boundary conditions for the simulations, we display 500-hPa geopotential height, absolute vorticity, SLP, 1000–500-hPa thickness, 850-hPa temperature, 10-m wind, and precipitation distributions for the ERA5-based, 21-case RCO ET composite (Fig. 4). The composite storm initiated near 28°N, 58°W and initially follows a northerly track over the North Atlantic (Fig. 5a); recall that we recenter the composite on the storm center at each composite time.
Synoptic overview of 21-ET-case composite based on ERA5 gridded data. (a)–(c) Absolute vorticity (10−5 s−1; shaded) superimposed by 500-hPa geopotential height (m; contours); (d)–(f) 10-m wind speed (m s−1; shaded) superimposed by sea level pressure (hPa; contours) and 850-hPa temperature (°C, red dashed contours); (g)–(i) as in (d)–(f), but for 3-h accumulated precipitation [mm (3 h)−1; shaded]. Valid at (a),(d),(g) ET − 30 h; (b),(e),(h) ET + 18 h; and (c),(f),(i) ET + 48 h. Coastlines are shown for reference.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
(a) Comparison of the ERA5 composite track of 21 RCO cases (blue) to the present-day simulation (red) with minimum SLP labels. (b) Time series of minimum sea level pressure (MSLP; hPa) for the ERA5-based 21 RCO ET case composite (blue contour) and present-day simulation ensemble average (red contour). Light blue and light pink shadings show composite and ensemble spread. Times shown relative to ET onset for the composite and simulation hour for the present-day simulation. Coastlines shown in (a) for reference.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
During its northeastward movement, the composite system underwent three distinct stages: (i) slight intensification, (ii) weakening, and (iii) reintensification (Fig. 5b; blue line). As expected, there are considerable differences in intensity and timing between the 21 cases within the ERA5 composite, and also between the composite and present-day WRF simulation ensemble mean; these result in part from the relatively coarse ERA5 resolution, from the dynamical evolution of the storm in the model, and from aforementioned timing differences in the composite. Still, the temporal trend of MSLP is consistent between the model and composite (Fig. 5b). After reaching a minimum SLP of 985-hPa at ET + 6 h, the ERA5 composite storm gradually weakens during the ET process. For the 30 h after ET onset, the composite system began to reintensify, attaining greater intensity than the original TC in the reanalysis. This is one of the typical pathways taken by a storm during the post-ET process (Kofron et al. 2010), suggesting that reintensification was a predominant pathway for the RCO ET cases during post-ET process over the composite sample. As the storm moved northward ahead of the upper-level trough to its northwest, the TC interacted with the approaching trough, featuring redistribution of the main precipitation region to the poleward side of the storm and expansion of the radius of maximum wind (Fig. 4). These changes in the distribution of precipitation and wind are representative signatures of TCs undergoing ET (e.g., Evans et al. 2017).
The characteristics that resemble observed TCs undergoing ET are even more clearly evident in the present-day simulation relative to the reanalysis-based composite. Figure 6 displays a similar synoptic overview to that in Fig. 4, except based on the present-day WRF ensemble mean. As the simulated TC translates into an increasingly baroclinic environment, it interacts with an upper-level trough, and the system exhibits modifications of its structure such as an expanded, asymmetric wind field (Figs. 6d–f). The notable alterations of the storm structure are manifest in a reduction in cold cloud-top temperatures in the southern quadrant of the storm, due to cold advection, as the storm begins to interact with the baroclinic zone (Figs. 6g–i). A comma-shaped cloud pattern, associated with warm and cold frontogenesis, has developed by simulation hour 78 (Fig. 6h). These structural changes in the simulated storm are typical characteristics to consider when subjectively detecting the transformation from a TC to an extratropical cyclone, increasing our confidence in the novel quasi-idealized method we are using: The simulated ET event, though initialized in an unconventional fashion using composites, nevertheless depicts a realistic transitioning TC and large-scale environment. The absence of small-scale synoptic and mesoscale features, as would characterize an individual case study, allows us to simulate a more representative “generic” ET evolution.
As in Fig. 4, but for the present-day WRF-simulated ensemble mean; (g)–(i) simulated cloud-top temperature (K; shaded as in legend). Valid at (a),(d),(g) 30; (b),(e),(h) 78; and (c),(f),(i) 108 h. These times correspond to the times shown in Fig. 5. Coastlines are shown for reference.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
4. Projected future changes in extratropical transition
a. Duration of ET and intensity
Based on the methods described in section 2b, we diagnose the onset and completion of ET in our high-resolution present-day and future ensemble simulations. Figure 7 shows simulated cloud top temperature during the transformation stage for both present-day and future ensemble mean storms. The storms exhibit a typical symmetric appearance of clouds and deep convection during the TC phase (Figs. 7a,d). As ET onset commences, deep convection begins to diminish in the southern and eastern quadrants, while a large expanse of multilayer clouds with embedded deep convection is expanding to the north and northeast of the storms (Figs. 7b,e). At the time of ET completion, determined by our experimental method, temperatures of the coldest cloud tops have warmed, and significant deep convection has diminished (Figs. 7c,f). Instead, the storms feature a broad area of multilayer clouds in the northern quadrant, associated with warm advection to the north and east of the storm center. Similar figures, derived from the convection-allowing inner domain, clearly depict a reduction in cloudiness and the appearance of dry slots in the southern quadrant due to the equatorward advection of cold air and warm frontogenesis in the northern quadrant at the time of diagnosed ET completion (Figs. 8 and 9). As expected, the future storm demonstrates colder cloud tops, indicating stronger convection over a wider region, especially during the TC phase (Fig. 8). It is worth noting that while all ensemble members undergo similar evolution until ET onset, ensemble members which simulate stronger reintensification complete the transition earlier than the other members (not shown). Given that deep convection has eroded in the western and southern areas relative to the storm center, and the development of a comma-shaped cloud pattern that commonly characterizes a TC that has completed ET (Klein et al. 2000; Jones et al. 2003; Evans et al. 2017), our method of ET diagnosis appears to capture the time of ET completion reasonably well; the threshold of the
Cloud-top temperature (°C, shaded) for the (a)–(c) present-day ensemble-mean simulation and (d)–(f) future counterpart. Valid (a),(d) at 24 h prior to diagnosed ET onset; (b),(e) at the diagnosed ET onset; and (c),(f) at the diagnosed ET completion. Coastlines are shown for reference.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
As in Fig. 7, but for inner-domain (4-km convection allowing). All panels are represented in storm-centered coordinates. Distance from the storm center in degrees is shown on the ordinate and abscissa with (0°, 0°) marking the storm center.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
As in Fig. 8, but for atmospheric temperature at 850-hPa (shaded; 0.5-K interval) and SLP (dashed contours; 4-hPa interval).
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
The diagnosed onset and completion of ET are shown in Fig. 10. Whereas Jung and Lackmann (2019) found that the duration of ET considerably extended in their future, warming simulations, no significant change in ET duration is evident here, with a modest increase of around 3 h in the future ensemble-mean simulation compared to its present-day counterpart. During the entire ET process, the track of the future storm is nearly identical to that of present-day system, which allows us to make a direct comparison without complications arising from track differences, unlike Jung and Lackmann (2019). Therefore, the results imply that even though the future changes in environmental fields are more favorable for ET, including a reduction in meridional SST gradient (Fig. 2) and reduced vertical wind shear along the track of the storm (not shown), these changes do not play a major role in extending future ET duration in these simulations. Michaelis (2019) found a similar result in high-resolution GCM simulations with a large number of ET cases, also consistent with unchanged ET duration in the North Atlantic.
Future ensemble comparison to present-day ensemble mean for (a) track and (b) minimum sea level pressure (hPa). The dark blue lines represent the present-day ensemble and the dark red solid lines represent the future ensemble. The onset and completion of ensemble ET are labeled ETB for the onset and ETE for the completion of the transition. The future minimum SLP is normalized by the background SLP changes. Coastlines are shown for reference.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
The intensity changes are measured by the minimum sea level pressure (SLP). The analyzed 10-m wind fields showed large variability in strength near the storm center in both the future and present-day ensemble simulations, which render them unrepresentative of the true storm intensity; thus we focus on SLP to represent intensity. To reflect the background environmental SLP changes in the future, the future minimum SLP is normalized by temporal average of SLP changes over the entire simulation; the temporal average of SLP change is approximately 3 hPa. Unless explicitly stated, all analyses for the future storm intensity are presented using the normalized SLP by the background environmental SLP changes. Comparison of the minimum sea level pressure (SLP) and maximum wind speed shows that the ensemble-mean future storm is stronger during the entire simulation period except for the last few simulation hours. Both future and present-day storms undergo reintensification after ET completion (Fig. 10b). The ensemble-mean minimum SLP reached by the future storm is ~10 hPa deeper before ET completion, but is almost weaker near the end of the simulations. The future storm undergoes weaker reintensification than its present-day counterpart during the post-ET process. In brief, the future ensemble-mean storm has greater intensity during the TC and entire ET stages, the present-day storm, meanwhile, goes through more rapid reintensification after ET completion.
In an effort to understand the causes of the differing intensity changes, we conducted an analysis of the eddy kinetic energy (Ke) budget along with surface latent heat flux for the future and present-day simulations as described in sections 2c and 2d. To examine the changes between baroclinic conversion and surface latent heat flux associated with the transitioning ensemble-mean storm during and after ET, the averages of the baroclinic conversion and surface latent heat flux within a 1000-km radius relative to storm center are used as described in section 2c. There is an obvious contrast between ensemble-mean baroclinic conversion and surface latent heat flux, illustrating the increasing trend of the baroclinic conversion and the decreasing trend of the surface latent heat flux over time in general (Fig. 11). These changes imply that while the storm is intensified and maintained by the surface diabatic heating mechanism until ET completion, reintensification post-ET is driven by baroclinic energy conversion. Another noticeable feature is a rapid rebound of the surface latent heat flux during the period from 96 to 120 h (Fig. 11). Examination of storm-centered 10-m wind speed distribution reveals an expansion of the area of elevated wind speed during this period in both future and present-day simulations, leading to the secondary increase in area-averaged surface latent heat flux during the middle and late stages of post ET process (Fig. 12).
Time series of storm-centered composites of minimum sea level pressure (hPa; solid lines), spatially averaged surface latent heat flux (W m−2; dashed lines), and baroclinic conversion (W m−2; dotted lines) for the future experiments (red lines) and present-day experiments (blue lines). The blue and red vertical lines represent the time of ET onset and ET completion as shown in Figs. 10b and 10c. The future minimum SLP is normalized by the background SLP change. The baroclinic conversion and surface latent heat flux are computed within a 1000-km radius relative to storm center.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
Storm-centered composite of 10-m wind speed derived from outer-domain (shaded as in legend; 1 m s−1 interval) for the (a),(c),(e) present-day ensemble-mean simulation and (b),(d),(f) future counterpart. At simulation hour (a),(b) 84 and (c),(d) 108. (e),(f) Differences in 10-m wind speed (shaded; 2 m s−1 interval) are shown between 108 and 84 h. Distance from the storm center in degrees is shown on the ordinate and abscissa with (0°, 0°) marking the storm center.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
The ensemble-mean surface latent heat flux for the future storm is ~19 W m−2 stronger with a maximum of ~23 W m−2 stronger than in the present-day ensemble mean through the time of ET completion, consistent with a more intense future storm from the beginning of the simulations until ET completion. On the other hand, the future ensemble-mean baroclinic conversion that is a crucial driver for the reintensification during the post-ET stage is ~8 W m−2 weaker on average, with a maximum of ~19 W m−2 weaker than the present-day storm during the reintensification process. A considerable reduction in baroclinic conversion associated with the future storm is seen mostly during the mid- and late stages of the post-ET process (simulation hours 84–126). This is consistent with the future storm undergoing less reintensification compared to the present-day system (Fig. 11). Differences in intensity between the future and present-day storms during the post-ET process is a trade-off between greater surface latent heat flux and weaker baroclinicity. A possible explanation for why the future storm has weaker baroclinic conversion during the post-ET stage is reductions in lower-tropospheric baroclinicity due to polar amplification (Fig. 2e), which could mitigate future reintensification owing to a weakened future meridional temperature gradient. In addition, a reduction of the available potential energy (APE) reservoir due to weaker production via a reduced meridional heat flux would hamper the conversion of APE to EAPE, and thus weaken the future reintensification process.
b. Precipitation
Several future projection studies with high-resolution models demonstrate that while simulated TCs in a warmed environment produce greater precipitation over the present-day systems, the greatest increase in precipitation rate is seen in the inner-core regions (e.g., 100-km radius or less) and the rate of increase diminishes with radius (e.g., Knutson and Tuleya 2004; Hill and Lackmann 2011). Inside a 100-km radius, precipitation increases exceed the rate of water vapor increase dictated by Clausius–Clapeyron relation. At larger radii (e.g., 200–500 km), the rate of precipitation increase was shown to be close to that anticipated from the Clausius–Clapeyron relation. This distribution of precipitation change is due to dynamical changes in response to warming (Knutson et al. 2013; Liu et al. 2020).
Here, we assess storm-scale changes in precipitation at various points in the ET life cycle (i.e., TC, ET, and post-ET), via temporal and spatial averaging of ensemble-mean precipitation (Table 1). The present-day storm produces less precipitation than its future counterpart over the entire simulation period, meeting expectations from prior studies. Specifically, the 3-hourly ensemble-mean precipitation rate of the storm increases by as little as ~23% at the averaging radius of 500 km during the TC phase and as much as ~50% at the averaging radius of 100 km during the ET and post-ET phases under a warmed environment (Table 1). In addition, an outward shift of the heaviest precipitation rate occurs during the life cycles of both future and present-day simulations, consistent with an expanded, asymmetric wind field (Table 1). While the heaviest precipitation rates are seen inside a radius of 100 km during the TC phase [~16.2 mm (3 h)−1 for the present, ~22.6 mm (3 h)−1 for the future), the heaviest precipitation rates shift to a radius of 300 km as the storm undergoes ET in both future and present-day simulations [~4.2 mm (3 h)−1 for the present, ~6.5 mm (3 h)−1 for the future]. For the future changes in precipitation at various radii, ensemble-mean precipitation increases at a super Clausius–Clapeyron rate in the inner- and midcore regions (i.e., within a 300-km radius). In the outer-core regions, precipitation increases at or below the rate of water vapor increase dictated by Clausius–Clapeyron relation (Table 1).
Temporally and spatially averaged changes and standard deviation (numbers following the
Jung and Lackmann (2019) demonstrated that a simulated increase in precipitation associated with the transitioning Hurricane Irene (2011) in a warmed environment was largely attributable to an increase in moisture convergence associated with increased water vapor in the atmosphere and a strengthened secondary circulation, linked in turn to an increase in storm intensity. These results are consistent with earlier studies (Braun 2006; Trenberth et al. 2007; Liu et al. 2020). Here, we employed the methods of Trenberth et al. (2007) and computed column-integrated moisture convergence along with mass convergence and specific humidity to examine dynamical and thermodynamic process associated with the transitioning TC (Table 2). Vertically integrated moisture convergence and specific humidity increase in the future simulations, regardless of ET stage and radius. Mass convergence, on the other hand, shows almost zero or negative changes in the outer-core regions (i.e., 500-km radius). This implies that the enhanced secondary circulation, associated with the intensified minimum SLP discussed in the previous section, transports more water vapor to inner-core regions via increased radial inflow, consistent with super Clausius–Clapeyron precipitation increases there.
Temporally and spatially averaged difference between the future and present-day simulations with regard to phase and radius for column-integrated moisture convergence (Moist conv), mass convergence (Mass conv), and specific humidity (q). Column integration was taken for moisture convergence and mass convergence between 925 and 100 hPa and for specific humidity between 925 and 700 hPa.
c. Downstream development
The present-day ensemble-mean TC intensifies to reach a minimum SLP of 976 hPa during the early stages of the ET process following a predominantly northwesterly track over the North Atlantic (Figs. 10a,c). As the ensemble-mean storm moves poleward, it approaches the inflection point between an upper-level trough and a downstream ridge of moderate amplitude (Fig. 13). Temperature advection and outflow associated with the TC likely contribute to an amplification and modification of the midlatitude Rossby wave train downstream of the TC (Figs. 13a,c,e). The future ensemble-mean TC exhibited a similar track to its present-day counterpart until the early stages of post-ET, as discussed in section 4a. Owing to the presence of the limited area domain in these simulations, the wave pattern near the boundaries is constrained by boundary conditions, and thus is not able to freely deviate as much as would be the case in a global model simulation.
500-hPa geopotential heights (contours; 60-m interval) and sea level pressure (shaded as in legend at right) at (a) 51, (c) 90, and (e) 120 h for the present-day ensemble. (b),(d),(f) The corresponding fields for the future ensemble. The location of surface minimum pressure center is indicated with a red dot or a blue dot. The future 500-hPa heights are normalized by subtracting the height difference between the future and present-day at the initial time. Coastlines are shown for reference.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
Careful inspection of Figs. 13e and 13f reveals greater amplification of the downstream wave pattern in the future simulation. Previous studies have examined the influence of ET on midlatitude wave packets via an eddy kinetic energy framework, demonstrating that transitioning TCs act as a source of Ke through latent heat release, contributing to wave amplification and energy dispersion (e.g., Harr and Dea 2009; Keller et al. 2014; Quinting and Jones 2016; Keller et al. 2019). In light of these previous studies, we hypothesize that the future moisture and precipitation increases documented in section 4b, and the associated increase in latent heating, could generate additional Ke and energy dispersion, supporting greater amplification of the wave pattern downstream of the future storm.
To test our hypothesis, we apply the eddy kinetic energy budget analysis described in section 2d with Hovmöller diagrams (averaged between 30° and 60°N), and display the ageostrophic geopotential flux divergence and baroclinic conversion. The divergence of the Ke fluxes with the total wind in Eq. (2) is associated with advection of Ke rather than downstream energy dispersion. Therefore, we focus on the ageostrophic geopotential flux divergence and baroclinic conversion terms to examine the downstream development in this study. Hovmöller diagrams are not only an effective way to identify the time evolution of Rossby wave packets, but they allow for an efficient comparison the relationship the surface storm may have with wave packets (Decker and Martin 2005; Keller 2012; Keller et al. 2014).
The amplification of the downstream wave train is measured using the amplitude of its associated Ke maxima in this study. The interaction between the transitioning storm and vertically averaged meridional wind component, representing the midlatitude wave train, is clearly evident over time in Figs. 14a and 14b. Both the future and present-day ensemble-mean storms moved ahead of a preexisting upper trough, maintaining their wave-relative location until the end of the simulations. Regions of strong wind speed appear as Ke maxima, with larger Ke maxima consistent with greater future storm intensity (Figs. 14a,b).
Hovmöller diagrams for column-integrated Ke budget terms derived from ensemble-mean data. (a),(c),(e) Present-day and (b),(d),(f) future, averaged between 30° and 60°N. (a),(b) Column-integrated Ke (shaded; 2 × 105 J m−2 interval) and vertically averaged meridional wind (contours; southerly in warm, northerly in cool colors, 2 m s−1 interval). (c),(d) Column-integrated Ke (shaded; 2 × 105 J m−2 interval) and baroclinic conversion (contours; 5 W m−2 interval from 5 W m−2). (e),(f) Column-integrated Ke (shaded; 2 × 105 J m−2 interval) and divergence of ageostrophic geopotential flux (contours; divergence in cool colors, 25, 35, 45, 55, and 65 W m−2). Surface position of TC marked as black dots.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
Baroclinic energy conversion associated with the ET events is mostly concentrated in the poleward quadrant of the storm, overlapping with regions of individual Ke maxima after the onset of ET (Figs. 14c,d). This would imply that the individual Ke maxima in both simulations are partially created by conversion of EAPE to Ke through a baroclinic conversion process associated with warm air rising along the frontogenesis region (Figs. 6h,k; i.e., poleward quadrant) of the transitioning storm over the baroclinic zone. Hsieh and Cook (2008) showed that baroclinic conversion of EAPE to Ke was directly compensated by the diabatic generation and conversion from APE to EAPE via changes in horizontal temperature gradient. Likewise, heavier precipitation and stronger latent heat release associated with the intensified future storm generates additional Ke in the vicinity of the transitioning TC through baroclinic conversion.
Divergence and convergence of the ageostrophic geopotential fluxes disperse Ke into downstream regions in both future and present-day simulations (Figs. 14e,f). While the initial Ke maximum continues to weaken in response to diverging ageostrophic geopotential flux, the individual Ke maxima grow in the region of flux convergence downstream (Keller et al. 2014 and references therein). The greater Ke maximum adjacent to the future storm, partly generated by enhanced baroclinic conversion during the first 3.5 days of the model simulation until the early stage of the post-ET process, and intensified ageostrophic geopotential flux convergence and divergence, result in stronger amplification of the downstream Ke maxima, and in turn, the wave train. This is especially evident near the rear flank of downstream trough between 7° and 20°E during the late stages of the simulation (Figs. 14e,f).
Although geopotential height anomaly at a single level does not represent a complete picture of downstream development, this allows us to estimate the development of the upper-level downstream wave train. Maxima of 300-hPa geopotential height anomaly overlap maxima of both Ke and ageostrophic geopotential flux divergence, indicating that Ke dispersion plays an important role in modifying the downstream ridge and trough in both future and present-day simulations (Figs. 14 and 15). While the geopotential height anomaly in the ridge immediately downstream of the transitioning system is actually weaker in the future system during the most of the ET and post-ET processes, it ends up being slightly stronger after 120 h. In addition, greater downstream trough amplification is found in the future as well (Fig. 15); the maximum strength of the downstream negative 300-hPa height anomaly decreases by more than 40 m by late in the simulation, consistent with our eddy kinetic energy budget analysis.
Hovmöller diagrams of geopotential anomalies at 300 hPa (contours; 10-m interval) for the (a) present-day ensemble and (b) future ensemble (averaged between 40° and 60°N). Surface position of TC marked as red dots for future and blue dots for present-day.
Citation: Journal of Climate 34, 11; 10.1175/JCLI-D-20-0543.1
In addition to the results from prior studies that examined future downstream ridge development (e.g., Jung and Lackmann 2019; Michaelis and Lackmann 2019), our simulations address the impact of the transitioning TC on the region further downstream, where potential high-impact weather events could occur through midlatitude baroclinic development. Despite limits imposed by lateral boundary conditions, this result is suggestive, and should be examined using global model simulations in future research.
5. Summary and conclusions
In this study, we extend previous research on the response of TCs undergoing ET to climate change in the North Atlantic basin using small six-member ensembles of convection-allowing, quasi-idealized simulations with varying initial and lateral boundary conditions. The high-resolution simulations are initialized from composites of a set of 15 randomly selected recurving, oceanic (RCO) ET cases, drawn from a 21-case sample. The simulation domain is oceanic. Future simulations are based on a pseudo–global warming method, using thermodynamic changes from an average of 20 CMIP5 RCP8.5 projections. To address difficulties with traditional ET identification methods when using high-resolution model simulations, we present a new, experimental method of ET completion based on energy sources.
The simulated present-day event evolves in a realistic fashion consistent with many prior ET studies, but without synoptic- and mesoscale complexities. The simplified domain and composite-based initial and lateral boundary conditions represent a “generic” North Atlantic ET event, and thus may yield results that are more representative than studies based on a single event, such as Jung and Lackmann (2019). Key findings include:
Despite the future environmental conditions supporting a shift toward slower transition, such as a reduction in meridional SST gradient and a reduced vertical wind shear, no significant changes in the duration of ET are found between present-day and future simulations.
The future ET event is substantially stronger during both the TC and ET phases. After ET completion, both the future and present-day storms undergo reintensification. During this stage, the future storm experiences less reintensification than its present-day counterpart.
A considerable reduction in baroclinic conversion, which is a crucial driver for the reintensification process, is evident during the mid- and late stages of the future post-ET process (simulation hours 84–126). This results in reduced reintensification of the future storm, consistent with weakened large-scale baroclinicity.
Arctic amplification could contribute to the reduction in large-scale baroclinicity, as the temperature change fields result in a weakening of the lower-tropospheric meridional temperature gradient in the vicinity of the ET event.
The future storm produces considerably heavier precipitation regardless of ET phase and radius. Specifically, the 3-hourly ensemble-mean precipitation rate increases by as little as ~23% and as much as ~50%, depending on ET phase and radius.
An outward shift of the heaviest precipitation from an average radius of 100–300 km is evident in both future and present-day simulations as the storms undergo ET. The future simulation exhibits a super Clausius–Clapeyron precipitation increase within a 300-km radius of the storm center, with increases at or below the Clausius–Clapeyron rate in the outer-core regions (>500-km radius).
An eddy kinetic energy budget analysis and Hovmöller diagrams demonstrate that the future transitioning storm creates greater additional Ke in the vicinity of the storm through intensified baroclinic conversion during the first 3.5 days of the model simulation. The larger Ke is dispersed into the downstream wave pattern via the ageostrophic geopotential flux, leading to stronger amplification of the downstream Ke maxima in the future simulation. Our limited-area simulation domain likely mutes this effect to some degree.
These results imply that future North Atlantic TCs undergoing ET could potentially pose a greater threat of extreme weather conditions to western Europe through both direct and remote processes. A surprising result is the reduced reintensification of the future ET system. Despite greater strength at earlier stages of development, reduced baroclinicity in the future North Atlantic environment may mitigate impacts to some extent. However, heavier precipitation and a stronger pre-ET system suggest an overall increase in storm impacts.
There are several limitations to our current study. These include the use of lateral boundary conditions in limited-area simulations, which constrain the model solution, especially near the edges of our domain. The compositing strategy used for initial conditions smooths out synoptic and mesoscale features, which is both advantageous but also less realistic. The climate change delta fields we apply are from a single emission scenario, and only account for changes due to increased anthropogenic greenhouse gases. The PGW method assumes that synoptic and SST patterns of today could plausibly repeat in the future. However, these limitations are for the most part consistent between future and present-day simulations, allowing us to consider any changes found in the future to be meaningful. The PGW method also precludes examination of the role of large-scale circulation changes on the ET process. Additional studies, using alternate experimental designs, are required to fully address the question of how ET will respond to climate change.
Acknowledgments
We acknowledge National Science Foundation (NSF) Grant AGS-1546743, awarded to North Carolina State University (NCSU). The ERA5 data were obtained from Climate Data Store (CDS; available through https://cds.climate.copernicus.eu/cdsapp#!/home). The WRF Model was made available by NCAR, which is sponsored by NSF. High-Performance Computing support from Henry2 cluster is provided by NCSU’s Office of Information and Technology. Figures for this paper were created using NCL software package from CISL at NCAR. We thank the Program for Climate Model Diagnosis and Intercomparison (PCMDI) for collecting and archiving the CMIP5 model output. Special thanks to Dr. Walter Robinson for providing insightful and constructive comments on an earlier version of this manuscript.
Data availability statement
Model output from the simulations presented in this manuscript is stored on the NCSU Henry2 Cluster. Please contact the corresponding author for details on access to these data. The ERA-5 and HURDAT2 datasets are available as discussed in the body of the paper.
REFERENCES
Allen, M. R., and W. J. Ingram, 2002: Constraints on future changes in climate and the hydrologic cycle. Nature, 419, 228–232, https://doi.org/10.1038/nature01092.
Ancell, B. C., A. Bogusz, M. J. Lauridsen, and C. J. Nauert, 2018: Seeding chaos: The dire consequences of numerical noise in NWP perturbation experiments. Bull. Amer. Meteor. Soc., 99, 615–628, https://doi.org/10.1175/BAMS-D-17-0129.1.
Archambault, H. M., D. Keyser, L. F. Bosart, C. A. Davis, and J. M. Cordeira, 2015: A composite perspective of the extratropical flow response to recurving western North Pacific tropical cyclones. Mon. Wea. Rev., 143, 1122–1141, https://doi.org/10.1175/MWR-D-14-00270.1.
Atallah, E. H., and L. F. Bosart, 2003: The extratropical transition and precipitation distribution of Hurricane Floyd (1999). Mon. Wea. Rev., 131, 1063–1081, https://doi.org/10.1175/1520-0493(2003)131<1063:TETAPD>2.0.CO;2.
Avila, L. A., and J. Cangialosi, 2011: Tropical cyclone report: Hurricane Irene (21–28 August 2011). NHC Tech. Rep. AL092011, 45 pp., http://www.nhc.noaa.gov/data/tcr/AL092011_Irene.pdf.
Baatsen, M., R. J. Haarsma, A. J. Van Delden, and H. De Vries, 2015: Severe autumn storms in future western Europe with a warmer Atlantic Ocean. Climate Dyn., 45, 949–964, https://doi.org/10.1007/s00382-014-2329-8.
Bieli, M., A. H. Sobel, S. J. Carmargo, H. Murakami, and G. A. Vecchi, 2020: Application of the cyclone phase space to extratropical transition in a global climate model. J. Adv. Model. Earth Syst., 12, e2019MS001878, https://doi.org/10.1029/2019MS001878.
Bove, M. C., J. B. Elsner, C. W. Landsea, X. Niu, and J. J. O’Brien, 1998: Effect of El Niño on U.S. landfalling hurricanes, revisited. Bull. Amer. Meteor. Soc., 79, 2477–2482, https://doi.org/10.1175/1520-0477(1998)079<2477:EOENOO>2.0.CO;2.
Braun, S. A., 2006: High-resolution simulation of Hurricane Bonnie (1998). Part II: Water budget. J. Atmos. Sci., 63, 43–64, https://doi.org/10.1175/JAS3609.1.
Colbert, A. J., and B. J. Soden, 2012: Climatological variations in North Atlantic tropical cyclone tracks. J. Climate, 25, 657–673, https://doi.org/10.1175/JCLI-D-11-00034.1.
Davis, C. A., 2018: Resolving tropical cyclone intensity in models. Geophys. Res. Lett., 45, 2082–2087, https://doi.org/10.1002/2017GL076966.
Decker, S. G., and J. E. Martin, 2005: A local energetics analysis of the life cycle differences between consecutive, explosively deepening continental cyclones. Mon. Wea. Rev., 133, 295–316, https://doi.org/10.1175/MWR-2860.1.
ECMWF, 2017: ERA5 Reanalysis. National Center for Atmospheric Research, Computational and Information Systems Laboratory, accessed 1 June 2018, https://doi.org/10.5065/D6X34W69.
Elsner, J. B., K. Liu, and B. Kocher, 2000: Spatial variations in major U.S. hurricane activity: Statistics and a physical mechanism. J. Climate, 13, 2293–2305, https://doi.org/10.1175/1520-0442(2000)013<2293:SVIMUS>2.0.CO;2.
Evans, C., and R. E. Hart, 2008: Analysis of the wind field evolution associated with the extratropical transition of Bonnie (1998). Mon. Wea. Rev., 136, 2047–2065, https://doi.org/10.1175/2007MWR2051.1.
Evans, C., and Coauthors, 2017: The extratropical transition of tropical cyclones. Part I: Cyclone evolution and direct impacts. Mon. Wea. Rev., 145, 4317–4344, https://doi.org/10.1175/MWR-D-17-0027.1.
Evans, J. L., and R. E. Hart, 2003: Objective indicators of the life cycle evolution of extratropical transition for Atlantic tropical cyclones. Mon. Wea. Rev., 131, 909–925, https://doi.org/10.1175/1520-0493(2003)131<0909:OIOTLC>2.0.CO;2.
Gentry, M. S., and G. M. Lackmann, 2010: Sensitivity of simulated tropical cyclone structure and intensity to horizontal resolution. Mon. Wea. Rev., 138, 688–704, https://doi.org/10.1175/2009MWR2976.1.
Haarsma, R. J., W. Hazeleger, C. Severijns, H. Vries, A. Sterl, R. Bintanja, G. J. van Oldenborgh, and H. W. van den Brink, 2013: More hurricanes to hit western Europe due to global warming. Geophys. Res. Lett., 40, 1783–1788, https://doi.org/10.1002/grl.50360.
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, https://doi.org/10.3178/hrl.2.61.
Harr, P. A., and R. L. Elsberry, 2000: Extratropical transition of tropical cyclones over the western North Pacific. Part I: Evolution of structural characteristics during the transition process. Mon. Wea. Rev., 128, 2613–2633, https://doi.org/10.1175/1520-0493(2000)128<2613:ETOTCO>2.0.CO;2.
Harr, P. A., and J. M. Dea, 2009: Downstream development associated with the extratropical transition of tropical cyclones over the western North Pacific. Mon. Wea. Rev., 137, 1295–1319, https://doi.org/10.1175/2008MWR2558.1.
Hart, R. E., 2003: A cyclone phase space derived from thermal wind and thermal asymmetry. Mon. Wea. Rev., 131, 585–616, https://doi.org/10.1175/1520-0493(2003)131<0585:ACPSDF>2.0.CO;2.
Hart, R. E., J. L. Evans, and C. Evans, 2006: Synoptic composites of the extratropical transition life cycle of North Atlantic tropical cyclones: Factors determining post transition evolution. Mon. Wea. Rev., 134, 553–578, https://doi.org/10.1175/MWR3082.1.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Hill, K. A., and G. M. Lackmann, 2011: The impact of future climate change on TC intensity and structure: A downscaling approach. J. Climate, 24, 4644–4661, https://doi.org/10.1175/2011JCLI3761.1.
Hsieh, J.-S., and K. H. Cook, 2008: On the instability of the African easterly jet and the generation of African waves: Reversals of the potential vorticity gradient. J. Atmos. Sci., 65, 2130–2151, https://doi.org/10.1175/2007JAS2552.1.
Jones, S. C., and Coauthors, 2003: The extratropical transition of tropical cyclones: Forecast challenges, current understanding, and future directions. Wea. Forecasting, 18, 1052–1092, https://doi.org/10.1175/1520-0434(2003)018<1052:TETOTC>2.0.CO;2.
Jung, C., and G. M. Lackmann, 2019: Extratropical transition of Hurricane Irene (2011) in a changing climate. J. Climate, 32, 4847–4871, https://doi.org/10.1175/JCLI-D-18-0558.1.
Kain, J., 2004: The Kain–Fritsch convective parameterization: An update. J. Appl. Meteor., 43, 170–181, https://doi.org/10.1175/1520-0450(2004)043<0170:TKCPAU>2.0.CO;2.
Keller, J. H., 2012: Diagnosing the downstream impact of extratropical transition using multimodel operational ensemble prediction systems. Ph.D. thesis, Karlsruhe Institute of Technology, 172 pp.
Keller, J. H., 2017: Amplification of the downstream wave train during extratropical transition: Sensitivity studies. Mon. Wea. Rev., 145, 1529–1548, https://doi.org/10.1175/MWR-D-16-0193.1.
Keller, J. H., S. C. Jones, and P. A. Harr, 2014: An eddy kinetic energy view of physical and dynamical processes in distinct forecast scenarios for the extratropical transition of two tropical cyclones. Mon. Wea. Rev., 142, 2751–2771, https://doi.org/10.1175/MWR-D-13-00219.1.
Keller, J. H., and Coauthors, 2019: The extratropical transition of tropical cyclones. Part II: Interaction with the midlatitude flow, downstream impacts, and implications for predictability. Mon. Wea. Rev., 147, 1077–1106, https://doi.org/10.1175/MWR-D-17-0329.1.
Klein, P. M., P. A. Harr, and R. L. Elsberry, 2000: Extratropical transition of western North Pacific tropical cyclones: An overview and conceptual model of the transformation stage. Wea. Forecasting, 15, 373–395, https://doi.org/10.1175/1520-0434(2000)015<0373:ETOWNP>2.0.CO;2.
Knutson, T. R., and R. E. Tuleya, 2004: Impact of CO2-induced warming on simulated hurricane intensity and precipitation: Sensitivity to the choice of climate model and convective parameterization. J. Climate, 17, 3477–3495, https://doi.org/10.1175/1520-0442(2004)017<3477:IOCWOS>2.0.CO;2.
Knutson, T. R., and Coauthors, 2013: Dynamical downscaling projections of twenty-first century Atlantic hurricane activity: CMIP3 and CMIP5 model-based scenarios. J. Climate, 26, 6591–6617, https://doi.org/10.1175/JCLI-D-12-00539.1.
Kofron, D. E., E. A. Ritchie, and J. S. Tyo, 2010: Determination of a consistent time for the extratropical transition of tropical cyclones. Part I: Examination of existing methods for finding “ET time.” Mon. Wea. Rev., 138, 4328–4343, https://doi.org/10.1175/2010MWR3180.1.
Kossin, J. P., S. J. Camargo, and M. Sitkowski, 2010: Climate modulation of North Atlantic hurricane tracks. J. Climate, 23, 3057–3076, https://doi.org/10.1175/2010JCLI3497.1.
Kossin, J. P., K. A. Emanuel, and G. A. Vecchi, 2014: The poleward migration of the location of tropical cyclone maximum intensity. Nature, 509, 349–352, https://doi.org/10.1038/nature13278.
Lackmann, G. M., 2015: Hurricane Sandy before 1900, and after 2100. Bull. Amer. Meteor. Soc., 96, 547–560, https://doi.org/10.1175/BAMS-D-14-00123.1.
Landsea, C. W., and J. L. Franklin, 2013: Atlantic hurricane database uncertainty and presentation of a new database format. Mon. Wea. Rev., 141, 3576–3592, https://doi.org/10.1175/MWR-D-12-00254.1.
Liu, M., and J. A. Smith, 2016: Extreme rainfall from landfalling tropical cyclones in the eastern United States: Hurricane Irene (2011). J. Hydrometeor., 17, 2883–2904, https://doi.org/10.1175/JHM-D-16-0072.1.
Liu, M., G. A. Vecchi, J. A. Smith, and H. Murakami, 2017: The present-day simulation and twenty-first-century projection of the climatology of extratropical transition in the North Atlantic. J. Climate, 30, 2739–2756, https://doi.org/10.1175/JCLI-D-16-0352.1.
Liu, M., L. Yang, J. A. Smith, and G. A. Vecchi, 2020: Response of extreme rainfall for landfalling tropical cyclones undergoing extratropical transition to projected climate change: Hurricane Irene (2011). Earth’s Future, 8, e2019EF001360, https://doi.org/10.1029/2019EF001360.
Lynch, P., and X.-Y. Huang, 1992: Initialization of the HIRLAM model using a digital filter. Mon. Wea. Rev., 120, 1019–1034, https://doi.org/10.1175/1520-0493(1992)120<1019:IOTHMU>2.0.CO;2.
Lynch, P., and X.-Y. Huang, 1994: Diabatic initialization using recursive filters. Tellus, 46A, 583–597, https://doi.org/10.3402/tellusa.v46i5.15646.
Mallard, M. S., G. M. Lackmann, A. Aiyyer, and K. A. Hill, 2013: Atlantic hurricanes and climate change. Part I: Experimental design and isolation of thermodynamic effects. J. Climate, 26, 4876–4893, https://doi.org/10.1175/JCLI-D-12-00182.1.
Meinshausen, M., and Coauthors, 2011: The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Climatic Change, 109, 213–241, https://doi.org/10.1007/s10584-011-0156-z.
Michaelis, A. C., 2019: Climate change effects on the extratropical transition of tropical cyclones in high-resolution global simulations. Ph.D. dissertation, North Carolina State University, 156 pp.
Michaelis, A. C., and G. M. Lackmann, 2019: Climatological changes in the extratropical transition of tropical cyclones in high-resolution global simulations. J. Climate, 32, 8733–8753, https://doi.org/10.1175/JCLI-D-19-0259.1.
Michaelis, A. C., and G. M. Lackmann, 2021: Storm-scale dynamical changes of extratropical transition events in present-day and future high-resolution global simulations. J. Climate, https://doi.org/10.1175/JCLI-D-20-0472, in press.
Neale, R. B., and B. J. Hoskins, 2000: A standard test for AGCMs including their physical parameterizations: I: The proposal. Atmos. Sci. Lett., 1, 101–107, https://doi.org/10.1006/asle.2000.0019.
Orlanski, I., and J. Katzfey, 1991: The life cycle of a cyclonic wave in the Southern Hemisphere. Part I: Eddy energy budget. J. Atmos. Sci., 48, 1972–1998, https://doi.org/10.1175/1520-0469(1991)048<1972:TLCOAC>2.0.CO;2.
Patricola, C. M., and M. F. Wehner, 2018: Anthropogenic influences on major tropical cyclone events. Nature, 563, 339–346, https://doi.org/10.1038/s41586-018-0673-2.
Pohorsky, R., M. Röthlisberger, C. M. Grams, J. Riboldi, and O. Martius, 2019: The climatological impact of recurving North Atlantic tropical cyclones on downstream extreme precipitation events. Mon. Wea. Rev., 147, 1513–1532, https://doi.org/10.1175/MWR-D-18-0195.1.
Pollard, R. T., P. B. Rhines, and R. O. Thompson, 1973: The deepening of the wind-mixed layer. Geophys. Fluid Dyn., 4, 381–404, https://doi.org/10.1080/03091927208236105.
Quinting, J. F., and S. C. Jones, 2016: On the impact of tropical cyclones on Rossby wave packets: A climatological perspective. Mon. Wea. Rev., 144, 2021–2048, https://doi.org/10.1175/MWR-D-14-00298.1.
Reed, K. A., A. M. Stansfield, M. F. Wehner, and C. M. Zarzycki, 2020: Forecasted attribution of the human influence on Hurricane Florence. Sci. Adv., 6, eaaw9253, https://doi.org/10.1126/sciadv.aaw9253.
Ritchie, E. A., and R. L. Elsberry, 2003: Simulations of the extratropical transition of tropical cyclones: Contributions by the midlatitude upper-level trough to reintensification. Mon. Wea. Rev., 131, 2112–2128, https://doi.org/10.1175/1520-0493(2003)131<2112:SOTETO>2.0.CO;2.
Ritchie, E. A., and R. L. Elsberry, 2007: Simulations of the extratropical transition of tropical cyclones: Phasing between the upper-level trough and tropical cyclones. Mon. Wea. Rev., 135, 862–876, https://doi.org/10.1175/MWR3303.1.
Schade, L. R., and K. A. Emanuel, 1999: The ocean’s effect on the intensity of tropical cyclones: Results from a simple coupled atmosphere–ocean model. J. Atmos. Sci., 56, 642–651, https://doi.org/10.1175/1520-0469(1999)056<0642:TOSEOT>2.0.CO;2.
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, https://doi.org/10.1029/96GL00265.
Sharmila, S., and K. J. E. Walsh, 2018: Recent poleward shift of tropical cyclone formation linked to Hadley cell expansion. Nat. Climate Change, 8, 730–736, https://doi.org/10.1038/s41558-018-0227-5.
Skamarock, W. C., and Coauthors, 2019: A description of the Advanced Research WRF Model version 4. NCAR Tech. Note NCAR/TN-556+STR, 145 pp., https://doi.org/10.5065/1dfh-6p97.
Soden, B. J., D. L. Jackson, V. Ramaswamy, D. Schwarzkopf, and X. Huang, 2005: The radiative signature upper tropospheric moistening. Science, 310, 841–844, https://doi.org/10.1126/science.1115602.
Steinfeld, D., and S. Pfahl, 2019: The role of latent heating in atmospheric blocking dynamics: A global climatology. Climate Dyn., 53, 6159–6180, https://doi.org/10.1007/s00382-019-04919-6.
Studholme, J., K. I. Hodges, and C. M. Brierley, 2015: Objective determination of the extratropical transition of tropical cyclones in the Northern Hemisphere. Tellus, 67A, 24474, https://doi.org/10.3402/tellusa.v67.24474.
Trenberth, K. E., L. Smith, T. Qian, A. Dai, and J. Fasullo, 2007: Estimates of the global water budget and its annual cycle using observational and model data. J. Hydrometeor., 8, 758–769, https://doi.org/10.1175/JHM600.1.
Wirth, V., M. Riemer, E. K. M. Chang, and O. Martius, 2018: Rossby wave packets on the midlatitude waveguide—A review. Mon. Wea. Rev., 146, 1965–2001, https://doi.org/10.1175/MWR-D-16-0483.1.
Xie, L., T. Yan, L. Pietrafesa, J. Morrison, and T. Karl, 2005: Climatology and interannual variability of North Atlantic hurricane tracks. J. Climate, 18, 5370–5381, https://doi.org/10.1175/JCLI3560.1.
Zarzycki, C. M., D. R. Thatcher, and C. Jablonowski, 2017: Objective tropical cyclone extratropical transition detection in high-resolution reanalysis and climate model data. J. Adv. Model. Earth Syst., 9, 130–148, https://doi.org/10.1002/2016MS000775.
We retained the default WRF mixed layer model namelist settings, including an initial mixed layer depth of 50 m.
