The Oxford Research Encyclopedia of Natural Hazard Science will be available via subscription on April 26. Visit About to learn more, meet the editorial board, learn about subscriber services.

Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, NATURAL HAZARD SCIENCE ( (c) Oxford University Press USA, 2019. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 22 April 2019

Modeling Tropical Cyclones in a Changing Climate

Summary and Keywords

Tropical cyclones (TCs) in their most intense expression (hurricanes or typhoons) are the main natural hazards known to humankind. The impressive socioeconomic consequences for countries dealing with TCs make our ability to model these organized convective structures a key issue to better understanding their nature and their interaction with the climate system. The destructive effects of TCs are mainly caused by three factors: strong wind, storm surge, and extreme precipitation. These TC-induced effects contribute to the annual worldwide damage of the order of billions of dollars and a death toll of thousands of people. Together with the development of tools able to simulate TCs, an accurate estimate of the impact of global warming on TC activity is thus not only of academic interest but also has important implications from a societal and economic point of view. The aim of this article is to provide a description of the TC modeling implementations available to investigate present and future climate scenarios.

The two main approaches to dynamically model TCs under a climate perspective are through hurricane models and climate models. Both classes of models evaluate the numerical equations governing the climate system. A hurricane model is an objective tool, designed to simulate the behavior of a tropical cyclone representing the detailed time evolution of the vortex. Considering the global scale, a climate model can be an atmosphere (or ocean)-only general circulation model (GCM) or a fully coupled general circulation model (CGCM). To improve the ability of a climate model in representing small-scale features, instead of a general circulation model, a regional model (RM) can be used: this approach makes it possible to increase the spatial resolution, reducing the extension of the domain considered. In order to be able to represent the tropical cyclone structure, a climate model needs a sufficiently high horizontal resolution (of the order of tens of kilometers) leading to the usage of a great deal of computational power.

Both tools can be used to evaluate TC behavior under different climate conditions. The added value of a climate model is its ability to represent the interplay of TCs with the climate system, namely two-way relationships with both atmosphere and ocean dynamics and thermodynamics. In particular, CGCMs are able to take into account the well-known feedback between atmosphere and ocean components induced by TC activity and also the TC–related remote impacts on large-scale atmospheric circulation.

The science surrounding TCs has developed in parallel with the increasing complexity of the mentioned tools, both in terms of progress in explaining the physical processes involved and the increased availability of computational power. Many climate research groups around the world, dealing with such numerical models, continuously provide data sets to the scientific community, feeding this branch of climate change science.

Keywords: tropical cyclones, hurricane model, climate model, climate variability, climate change

Tropical Cyclone Interaction with the Climate System

Tropical cyclones (TCs) are low-pressure systems developing over warm waters of the tropical oceans and have organized convection and an associated intense cyclonic surface wind circulation. TCs get energy from the ocean through the evaporation flux, then from the associated condensation of water in upward air masses near their center. This energy release in the troposphere originates the well-known “warm core,” causing the extraordinary strong winds at the surface. These systems differ from mid-latitude cyclones, since the mechanism used to obtain energy is completely different. In fact, mid-latitude cyclones rely on the strong horizontal temperature gradients in the atmosphere as a source of energy and do not have any warm core within their vertical structure (Holland, 1993). The very strong winds associated with TCs near the surface cause damage to coastal regions through extreme wind, intense precipitation, and storm surge. On the other hand, these organized systems are important components of the earth’s climate, and their activity is modulated at different time scales by different climate forcings.

The annual number of TCs worldwide is approximately 90. The less intense events, called tropical depressions (when a 10-m sustained wind speed of 17.5 m/s is not reached), are not considered in this number, just tropical storms (wind speed greater than 17.5 m/s) and hurricanes/typhoons. The most intense TCs are classified following the Saffir–Simpson scale as “hurricanes” or “typhoons” depending on the location where they form: the term “hurricane” is used in the Atlantic and Northeast Pacific, “typhoon” is used in the Northwest Pacific, and “cyclone” is used in the Southern Hemisphere. Hurricane classes range from category 1 (at least 33 m/s) to category 5 (greater than 70 m/s).

The globally averaged annual variation of TC occurrence is of the order of 10%. It is well known that the temporal variability and the spatial distribution of TC count are modulated by natural phenomena such as the El Niño Southern Oscillation (ENSO; Pielke & Landsea, 1999), the Atlantic Multidecadal Oscillation (AMO; Goldenberg, Landsea, Mestas-Nunez, & Gray, 2001), the Atlantic Meridional Model (AMM; Kossin & Vimont, 2007), the Madden-Julian Oscillation (MJO; Klotzbach, 2014), and the North Atlantic Oscillation (NAO; Elsner et al., 2000). When investigating the TC count temporal variability at the regional scale, more or less pronounced values, compared to the global TC number, can be found. An important issue to be considered when studying the variability of TC count in the past is that different techniques, approaches, and dedication have been applied in collecting the different regional data sets, thus producing different basin-dependent biases (Nicholls, Landsea, & Gill, 1998). Figure 1 illustrates the observed TC tracks over the period 2005–2010 from the NOAA’s Tropical Prediction Center (Atlantic and eastern North Pacific tracks), and from the U.S. Navy’s Joint Typhoon Warning Center.

Modeling Tropical Cyclones in a Changing ClimateClick to view larger

Figure 1. Observed tropical cyclone tracks in the 2005–2010 period. (IBTrACS data from Knapp et al., 2010)

Notably, the major number of TCs occur in the Northern Hemisphere, mainly due to the absence of tropical storms over the South Atlantic basin and Southeastern Pacific basin. The absence of TCs in the 5S–5N latitude belt is due to the necessity of a not negligible Coriolis parameter for TC genesis.

In addition to a sufficiently high Coriolis parameter, the environmental conditions favorable for cyclogenesis are large values of low-level relative vorticity, high sea surface temperatures, high values of relative humidity in the lower and middle troposphere, weak vertical wind shear, and conditional instability in the deep atmospheric layer (Gray, 1979). Despite the mentioned conditions considered necessary but not sufficient for genesis to occur, a series of TC genesis parameters (GPs) have been developed to verify potential changes in the TC count under different climate conditions (Camargo et al., 2007). This approach makes it possible to use climate model outputs, even at coarse (monthly) time resolutions, to investigate spatial patterns of regions prone to TC formation.

Many studies have been done in the past to describe the effects that climate change may have on TC activity (Gualdi, Scoccimarro, & Navarra, 2008; Knutson, Landsea, & Emanuel, 2010). A range of techniques has been used to understand how TC behavior is affected by natural or anthropogenic climate change. The base approach is to quantify changes in TC activity due to past climate change, based on historical climate and tropical cyclones records, but this approach is limited by the short history of tropical storm records and their reliability when investigating the pre-satellite period (Emanuel, Sundararajan, & Williams, 2008; Landsea, 2007). The North Atlantic region is the one with the major number of TC-related information, due to the dense ship tracks since the pre-satellite era, leading to the existence of an official archive extending back to 1851 (Landsea et al., 2004). Despite no significant trends having been identified in observed TC activity in the Atlantic basin since the 19th century, significant trends in TC numbers and intensities have occurred in this basin over the past few decades. In addition, significant trends in other basins have been registered, but the relative confidence is low due to the lack of observations, with particular uncertainties over the East North Pacific and the whole Southern Hemisphere.

A theoretical basis for maximum TC intensity built on Emanuel (1987) is well established, but a climate theory of TC formation remains elusive (Walsh et al., 2015): the reason why about 90 TCs generate worldwide every year is still to be explained. In addition, a better understanding of the role of TCs in the mean climate, and vice versa, is needed.

There is evidence that TCs modulate meridional heat transport through their interplay with the oceans (Emanuel, 2001; Scoccimarro et al., 2011), affecting the climate system (Hu & Meehl, 2009). The interaction between TCs and the ocean is in fact a major mechanism responsible for energy exchange between the atmosphere and the ocean (Bueti et al., 2014). TCs affect the thermal and dynamical structure of the ocean through surface wind stresses: the wind structure associated with TCs is responsible for two important atmosphere–ocean feedbacks. The first positive feedback is driven by the latent heat associated with the increased evaporation rate and leads to an enhancement of the available energy for TC development. This first feedback is responsible for the auto-fueling mechanism behind the process. The second feedback—negative—is due to the cold water upwelling induced by the increased wind stress at the ocean surface and by the cooling of surface layers due to the shear-induced mixing at the base of the mixed layer (Sriver et al., 2010). The induced cooling of the sea surface leads to a weakening of the cyclone intensity due to the reduction of the total heat flux into the atmosphere.

This cooling effect is a function of the initial ocean condition, TC intensity, TC traveling speed, and TC size (Huang, Lin, Chou, & Huang, 2015). The mixing at the base of the mixed layer also provides heat to the subsurface, below the surface thermocline. In addition, after the storm passage, the sea surface temperature (SST) is restored to the climatological values within a few weeks, transferring additional heat to the atmospheric component. A potential role of the described feedbacks in modulating the ocean mean state highlights the importance of a good representation of the atmosphere–ocean interaction associated with TC activity.

The TC effects on the climate system, through their interplay with the ocean, are local and affect a wide area around the tropical cyclone track. On the other hand, there are also non-local effects linking the TCs activity to the large-scale atmospheric circulation (Scoccimarro, Gualdi, & Navarra, 2012) and affecting extratropical latitudes. This is the case of extreme precipitation events over Europe, linked to the moist air transport from the Atlantic basin, also fueled by intense hurricanes (Krichack, Barkan, Breitgand, Gualdi, & Feldstein, 2014).

The emerged picture of tropical cyclones as key players within the earth’s climate machine makes TC modeling a promising field to use for climate change studies.

Tropical Cyclone Modeling

Modeling tropical cyclones is important improve our knowledge of the physical processes involved in such events and to better understand their interplay with the climate system (Camargo & Wing, 2016). Also, a good representation of TC development within the climate system leads to an improvement of our ability to forecast their temporal evolution, in terms of both intensity and tracks, particularly important when investigating potential damage along the coastal regions. TC frequency and intensity are modulated by environmental factors, such as the large-scale circulation, vertical wind shear, atmospheric water content, and availability of energy at the surface boundary. Significant changes in these environmental factors, associated with different climate conditions (i.e., preindustrial conditions vs. present climate) have been demonstrated to alter TC activity in terms of both count and intensity (Camargo, Ting, & Kushnit, 2013; Emanuel, 2005; Gualdi et al., 2008; Knutson et al., 2010; Scoccimarro et al., 2011, 2014; Zhao, Held, Lin, & Vecchi, 2009). State of the art general circulation models start to be sufficiently resolved (order of tens of degrees both in latitude and longitudes) to realistically represent TCs, even in terms of intensity (Murakami et al., 2015). Also, such global dynamical models are already used by many research centers around the world (UK-MetOffice, ECMWF, GFDL, GFS), but the limitation posed by the computational cost is still a major issue. The scientific community is now looking forward to the next Coupled Model Intercomparison Project (CMIP) as a provider of significantly improved results compared to the ones already available (CMIP3—Meehl et al., 2007; CMIP5—Taylor, Stouffer, & Meehl, 2012), even on a climate perspective. These models take several hours to run a year of simulation on the world’s most advanced supercomputers. For this reason, the horizontal resolution adopted in recent climate modeling efforts (CMIP3, CMIP5) is not comparable to the one used in the short-term (days) forecast framework. The resolution-induced deficiency in representing TCs in CMIP3 and CMIP5 general circulation models prevents these tools from simulating major storms at their observed intensities and providing a reasonable distribution of intense storms. In order to overcome the low resolution of such general circulation models, downscaling strategies to connect the large-scale circulation to a smaller spatial scale have been developed. Three different downscaling technics are available: dynamical, statistical, and hybrid statistical-dynamical. Dynamical downscaling has been shown to be a very useful tool for understanding climate projections of TC activity and can increase the spatial resolution through regional climate models (e.g., Knutson, Sirutis, Garner, Held, & Tuleya, 2007) and hurricane models (Bender, Tuleya, Thomas, & Marchok, 2007; Emanuel, 1995). On the other hand, statistical downscaling (Elsner & Jagger, 2006) is much cheaper in terms of computational power since it is based on historical relationships between TC-specific information and the behavior of historical hurricanes. A mixed approach is also a viable tool to model TC activity. For example, in the seasonal prediction of the North Atlantic hurricane activity contest, Vecchi et al. (2011) used the Atlantic main development region SST and the global tropical-mean SST from high-resolution general circulation models as predictors in a Poisson-based regression.

In the following paragraphs the methodology developed in recent years to model TCs are described with a special focus on dynamical modeling only: hurricane models (specifically designed to simulate TCs) and climate models (both general circulation models and regional models). The TC modeling tools described hereafter are also used for seasonal and near-term predictions (Vecchi & Villarini, 2014; Vecchi et al., 2013), but in the present work we will use them to investigate long-term (interdecadal) projections only.

Hurricane Models

We use the term “hurricane model” to indicate a dynamical numerical code specifically designed to represent tropical cyclones. In addition to models devoted to the forecast of TC trajectories, represented by simplified dynamical track models (beta and advection models), there are local models able to represent the TC intensity evolution. This is the case of highly resolved hurricane intensity coupled models, formulated in angular momentum coordinates (Emanuel, 1995, 2004). In addition, there are more complex systems able to represent TC evolution in time and space over a wide (regional scale) area: Figure 2 shows a gridded computational mesh structure, where different grids are nested together with increasingly finer grid-point spacing in each mesh (up to three in the GFDL hurricane model; Bender et al., 2007).

Modeling Tropical Cyclones in a Changing ClimateClick to view larger

Figure 2. Representation of a modeling system based on nested computational meshes.

The atmospheric component of these hurricane models permits exceptionally high horizontal resolution in the eyewall, where it is needed, leading to a realistic simulation of the intensity evolution of the storm moving over the ocean. A basic assumption of the first class of models (Emanuel, 1995) is that the cyclone is axisymmetric, thus hurricane vortex is assumed to be steady and circularly symmetric about its axis of rotation and the airflow is in hydrostatic and gradient wind balance (Emanuel, 2004). The main advantage of these models, within the bunch of dynamical-like models, is its computational speed: a storm can be simulated in a few minutes on a typical computer workstation, thus no supercomputers are needed to use such tools to investigate the TC activity dependence on climate variability. The sea surface temperature seen by the TC in the coupled hurricane model can also vary in time and space, considering the coupling to the ocean model component. One of the limitations of axisymmetric models is that the direct influence from environmental wind shear, which is known to be a major factor inhibiting tropical cyclone intensification, is precluded (Emanuel, 2004) and needs to be parameterized. In recent years many improvements have been made in representing hurricane evolution formulation (Persing, Montgomery, McWilliams, & Smith, 2013) also highlighting the importance of three-dimensional models with respect to axisymmetric configurations. Three-dimensional models better represent the intensification problem, mainly related to the emerging differences in convective organization in the two formulations, particularly relevant at forecast timescales.

The axisymmetric model (Emanuel, 2004) has been optimized over the past 10 years demonstrating skillful hurricane intensity forecast ability (see also the coupled hurricane intensity prediction system developed at the National Center for Atmospheric Research, NCAR), and it is also used to investigate the effect of climate change on hurricane activity (Emanuel, 2013, 2015) based on multiple TC intensity simulations, initiated by random seeding in space and time. These simulations are driven by boundary conditions derived from climate model simulations over the different climate periods. The intensity model is integrated along each synthetic track, and the survivors represent the tropical cyclone climatology of the model for the corresponding climate period (defined by the climate model simulation). These random “seeds” are seeded everywhere and at all times, regardless of latitude, sea surface temperature, or other factors; the only limit regards the fact that storms are not allowed to form equatorward within a few degrees of latitude (Emanuel, 2013). Although the seeds are done everywhere, not all seeds survive, and the survival of each seed is indeed dependent on the environmental conditions at the place of the seeding. Furthermore, at each step the environmental conditions determine how long the surviving storms last and how intense they become.

As already mentioned another class of hurricane models is represented by the one developed for instance by the Geophysical Fluid Dynamics Laboratory (GFDL) hurricane prediction group, characterized by computational meshes nested together with increasingly finer grid-point spacing in each mesh, up to a few kilometers in the inner mesh (Bender et al., 2010). The inner mesh of the model is automatically relocated to follow the path of the investigated tropical cyclone. This system is coupled to an ocean model to take into account the TC dependence on the upper ocean as dynamically varying source of energy. The GFDL hurricane model was adopted by the U.S. National Weather Service in 2006, and an updated version of the same model (named GFDN) has been used by the U.S. Nav, since 2008. The differences between the two mentioned versions of the GFDL hurricane model can be found in Knutson et al. (2013).

Hurricane model downscaling is used to investigate climate change effects on tropical cyclone frequency and intensity (Emanuel, Sundararajan, & Williams, 2008; Emanuel, 2013; Knutson et al., 2010) as an alternative to the assessments based on general circulation models (GCMs) or regional climate models (RCMs) described in the following section.

Notably, hurricane models need boundary conditions can be obtained by GCMs, CGCMs, RCMs, or observational/reanalysis datasets.

Tropical Cyclone Representation in Climate Models

The theory behind the control of the global number and distribution of tropical cyclones worldwide under the current climate, and potential changes under different climate conditions, is still to be defined. To address this kind of investigation it is necessary to model as finely as possible not only the TC itself but also all the spatial and temporary scales associated with the physical processes of tropical cyclone genesis: high-resolution, fully coupled general circulation models are the best tool available to rely on.

The evolution of climate models started more than 40 years ago (Figure 3).

Modeling Tropical Cyclones in a Changing ClimateClick to view larger

Figure 3. Time evolution from the 1970s to present of climate models.

Starting from atmosphere-only general circulation models, fully coupled general circulation models, and even more complex systems (earth system models, ESMs) have been now designed, where also aerosols, carbon cycle, and atmospheric chemistry are considered. The time evolution of each component went together with the increase of the spatial resolution and the improvement of its coupling feasibility to the other components of the climate system.

As an example, Figure 4 shows a subsample of the land sea mask and the orography used by the atmospheric component of GCMs at different horizontal resolution, starting from 400 km (typical resolution used in the 90s), up to 25 km (frontier GCMs in a climate experiments perspective).

Modeling Tropical Cyclones in a Changing ClimateClick to view larger

Figure 4. Model orography in the Western Pacific at different horizontal resolutions: 400 km and 300 km in the upper panels, 200 km and 100 km in the central panel, 50 km and 25 km in the lower panel. White patterns indicate the ocean portion. Colorbar saturates over Tibetan Plateau.

The shown domain covers the Western Pacific basin. A certain portion of land is completely omitted (i.e., Taiwan) at resolution lower than 100 km, and also the model orography is strongly impacted: the maximum orography over Australia is almost doubled (and closer to the observations) after moving from 400 to 50 km. Bearing in mind that the evolution of the horizontal resolution involves also the ocean component, these differences have an effect on the ability of the model to represent tropical cyclones structure, intensity, and formation rate and also their interaction with the surrounding environment.

Since it is crucial that a climate model used for the prediction of future climate give a good simulation of the current climate (DelSole & Shukla, 2006), the quality of CGCMs representation of the current TC climatology is important. Many factors in the real climate modulate the tropical cyclone count rate: the spatial and temporal variations in sea surface temperature (Vecchi & Soden, 2007), the vertical wind shear (McBride & Zehr, 1981), the presence of pre-existing convective development (Hendricks, Montgomery, & Davis, 2004) and the amount of mid-tropospheric relative humidity (Bister & Emanuel, 1997). The ability of the climate models to represent these variables is relevant to a realistic representation of TC genesis. Many “genesis indexes” have been developed based on climatological values of the mentioned fields in a semi-empirical framework with the aim to best fit the observed tropical cyclone formation rate (GPIs; Camargo, Wheeler, & Sobel, 2009; Menkes et al., 2012). In any case, a good representation of the mentioned fields is not sufficient to ensure a good representation of TCs in the models: additional model-dependent factors can influence formation rates, such as the horizontal resolution and convective parameterization (Vitart, Anderson, Sirutis, & Tuleya, 2001). For this reason, the agreement of modeled GPIs with observed values does not always correspond to what is obtained with TC tracking technics. TC detection and tracking in GCMs has been performed following different schemes (e.g., Camargo & Zebiak, 2002; Walsh et al., 2013a; Zhao et al., 2009). These algorithms investigate GCM output data mapping where and when TC criteria (wind velocity, surface pressure minimum, high vorticity, high temperature in the warm core region of the TC, etc.) are met. The threshold defined for each variable representative as TC-prone condition can be based either on absolute values or as anomaly with respect to the mean in that model or region. The usage of absolute values, defined based on observations, can be a limiting factor, especially when the horizontal resolution of the models is not sufficiently high to represent TC-like vortex with magnitude comparable to real TCs. For this reason the usage of relative thresholds that are adjusted model to model (or even basin to basin) is preferred, especially at low resolution. In addition, different tracking schemes rely on different vertical levels and different averaging domains to evaluate TC prone parameters. In general it is found that differences in responses between different tracking schemes are mainly due to the different thresholds used and not to the detection scheme design (Horn et al., 2014). It has been also found that model biases influence model TC projections, and the influence of the biases strongly depends on model physics and choice of parameterization, not only on horizontal resolution (Murakami et al., 2014).

Despite the deficiency in representing a realistic TC structure, state-of-the-art general circulation models and coupled general circulation models have shown considerable skill at least in simulating the climatological observed count of tropical cyclones at the global scale, based on tracking detection investigations (Knutson et al., 2010; Murakami et al., 2015; Scoccimarro et al., 2011; Walsh et al., 2013a; Zhao et al., 2009). Less agreement is found in the basin-by-basin comparison with the observations (Walsh et al., 2013a). Noteworthy, the higher resolution models have also demonstrated an ability to generate a realistic distribution of tropical cyclone intensity within the five Saffir-Simpson hurricane classes (Murakami et al., 2015; Reed et al., 2015; Wehner et al., 2014, 2015). Climate models with horizontal resolution higher than 50 km are able to represent TCs with surface wind velocity greater than 70 m/s (category 5 hurricane in the Saffir-Simpson scale): Figure 5 shows a category 5 typhoon in the West North Pacific, as represented by a GCM (CAM5 model developed at National Center for Atmospheric Research) in its configuration with 25 km as horizontal resolution.

Modeling Tropical Cyclones in a Changing ClimateClick to view larger

Figure 5. Tropical cyclone representation in a 25 km GCM: a Cat 5 typhoon developed over the West North Pacific. The upper panel represents the maximum surface wind intensity reached within a 6-hourly period, and the lower panel represents the averaged precipitation flux over the same time period. Units are [m/s] and [mm/day], respectively.

It is interesting to note that this level of detail in the frontier GCMs (25 km) makes it possible to represent also the well-known asymmetry in the precipitation pattern (Figure 5, lower panel). On the other hand, coarse resolution (lower than 100 km) GCMs have only a limited ability to simulate tropical cyclone intensity, but they still have a reasonable performance in terms of interannual TC count variation (Vitart & Anderson, 2001; Zhao et al., 2009).

Many TC-permitting climate models are atmosphere-only models, using prescribed sea surface temperature, not able to represent TC–ocean feedbacks. As already mentioned when discussing the TC relationship with the climate system, omitting TC–ocean interaction is a strong limitation, especially in terms of TC intensity representation.

This deficiency can be solved using coupled general circulation models (CGCMs), where an ocean climate model is coupled to an atmosphere climate model, without any external forcing (model boundary conditions), but solar constant, greenhouse gasses (GHGs) and aerosols concentrations in the atmosphere (also provided to atmosphere-only simulations). In a typical CGCM the ocean model provides to the atmosphere model the surface information necessary to compute heat, water, and momentum fluxes, such as sea surface temperature, sea ice cover, and ocean currents; the computed fluxes are then provided to the ocean component by the atmospheric model. These operations are performed every coupling time step, as defined by the modelers. A good representation of TC–ocean feedback needs high coupling frequency, also depending on the horizontal resolution of the atmosphere and ocean components. In fact, high horizontal resolution in the atmosphere permits to represent intense TCs, moving fast from one grid point to another. Thus, the information between the atmosphere and the ocean must be exchanged at the higher frequency available, coherently with the time stepping of both climate model components. The typical coupling frequency for a high-resolution coupled general circulation model (a few tens of kilometers as in the example shown in Figures 5 and 6), aiming at representing TC–ocean feedback, is of the order of one hour. Figure 6 shows the surface cooling induced by a category 5 typhoon (the one shown in Figure 5) after two days, as simulated by a coupled general circulation model: a TC-induced SST negative anomaly up to 7 degrees is made apparent thanks to the coupling of the atmospheric component to an ocean model able to represent the aforementioned feedbacks.

Modeling Tropical Cyclones in a Changing ClimateClick to view larger

Figure 6. Sea surface temperature patterns associated to the cat5 typhoon shown in Figure 5. The upper panel represents the SST averaged over a 6-hourly period when TC was located as indicated by the black circle (the same step considered in Figure 5). The lower panel represents the SST averaged over the 6-hourly period just before TC landfall, thus two days (eight 6-hourly steps) later. Units are [°C]. Black dots represent TC track as 6-hourly steps.

The absence of TC-induced surface cooling induces an overestimation of TC intensification: this is the case of tropical cyclones generated by high-resolution atmosphere-only GCMs. Such a deficiency appears only at horizontal resolution sufficient to represent realistic TCs: atmosphere GCMs with a resolution lower than 50 km tend to underestimate TC intensity, thus the importance of the described TC self-induced intensity reduction is less evident.

Future Projection of Tropical Cyclone Activity

Tropical cyclone landfall over tropical and sub-tropical coastal regions provide high winds, storm surge, and heavy precipitation, leading to different societal impacts also depending on geophysical exposure and additional societal factors such as state of development of the region and local economy. Moreover, TC role in modulating high-impact weather events at higher latitudes (such as floods over Europe), suggest that their impact upon society is not only confined to the tropical belt. For these reasons the possibility to investigate TC behavior under different potential climate scenarios is of a great interest for our society.

To this aim we first need a tool to represent the potential future climate scenarios.

Modeling Tropical Cyclones in a Changing ClimateClick to view larger

Figure 7. Atmospheric CO2 concentration time series from the beginning of the industrial period (here referred as historical, HIST) to the end of the current century, following different suggested future scenarios. A1B and A2 refer to CMIP3 experiments. RCP4.5 and RCP8.5 refer to CMIP5 experiments. Units are [ppmv].

Modeling Tropical Cyclones in a Changing ClimateClick to view larger

Figure 8. Same as Figure 7, but for atmospheric CH4 concentration. Units are [ppbv].

As mentioned in the previous paragraph, since fully coupled general circulation models (and also their evolution in earth system models) are able to represent the climate system and the interaction between its main components without external provision of boundary conditions, they are eligible to be used in this context. The only information needed is the radiative forcing represented by the solar constant and the GHGs and aerosol concentration or emission rates.

In recent years, under the coordination of the Intergovernmental Panel on Climate Change (IPCC), potential changes in future GHGs and aerosol concentration, induced by the human activity, have been hypothesized under different population and economic growth and climate mitigation policies. The resulting GHGs and aerosols have been then used (in terms of emissions or atmospheric concentration) to model the different potential climates we expect, for example, to the end of the current century, based on CGCM simulations. Figure 7 shows the temporal evolution of the carbon dioxide (CO2) atmospheric concentration in the past (black line) and in some different potential future climate scenarios, as prepared for the last two phases of the Coupled Model Intercomparison Project: A1B and A2 (CMIP3 scenarios) and RCP4.5 and RCP8.5 (CMIP5 scenarios). The same scenario projections in terms of methane atmospheric concentration are shown in Figure 8. These are just few of the available scenarios, illustrating different potential paths in terms of concentrations, depending on the different actions we will define to tackle global warming in the long-term (decades) period. Not only GHG concentration changes have been linked to TC activity changes, but also aerosol concentration changes. The SST increase associated to the recent reduction of aerosols pollutant over the North Atlantic basin, and the reduced vertical wind shear over the northern Indian Ocean, due to the recent increase of such pollutants, are both related to an increase of TC activity (IPCC, 5th Assesment Report).

Four important components of TC activity directly bear on the societal impacts generated from such systems: TC frequency, intensity, precipitation amount, and storm surge. The associated hazards of wind, excessive rainfall, and storm surge are important at present, but under differing climate conditions could increase the societal impacts of TC activity in many regions. In the following section, potential changes in tropical cyclone activity under future climate scenarios are discussed based on the modeling tools described earlier.

Changes in TC Frequency

The investigation of tropical cyclone count in climate models can be performed based on different detection methods (i.e., tracking schemes) or based on Genesis Potential indexes. Recent works suggest that the responses of TC numbers to climate perturbation experiments vary moderately between tracking schemes (Horn et al., 2014), increasing the reliability of the projection obtained by different climate models, developed by different climate research centers leveraging on different tracking schemes. On the other hand, other analyses show no strong relationship between the Genesis Potential indexes in a given model and the annual number of TCs (Camargo, 2013; Camargo et al., 2007; Walsh et al., 2013a), especially at the low horizontal resolution.

Many previous studies have suggested a TC count reduction in a warmer climate. The prediction of a decrease in TC count comes from both fully coupled general circulation models and hurricane model studies (Emanuel et al., 2008; Gualdi et al., 2008; Knutson et al., 2010, 2013), where fewer TCs in a warmer climate are found compared to the present climate, with a few exceptions (e.g., Emanuel, 2013). One of the reasons behind the general TC number reduction in a warmer climate seems to be related to more stable projected atmospheric conditions (De Maria, Knaff, & Connell, 2001), leading to an environment less prone to the development of organized convective systems (Gualdi et al., 2008). A more systematic predicted decreases in Southern Hemisphere TC count has been also made evident (Knutson et al., 2010) and linked to a more pronounced decrease in mid-tropospheric vertical velocities in the Southern Hemisphere than in the Northern Hemisphere(Walsh et al., 2013b). In fact, the projected TC reduction is not uniform in space, with basins-by-basin differences (Gualdi et al., 2008) and also intra-basin TC count reduction (or increase) due to projected shifts of main TC development regions, depending on the new climate conditions characterizing the considered scenarios (Yokoi, Yukari, & Murakami, 2013). As stated in the last IPCC Assessment Report (AR5, Working Group 1), the tropical cyclone activity is likely to vary by region in a warmer climate, but there is low confidence in region-specific projections. This is also confirmed by Camargo (2013), finding no robust signal across the CMIP5 models in regional TC count in future scenarios. The preponderance of projected negative changes in TC count, in a warmer climate, significant at least at the global scale, is mainly related to weak TCs, while the strongest storms, such as category 4 and 5 storms (winds greater than 59 m/s) tend to be more frequent. For instance, based on dynamical downscaling through the GFDL hurricane model, in the Atlantic basin, the number of annual TCs is projected to significantly decrease (ranging from -30% to -60%) at the end of the current century, compared to the historical period. On the other hand, the number of intense TCs (categories 4 and 5 storms) is projected to double in more than one model (Knutson et al., 2013). These results are based on hurricane model outcomes obtained in a multi-model framework: boundary conditions are provided by CGCMs taking part to the CMIP3 and CMIP5 efforts. In order to have similar information regarding potential change in the shape of the TC count distribution within the intensity scale, directly from CGCMs (without any downscaling additional effort), climate scientists are looking forward to the next Coupled Model Intercomparison Project, CMIP6: a subset of simulations will be performed through very high-resolution models within the HiResMIP project developed in the contest of the World Climate Research Programme.

Changes in TC Intensity

The already discussed poor representation of intense TCs in currently available long-term (decades) climate simulations emphasizes the need for dynamical downscaling through regional climate models or hurricane models to leverage on to obtain reliable future projections of TC intensity.

Theoretical arguments indicate that tropical cyclone winds should increase with increasing ocean temperatures (Emanuel, 1991; Holland, 1997), which seems confirmed by the modeling results described next. In the previous section the projected increased number of the most intense storms (categories 4 and 5) in a warmer climate over the Atlantic basin was partially discussed. Although these two classes of TCs account for just 15% of the total number of landfalling TCs over the United States, it has been found that they are responsible for about 50% of the coastal damage induced by TCs (Pielke et al., 2008). Thus, even small changes in the number of very intense hurricanes may have a strong impact on our society. The trend to a world with more intense TCs has been recently confirmed (Knutson et al., 2015) considering all TC genesis basins. An increase in the projected averaged lifetime at the maximum intensity of Atlantic and Pacific hurricanes and typhoons is also expected, even more pronounced for strong hurricanes (Knutson et al., 2013, 2015). This is consistent with upward trends in the estimated lifetime-maximum wind speeds of the very strongest tropical cyclones (99th percentile) found in the past years and linked to the historical warming trend (Elsner, Kossin, & Jagger, 2008). Another metric to investigate future changes in storm intensity is the power dissipation index (PDI). The PDI (Emanuel, 2005) is a well-recognized indicator of tropical cyclone activity considering storm duration, frequency, and intensity: it is computed as the time integral of the third power of the maximum sustained wind speed (at 10 m) associated to each TC, then accumulated considering all the TCs in a region or/and period. Global PDI is projected to increase in a warmer climate when normalized by TC count, mainly due to the contribution of intense storms (Emanuel, 2013; Knutson et al., 2015), and this can be related to a projected increase in the surface energy available for TC development. A different approach to verify potential changes in the intensity that can be reached by a TC under different climates is based on the maximum potential intensity (MPI)theories, defined as an upper bound of intensity that a TC may reach under a given suite of environmental conditions. Different approaches (Bister & Emanuel, 2002; Holland, 1997) have been designed to determine the upper limit of tropical cyclone intensity, with a different emphasis of the convective available potential energy (CAPE) in their formulation. A recent work (Tsuboki et al., 2015) based on dynamical downscaling indicates that the supertyphoon (category 4 and 5 typhoons) intensity increase, projected in a warmer climate over the West North Pacific, can be also associated to no significant projected changes in MPI. A significant future increase of supertyphoon intensity is expected with an increase of SST by 2°C. This result shows that the upper limit of supertyphoon intensity could increase substantially in the future climate, since an increase of the sea surface temperature higher than 2°C at the end of the current century (compared to the historical period) is projected by the majority of the potential future GHG scenarios proposed within both CMIP3 and CMIP5 experiments.

Changes in TC-Associated Precipitation

In terms of amount of precipitation reaching the surface of the planet in a year, the amount induced by TCs reaches 40% of the total amount over some ocean regions (e.g., Eastern Pacific basin and Southeastern Indian Ocean; Scoccimarro et al., 2014). This is relevant even over land: for instance, Larson, Zhou, and Higgins (2005) found that rainfall associated to TCs contributes up to 20% of the total precipitation over the coast of Mexico. Moreover, focusing on a few case studies, some TCs accounted for more than 90% of the summer rainfall experienced in some regions, such as Australia (Dare, Davidson, & McBride, 2012). Extreme precipitation associated with tropical cyclones is responsible for a large number of fatalities and economic damage worldwide, even due to the associated flooding events (Mendelsohn et al., 2012; Peduzzi et al., 2012). In fact, freshwater flooding has been suggested as the largest threat to human lives due to TCs, at least in the United States (Rappaport, 2000). Due to the large size of TCs (radius of the order of few hundreds of kilometers), the associated precipitation usually affects large domains, even after landfall, including locations located far from the coast (Villarini, Goska, Smith, & Vecchi, 2014). Because of the societal importance of this hazard, potential changes in heavy rainfall associated with tropical cyclones in a warmer climate have been performed in the past years based on the already described dynamical models (Gualdi et al., 2008; Hasegawa & Emori, 2005; Knutson et al., 2013, 2015; Scoccimarro et al., 2014; Villarini et al., 2014). There is an overall agreement in projecting a general increase in the rainfall amount associated to TCs, based on climate and hurricane models. Depending on the chosen future 21st- century scenario pathway and considering the differences in projections between different regions, the projected increase varies between 5% and 20% compared to the present climate. Intermodel differences in regional projections highlight the particularly low confidence in making projections basin by basin. The magnitude of the projected increase also depends on the averaging radius considered when accumulating TC-related precipitation. In particular, a large sensitivity is found for small accumulation radii (Knutson et al., 2010, 2015). The enhanced tropospheric water vapor projected in a warmer climate, inducing an increase of the moisture convergence, has been proposed as a possible cause of the projected rainfall increase (Knutson et al., 2015; Wang, Lin, Chen, & Lo, 2015). In fact, climate models are unanimously suggesting an increase of the vertically integrated water content in the tropics, associated to a warmer atmosphere (Scoccimarro et al., 2014), in agreement with the Clausius-Clapeyron law. Due to the water content increase in the tropical atmosphere, the moisture convergence for a given amount of mass convergence is enhanced (Knutson et al., 2010). This is expected to be linked to an increase of TC-related precipitation, since moisture convergence is a primary factor in determining the amount of precipitation associated to these systems.

Changes in Storm Surge

Another important natural hazard linked to TC activity is storm surge. Storm surges are a highly destructive aspect of TCs in the current climate (Grinsted, Moore, & Jevrejeva, 2012; Woodruff, Irish, & Camargo, 2013): Hurricane Katrina (2005), for instance, produced the greatest coastal flood heights ever recorded in the United States, causing more than $100 billion in losses and resulting in about 2,000 fatalities. A question of increasing concern is whether such devastating surge events will become more frequent in a warmer world, even due to the constantly increasing sea level rise. Since storm surge is a rise of coastal shallow water driven by a storm’s surface wind and pressure gradient forces and the relative magnitude is determined by the characteristics of the storm in conjunction with geometry and bathymetry of the coast (Lin, Emanuel, Oppenheimer, & Vanmarcke, 2012), its severity cannot be inferred directly from storm intensity. A very complex modeling framework is needed. Unfortunately, the actual resolution of CGCM ocean component and the absence of wave and inland penetration representation in the system make a direct assessment of storm surge changes in a warmer world very difficult to obtain.

Another important process not included in state-of-the-art CGCMs is land ice melting. The resulting release of freshwater strongly impacts the mean sea level projections and consequently the possibility to realistically represent the combined effect of projected sea level rise and projected changes in the frequency and intensity of tropical cyclones.

Combining the available hurricane modeling tools to surge models, interesting results have been achieved. For instance, based on the combined use of a CGCM-driven hurricane model and a hydrodynamic surge model, Lin et al. (2012) found that some climate models, within the CGCMs used to create the ensemble of storm simulations, predict an increase of the surge level due to the change of storm climatology at the end of the current century comparable to the projected sea level rise for the same period (order of 1 m around New York City). In addition, joint projections of sea level rise and storm surge have been performed using CGCMs results: the combined influence of PDI and sea level rise on the coastal flood hazard have been summarized in a flood index (Little et al., 2015), driven by both factors, based on a peak-over-threshold approach. The obtained projections for the end of the century suggest that the U.S. East Coast flood hazard will increase substantially, even with low emission scenarios.

A recent work (Lin & Emanuel, 2016) based on a chain of modeling efforts (statistical–deterministic TC model to generate TC tracks + coupled hurricane intensity model + storm surge model [ADCIRC; Luettich, Westreink, & Scheffner, 1992]) suggests a non-negligible probability of huge storm surges, of the order of more than 7 m in some regions (e.g., Tampa and Dubai), especially toward the end of the current century. It is expected that the combined effect of the projected sea level rise and the increase of intense storm number will greatly shorten the surge flooding return periods over certain regions of the globe. As already mentioned, storm surge potential damages are strongly related to the sea level rise. Thus, the higher confidence in the projections of sea level rise, compared to the projections of changes in TC count and intensity, make the storm surge increase the most reliable projected change in TC-related hazards (Walsh et al., 2016).

Conclusions and Perspectives

Dealing with tropical cyclones is a major issue for people living near the coasts of tropical and subtropical regions. For this reason, increasing effort has been put in trying to better understand their nature and statistics, especially during the last part of the 20th century, when the availability of satellite technology (since 1979) made possible a great step in obtaining TC-related information. Together with a growing statistic relative to the past TCs, an increasing ability in modeling their structure and evolution through numerical codes emerged in recent decades. Until now it has been necessary to merge the available technical resources, putting together climate and hurricane models: in order to have a detailed representation of the TC, dynamical downscaling techniques have been used. A TC picture with a realistic structure and intensity cannot be reached directly through state-of-the-art climate models, especially general circulation models. In fact, the horizontal resolution of fully coupled GCMs reached within the last IPCC effort (the CMIP5 experiment) is of the order of hundreds of kilometers, while a few kilometers resolution is needed for a realistic representation of such organized convective systems. Relying on all of the mentioned tools, a reasonably realistic TC modeling framework has been reached, both in terms of counting and intensity.

A general agreement in suggesting a decrease of tropical cyclone count at the global scale under warmer conditions has been reached. Also an increase in intense hurricanes/typhoons has been found, suggesting a potential change of the shape of the distribution of TC count within the different intensity classes. Robust results also suggest an increase in TC- associated precipitation and surge in a warmer climate.

Not negligible uncertainties are still present in the quantification of the magnitude of the mentioned TC activity changes projected in a warmer climate, more pronounced when investigated at the regional scale. This is due partially to the uncertainty cascade through the different investigation tools needed, but also to the level of confidence we put in future climate model projections, at least in their ability to represent the mean patterns of change of regional SSTs. Therefore, it is of fundamental importance to improve climate models with the aim to reduce uncertainties in the 21st-century projections of the mean regional SST patterns and vertical structure of the atmosphere in terms of wind, water content, and temperature (Walsh et al., 2016).

Focusing on the ability of a climate model to represent the TC structure, promising results have been recently presented (Murakami et al., 2015), suggesting the new-generation, very high-resolution (order of few tens of degrees) general circulation models as a potential tool to avoid dynamical downscaling in future generation climate projection analysis. To this aim many modeling research groups are planning to participate to the HiResMIP project within the next Coupled Model Intercomparison Project (CMIP6). The ability to represent category 5 tropical cyclones by the involved GCMs and the possibility to investigate their relationship with the climate system under different climate conditions without any additional downscaling effort is also due to the growing availability of computational power. Bearing in mind the huge computational cost needed to run a very high-resolution climate model simulation (hundreds of years) and considering the necessary model spin-up, we do expect to have exciting multi-model results in the next couple of years. The ability to represent intense hurricanes and supertyphoons within their evolving environment is particularly relevant to investigating TC–ocean feedbacks and the TC role in determining the mean climate conditions. For example, different TC–ocean interactions are expected (Emanuel, 2015; Lin, Pun, & Lien, 2014) under different climates, and the possibility to dynamically represent TC–ocean feedbacks under evolving climate conditions is of great interest for the TC community. Together with the increase of the horizontal and vertical resolution of model components, the inclusion of land ice melting, ocean waves, and inland water penetration processes is expected in future generation fully coupled general circulation models. Such improved tools will make it possible to investigate and quantify potential changes in TC activity, including in terms of storm surges, with increasing reliability.

Selected Readings

Bender, M. A., Tuleya, R. E., Thomas, B., & Marchok, T. (2007). The operational GFDL coupled hurricane–ocean prediction system and a summary of its performance. Monthly Weather Reviews, 132, 3965–3989.Find this resource:

Emanuel, K. A. (2001). Contribution of tropical cyclones to meridional heat transport by the oceans. Journal of Geophysical Research, 106(D14), 14, 771–14, 781.Find this resource:

Gualdi, S., Scoccimarro, E., & Navarra, A. (2008). Changes in tropical cyclone activity due to global warming: Results from a high-resolution coupled general circulation model. Journal of Climate, 21, 5204–5228.Find this resource:

Holland, G. J. (1997). The maximum potential intensity of tropical cyclones. Journal of Atmospheric Sciences, 54, 2519–2541.Find this resource:

Knutson, T. R., et al. (2015). Global projections of intense tropical cyclone activity for the late twenty-first century from dynamical downscaling of CMIP5/RCP4.5 scenarios. Journal of Climate, 28(18).Find this resource:

Landsea, C. W. (2007). Counting Atlantic tropical cyclones back to 1900. Eos, Trans. Amer. Geophys. Union, 88, 197–202.Find this resource:

Lin, I.-I., Pun, I.-F., & Lien, C.-C. (2014). “Category-6” supertyphoon Haiyan in global warming hiatus: Contribution from subsurface ocean warming. Geophysical Research Letters, 41.Find this resource:

Mendelsohn, R., Emanuel, K., Chonabayashi, S., & Bakkensen, L. (2012). The impact of climate change on global tropical cyclone damage. Nature Climate Change, 2, 205–209.Find this resource:

Murakami, H., Vecchi, G. A., Underwood, S., Delworth, T. L., Wittenberg, A. T., Anderson, W. G., et al. (2015). Simulation and prediction of category 4 and 5 hurricanes in the high-resolution GFDL HiFLOR coupled climate model. Journal of Climate, 28(23), 9058–9079.Find this resource:

Pielke, R. A., Gratz, J., Landsea, C. W., Collins, D., Saunders, M. A., & Musulin, R. (2008). Normalized hurricane damage in the United States: 1900–2005. Natural Hazards Reviews, 9, 29–42.Find this resource:

Sriver, R. L., Goes, M., Mann, M. E., & Keller, K. (2010). Climate response to tropical cyclone-induced ocean mixing in an Earth system model of intermediate complexity. Journal of Geophysical Research, 115, C10042.Find this resource:

Taylor, K. E., Stouffer R. J., & Meehl, G. A. (2012). An overview of CMIP5 and the experiment design. Bulletin of the American Meteorological Society, 93, 485–498.Find this resource:

Tsuboki, K., Yoshioka,, M. K., Shinoda, T., Kato, M., Kanada, S., & Kitoh, A. (2015). Future increase of supertyphoon intensity associated with climate change. Geophysical Research Letters, 42, 646–652.Find this resource:

Vecchi, G. A., & Soden, B. J. (2007). Effect of remote sea surface temperature change on tropical cyclone potential intensity. Nature, 450, 1066–1070.Find this resource:

Walsh, K. J. E., Camargo, S. J., Vecchi, G. A., Daloz, A. S., Elsner, J., Emanuel, K., Horn, M. et al. (2015). Hurricanes and climate: The U.S. CLIVAR Working Group on Hurricanes. Bulletin of the American Meteorological Society, 96, 1440.Find this resource:


Bender M. A., Knutson, T. R., Tuleya, R. E., Sirutis, J. J., Vecchi, G. A., Garner, S. T., et al. (2010). Modeled impact of anthropogenic warming of the frequency of intense Atlantic hurricanes. Science, 327, 454–458.Find this resource:

Bender M. A., Tuleya, R. E., Thomas, B., & Marchok, T. (2007). The operational GFDL coupled hurricane–ocean prediction system and a summary of its performance. Monthly Weather Review, 132, 3965–3989.Find this resource:

Bister, M., & Emanuel, K. (1997). The genesis of Hurricane Guillermo: TEXMEX analyses and a modelling study. Monthly Weather Review, 125, 1397–1413.Find this resource:

Bister, M., & Emanuel, K. (2002). Low frequency variability of tropical cyclone potential intensity 1. Interannual to interdecadel variability. Journal of Geophysical Research, 107, 4801.Find this resource:

Bueti, M. R., et al. (2014).Tropical cyclone–induced thermocline warming and its regional and global impacts. Journal of Climate, 27, 6978–6999.Find this resource:

Camargo, S. J. (2013). Global and regional aspects of tropical cyclone activity in the CMIP5 models. Journal of Climate, 26(24), 9880–9902.Find this resource:

Camargo, S. J., Ting, M., & Kushnir, Y. (2013). Influence of local and remote SST on North Atlantic tropical cyclone potential intensity. Climate Dynamics, 40, 1515–1529.Find this resource:

Camargo, S. J., Wheeler, M. C., & Sobel, A. H. (2009). Diagnosis of the MJO modulation of tropical cyclogenesis using an empirical index. Journal of Atmospheric Sciences, 66, 3061–3074.Find this resource:

Camargo, S. J., & Wing, A. A. (2016). Tropical cyclones in climatemodels. WIREs Climate Change, 7, 211–237.Find this resource:

Camargo, S. J., & Zebiak, S. E. (2002). Improving the detection and tracking of tropical cyclones in atmospheric general circulation models. Weather Forecasting, 17, 1152–1162.Find this resource:

Camargo, S. J., et al. (2007). Tropical cyclone genesis potential index in climate models. Tellus, 59A, 428–443.Find this resource:

Dare, R. A., Davidson, N. E., & McBride, J. L. (2012). Tropical cyclone contribution to rainfall over Australia. Monthly Weather Reviews, 140, 3606–3619.Find this resource:

De Maria, M., Knaff, J.A., & Connell, B. H. (2001). A tropical cyclone genesis parameter for the tropical Atlantic. Weather Forecasting, 16, 219–233.Find this resource:

DelSole, T., & Shukla, J. (2006). Model fidelity versus skill in seasonal forecasting. Journal of Climate, 23, 4794–4806.Find this resource:

Elsner, J. B., & Jagger, T. H. (2006). Prediction models of U.S. hurricane activity. Journal of Climate, 19, 2935–2952.Find this resource:

Elsner, J. B., Kossin, P, & Jagger, T. H. (2008). The increasing intensity of the strongest tropical cyclones. Nature, 455, 92–95.Find this resource:

Elsner, J. B., Liu, K. B., & Kocher, B. (2000). Spatial variations in major U.S. hurricane activity: Statistics and a physical mechanism. Journal of Climate, 13, 2293–2305.Find this resource:

Emanuel, K. (2015). Effect of upper-ocean evolution on projected trends in tropical cyclone activity. Journal of Climate, 28, 8165–8170.Find this resource:

Emanuel, K. A. (1987). The dependence of hurricane intensity on climate. Nature, 326, 483–485.Find this resource:

Emanuel, K. A. (1991). The theory of hurricanes. Annual Review of Fluid Mechanics, 23, 179–196.Find this resource:

Emanuel, K. A. (1995). The behavior of a simple hurricane model using a convective scheme based on subcloud-layer entropy equilibrium. Journal of Atmospheric Sciences, 52, 3959–3968.Find this resource:

Emanuel, K. A. (2001). Contribution of tropical cyclones to meridional heat transport by the oceans. Journal of Geophysical Research, 106(D14), 14771–14781.Find this resource:

Emanuel, K. A. (2004). Tropical cyclone energetics and structure. In Atmospheric turbulence and mesoscale meteorology (pp. 165–191). Cambridge, U.K. Cambridge University Press.Find this resource:

Emanuel, K. A. (2005). Increasing destructiveness of tropical cyclones over the past 30 years. Nature, 436, 686–688.Find this resource:

Emanuel, K. A. (2006). Climate and tropical cyclone activity: A new model downscaling approach. Journal of Climate, 19, 4797–4802.Find this resource:

Emanuel, K. A. (2013). Downscaling CMIP5 climate models shows increased tropical cyclone activity over the 21st century. Proc. Natl. Acad. Sci. USA, 110, 12219–12224.Find this resource:

Emanuel, K. A., Sundararajan, R., & Williams, J. (2008). Hurricanes and global warming: Results from downscaling IPCC AR4 simulations. Bulletin of the American Meteorological Society, 89, 347–367.Find this resource:

Goldenberg, S. B., Landsea, C., Mestas-Nunez, A. M., & Gray, W. M. (2001). The recent increase in Atlantic hurricane activity. Science, 293, 474–479.Find this resource:

Gray, W. M. (1979). Hurricanes: Their formation, structure and likely role in the tropical circulation. In D. B. Shaw (Ed.), Meteorology over the tropical oceans (pp. 155–218). Bracknell, U.K.: Royal Meteorological Society.Find this resource:

Grinsted, A., Moore, J. C., & Jevrejeva, S. (2012). Homogeneous record of Atlantic hurricane surge threat since 1923. Proceedings of the National Academy of Sciences, 109(4), 19601–19605.Find this resource:

Gualdi, S., Scoccimarro, E., & Navarra, A. (2008). Changes in tropical cyclone activity due to global warming: Results from a high-resolution coupled general circulation model. Journal of Climate, 21, 5204–5228.Find this resource:

Hasegawa, A., & Emori, S. (2005). Tropical cyclones and associated precipitation over the western North Pacific: T106 atmospheric GCM simulation in present and doubled CO2 climates. SOLA, 1, 145–148.Find this resource:

Hendricks E.A., Montgomery, M. T., & Davis, C. A. (2004). The role of “vortical” hot towers in the formation of tropical cyclone Diana (1984). Journal of Atmospheric Sciences, 61, 1209–1232.Find this resource:

Holland, G. J. (1993). Tropical cyclone motion. In G. J. Holland (Ed.), Global guide to tropical cyclone forecasting WMO/TD-560 (pp. 3.1–3.46). Geneva, Switzerland: World Meteorological Organization.Find this resource:

Holland, G. J. (1997). The maximum potential intensity of tropical cyclones. Journal of Atmospheric Sciences, 54, 2519–2541.Find this resource:

Horn, M., Walsh, K., Zhao, M., Camargo, S. J., Scoccimarro, E., Murakami, H., et al. (2014). Tracking scheme dependence of simulate tropical cyclone response to idealized climate simulations. Journal of Climate, 27, 9197–9213.Find this resource:

Hu, A., & Meehl, G. A. (2009). Effect of the Atlantic hurricanes on the oceanic meridional overturning circulation and heat transport. Geophysical Research Letters, 36, L03702.Find this resource:

Huang, P., Lin, I.-I., Chou, C., & Huang, R.-H. (2015). Change in ocean subsurface environment to suppress tropical cyclone intensification under global warming. Nature Communications, 6, 7188.Find this resource:

Klotzbach, P. J. (2014). The Madden-Julian Oscillation’s impacts on worldwide tropical cyclone activity. Journal of Climate, 27, 2317–2330.Find this resource:

Knapp, K. R., Kruk, M. C., Levinson, D. H., Diamond, H. J., & Neumann, C. J. (2010). The international best track archive for climate stewardship (IBTrACS): Unifying tropical cyclone best track data. Bulletin of the American Meteorological Society, 91, 363–376.Find this resource:

Knutson, T. R., Landsea, C., & Emanuel, K. A. (2010). Tropical cyclones and climate change: A review. In Global perspectives on tropical cyclones: From science to mitigation (pp. 243–284). Singapore: World Scientific Publishing Company.Find this resource:

Knutson, T. R., Sirutis, J. J., Garner, S. T., Held, I. M., & Tuleya, R. E. (2007). Simulation of the recent multidecadal increase of Atlantic hurricane activity using an 18-km-grid regional model. Bulletin of the American Meteorological Society, 88(10), 1549–1565.Find this resource:

Knutson, T. R., Sirutis, J. J., Vecchi, G. A., Garner, S. T., Zhao, M., Kim, H. S., Bender, M., et al. (2013). Dynamical downscaling projections of 21st century Atlantic hurricane activity: CMIP3 and CMIP5 model-based scenario. Journal of Climate, 26(17).Find this resource:

Knutson, T. R., Tuleya, R. E., Bender, M., Villarini, G., & Chaves, D. (2015). Global projections of intense tropical cyclone activity for the late twenty-first century from dynamical downscaling of CMIP5/RCP4.5 scenarios. Journal of Climate, 28(18).Find this resource:

Kossin, J. P., & Vimont, D. J. (2007). A more general framework for understanding Atlantic hurricane variability and trends. Bulletin of the American Meteorological Society, 88, 1767–1781.Find this resource:

Krichak, S. O., Barkan, J., Breitgand, J. S., Gualdi, S., & Feldstein, S. B. (2014). The role of the export of tropical moisture into midlatitudes for extreme precipitation events in the Mediterranean region. Theoretical and Applied Climatology, 121, 499.Find this resource:

Landsea, C. W., (2007). Counting Atlantic tropical cyclones back to 1900. Eos, Trans. Amer. Geophys. Union, 88, 197–202.Find this resource:

Landsea, C. W., et al. (2004). The Atlantic hurricane database re-analysis project: Documentation for the 1851–1910 alterations and additions to the HURDAT database. In R. J. Murnane & K.-B. Liu (Eds.), Hurricanes and typhoons: Past, present and future (pp. 177–221). New York: Columbia University Press.Find this resource:

Larson, J., Zhou, Y., & Higgins, R. W. (2005). Characteristics of landfalling tropical cyclones in the United States and Mexico: Climatology and interannual variability. Journal of Climate, 18, 1247–1262.Find this resource:

Lin, I.-I., Pun, I.-F., & Lien, C.-C. (2014). “Category-6” supertyphoon Haiyan in global warming hiatus: Contribution from subsurface ocean warming. Geophysical Research Letters, 41.Find this resource:

Lin, N., & Emanuel, K. (2016). Grey swan tropical cyclones. Nature Climate Change, 6, 106–111.Find this resource:

Lin, N., Emanuel, K., Oppenheimer, M., & Vanmarcke, E. (2012). Physically based assessment of hurricane surge threat under climate change. Nature Climate Change, 2, 462–467.Find this resource:

Little, C. M., Horton, R. M., Kopp, R. E., Oppenheimer, M., Vecchi, G. A., & Villarini, G. (2015). Joint projections of US East Coast sea level and storm surge. Nature Climate Change, 5, 1114–1120.Find this resource:

Luettich, R. A., Westerink, J. J., & Scheffner, N. W. (1992). ADCIRC: An advanced three-dimensional circulation model for shelves, coasts and estuaries, Report 1: Theory and methodology of ADCIRC-2DDI and ADCIRC-3DL. DRP Technical Report DRP-92-6. Department of the Army, US Army Corps of Engineers, Waterways Experiment Station, Washington, D.C.Find this resource:

McBride, J. L., & Zehr, R. (1981). Observational analysis of tropical cyclone formation. Part II: Comparison of non-developing versus developing systems. Journal of Atmospheric Sciences, 38, 1132–1151.Find this resource:

Meehl, G., Covey, D., Delworth, T., Latif, M., McAvaney, B., Mitchell, J., et al. (2007). The WCRP CMIP3 multimodel dataset. Bulletin of the American Meteorological Society, 88, 1383–1394.Find this resource:

Mendelsohn, R., Emanuel, K., Chonabayashi, S., & Bakkensen, L. (2012). The impact of climate change on global tropical cyclone damage. Nature Climate Change, 2, 205–209.Find this resource:

Menkes, C. E., Lengaigne, M., Marchesiello, P., Jourdain, N. C., Vincent, E. M., Lefevre, J., et al. (2012). Comparison of tropical cyclonegenesis indices on seasonal to interannual timescales. Climate Dynamics, 38, 301–321.Find this resource:

Murakami, H., Hsu, P. C., Arakawa, O., & Li, T. (2014). Influence of model biases on projected future changes in tropical cyclone frequency of occurrence. Journal of Climate, 27, 2159–2181.Find this resource:

Murakami, H., Vecchi, G. A., Underwood, S., Delworth, T. L., & Wittenberg, A. T. (2015). Simulation and prediction of category 4 and 5 hurricanes in the high-resolution GFDL HiFLOR coupled climate model. Journal of Climate, 28(23), 9058–9079.Find this resource:

Nicholls, N., Landsea, C. W., & Gill, J. (1998). Recent trends in Australian region tropical cyclone activity. Meteorology and Atmospheric Physics, 65, 197–205.Find this resource:

Peduzzi, P., Chatenou, B., Dao, H., De Bono, A., Herold, C., Kossin, J., et al. (2012). Global trends in tropical cyclone risk. Nature Climate Change, 2, 289–294.Find this resource:

Persing, J., Montgomery, M. T., McWilliams, J., & Smith, R. K. (2013). Asymmetric and axisymmetric dynamics of tropical cyclones. Atmospheric Chemistry and Physics, 13, 12299–12341.Find this resource:

Pielke, R. A., Gratz, J., Landsea, C. W., Collins, D., Saunders, M. A., & Musulin, R. (2008). Normalized hurricane damage in the United States: 1900–2005. Natural Hazards Review, 9, 29–42.Find this resource:

Pielke, R. A., & Landsea, C. W. (1999). La Niña, El Niño, and Atlantic hurricane damages in the United States. Bulletin of the American Meteorological Society, 80, 2027–2033.Find this resource:

Rappaport, E. N. (2000). Loss of life in the United States associated with recent Atlantic tropical cyclones. Bulletin of the American Meteorological Society, 81, 2065–2074.Find this resource:

Reed, K. A., Bacmeister, J. T., Rosenbloom, N. A., Wehner, M. F., Bates, S. C., Lauritzen, P. H., et al. (2015). Impact of the dynamical core on the direct simulation of tropical cyclones in a high-resolution global model. Geophysical Research Letters, 42.Find this resource:

Scoccimarro E., Gualdi, S., & Navarra, A. (2012). Tropical cyclone effects on Arctic sea ice variability. Geophysical Research Letters, 39, L17704.Find this resource:

Scoccimarro, E., Gualdi, S., Villarini, G., Vecchi, G., Zhao, M., Walsh, K., et al. (2014). Intense precipitation events associated with landfalling tropical cyclones in response to a warmer climate and increased CO2. Journal of Climate, 27, 4642–4654.Find this resource:

Scoccimarro E., et al. (2011). Effects of Tropical Cyclones on Ocean Heat Transport in a High-Resolution Coupled General Circulation Model. Journal of Climate, 24, 4368–4384.Find this resource:

Sriver, R. L., Goes, M., Mann, M. E., & Keller, K. (2010). Climate response to tropical cyclone-induced ocean mixing in an earth system model of intermediate complexity. J Geophys Res, 115.Find this resource:

Taylor, K. E., Stouffer, R. J., & Meehl, G. A. (2012). An overview of CMIP5 and the experiment design. Bulletin of the American Meteorological Society, 93, 485–498.Find this resource:

Tsuboki, K., Yoshioka, M. K., Shinoda, T., Kato, M., Kanada, S., & Kitoh, A. (2015). Future increase of supertyphoon intensity associated with climate change. Geophysical Research Letters, 42, 646–652.Find this resource:

Vecchi, G.A., & Soden, B. J. (2007). Effect of remote sea surface temperature change on tropical cyclone potential intensity. Nature, 450, 1066–1070.Find this resource:

Vecchi, G. A., & Villarini, G. (2014). Next season’s hurricanes. Science, 343, 618–619.Find this resource:

Vecchi, G. A., Zhao, M., Wang, H., Villarini, G., Rosati, A., Kumar, A., et al. (2011). Statistical–dynamical predictions of seasonal North Atlantic hurricane activity. Monthly Weather Reviews, 139, 1070–1082.Find this resource:

Vecchi, G. A., Msadek, R., Anderson, W., Chang, Y. S., Delworth, T., Dixon, K., et al. (2013). Multi-year predictions of North Atlantic hurricane frequency: Promise and limitations. Journal of Climate, 26, 5337–5357.Find this resource:

Villarini, G., Goska, R., Smith, J.,A., & Vecchi G. A. (2014). North Atlantic tropical cyclones and U.S. flooding. Bulletin of the American Meteorological Society, 95, 1381–1388.Find this resource:

Villarini, G., & Vecchi, G. A. (2012). Projected increases in North Atlantic tropical cyclone intensity from CMIP5 models. Journal of Climate, 26, 3231–3240.Find this resource:

Vitart F., & Anderson, J. L. (2001). Sensitivity of Atlantic tropical storm frequency to ENSO and interdecadal variability of SSTs in an ensemble of AGCM integrations. Journal of Climate, 14, 533–545.Find this resource:

Vitart, F, Anderson, J. L., Sirutis, J., & Tuleya, R. E. (2001). Sensitivity of tropical storms simulated by a general circulation model to changes in cumulus parameterization. Quarterly Journal of the Royal Meteorological Society, 127, 25–51.Find this resource:

Walsh, K. J. E., Horn, M., Camargo, S., Muakami, H., Wang, H., & Scoccimarro, E. (2013b). Changes in future Southern Hemisphere tropical cyclone numbers. U.S. Clivar Variations, 11(3).Find this resource:

Walsh, K. J. E., Lavender, S., Scoccimarro, E., & Murakami, H. (2013a). Resolution dependence of tropical cyclone formation in CMIP3 and finer resolution models. Climate Dynamics, 40, 585–599.Find this resource:

Walsh, K. J. E., Camargo, S. J., Vecchi, G. A., Daloz, A. S., Elsner, J., Emanuel, K., et al. (2015). Hurricanes and climate: The U.S. CLIVAR Working Group on Hurricanes. Bulletin of the American Meteorological Society, 96, 1440.Find this resource:

Walsh, K. J. E., McBride, J. L., Klotzbach, P. J., Balachandran, S., Camargo, S. J., Holland, G. et al. (2016). Tropical cyclones and climate change. WIREs Climate Change, 7, 65–89.Find this resource:

Wang, C.-C., Lin, B.-X., Chen, C.-T., & Lo, S.-H. (2015). Quantifying the effects of long-term climate change on tropical cyclone rainfall using a cloud-resolving model: Examples of two landfall typhoons in Taiwan. Journal of Climate, 28, 66–85.Find this resource:

Wehner, M. F., et al. (2014). The effect of horizontal resolution on simulation quality in the Community Atmospheric Model, CAM5.1. Journal of Advances in Modeling Earth Systems, 6, 980–997.Find this resource:

Wehner, M. F., et al. (2015). Resolution dependence of future tropical cyclone projections of CAM5.1 in the U.S. CLIVAR Hurricane Working Group idealized configurations. Journal of Climate, 28(10), 3905–3925.Find this resource:

Woodruff, J. D., Irish, J. L., & Camargo, S. J. (2013). Coastal flooding by tropical cyclones and sea-level rise. Nature, 504, 44–52.Find this resource:

Yokoi, S., Yukari, N. T., & Murakami, H. (2013). Attribution of projected future changes in tropical cyclone passage frequency over the western North Pacific. Journal of Climate, 26, 4096–4111.Find this resource:

Zhao, M., Held, I. M., Lin, S.-J., & Vecchi, G. A. (2009). Simulations of global hurricane climatology, interannual variability, and response to global warming using a 50km resolution GCM. Journal of Climate, 22, 6653–6678.Find this resource: