Association between tropospheric temperature and tropical cyclone peak intensity over the North Pacific and North Atlantic

Zhenxi Zhng , , Cong Zhng , Wen Zhou

a College of Energy and Power Engineering, Inner Mongolia University of Technology, Hohhot, China

b School of Energy and Environment, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong, China

Keywords:Tropical cyclone Carnot cycle Potential intensity theory

ABSTRACT As natural weather disasters, tropical cyclones (TCs) are destructive in proportion to their peak intensity. This study investigates the association of North Atlantic, western North Pacific, and eastern North Pacific TC peak intensities with tropospheric air temperature, respectively, by applying NCEP—NCAR and MERRA reanalysis data.Both the correlation between TC peak intensity and air temperature and the difference in air temperature between strong and weak TC peak intensity conditions reveal that significant cooling of the tropopause and uppertropospheric warming are accompanied by strengthening TC peak intensity for North Atlantic TCs, suggesting an important effect of upper-tropospheric static stability on TC peak intensity. However, warming in the lower troposphere is associated with strong TC peak intensity for eastern North Pacific TCs, indicating a major effect of lower-tropospheric static stability on TC peak intensity. The peak intensity of western North Pacific TCs is mainly affected by vertical wind shear, not the atmospheric temperature.

1. Introduction

Tropical cyclone (TC) intensity is regulated by several thermodynamic (sea surface temperature and free-troposphere humidity) and dynamic (vertical wind shear and vorticity) conditions ( Bosart et al., 1999 ;Emanuel, 2013 ; Lin et al., 2015 ). As suggested by the potential intensity theory ( Emanuel, 1986 , 1995 , 1997 , 2012 ), TC intensity is closely related to TC energy. A steady TC may be regarded as a Carnot cycle that acquires moist entropy from ocean—air heat and moisture fluxes. The work done through this cycle is responsible for driving the TC’s winds.

The original assumption for atmospheric stability in the potential intensity theory is neutrality ( Emanuel, 1986 ), meaning that convective heating is completely balanced by adiabatic cooling ( Kieu, 2015 ). However, the neutrality condition is not appropriate for real TC development because the middle level becomes warmer ( Ooyama, 1969 ; Vigh and Schubert, 2009 ). Kieu and Wang (2017a , 2017b ) modified the neutrality assumption and revealed the dependence of TC potential intensity on static stability. Their finding was confirmed ( Kieu and Zhang, 2018 )and coincides with previous modeling studies ( Shen et al., 2000 ; Hill and Lackmann, 2011 ; Tuleya et al., 2016 ).

The nature of tropospheric static stability is the relative change in atmospheric temperature, especially the temperature in the upper troposphere, where the rate of change is larger than in the lower troposphere ( Hill and Lackmann, 2011 ; Sohn et al., 2016 ; Plesca et al., 2018 ).Therefore, upper-tropospheric temperature plays a critical role in the variation of TC maximum potential intensity (MPI) ( Emanuel et al.,2013 ), as documented by many modeling and observational studies( Ramsay, 2013 ; Wang et al., 2014 ; Wing et al., 2015 ; Kossin, 2015 ;Kieu et al., 2016 ; Ferrara et al., 2017 ; Moon and Kieu, 2016 ). According to the potential intensity theory, the impact of the upper-tropospheric temperature on MPI is expected, because MPI depends on the energy conversion efficiency of the Carnot cycle in a TC, which converts more enthalpy from the sea surface to TC kinetic energy in a colder upper troposphere.

The category of a TC is determined by the maximum intensity attained by this TC during its lifetime, which is called the peak intensity.But what factors affect the peak intensity of real TCs? Does the peak intensity of real TCs depend on tropospheric static stability and uppertropospheric temperature, the same as TC intensity? In this study, we investigate the influence of air temperature on the peak intensity of real TCs over the North Atlantic, western North Pacific, and eastern North Pacific.

2. Datasets

TC track data for the Atlantic and eastern North Pacific were obtained from NOAA’s Tropical Prediction Center. TC track data for the western North Pacific were obtained from the U.S. Navy’s Joint Typhoon Warning Center. All TC track data record the location and intensity of TCs at 6-h intervals, including maximum wind speed, longitude, latitude, time (month, day of the month, and hour (GMT)), and central surface pressure.

Monthly atmospheric temperature data used in this study were obtained from the reanalysis product of the National Centers for Environmental Prediction—National Center for Atmospheric Research (NCEP—NCAR) ( Kalnay et al., 1996 ), which is currently available starting in January 1948. Besides the NCEP—NCAR data, atmospheric temperature fields from the Modern Era-Retrospective Analysis for Research and Applications (MERRA) reanalysis for 1979—2015 are also used in this study,which include monthly data with a resolution of 1.25° longitude by 1.25°latitude, and 42 vertical pressure levels. As a reanalysis dataset, MERRA combines model fields from version 5 of the Goddard Earth Observing System atmospheric model and Data Assimilation System with observations from the satellite era ( Rienecker et al., 2011 ).

3. Results

3.1. Characteristics of TC peak intensity

As shown by the seasonal distribution of the number of TCs in the North Atlantic, western North Pacific, and eastern North Pacific during the period 1979—2015 ( Fig. 1 ), the storm season is from July to October(JASO), when there are more TCs than in other months. The stormseason mean peak intensity of TCs in a given year is calculated by the following formula:

whereVis the peak intensity of a TC;nis the number of TCs occurring in the storm season (JASO) of a given year; and the subscriptmis the ordinal number of a TC in the storm season (JASO) of a given year.

Fig. 1 shows the time series of the storm-season mean peak intensity of TCs over the North Atlantic, western North Pacific, and eastern North Pacific from 1979 to 2015. Statistics for the storm-season mean peak intensity of TCs are summarized in the box plots, including medians, lower and upper quartiles, minimum and maximum of normal values, and outliers. Standardization of the time series is calculated by subtracting the mean and dividing the resultant deviations from the mean by the standard deviation of the respective time series.

The horizontal distribution of storm-season TC intensity is computed by binning the JASO TC intensity (best-track maximum wind speed) of a given year into the 2° grid cells. Fig. S1 shows the difference in the horizontal distribution of storm-season TC intensity between strong and weak TC peak intensity conditions. In this study, the composite patterns for strong and weak TC peak intensity are constructed using two methods. In the first method (named “Qua ”), the composite patterns for strong and weak TC peak intensity are based on the conditions when the storm-season mean peak intensity of TCs is larger/smaller than the upper/lower quartile ( Fig. 1 ). In the second method (named “Std ”), the composite patterns for strong and weak TC peak intensity are based on years with larger/smaller storm-season mean peak intensity of TCs,while its standardized value is more/less than 1/ ? 1 ( Fig. 1 ). Generally,all differences over the North Atlantic, western North Pacific, and eastern North Pacific are more obvious in the second method than in the first method. The peak intensity of TCs occurs mainly over the region with large TC intensity difference. As shown by the results from both methods, the peak intensity of the North Atlantic TCs occurs mainly over the region of 20°—30°N and 60°—80°W; the peak intensity of the western North Pacific TCs occurs mainly over the region of 15°—25°N and 120°—140°E; and the peak intensity of the eastern North Pacific TCs occurs mainly over the region of 10°—20°N and 100°—130°W. Over the region with TC peak intensity, the change in the western and eastern North Pacific TC intensity is larger than that of the North Atlantic TCs.

3.2. Relationship of TC peak intensity with air temperature

3.2.1.Verticaldistribution

The long-term statistical relationship between TC peak intensity and air temperature in the troposphere is studied by computing the interannual correlation between the storm-season (JASO) mean peak intensity of TCs and air temperature at each grid point. The air temperature comes from NCEP—NCAR and MERRA data (1979—2015), respectively. Correlation analysis is implemented for the North Atlantic, western North Pacific, and eastern North Pacific storm-season mean peak intensity of TCs, respectively.

As shown by Fig. S1, the storm-season peak intensities of the North Atlantic, western North Pacific, and eastern North Pacific TCs occur mainly between 20°N and 30°N, 15°N and 25°N, and 10°N and 20°N,respectively. Therefore, the JASO meridional mean air temperature between 20°N and 30°N (North Atlantic), between 15°N and 25°N (western North Pacific), and between 10°N and 20°N (eastern North Pacific)is calculated first. The correlation between the storm-season mean peak intensity of TCs and JASO meridional mean air temperature from NCEP and MERRA (1979—2015) is displayed in Fig. 2 .

Both the NCEP and MERRA results show that, accompanied by strong peak intensity of the North Atlantic TCs, significant cooling at 100 hPa (T100) appears over the longitudinal position of the peak intensity (60°—80°W), with correlations smaller than ? 0.4. Moreover, obvious warming at 300 hPa (T300) between 60°W and 80°W is accompanied by strengthening peak intensity, with correlations reaching 0.4.

As shown by the correlations with NCEP—NCAR, the peak intensity of the western North Pacific TCs is positively associated with air temperature at 600 hPa (T600) over the longitudinal position of peak intensity (120°—140°E). This is consistent with the finding of Downs and Kieu (2019) : the maximum intensity of the western North Pacific TCs is positively correlated with the lower-tropospheric (500—1000 hPa)stability, while the increasedT600can enhance the lower-tropospheric(500—1000 hPa) stability. Although the peak intensity of the western North Pacific TCs is negatively associated with the NCEP—NCAR air temperature at 70 hPa between 120°E and 140°E, it is not correlated with the MERRA air temperature at 70 hPa at the 95% confidence level. The peak intensity of the eastern North Pacific TCs is correlated with air temperature at 925 hPa (T925) over the longitudinal position of peak intensity(100°—130°W), with NCEP—NCAR and MERRA correlations reaching 0.4.

To confirm the results of the correlation analysis, Fig. 2 shows the vertical and zonal distribution of the JASO meridional mean air temperature differences between strong and weak TC peak intensity conditions. Accompanied by strong peak intensity of the North Atlantic TCs,between 60°W and 80°W, cold (with magnitude larger than 1.5 K) and warm (with magnitude larger than 0.75 K) temperatures occur at 100 hPa and 200—400 hPa, respectively. Associated with strong peak intensity of the western North Pacific TCs, theT600between 120°E and 140°E becomes warmer, with magnitude larger than 0.15 K. For the eastern North Pacific TCs, lower troposphere (900—1000 hPa) warming (with magnitude larger than 0.45 K) between 100°W and 130°W is accompanied by strengthening peak intensity.

Fig. 1. Statistics of the North Atlantic (NA), western North Pacific (WP), and eastern North Pacific (EP) TC number (seasonal distribution) and peak intensity(including time series, box plot, and standardized time series) during the period 1979—2015. TC peak intensity is the storm-season (July—October) mean. The red and blue lines in the time series of the TC peak intensity plots represent the median and the lower and upper quartiles. The eastern North Pacific TCs are those originating from the Pacific east of 150°W.

3.2.2.Horizontaldistribution

According to the analysis of Fig. 2 , we further compare the horizontal variation in air temperature associated with the North Atlantic, western North Pacific, and eastern North Pacific TC peak intensity changes.

Fig. 3 presents the grid-point correlation between the storm-season mean peak intensity of TCs and the JASO mean air temperature from NCEP—NCAR and MERRA (1979—2015). For North Atlantic TCs, the storm-season mean peak intensity has statistically significant negative correlations withT100over the region with TC peak intensity (20°—30°N,60°—80°W). Correlations with both NCEP—NCAR and MERRA are smaller than ? 0.6. On the other hand, the storm-season mean peak intensity has obviously positive correlations withT300over the same region. Correlations with both NCEP—NCAR and MERRA are larger than 0.5. Western North Pacific TC peak intensity is strongly correlated withT600over the region with TC peak intensity (15°—25°N, 120°—140°E). Correlations with both NCEP—NCAR and MERRA reach 0.6. The storm-season mean peak intensity of the eastern North Pacific TCs is highly correlated withT925over the region with TC peak intensity (10°—20°N, 100°—130°E). Correlation with NCEP—NCAR is larger than 0.5, while that with MERRA is larger than 0.4.

Fig. 4 shows the difference between strong and weak TC peak intensity conditions of JASO meanT100,T300,T600, andT925from NCEP and MERRA data (1979—2015). Strong North Atlantic TC peak intensity is associated with a decrease inT100and an increase inT300over the region of 20°—30°N and 60°—80°W, respectively. The largest changes inT100between strong and weak TC peak intensity conditions are more than 1.75 K, occurring over the region where North Atlantic TCs reach their peak intensities (20°—30°N, 60°—80°W). Over this region, the change inT100from the NCEP—NCAR data is higher than that from the MERRA data,regardless of whether the “Qua ” or “Std ” calculation method is used.Both the NCEP—NCAR and MERRA data show that the largest changes inT300between strong and weak TC peak intensity conditions occur over the region with North Atlantic TC peak intensities, and the changes are 1.25 K and 1.5 K for the “Qua ” and “Std ” calculation method, respectively. Between strong and weak western North Pacific TC peak intensity conditions, the differences inT600from both the NCEP—NCAR and MERRA data are more than 0.25 K for the “Qua ” calculation method,and more than 0.5 K for the “Std ” calculation method, over the region with TC peak intensities (15°—25°N, 120°—140°E). The difference inT925between strong and weak eastern North Pacific TC peak intensity conditions presents a dipole pattern, as shown by the MERRA data. At the higher and lower latitudes of the region with TC peak intensities (10°—20°N, 100°—130°W), there exists warming with magnitude greater than 1.25 K. However, this dipole pattern shifts southwards, as shown by the NCEP—NCAR data, and becomes weak in the “Std ” calculation method.

Fig. 2. Top two rows: Grid-point correlation between the storm-season mean peak intensity of TCs and JASO meridional mean air temperature (North Atlantic (NA):20°—30°N meridional mean; western North Pacific (WP): 15°—25°N meridional mean; eastern North Pacific (EP): 10°—20°N meridional mean) from NCEP—NCAR and MERRA (1979—2015), respectively. The solid blue contour lines represent values over the 95% confidence level. Bottom four rows: Difference between strong and weak TC peak intensity conditions (constructed with the Qua and Std methods (described in section 3.1 ), respectively) of the meridional mean air temperature (units:K) from NCEP—NCAR and MERRA data (1979—2015).

One important point is that TC peak intensity is not only controlled by temperature, while larger vertical wind shear also weakens the intensity of a TC. A detailed discussion of the relationship between TC peak intensity and vertical wind shear can be found in the Supplementary Material.

4. Summary and discussion

By applying NCEP—NCAR and MERRA reanalysis data, the respective associations of North Atlantic, western North Pacific, and eastern North Pacific TC peak intensities with regional air temperature were investigated in this study. Based on the correlation between TC peak intensity and air temperature and the difference in air temperature between strong and weak TC peak intensity conditions, we have discussed the impacts of tropospheric temperature and static stability on the peak intensity of real TCs.

Emanuel et al. (2013) studied North Atlantic TCs and found that cooling of the tropopause has an effect on TC MPI, which is consistent with our results for North Atlantic real TC peak intensity. However, this cooling of the tropopause accompanied by strengthening TC peak intensity does not occur for either the western or eastern North Pacific TCs. On the other hand, warming in the troposphere associated with strong TC peak intensity occurs for all North Atlantic and eastern North Pacific TCs. Specifically, it is in the upper troposphere for the North Atlantic TCs, but in the lower troposphere for the eastern North Pacific TCs. The physical mechanism for lower-tropospheric temperature affecting the intensity of a TC differs from that of tropopause temperature. Lower-tropospheric temperature (e.g., at 925 hPa) may mainly regulate the thermodynamic disequilibrium at the sea surface and hence TC intensity. Lower-tropospheric warming could promote a conditionally unstable atmosphere, encouraging moist adiabatic ascent through a deep layer of the atmosphere. On the other hand, tropopause temperature largely affects TC intensity via thermodynamic efficiency. More sea surface enthalpy supply can be converted to TC kinetic energy, increasing TC intensity, under a colder tropopause. The change in the eastern North Pacific TC peak intensity is larger than that of the North Atlantic TCs (Fig. S1), indicating that lower-tropospheric warming could induce a larger increase in TC peak intensity compared to tropopause cooling.Therefore, lower-tropospheric warming will lead to more occurrence of intense TCs.

Fig. 3. Grid-point correlation between the storm-season mean peak intensity of North Atlantic (NA), western North Pacific (WP), and eastern North Pacific (EP) TCs and JASO mean air temperature at 100, 300, 600, and 925 hPa ( T 100 , T 300 , T 600 , and T 925 ) from NCEP—NCAR and MERRA for the period 1979—2015. The solid blue contour lines represent values over the 95% confidence level.

Fig. 4. Difference between strong and weak TC peak intensity conditions (constructed with the Qua and Std methods (described in section 3.1 ), respectively) of JASO mean air temperature at 100, 300, 600, and 925 hPa ( T 100 , T 300 , T 600 , and T 925 ; units: K) from NCEP—NCAR and MERRA data (1979—2015).

Over the region where North Atlantic TCs reach their peak intensities (20°—30°N, 60°—80°W), colderT100and warmerT300associated with strengthening TC peak intensity appear at the same time. Therefore, the upper-atmospheric (50—400 hPa) stability has a great influence on the peak intensity of the North Atlantic TCs. This is consistent with the conclusions obtained by Downs and Kieu (2019) . They found that the impact of upper-tropospheric (200—500 hPa) stability is dominant compared to lower-tropospheric (500—1000 hPa) stability, for North Atlantic TCs. However, they did not examine eastern North Pacific TCs. In our study, warmerT925was found to accompany the intensifying peak intensity of the eastern North Pacific TCs appearing over the region 10°—20°N and 100°—130°W. Therefore, the lower-tropospheric(700—1000 hPa) stability has a great influence on the peak intensity of eastern North Pacific TCs. TC peak intensity is not only controlled by the atmospheric temperature, while larger vertical wind shear also weakens the intensity of a TC. It was found that the main factor affecting the peak intensity of western North Pacific TCs is not the atmospheric temperature, but the vertical wind shear, because the increase in the TC peak intensity strongly corresponds to the weakening of vertical wind shear between 200 and 850 hPa, compared to the atmospheric temperature.

Funding

This work was supported by the National Natural Science Foundation of China [grant numbers 41675062 and 41375096] and the Research Grants Council of the Hong Kong Special Administrative Region, China[project numbers CityU 11306417 and 11335316].

Acknowledgments

The MERRA data were obtained from the Goddard Earth Sciences Data and Information Services Center. The NCEP—NCAR reanalysis data were obtained from the Physical Sciences Laboratory of NOAA’s NCEP. The TC track data (1979—2015) used in this study are from a global TC dataset archived by the Massachusetts Institute of Technology as a related resource for the open course “Tropical Meteorology ” ( ftp://texmex.mit.edu/pub/emanuel/HURR/tracks/ ). In this dataset, the North Atlantic and eastern North Pacific TC track data were obtained from NOAA’s Tropical Prediction Center, and the western North Pacific TC track data were obtained from the U.S. Navy’s Joint Typhoon Warning Center.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi: 10.1016/j.aosl.2021.100117 .