TheDependenceofNorthwestPacificTropicalCyclone …downloads.hindawi.com/journals/amete/2019/9456873.pdf · 2019. 12. 12. · ResearchArticle TheDependenceofNorthwestPacificTropicalCyclone

Research ArticleThe Dependence of Northwest Pacific Tropical CycloneIntensification Rates on Environmental Factors

Xinyan Lyu12 Xuguang Wang 1 and Lance M Leslie 13

1School of Meteorology University of Oklahoma Norman Oklahoma 73072 USA2National Meteorological Center Beijing 100081 China3School of Mathematical and Physical Sciences University of Technology Sydney Sydney Australia

Correspondence should be addressed to Xuguang Wang xuguangwangouedu

Received 19 June 2019 Revised 18 September 2019 Accepted 22 October 2019 Published 7 December 2019

Academic Editor Budong Qian

Copyright copy 2019 Xinyan Lyu et al is is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

e Northwest Pacific tropical cyclone (TC) intensification is classified into rapid intensification (RI) normal intensification (NI)and slow intensification (SI) categoriese initial location and intensity the preceding intensity change the motion direction theoccurrence month and the intensification duration time are all found to differ for RI cases compared with NI and SI cases edependence of RI NI and SI on environmental conditions is further examined statistically by using the intensification rates ofnamed TCs for the 21-year period 1995ndash2015 obtained from JTWC best track data and the environmental conditions derivedfrom the ERA-Interim reanalysis data and GODAS high-resolution global ocean analysis data It was found that deep-layer andupper-mid vertical wind shear (VWS) upper-level outflow sea surface temperature (SST) and ocean heat content (OHC) arestatistically different among RI NI and SI both before and during intensification RI is enhanced by weaker and decreasing VWSwarmer oceans and stronger and increasing outflow In contrast SI typically occurs with larger and increasing VWS cooleroceans and weaker static outflowe impacts of low-level VWS and net moisture inflow are only significantly different betweenRI and SI and between NI and SI but not between RI and NI Another key finding is that increased upper-level outflow anddecreased VWS are important precursors and hence are possible predictors of RI onset e direction of upper-level outflowaffects TC intensification with NW and NE outflow being more favorable for TC RI than SE and SW outflow

1 Introduction

e operational prediction of tropical cyclone (TC) in-tensity change remains a major challenge Operationalforecasts often are required to distinguish rapid in-tensification (RI) and slow intensification (SI) from normalintensification (NI) [1ndash3] However during RI the forecastintensification rate (IR) typically is far lower than theobserved IR whereas it is much higher than the observedIR in the SI cases [3 4] In some operational TC forecastcenters (Figure 1) when forecasters are unsure of howrapidly TCs will intensify the default forecast issued is thatthe TC will undergo NI [3] e difficulty of distinguishingbetween RI NI and SI during the forecast is due partly tothe challenge of quantifying the relationship between thepredictors and the IR Most previous studies focused solelyon comparing environmental conditions between RI and

non-RI cases Some of these studies revealed statisticallysignificant differences in the environmental conditionsbetween RI and non-RI cases [5 6] However in thesestudies the non-RI category does not distinguish betweenNI SI and weakening storms Hendricks et al [7] classifiedthe intensification into two categories rapidly intensifyingand intensifying and showed that the environmentalconditions including sea surface temperature (SST) andother environmental parameters associated with these twocategories are similar e present study further exploresthe relationship between environmental conditions and TCIR and in contrast to previous studies also classifies the IRinto the three categories of RI NI and SI e mainmotivation for this study is toward addressing the need forand challenges of distinguishing between the various in-tensification phases present during the operational forecastperiod

HindawiAdvances in MeteorologyVolume 2019 Article ID 9456873 18 pageshttpsdoiorg10115520199456873

Most of the earlier studies have focused on hurricanes inthe North Atlantic or the East Pacific basins [5 6 8]However this study concentrates on TCs over the NorthwestPacific Basin (NWP) e NWP is controlled by the mon-soon background with about 70 of all TC genesis oc-curring in the monsoon trough [9 10] is is very differentfrom other ocean basins For example in the East Pacific andNorth Atlantic basins the main genesis source is AfricanEasterly Waves [11] Hence the relative importance of theindividual predictors for TC intensification is likely to bedifferent in various basins [6 12] Rios-Berrios and Torn [13]also suggested it is problematic to consider all basins jointlybecause the relative importance of environmental conditionsfor TC intensity change depends on the ocean basin

e present study seeks to address an important but asyet unanswered question ldquoDo significant differences inenvironmental factors including atmospheric conditionsand ocean heat conditions exist between rapidly normallyand slowly intensifying TCs in the NWPrdquo In this study arange of environmental factors is examined includingvertical wind shear (deep-layer upper-mid mid-low andlow-level) net moisture inflow upper-level outflow(strength and direction) and ocean heat conditions (bothSST and ocean heat content (OHC)) Earlier limited studiesof the NWP TC basin either examined different TC cate-gories [14] or focused only on one environmental factor [12]

Specifically the following environmental factors andtheir relationships with RI NI and SI are assessed Envi-ronmental vertical wind shear (VWS) is one of the mostimportant factors influencing TC intensity change and it iswidely accepted that strong VWS generally inhibits TCintensification [5 7 12 15ndash18] VWS also provides prob-ability estimates of RI in dynamical-statistical models[5 6 16 19] and can assist in determining if and when RIwill occur [20ndash22] e deep layer 200ndash850 hPa VWS iscommonly used whereas other studies showed that theVWS at different levels can also impact TC intensity change[12 23ndash25] Hence it is too simplistic to assess the impact ofVWS on TC intensity change by using a VWS calculatedfrom just two levels [26ndash29] In this study the VWS between

different vertical levels are used to analyze the impact ofVWS on the three TC intensification rates RI NI and SI

Environmental moisture is also suggested as a possibleimportant factor for TC intensification Both theoretical andmodeling studies showed that high environmental moisturefavors TC intensification [30 31] whereas dry air intrusioncan cause TC decay [21 32] e environmental relativehumidity (RH) of the lower and middle troposphere ishigher for RI cases than non-RI cases [5ndash7] However otherstudies suggested that moisture might have a negative or nosignificant impact on TC intensification [33ndash35] ereforethe impact of environmental moisture on TC IR requiresfurther more detailed examination In this study the im-pact of environmental moisture on the TC IR is studied byusing the low-level to mid-level net moisture inflow definedas the sum of net moisture inflow from all sides of a boxsurrounding the TC center and extending from the surfaceto 500 hPa is quantity is used to represent the total low-level to mid-level moisture transport from the environmentinto the TC core

Due to the asymmetric structure [36 37] lowRichardson number [38 39] and weaker inertial instability[40ndash43] of the TC upper-level outflow the outflow can easilyinteract with the upper-level larger-scale environment andthe inner-core of a TC thereby playing a mediating rolebetween the environment and the storm core [41] As such itcan affect the secondary circulation and consequently theTC intensity [37 41 43ndash49] e upper-level radial outflowis much stronger [6 37 43] and usually is more concentratedclose to the TC center for intensifying cases than for non-intensifying cases [43] In addition TC intensification isassociated with the multiple upper-level outflow channelsand TCs usually weaken when one of the efficient upper-level outflow channels is cut off [44 50ndash53] Kaplan andDeMaria [5] and Shu et al [14] noted that upper-tropo-spheric flows for RI cases are more easterly than for non-RIcases Rappin et al [41] showed that the ventilation ofoutflow in north and northeast directions minimizes thehurricane energy expenditure leading to larger in-tensification rates Hence the direction of TC outflowpossibly modifies the effect of outflow on TC intensitychange In this study the significant differences in theimpacts of both the strength and direction of outflow areexamined for all three different TC intensification cate-gories As far as the authors are aware this relationshipbetween outflow direction and TC IR has not previouslybeen statistically analyzed

e effect of the ocean heat conditions on TC intensitychange has been addressed for decades [54ndash57] SST iswidely known to affect TC intensification [5 6 58] Nu-merous supporting case studies exist [21 59ndash61] Statisticalanalyses [8 12] showed that TCs typically intensify (weaken)over warmer (colder) ocean eddy regions with higher(lower) SST Some studies have shown that the OHCmeasured from the surface to the depth of the 26degC isothermis important for TC intensity change [59] and that SSTaloneis not always a good measure of the total heat content thataffects TC intensity change [62] is hypothesis is furthersupported by studies showing that the use of OHC can help

0

10

20

30

40

50

60

70

80

90

100

0

5

10

15

20

25ndash7

0ndash6

5ndash6

0ndash5

5ndash5

0ndash4

5ndash4

0ndash3

5ndash3

0ndash2

5ndash2

0ndash1

5ndash1

0 ndash5 0 5 10 15 20 25 30 35 40 45 50 55 60

Cum

ulat

ive f

requ

ency

()

Freq

uenc

y (

)

TC intensification rate (unit kt 12hrndash1)

85

RINI

SI

95

Figure 1 Frequency distribution and cumulative frequency dis-tribution of the 12 hr intensity changes of all TCs in the 1995ndash2015in Northwest Pacific basin (10156 cases)

2 Advances in Meteorology

reduce TC intensity forecast errors in the statistical modelSHIPS [63 64] In the present study both SSTand OHC areused to determine the impact of ocean heat conditions on RINI and SI

In summary the main question that needs to beaddressed in this study is whether there are significantdifferences in large-scale atmospheric environmental con-ditions such as VWS at different vertical levels net envi-ronmental moisture inflow strength and direction of upper-level outflow and ocean heat conditions (SST and OHC)between RI NI and SI for NWPTCs Additionally a uniqueaspect of this study is the use of a relatively long period ofdata to derive statistically robust results as all named NWPTCs during the 21-year period from 1995 to 2015 are in-cluded e findings from this study are relevant to im-proving TC forecast skill and to increasing ourunderstanding of the TC intensification processes

Section 2 describes the data and provides definitions ofthe environmental conditions Section 3 defines the RI NIand SI categories and shows the distribution of TC intensitychanges in the NWP basin and Section 4 describes anddiscusses the main results Finally Section 5 summarizes thefindings of this study

2 Data and Methodology

21 Data e Joint Typhoon Warning Center (JTWC) besttrack data including the TC center position and maximum1min mean sustained 10m wind speed (Vmax) at 6 h in-tervals for all named TCs over the NWP basin from 1995 to2015 were used to calculate TC intensity change and todefine the TC environment

e 6-hourly wind and specific humidity fields and SSTdata from 1995 to 2015 are from the European Center forMedium-Range Weather Forecasts (ECMWF) interim re-analysis website (httpswwwecmwfintenforecastsdatasetsreanalysis-datasetsera-interim) [65] e sameECMWF reanalysis data were widely used in publishedstatistical studies to investigate the relationship between theenvironmental conditions and the tropical cyclone intensitychange (eg [25 66]) Schenkel and Hart show that for theECMWF interim reanalysis most of the TC position errorsare less than 150 km in the Northwest Pacific basin [67] estudy of Hodges et al showed that for the ECMWF interimreanalysis [68] the majority of the TCs in the reanalysis havea mean position difference from the best track that is lessthan 2deg (sim220 km) ey also showed that the peak of theposition error distribution is at about 110 km Hence mostof the TC position errors in the ECMWF reanalysis are lessthan 110 km [68] In the studies of Schenkel and Hart [67]and Hodges et al [68] the earlier ECMWF interim re-analysis was evaluated In our study we used the most recent20-year ECMWF reanalysis (1995ndash2015) Because this pe-riod is more recent and therefore the quality of the ECMWFinterim reanalysis data should be improved the TC positionerrors of the ECMWF reanalysis in our study should besmaller than that of Schenkel and Hart [67] and Hodges et al[68] In addition the horizontal grid spacing of the ECMWFinterim reanalysis is 05deg longitudetimes 05deg latitude In other

words the smallest resolvable scale is about sim100 kmcomparable to the peak position error in Hodges et alrsquosstudy [68] Moreover the computational domain which isused to calculate the environmental conditions in this studyis relatively large compared to the TC position error(sim110 km) Based on all of the above the TC position error inthe ECMWF interim analysis is not expected to impact ourresults So the reanalysis data are used to calculate differentlevel VWS upper-level outflow net moisture inflow andSSTaround the TC center during the intensification period

To study the impact of OHC the high-resolution globalocean analysis sea temperature data (34 levels up to 5000m12 km grid spacing) were provided by US GODAE (USGlobal Ocean Data Assimilation Experiment) (httpusgodaeorg) GODAE was shown by early studies toprovide high-quality ocean analysis data [69] and was usedto analyze the upper-ocean heat condition (eg [70]) esedata are used to calculate the OHC around the TC center inthis study Although the GODAE data have high spatialresolution they are only available after 2007 erefore onlya 9-year worth of data (2007ndash2015) were used Given thestatistical relationship between intensification rates andother environmental variables was studied using 20-yeardata caution should be taken to intercompare the OHCimpact with other environmental variables

22 Definitions of TC Environmental Conditions

221 Vertical Wind Shear e environmental vertical windshear VWS is defined as the wind vector difference betweentwo levels as shown in the following equation

VWSlev1minus lev2 Vlev1 minus Vlev21113868111386811138681113868

1113868111386811138681113868 (1)

where the Vlev1 and Vlev2 variables are annular-averagedhorizontal wind vector at level 1 and level 2 In this cal-culation the annular is between 200 and 800 km from thestorm centererefore the inner-core region of the storm isremoved or filtered when the environmental shear is cal-culated e subscripts ldquolev1rdquoand ldquolev2rdquo indicate two pres-sure levels is method of defining and calculating thevertical wind shear follows the studies [14 19] We calculate4 types of wind shear VWS200minus 850 VWS200minus 500 VWS500minus 850and VWS850minus 1000 as follows e deep-layer shearVWS200minus 850 is calculated It is commonly used to diagnosethe relationship between TCs intensity change and envi-ronmental vertical wind shear (eg [18 19]) 850 hPa500 hPa and 200 hPa are commonly used to analyze the low-level mid-level and upper-level circulations of tropicalcyclone erefore in addition to VWS200minus 850 VWS200minus 500and VWS500minus 850 are also used because early studies suggestthese shears can also impact TC intensity change [19 71]Moreover both the statistical and idealized simulationstudies found that the low-level shear could have a negativeeffect on TC intensification [12 24] In particular the studyby Wang et al [12] showed that the low-level shear between850 and 1000 hPa can have a larger negative impact on TCintensity change than VWS at other levels during activeseasons So the low-level shear is also calculated in this study

Advances in Meteorology 3

222 Net Moisture Inflow e net inflow of environmentalmoisture is used to assess the impact of environmentalmoisture on TC intensity change e net moisture inflowflux is defined as

Quv 1113946500hpa

surface(qe + qs + qw + qn)dp (2)

where qe qs qw and qn are the net moisture inflows fromthe eastern southern western and northern boundariesrespectively of a 8deg times 8deg latitude-longitude box around theTC center Hence Quv is the total net moisture transport tothe TC core from the surface to 500 hPae units ofQuv arekgmminus 2 sminus 1 In equation (2)

qe minus1g

qu

qs 1g

qv

qw 1g

qu

qn minus1g

qv

(3)

where q is the specific humidity u and v are the zonal windand meridional wind components respectively and g is98m sminus 2

223 Upper-Level Outflow e strongest TC outflow is onaverage concentrated at about 200 hPa and around 500 kmfrom the TC center [38 43] In this study

∯D

nablaF dS 1113929 Fn ds (4)

where from Greenrsquos law the upper-level outflow strength isdefined as the area-averaged divergence over the 8deg times 8deglatitude-longitude box around the TC center at 200 hPa

For simplicity the impact of outflow direction on IR isinvestigated only for cases with a single-outflow channel Asingle-outflow channel is defined as occurring when thequadrant-mean wind within 500 km from the storm center isoutward and greater than 6m sminus 1 In the Northern Hemi-sphere the anticyclonic outflow of TCs mainly has fourradial directions northwest (NW) northeast (NE) south-west (SW) and southeast (SE) erefore in this study onlythe NW NE SW and SE quadrants are considered

224 SST and OHC e OHC definition follows Buckleyet al [72] except that we choose the 26degC isotherm depth asthe integral depth during the vertical integration is re-placement was according to early studies (eg [73]) whichshowed that the TC genesis andor TC intensity change canbe sensitive to the depth of the 26degC isotherm Accordinglythe areal density of the OHC is defined as

OHC ρcp 11139460

h T26∘C( )T(z)dz (5)

where ρ is the seawater density cp is the specific heat of seawater h is the depth of the 26degC sea temperature contourand T(z) is the sea temperature at the depth z e unit of theareal density of ocean heat content is kJ cmminus 2 Hence OHCis the total ocean heat content per square centimeter fromthe depth of the 26degC contour sea temperature to the seasurface In this study the area-averaged SST and OHC overthe 2deg times 2deg square box around the TC center is used tocalculate the ocean heat conditions

23 Significant Difference Tests e testing of significantdifferences between mean values was carried out using astandard bootstrapping method [74 75] is method wasused to obtain confidence levels for the differences in thesample means of the various parameters in this study As anexample it was used to compare the mean intensificationrates for RI and SI in deep-layer VWS environments Hereconfidence levels of 90 and 95 are used and the numberof resamples in this study is 10000 e bootstrap techniquewas used in early studies by Wang and Bishop [75] andHamill et al [76]

3 Definitions and Distributions of RI NI andSI in the NWP Basin

31 Definitions of RI NI and SI e frequency and cu-mulative frequency distributions of 12 h intensity changesfor all 499 named TCs over the NWP basin during 1995ndash2015 are shown in Figure 1 As this study focuses on TCintensification only intensifying cases are chosen e IR isclassified into 3 categories rapid intensification (RI) normalintensification (NI) and slow intensification (SI) ese areformally defined in Table 1

Note that in earlier studies RI is usually defined as the95th percentile of all 24 h TC intensity changes ie ap-proximately 30 kt or greater [5 12 14] ese previousdefinitions did not include the 6 h TC intensity change Asshown later the average intensification duration time of thecontinuous RI process found in this study is 188 hrerefore using a 24 h period to define the intensificationperiod might also include a weakening phase Hence in thisstudy a 12 h period is used to define RI In addition toensure the selection of continuously intensifying caseswhere no weakening phases are included the definition ofintensification category is also constrained by the 6 h in-tensity change Specifically as shown in Figure 1 RI is de-fined as a 12 h intensity change ge20 kt (ie the 95thpercentage of TC intensity change for 12 hr) and a 6 h in-tensity change ge5 kt NI is defined as a 12 h intensity changebetween 10 kt and 20 kt (ie between 85th and 95th per-centage of TC intensity change for 12 hr) and a 6 h intensitychange is between 5 and 10 kt SI requires that the 12 hintensity change must lie between 0 kt and 10 kt (ie be-tween 60th and 85th percentage of TC intensity change for12 hr) and the 6 h intensity change between 0 kt and 5 ktFinally the total numbers of RI NI and SI cases examined inthis study are 763 1345 and 1218 respectively


32 Distributions of RI NI and SI In this section the sta-tistical distribution of RI NI and SI with respect to initiallocation initial intensity moving direction occurrencemonth and intensification duration time are discussed einitial TC center locations and intensities of the three in-tensity categories are shown in Figure 2 Hereafter TC in-tensities are classified into six categories tropical depression(TD 108m sminus 1leVmaxle 171m sminus 1) tropical storm (TS172m sminus 1leVmaxle 244m sminus 1) severe tropical storm (STS245m sminus 1leVmaxle 326m sminus 1) typhoon (TY 327m sminus 1leVmaxle 414m sminus 1) strong typhoon (STY 415 m sminus 1leVmaxle509m sminus 1) and super typhoon (Super TY Vmaxge 51msminus 1)based on the CMA (China Meteorological Administration)criteria of 2006 [77] Only named TCs are chosen in this studyand TDs are not included

From Figure 2 the TC data reveal that most RI casesoccur at latitudes between 10degN and 25degN whereas most NIcases occur between 5degN and 30degN and SI cases occur in aneven larger area between 5degN and 35degN RI cases mainly areconcentrated east of the Philippines and in the northernSouth China Sea For each intensification category thedistribution of initial intensity as a function of TC intensitygrade is shown in Figure 3 For RI cases the initial intensitymost frequently falls in the TY grade For NI and SI casesthe initial intensity most commonly falls in the TS intensitycategory e combined frequency of TY STY and superTY categories is about 64 for RI but only around 33 forboth NI and SI Overall the initial intensity of the RI casesgenerally is much greater than that of NI and SI However theRI distribution over the TY STY and super TY intensities inFigure 3 shows that fewer RI cases are initiated from the STYand super TY intensities than from TYs is possibly isbecause TYs are farther from their MPI (maximum potentialhurricane intensity [78]) than STYs and super TYs and hencehave greater potential to intensify [79]

e frequency distributions of the TC moving directionsfor the three intensification rate categories (SI NI and RI)are shown in Figure 4 Overall most of the intensificationcases (about 75) move northwest (including WNW andNNW) whereas only about 16 and 7 of the in-tensification cases move northeast (including NNE andENE) and west-southwest and very few cases movesouthward (including SSE and SSW) or east-southeastwardIt is also found that the percentage of west-northwestwardmotion is higher for RI cases than NI and SI cases whereasthe percentage of north-northwestward moving TCs is lowerfor RI cases than NI and SI cases at is RI cases tend tomove WNW more than NI and SI cases

Most of the RI cases (about 61) occur in AugustSeptember and October with a peak in September (Figure 5)

but about 56 of NI and SI cases occurmainly in July Augustand September with an earlier peak in August As expectedvery few TC intensification cases occur in January Februaryand March

As shown in Figure 6 the mean intensification durationtimes for a continuous RI NI and SI process are 188 hours183 hours and 146 hours respectively About 70 of theRI and NI cases intensify for 12ndash18 hours and only about30 of the RI and NI cases continue for 24 hours or longer(Figure 6) Very few TCs can intensify rapidly or normallyfor longer than 36 hours Most SI cases (about 71) in-tensify only for 12 hours Overall RI has the longest in-tensification period and SI has the shortest As discussedearlier these results form the basis to define different in-tensification categories using the 12 h period rather thanthe 24 h period

4 Results of Dependence of IR onEnvironmental Factors

Section 32 shows the statistical distributions of initial lo-cation initial intensity moving direction occurrence monthand the intensification duration time which are differentamong the RI NI and SI cases is finding warrants anassessment of their dependence on environmental conditions

As discussed in Section 1 the selection of the envi-ronmental conditions is based on factors that are known tostrongly influence TC intensity change e variables se-lected are deep-layer shear VWS200minus 580 the upper-mid shearVWS200minus 500 and mid-low- shear VWS500minus 850 and low-levelshear VWS850minus 1000 the net environmental moisture inflowthe upper-level outflow and ocean heat condition (includingboth SST and OHC)

41 Environmental Vertical Wind Shear Figure 7 shows theVWS before during and after the intensification for the threeintensification categories e significant differences of themean magnitudes between the three intensification categoriesare tested using a bootstrap technique Overall the deep-layershear and the upper-mid shear are significantly (greater than95 confidence level) lower before during and after the RIprocess than the NI and SI processes is is not the case forthe low-level shears Notably though not statistically signif-icantly different the average low-level shear VWS850minus 1000 isslightly higher for RI than NI For mid-low VWS500ndash850consistent significant differences exist between RI and SI butthe differences between RI and NI are significant only starting12 hours before the onset of the intensification and during theintensification (greater than 90 confidence level) ese

Table 1 Definitions and sample sizes of the three intensification categories

Category ΔV12hmax ΔV6h

max Num Percentage

SI 0ltΔV12hmax lt 10 kt 0leΔV6h

max le 5 kt 1218 366NI 10leΔV12h

max lt 20 kt 5ltΔV6hmax lt 10 kt 1345 405

RI ΔV12hmax ge 20 kt ΔV6h

max ge 5 kt 763 229ΔV6h

max and ΔV12hmax are TC intensity changes for 6 hr and 12 hr periods respectively


results suggest that RI is distinguished fromNI and SI primarilyby deep-layer upper-level shear andmid-lowVWS rather thanthe low-level VWS In addition VWS200minus 850 VWS200minus 500VWS500ndash850 and VWS850minus 1000 are also significantly (greaterthan 90 confidence level) lower during the NI processes thanthe SI processes except during the onset period forVWS500ndash850 ese results show that SI can be distinguishedfrom NI by all VWS

On average there are clear decreasing trends of VWSbefore the onset of RI except for VWS850minus 1000 (Figure 7)edecreasing trends commence at least 24 hours prior to the

onset of RI For example VWS200ndash850 and VWS200ndash500 reachtheir minima at the onset of RI (Figures 7(a) and 7(b))VWS500minus 850 decreases to a minimum at 12 hours prior to theonset of RI (Figure 7(c)) Further diagnostics (not shown)reveal that the decreasing trends are accompanied by areduction in vortex tilt

For NI and SI the VWS usually exhibits slight changesbefore the onset of RI In particular VWS500ndash850 shows aslight increasing trend for NI and SI before the onset of RIIn other words in addition to themagnitude the variation ofVWS differs for the three TC intensification categories and

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ

e typhoon center of RI

TSSTSTY

STYSuperTY

(a)

e typhoon center of NI

TSSTSTY

STYSuperTY

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ

(b)

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E

e typhoon center of SI

TSSTSTY

STYSuperTY

140E 150E 160E 170E 180EQ

(c)

Figure 2 e location distribution of (a) RI cases (b) NI cases and (c) SI cases TC intensities over the Northwest Pacific basin are dividedinto five categories tropical storm (TS 172m sminus 1leVmaxle 244m sminus 1) severe tropical storm (STS 245m sminus 1sle Vmaxle326m sminus 1) typhoon (TY 327m sminus 1leVmaxle 414m sminus 1) strong typhoon (STY 415m sminus 1leVmaxle 509m sminus 1) and super typhoon(SuperTY Vmaxge 510m sminus 1)


therefore can serve as one of the preceding signals forpredicting the three intensification categories

Figure 8 shows the frequency distributions of the threeintensification categories as a function of VWS ForVWS850ndash1000 the percentage peak for all three intensificationcategories corresponds to the same shear value of 1-2m sminus 1For shear at other levels VWS200ndash850 VWS200ndash500 andVWS500ndash850 the percentage peak of RI corresponds to asmaller VWS shear than NI and SI Specifically the per-centage peak of RI is at 2ndash6m sminus 1 for VWS200ndash850 2ndash4m sminus 1

for VWS200ndash500 and about 2-3m sminus 1 for VWS500ndash850 Incomparison the percentage peak for NI and SI is about

2-3m sminus 1 larger than RI NI and SI are also further com-pared For VWS200ndash850 and VWS200ndash500 the percentage peakfor NI is at a smaller shear than SI whereas for VWS500ndash850and VWS850ndash1000 they both peak at the same shear Figure 8also shows that the shape of the frequency distribution as afunction of VWS is different among all three intensificationcategories e distribution of RI is sharper and has a higherpeak value than NI and SI for VWS200ndash850 VWS200ndash500 andVWS500ndash850 followed by NI However for VWS850ndash1000 thedistributions among the three categories are in generalsimilar And for VWS500ndash850 the percentage peaks of NI andSI are both at 2-3m sminus 1 the differences just are that that thepercentage and peak value is slightly higher for NI than SIwhen VWS500ndash850 is less than 4m sminus 1 ese results furthersuggest that deep-layer and upper-mid VWS of the threeintensification categories are more significantly different andplay more important roles in distinguishing the three in-tensification categories

Figure 9 shows the relative occurrence probability ofthe three intensification categories in each VWS bin Ev-idently the RI occurrence probability decreases linearlywith increasing deep-layer VWS (VWS200minus 850) and upper-mid VWS (VWS200minus 500) It decreases only marginally withincreasing mid-low VWS (VWS200minus 850) RI probabilitydistribution is almost flat when the low-level shear(VWS850minus 1000) is less than 3m sminus 1 and decreases with in-creasing low-level shear when VWS850minus 1000 is greater than3m sminus 1 is result suggests that the RI occurrenceprobability is more sensitive to the deep layer and mid-upper VWS e NI mean probability decreases only withincreasing VWS850minus 1000 and with the increasing VWS500minus 850if the VWS500minus 850 is higher than 4m sminus 1 It is almost flatacross the VWS200minus 850 bin However it increases slightlywith the increasing VWS200minus 500 ese results indicate thatthe NI occurrence probability is more sensitive to theupper-mid and low-level VWS than VWS at other levels

0

10

20

30

40

TS STS TY STY SuperTY

Freq

uenc

y (

)

SINIRI

Figure 3 e frequency distribution of the initial intensity of thedevelopment in the three intensification categories (RI NI and SI)for the period 1995ndash2015 e definition of TC intensities cate-gories is the same as in Figure 2

Freq

uenc

y (

)

0

10

20

30

40

50

60

NNE ENE ESE SSE SSW WSW WNW NNWMoving direction

SINIRI

Figure 4 e frequency distribution of the moving direction forthe three intensification categories (RI NI and SI)

Freq

uenc

y (

)

0

5

10

15

20

25

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

SINIRI

Figure 5 e frequency distribution of the initial month of thedevelopment for the three intensification categories (RI NI and SI)in the Northwest Pacific basin


e mean occurrence probability of SI increases especiallywith an increasing deep-layer and low-level VWS sug-gesting its occurrence is more sensitive to the variation ofthe shear at these levels

Notably the error bar range is much larger when VWS isstronger especially for VWS200minus 850 VWS200minus 500 andVWS850minus 1000 is result shows that the occurrence proba-bilities of SI NI and RI at large shear are with high un-certainty suggesting that the probabilities of SI NI and RIare more difficult to estimate under strong environmentalVWS ese results are consistent with the study of Zhangand Tao [20] that suggests that the larger the VWS the largerthe uncertainty of the predictability of TC intensity Al-though the RI occurrence probability shows a decreasingtrend for larger shear values for all levels and the error barsof occurrence probabilities are much larger in low-level thanother levels especially when low-level shear is larger than4msminus 1 is result together with those from Figures 7 and 8show that RI NI and SI are overall better delineated by thedeep-layer and upper-mid shears than the mid-low and low-level shears e role of mid-low and low-level shears ondistinguishing the three intensification categories is mixedese results are in general consistent with earlier studiesshowing the VWS at different levels can impact TC intensitychange differently [12 23ndash25]

42 Environmental Moisture Condition Earlier studies onthe impact of environmental moisture on the TC in-tensification produce mixed results [6 7 33 35] is studyreexamines the impact of the net environmental moisture indistinguishing RI NI and SI e net moisture inflow (Quv)which measures the moisture transport from the environ-ment to the TC core is used For each intensification cat-egory (SI NI and RI) the net moisture inflow changesbefore and during the intensification are shown in

Figure 10(a) e net moisture inflow apparently increasesbefore during and after the RI and NI processes e in-creasing rate of the net moisture inflow is much smaller forSI On average the net moisture inflow is slightly higher forRI followed by NI and then SI e difference in the netmoisture inflow between SI and RI (NI) is significant at the90 level just after the onset of the intensification Howeverthe difference in net moisture inflow between RI and NI isnot statistically significant

Figure 10(b) shows the percentage distribution of thethree intensification categories as a function of net moistureinflow e percentage peak of RI corresponds to a netmoisture inflow of 20ndash30 kg mminus 2 sminus 1 about 10 kg mminus 2 sminus 1

larger than that of SI As expected the distribution of NI is inbetween showing a near flat top between 10 and 30 kgmminus 2 sminus 1 Results from Figure 10 are in general consistent withearly studies which show that the environmental moisture ofthe lower and middle troposhere is higher for RI cases thannon-RI cases [6 7] However when the intensification casesare further delineated among RI SI and NI this study findsthe net moisture inflow before during and after the in-tensification is not significantly higher for RI thanNI and themoisture inflow of RI and NI is only significantly higher thanSI after the intensification

43 Upper-Level Outflow

431 Strength of Upper-Level Outflow e upper-leveloutflow strength is defined as the area-averaged divergenceover a (8deg times 8deg) box around the TC center at 200 hPaFigure 11(a) shows the variation of the mean upper-leveloutflow strength 24 hours before and after the onset of theintensification for the three intensification categories Onaverage the upper-level outflow strength increases prior tothe RI and NI onset and continues strengthening thereafterIn contrast there is little variation of outflow strength for SIcases Moreover the outflow is strongest in RI followed byNI e outflow strength difference between RI and NI isstatistically significant at the 90 level SI shows the weakestoutflow with SI being statistically significant weaker than RIand NI at the 95 level

Figure 11(b) shows the frequency distribution of thethree intensification categories as a function of outflowstrength e frequency peak of RI occurs at a larger outflowstrength value than NI and SI Specifically the frequencypeak of RI is at 09times10minus 5 sminus 1 while those of NI and SI are atabout 06times10minus 5 sminus 1 Although NI and SI have similar peakoutflow strength the frequency distribution suggests that SIoccurs more frequently than NI when the outflow strength issmaller than 12times10minus 5 sminus 1 and vice versa Consistent withFigure 11(a) these results further suggest that the occur-rence of RI is associated with stronger outflow strength thanboth NI and SI and SI usually occurs under the environmentwith the weaker outflow In other words the three in-tensification categories occur under significantly differentoutflow strengths ese results are consistent with theprevious studies that the upper-level radial outflow is muchstronger for intensifying cases than nonintensifying cases

Freq

uenc

y (

)

0

10

20

30

40

50

60

70

12 18 24 30 36 42 48 54 60 66 72The intensification duration time (unit hour)

SI (146 hr)NI (183 hr)RI (188 hr)

Figure 6e frequency distribution of the intensification durationtime of the three intensification categories (RI NI and SI) for1995ndash2015


[43] But our results focus only on the intensification caseswithout including weakening cases and examine the asso-ciation of the occurrence of difference intensification rateswith the strength of the outflow

432 Direction of Upper-Level Outflow For each in-tensification category the distributions of the occurrenceprobabilities in the four outflow directions (NW SW NEand SE) are shown in Figure 12(a) In brief different outflowdirections favor different intensification rates RI occursmore frequently under NW and NE outflow environmentsthan under SW and SE outflow environments In contrastNI occurs more frequently under NW and SW outflow

environments and occurs least frequently under a SE outflowenvironment SI most likely occurs within SE outflow fol-lowed by the SW NE and NW outflow

Figure 12(b) shows the variation of the mean in-tensification rate in the four outflow directions whichfurther underlines the sensitivity of the intensification rate toboth the outflow direction and strength For outflowstrength greater than 8ndash10m sminus 1 the intensification rateassociated with the NW and NE outflows is higher than thatwith the SW and SE outflows e intensification rate withSW outflow is the second highest and the lowest in-tensification rates are under SE outflow For the weakoutflow strength (6ndash8m sminus 1) intensification rates associatedwith the NW SW and NE outflows are similar and higher

ndash24h ndash18h ndash12h ndash6h 0 6h 12h 18h 24h45

5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)

Figure 7e change in environmental vertical wind shear (VWS m sminus 1) (a) deep-layer shear VWS200minus 850 (b) upper-mid shear VWS200minus 500(c) mid-low shear VWS500minus 850 (d) low-level shear VWS850minus 1000 of RI (green) NI (red) and SI (blue) process 0 of x-axis is the onset time ofRI and minus 24 h minus 18 h minus 12 h and minus 6 h are 24 hours 18 hours 12 hours and 6 hours before intensification e thick and thin error bars arethe 95 and 90 confidence intervals respectively


than that of the SE outflow In short the larger in-tensification rate is associated with the NW and NE outflowand a slower intensification rate is associated with SE out-flow Additional calculation (not shown) also shows that theNE and NW outflow environment is often associated with amuch lower mean VWS than the other two outflow di-rections e SE outflow is accompanied with the largestmean VWS In addition however the mean VWS decreasesprior to the onset of intensification in the NE and NWoutflow environment and the mean VWS for the SE outflowincreases In brief significant outflow differences exist

between the three intensification categories For the sameoutflow strength faster intensification is most favored in theNW and NE outflows and least favored in the SE outflow Asfar as the authors are aware this study is the first publishedwork that provides the statistic relationship between di-rection of outflow and intensification categories

e relative eddy momentum flux convergence (REFC)has been used as a diagnostic to identify andor measure theupper-level TC-environment interaction [80ndash83] In thisstudy the REFC was calculated at 200 hPa over 300ndash600 kmradially from the TC center for all three TC IR categories

SINIRI

0

5

10

15

20

25

30

0ndash2 2ndash4 4ndash6 6ndash8 8ndash10 10ndash12 12ndash14 ge14

Freq

uenc

y (

)

VWS 200ndash850hPa (unit m sndash1)

(a)

SINIRI

0

5

10

15

20

25

30

35

40

0ndash2 2ndash4 4ndash6 6ndash8 8ndash10 10ndash12 ge12

Freq

uenc

y (

)

VWS 200ndash500hPa (unit m s ndash1)

(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)

Figure 8e frequency distribution of the three intensification categories (RI NI and SI) as a function of vertical wind shear (VWS m sminus 1)(a) deep-layer shear VWS200minus 850 (b) upper-mid shear VWS200minus 500 (c) mid-low shear VWS500minus 850 (d) low-level shear VWS850minus 1000


However consistent with Shu et al [14] no significantdifferences were found between the categories (not shown)

44 Ocean Conditions is section studies the statisticalrelationship between ocean conditions and intensificationrates Note as stated at the end of Section 21 the OHC isbased on data for a shorter 9-year period (2007ndash2015) due tothe availability of the GOADE analysis data Figure 13 showsthe variation of SST and OHC 24 hours before and after theonset of intensification For all three intensification cate-gories both SST and OHC decrease during the 48-hourperiod As shown in Section 32 most intensification casesmove from warm pool toward northwest e ocean is

generally warmer in the tropical eastern and southern areascompared with the tropical western and northern NWPbasin Figure 13 also shows that SST and OHC are thehighest for RI followed by NI and then SI eir differencesare statistically significant at least at the 90 confidenceinterval

Figure 14 shows the frequency distributions of the threedifferent intensification categories as functions of SST andOHC For SST the frequencies of the three intensificationcategories all peak at 295degC (Figure 14(a)) but the frequencpeaks of RI and NI are higher than SI Notably the in-tegrated frequency of RI is higher than NI and SI and theintegrated frequency of SI is the lowest when the SSTis above29degC In contrast the integrated frequency of RI is lower

0ndash2 2ndash4 4ndash6 6ndash8 8ndash10 10ndash12 12ndash14 gt14VWS 200ndash850hPa (unit m sndash1)

0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)

0ndash2 2ndash4 4ndash6 6ndash8 8ndash10 10ndash12 gt12VWS 200ndash500hPa (unit m sndash1)

0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 gt8VWS 500ndash850hPa (unit m sndash1)

0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)

0-1 1-2 2-3 3-4 4-5 gt5VWS 850ndash1000hPa (unit m sndash1)

0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)

Figure 9e probability distribution of the three intensity change categories (SI blue NI red and RI green) in each VWS bin (m sminus 1) (a)deep-layer shear VWS200minus 850 (b) upper-mid shear VWS200minus 500 (c) mid-low shear VWS500minus 850 (d) low-level shear VWS850minus 1000 e thickand thin error bars are the 95 and 90 confidence intervals respectively


than NI and SI and the frequency of SI is the highest whenthe SST is below 28degC is result further suggests that fasterintensification usually occurs over the warmer SST In ad-dition Figure 14(a) shows that only a small percentage(about 5) of RI cases occurs with SST below 28degC isvalue is far higher than the threshold (265degC) for TC genesisand intensification [84]

For OHC the frequency distributions of the three in-tensification categories are also different (Figure 14(b)) efrequency of SI is the highest followed by NI and then RI

when OHC is less than 40 kJ cmminus 2 However the ranking isreversed when the OHC is above 70 kJ cmminus 2 More than 50of the RI cases occur over the deeper and warmer oceanwhere the OHC is above 70 kJ cmminus 2 but less than 25 SIcases occur there e distribution of NI lies between SI andRI is result further suggests that faster intensificationusually occurs more frequently under deeper warmer oceanconditions

e occurrence probabilities of the three intensificationcategories under the same ocean heat condition are shown in


21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40

lt10 10ndash20 20ndash30 30ndash40 40ndash50 ge50

Freq

uenc

y (

)

Quv flux (kg mndash2sndash1)

SINIRI

(b)

Figure 10 (a) As in Figure 7 but for net moisture inflow (kg mminus 2 sminus 1) (b) As in Figure 8 but for net moisture inflow

SINIRI

ndash24h ndash18h ndash12h ndash6h 0 6h 12h 18h 24h

09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25

03 06 09 12 15 18 21 gt21Outflow strenghth (10ndash5 Sndash1)

Freq

uenc

y (

)

(b)

Figure 11 (a) As in Figure 7 but for the change in upper-level outflow strength (10minus 5 sminus 1) averaged over a 8deg times 8deg square around TC center(b) As in Figure 8 but for upper-level outflow strength


Figure 15 First the occurrence of RI is rarer than NI and SIfor all SST e occurrence probability of RI increases whenthe SST is above 27degC (Figure 15(a)) e occurrenceprobability of SI decreases with an increasing SST whereasthe occurrence probability of NI does not vary with the SSTe error bars are larger when the SST is lower ereforewhen the ocean is relatively cold the occurrence of the threedifferent intensification cases is more uncertain than when

the ocean is warmer e occurrence probability as afunction of OHC is in general similar to SST (Figure 15(b))But the occurrence probability of RI increases more sig-nificantly with increase of OHC when OHC is less than 90 kJcmminus 2 and decreases slightly when OHC is more than 90 kJcmminus 2

erefore the warmer the ocean is the higher occur-rence probability of RI is and the lower occurrence

10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35

6ndash8 8ndash10 10ndash12 12ndash14 gt14Outflow strength ( m sndash1)

NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)

Figure 12 (a)e frequency of RI NI and SI in four outflow directions e number of the below direction is the total number of the casesfor each single outflow direction (b) e average intensification rate for 12 hr of the four outflow directions in each single outflow strengthbin (m sminus 1) NW NE SE SW is northwest northeast southeast and southwest respectively

SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)

Figure 13 As in Figure 7 but for (a) SST (degC) and (b) OHC (kJ cmminus 2)


probability of SI is To some extent this result is consistentwith Xu and Wang that shows that the maximum potentialintensification rate increases with increasing SST [79] isstudy however emphasizes the dependence of RI NI and SIon ocean conditions

5 Conclusions

Using statistical approach this study aims to addressthe following question ldquoDo significant differences in

environmental factors including atmospheric conditions andocean heat conditions exist between rapidly normally andslowly intensifying TCs in the North West Pacific BasinrdquoBased on the JTWC best track data the intensification rates ofall named TCs over the NWP for the 21-year period 1995ndash2005 were categorized into three intensity categories RI NIand SI e initial location and intensity the preceding in-tensity change the moving direction the occurrence monthand the intensification duration time of TC intensification areall found to differ for RI cases compared with NI and SI cases

SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25

lt30 30ndash40 40ndash50 50ndash60 60ndash70 70ndash80 80ndash90 90ndash100OHC (kJ cmndash2)

Freq

uenc

y (

)(b)

Figure 14 As in Figure 8 but for ocean heat condition (a) SST (degC) (b) OHC (kJ cmminus 2)

SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI

lt30 30ndash40 40ndash50 50ndash60 60ndash70 70ndash80 80ndash90 ge90OHC (kJ cmndash2)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)

Figure 15 e same as in Figure 9 but for ocean heat condition (a) SST (degC) and (b) OHC (kJ cmminus 2)


Most RI cases occur east of the Philippines and over thenorthern South China Sea e mean initial intensity of RIcases is stronger than that of both NI and SI e precedingintensity change is much faster for RI than for NI and SI andRI cases tend to move WNW more than the NI and SI TCse mean intensification duration time of both RI and NI isfound to be about 18hr whereas it is only about 14hr for SIe peak active month for RI is found to be September but isAugust for NI and SI Using the ERA-Interim reanalysis dataand the high-resolution global ocean analysis data the dif-fering impacts of the large-scale environmental atmosphericand ocean heat conditions between RI NI and SI were ex-amined Overall there are statistically significant differencesbetween RI and NI or RI and SI responses to deep-layerVWS200minus 850 upper-mid VWS200minus 500 andmid-low VWS500minus 850upper-level outflow SST and OHCe difference between RIand SI is significant for low-level VWS850minus 1000 For netmoisture inflow the difference between RI and SI issignificant only after intensification onset ere is nostatistically significant difference between RI and NI inthe net moisture inflow and the low-level VWS850minus 1000e differences between NI and SI significantly exist indeep-layer VWS200minus 850 upper-mid VWS200minus 500 and low-level VWS850minus 1000 upper-level outflow and SSTOHCbefore during and after the intensification but only existin mid-low VWS500minus 850 and net moisture inflow afterintensification onset ese results are summarized inTable 2

In summary RI usually occurs under lower and de-creasing VWS enhanced and higher net moisture inflowdeeper warm ocean conditions and stronger and increasingupper-level outflow In contrast SI often occurs in higherand increasing VWS slightly increasing and lower netmoisture inflow colder ocean conditions and weaker andstatic upper-level outflow NI easily occurs in the envi-ronmental conditions that lie in between those of SI and RIe deep-layer VWS and upper-mid layer VWS are moreimpactful in distinguishing the three intensification ratecategories than mid-low and low-level VWS e analysisalso reveals that the direction strength and the increasingtrend of outflow all can impact the rate of TC intensificationFor example both the NW andNE outflows favor RI but theSE outflow is often associated with slower intensification Inaddition the differing impacts of OHC were found to behighly significant and more clearly distinguish the threedifferent intensification categories than with only the SST

is study is only limited to examine the statistical re-lationship between RI NI and SI and the individual large-scale environmental factors It is possible that favorableconditions sometimes can cancel the negative impact ofunfavorable conditions on TC intensification For example inthe situation of weak vertical wind shear and the high upper-ocean heat content TC may not intensify if other environ-mental conditions are not favorable In other words the TCintensity change largely depends on the integrated impact ofall the environmental factors e results of this study indeedverify that VWS outflow ocean heat condition and moistureflux can all impact TC intensification In other words theresult of the current study also suggests that one should

consider all these environmental factors when determiningand predicting the TC intensification categories Although thefocus of this paper is not to directly improve SHIPS (eg[64]) the results of the study suggest SHIPS may benefit fromusing additional predictors such as the trend of the VWS thetrend of the upper-level outflow strength and the direction ofoutflow during the TC intensity forecast

It is also found that both the strength and direction ofupper-level outflow can impact TC intensification rates Pre-cisely how the outflow direction and how the interaction ofoutflow and the wind shear affect intensity intensification ratesare still not clear which will require further modeling studiesDifference of moisture conditions between the RI and NIcategories is not significant However this does not imply thatthe moisture conditions are unimportant for TC in-tensification To some extent the monsoon background overNWP basin usually provides favorable moisture condition forTC intensification during an active TC season It is also foundthat the occurrence probability of both RI and the mean in-tensification rate decreases with an increase ofQuvwhenQuv ishigher than 40kg mminus 2 sminus 1 (not shown) suggesting that ex-cessivemoisture is not favorable for TCRI In addition satelliteobservations showed that the impact of moisture on TC in-tensity change depends on its quadrant distribution or relativelocation to the TC motion [35 85] is study uses statisticalapproach to associate the three intensification categories withthe environmental factors Additional modeling studies shouldbe planned to understand the dynamical and thermodynamicalprocesses that explain the relationship between the large-scaleenvironment and the three TC intensification rates

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Acknowledgments

e research documented in this paper was supported byONR grants N00014-14-1ndash0125 and N000141712111

Table 2 Significant differences between the three intensificationcategories

Environmental factors RI and NI RI and SI SI and NIDeep-layer VWS lowastlowastlowast lowastlowastlowast lowastlowastlowastUpper-mid VWS lowastlowastlowast lowastlowastlowast lowastlowastlowastMid-low VWS lowastlowast lowastlowastlowast lowastlowastLow-level VWS times lowastlowast lowastlowastlowastNet moisture inflow times lowastlowast lowastlowastUpper-level outflow lowastlowast lowastlowastlowast lowastlowastlowastSST lowastlowastlowast lowastlowastlowast lowastlowastOHC lowastlowastlowast lowastlowastlowast lowastlowastlowastlowastlowastlowast and lowastlowast denote significant differences with 95 and 90 confidenceintervals respectively times shows that no significant difference exists


References

[1] R L Elsberry T D B Lambert and M A Boothe ldquoAccuracyof Atlantic and eastern North Pacific tropical cyclone intensityforecast guidancerdquo Weather and Forecasting vol 22 no 4pp 747ndash762 2007

[2] R Gall J Franklin F Marks E N Rappaport and F Toepferldquoe hurricane forecast improvement projectrdquo Bulletin of theAmerican Meteorological Society vol 94 no 1 pp 329ndash3432013

[3] N Wei J L McBride X Zhang and Y Duan ldquoUn-derstanding biases in tropical cyclone intensity forecast er-rorrdquo Weather and Forecasting vol 33 no 1 pp 129ndash1382018

[4] A Simon A B Penny M DeMaria et al ldquoA description ofthe real-time HFIP corrected consensus approach (HCCA)for tropical cyclone track and intensity guidancerdquo Weatherand Forecasting vol 33 no 1 pp 37ndash57 2018

[5] J Kaplan and M DeMaria ldquoLarge-scale characteristics ofrapidly intensifying tropical cyclones in the north AtlanticbasinrdquoWeather and Forecasting vol 18 no 6 pp 1093ndash11082003

[6] J Kaplan M DeMaria and J A Knaff ldquoA revised tropicalcyclone rapid intensification index for the atlantic and easternnorth pacific basinsrdquo Weather and Forecasting vol 25 no 1pp 220ndash241 2010

[7] E A Hendricks M S Peng B Fu and T Li ldquoQuantifyingenvironmental control on tropical cyclone intensity changerdquoMonthly Weather Review vol 138 no 8 pp 3243ndash3271 2010

[8] J Xu Y Wang and Z-M Tan ldquoe relationship between seasurface temperature and maximum intensification rate oftropical cyclones in the north atlanticrdquo Journal of the At-mospheric Sciences vol 73 no 12 pp 4979ndash4988 2016

[9] J L McBride ldquoTropical cyclone formationrdquo in Chapter 3 ofGlobal Perspectives on Tropical Cyclones R Elsberry Edpp 63ndash105 World Meteorological Organization GenevaSwitzerland 1995

[10] E A Ritchie and G J Holland ldquoLarge-scale patterns asso-ciated with tropical cyclogenesis in the western PacificrdquoMonthlyWeather Review vol 127 no 9 pp 2027ndash2043 1999

[11] C orncroft and K Hodges ldquoAfrican easterly wave vari-ability and its relationship to atlantic tropical cyclone activ-ityrdquo Journal of Climate vol 14 no 6 pp 1166ndash1179 2001

[12] Y Wang Y Rao Z-M Tan and D Schonemann ldquoA sta-tistical analysis of the effects of vertical wind shear on tropicalcyclone intensity change over the western north pacificrdquoMonthly Weather Review vol 143 no 9 pp 3434ndash3453 2015

[13] R Rios-Berrios and R D Torn ldquoClimatological analysis oftropical cyclone intensity changes under moderate verticalwind shearrdquo Monthly Weather Review vol 145 no 5pp 1717ndash1738 2017

[14] S Shu J Ming and P Chi ldquoLarge-scale characteristics andprobability of rapidly intensifying tropical cyclones in thewestern north Pacific basinrdquoWeather and Forecasting vol 27no 2 pp 411ndash423 2012

[15] T A Cram J Persing M T Montgomery and S A BraunldquoA Lagrangian trajectory view on transport and mixingprocesses between the eye eyewall and environment using ahigh-resolution simulation of hurricane bonnie (1998)rdquoJournal of the Atmospheric Sciences vol 64 no 6 pp 1835ndash1856 2007

[16] M DeMaria and J Kaplan ldquoAn updated statistical hurricaneintensity prediction scheme (SHIPS) for the Atlantic and

eastern North Pacific basinsrdquo Weather and Forecastingvol 14 no 3 pp 326ndash337 1999

[17] M Riemer M T Montgomery and M E Nicholls ldquoA newparadigm for intensity modification of tropical cyclonesthermodynamic impact of vertical wind shear on the inflowlayerrdquo Atmospheric Chemistry and Physics vol 10 no 7pp 3163ndash3188 2010

[18] Z Zeng Y Wang and C-C Wu ldquoEnvironmental dynamicalcontrol of tropical cyclone intensity-an observational studyrdquoMonthly Weather Review vol 135 no 1 pp 38ndash59 2007

[19] J A Knaff C R Sampson and M DeMaria ldquoAn operationalstatistical typhoon intensity prediction scheme for the westernnorth pacificrdquo Weather and Forecasting vol 20 no 4pp 688ndash699 2005

[20] F Zhang and D Tao ldquoEffects of vertical wind shear on thepredictability of tropical cyclonesrdquo Journal of the AtmosphericSciences vol 70 no 3 pp 975ndash983 2013

[21] D Tao and F Zhang ldquoEffect of environmental shear sea-surface temperature and ambient moisture on the formationand predictability of tropical cyclones an ensemble-meanperspectiverdquo Journal of Advances in Modeling Earth Systemsvol 6 no 2 pp 384ndash404 2014

[22] E B Munsell F Zhang J A Sippel S A Braun and YWengldquoDynamics and predictability of the intensification of hur-ricane edouard (2014)rdquo Journal of the Atmospheric Sciencesvol 74 no 2 pp 573ndash595 2017

[23] S Shu Y Wang and L Bai ldquoInsight into the role of lower-layer vertical wind shear in tropical cyclone intensificationover the western North Pacificrdquo Acta Meteorologica Sinicavol 27 no 3 pp 356ndash363 2013

[24] P M Finocchio S J Majumdar D S Nolan andM Iskandarani ldquoIdealized tropical cyclone responses to theheight and depth of environmental vertical wind shearrdquoMonthly Weather Review vol 144 no 6 pp 2155ndash2175 2016

[25] M J Onderlinde and D S Nolan ldquoEnvironmental helicityand its effects on development and intensification of tropicalcyclonesrdquo Journal of the Atmospheric Sciences vol 71 no 11pp 4308ndash4320 2014

[26] C S Velden and L M Leslie ldquoe basic relationship betweentropical cyclone intensity and the depth of the environmentalsteering layer in the Australian regionrdquo Weather and Fore-casting vol 6 no 2 pp 244ndash253 1991

[27] Z Zeng Y Wang and L-S Chen ldquoA statistical analysis ofvertical shear effect on tropical cyclone intensity change in theNorth Atlanticrdquo Geophysical Research Letters vol 37 no 2Article ID L02802 2010

[28] C S Velden and J Sears ldquoComputing deep-troposphericvertical wind shear analyses for tropical cyclone applicationsdoes the methodology matterrdquo Weather and Forecastingvol 29 no 5 pp 1169ndash1180 2014

[29] P M Finocchio and S J Majumdar ldquoA statistical perspectiveon wind profiles and vertical wind shear in tropical cycloneenvironments of the northern Hemisphererdquo MonthlyWeather Review vol 145 no 1 pp 361ndash378 2017

[30] K Emanuel C DesAutels C Holloway and R KortyldquoEnvironmental control of tropical cyclone intensityrdquo Journalof the Atmospheric Sciences vol 61 no 7 pp 843ndash858 2004

[31] X Lu K K W Cheung and Y Duan ldquoNumerical study onthe formation of typhoon ketsana (2003) Part I roles of themesoscale convective systemsrdquo Monthly Weather Reviewvol 140 no 1 pp 100ndash120 2012

[32] S A Braun J A Sippel and D S Nolan ldquoe impact of drymidlevel air on hurricane intensity in idealized simulations


with no mean flowrdquo Journal of the Atmospheric Sciencesvol 69 no 1 pp 236ndash257 2012

[33] Y Wang ldquoHow do outer spiral rainbands affect tropicalcyclone structure and intensityrdquo Journal of the AtmosphericSciences vol 66 no 5 pp 1250ndash1273 2009

[34] K A Hill and G M Lackmann ldquoInfluence of environmentalhumidity on tropical cyclone sizerdquo Monthly Weather Reviewvol 137 no 10 pp 3294ndash3315 2009

[35] L Wu H Su R G Fovell et al ldquoImpact of environmentalmoisture on tropical cyclone intensificationrdquo AtmosphericChemistry and Physics Discussions vol 15 no 24pp 16111ndash16139 2015

[36] P G Black and R A Anthes ldquoOn the asymmetric structure ofthe tropical cyclone outflow layerrdquo Journal of the AtmosphericSciences vol 28 no 8 pp 1348ndash1366 1971

[37] R T Merrill ldquoEnvironmental influences on hurricane in-tensificationrdquo Journal of the Atmospheric Sciences vol 45no 11 pp 1678ndash1687 1988

[38] J Molinari P Duran and D Vollaro ldquoLow Richardsonnumber in the tropical cyclone outflow layerrdquo Journal of theAtmospheric Sciences vol 71 no 9 pp 3164ndash3179 2014

[39] P Duran and J Molinari ldquoUpper-tropospheric lowRichardson number in tropical cyclones sensitivity to cycloneintensity and the diurnal cyclerdquo Journal of the AtmosphericSciences vol 73 no 2 pp 545ndash554 2016

[40] G J Holland and R T Merrill ldquoOn the dynamics of tropicalcyclone structural changesrdquo Quarterly Journal of the RoyalMeteorological Society vol 110 no 465 pp 723ndash745 1984

[41] E D Rappin M C Morgan and G J Tripoli ldquoe impact ofoutflow environment on tropical cyclone intensification andstructurerdquo Journal of the Atmospheric Sciences vol 68 no 2pp 177ndash194 2011

[42] B S Barrett E R Sanabia S C Reynolds J K Stapleton andA L Borrego ldquoEvolution of the upper-tropospheric outflowin hurricanes Iselle and Julio (2014) in the navy global en-vironmental model (NAVGEM) analyses and in satellite anddropsonde observationsrdquo Journal of Geophysical ResearchAtmospheres vol 13 no 10 pp 273ndash286 2016

[43] W A Komaromi and J D Doyle ldquoTropical cyclone outflowand warm core structure as revealed by HS3 dropsonde datardquoMonthly Weather Review vol 145 no 4 pp 1339ndash1359 2017

[44] L Chen and W M Gray ldquoGlobal view of the upper leveloutflow patterns associated with tropical cyclone intensitychanges during FGGErdquo Colorado State University Dept ofAtmospheric Science Paper 392 Colorado State UniversityFort Collins Colorado p 126 1985

[45] J Molinari and D Vollaro ldquoExternal influences on hurricaneintensity Part I outflow layer eddy angular momentumfluxesrdquo Journal of the Atmospheric Sciences vol 46 no 8pp 1093ndash1105 1989

[46] R L Pfeffer and M Challa ldquoe role of environmentalasymmetries in Atlantic hurricane formationrdquo Journal of theAtmospheric Sciences vol 49 no 12 pp 1051ndash1059 1992

[47] C-C Wu and K A Emanuel ldquoOn hurricane outflowstructurerdquo Journal of the Atmospheric Sciences vol 51 no 13pp 1995ndash2003 1994

[48] W M Frank and E A Ritchie ldquoEffects of environmental flowupon tropical cyclone structurerdquo Monthly Weather Reviewvol 127 no 9 pp 2044ndash2061 1999

[49] J Feng and X Wang ldquoImpact of assimilating upper-leveldropsonde observations collected during the TCI field cam-paign on the prediction of intensity and structure of Hurri-cane Patricia (2015)rdquoMonthlyWeather Review vol 147 no 8pp 3069ndash3089 2019 In press

[50] L F Bosart C S Velden W E Bracken J Molinari andP G Black ldquoEnvironmental influences on the rapid in-tensification of hurricane opal (1995) over the gulf of MexicordquoMonthly Weather Review vol 128 no 2 pp 322ndash352 2000

[51] J C Sadler ldquoA role of the tropical upper tropospheric troughin early season typhoon developmentrdquo Monthly WeatherReview vol 104 no 10 pp 1266ndash1278 1976

[52] R T Merrill and C S Velden ldquoA three-dimensional analysisof the outflow layer of Supertyphoon Flo (1990)rdquo MonthlyWeather Review vol 124 no 1 pp 47ndash63 1996

[53] J D Ventham and B Wang ldquoLarge-scale flow patterns andtheir influence on the intensification rates of western NorthPacific tropical stormsrdquo Monthly Weather Review vol 135no 4 pp 1110ndash1127 2007

[54] H R Byers General Meteorology McGraw-Hill New YorkUSA 1944

[55] B I Miller ldquoOn the maximum intensity of hurricanesrdquoJournal of Meteorology vol 15 no 2 pp 184ndash195 1958

[56] A Wada ldquoReexamination of tropical cyclone heat potential inthe western north pacificrdquo Journal of Geophysical ResearchAtmospheres vol 121 no 12 pp 6723ndash6744 2016

[57] A Wada ldquoVerification of tropical cyclone heat potential fortropical cyclone intensity forecasting in the Western NorthPacificrdquo Journal of Oceanography vol 71 no 4 pp 373ndash3872015

[58] J S Malkus and H Riehl ldquoOn the dynamics and energytransformations in steady-state hurricanesrdquo Tellus vol 12no 1 pp 1ndash20 1960

[59] L K Shay G J Goni and P G Black ldquoEffects of a warmoceanic feature on hurricane opalrdquoMonthly Weather Reviewvol 128 no 5 pp 1366ndash1383 2000

[60] A Lowag M L Black and M D Eastin ldquoStructural andintensity changes of hurricane bret (1999)mdashpart I environ-mental influencesrdquoMonthly Weather Review vol 136 no 11pp 4320ndash4333 2008

[61] J C L Chan Y Duan and L K Shay ldquoTropical cycloneintensity change from a simple ocean-atmosphere coupledmodelrdquo Journal of the Atmospheric Sciences vol 58 no 2pp 154ndash172 2001

[62] P C Meyers L K Shay and J K Brewster ldquoDevelopmentand analysis of the systematically merged Atlantic regionaltemperature and salinity climatology for oceanic heat contentestimatesrdquo Journal of Atmospheric and Oceanic Technologyvol 31 no 1 pp 131ndash149 2014

[63] M Mainelli M DeMaria L K Shay and G Goni ldquoAppli-cation of oceanic heat content estimation to operationalforecasting of recent atlantic category 5 hurricanesrdquo Weatherand Forecasting vol 23 no 1 pp 3ndash16 2008

[64] M Yamaguchi H Owada U Shimada et al ldquoTropical cy-clone intensity prediction in the western North Pacific basinusing SHIPS and JMAGSMrdquo SOLA vol 14 pp 138ndash1432018

[65] D P Dee S Uppala A J Simmons et al ldquoe ERA Interimreanalysis configuration and performance of the data as-similation systemrdquo Journal of the Royal Meteorological Societyvol 137 no 656 pp 533ndash597 2011

[66] N Lin R Jing Y Wang E Yonekura J Fan and L Xue ldquoAstatistical investigation of the dependence of tropical cycloneintensity change on the surrounding environmentrdquo MonthlyWeather Review vol 145 no 7 pp 2813ndash2831 2017

[67] B A Schenkel and R E Hart ldquoAn examination of tropicalcyclone position intensity and intensity life cycle withinatmospheric reanalysis datasetsrdquo Journal of Climate vol 25no 10 pp 3453ndash3475 2012


[68] K Hodges A Cobb and P L Vidale ldquoHow well Are tropicalcyclones represented in reanalysis datasetsrdquo Journal of Cli-mate vol 30 no 14 pp 5243ndash5264 2017

[69] M J Bell A Schiller P Y L Traon N R SmithE Dombrowsky and K Wilmer-Becker ldquoAn introduction toGODAE ocean viewrdquo Journal of Operational Oceanographyvol 8 no 1 pp 2ndash11 2015

[70] B Huang Y Xue D Zhang A Kumar and M J McPhadenldquoe NCEP GODAS ocean analysis of the tropical pacificmixed layer heat budget on seasonal to interannual timescalesrdquo Journal of Climate vol 23 no 18 pp 4901ndash4925 2010

[71] L N Bai and Y Wang ldquoEffect of vertical wind shear ontropical cyclone intensity changerdquo Journal of Tropical Mete-orology vol 22 no 1 pp 11ndash18 2016

[72] M W Buckley R M Ponte G Forget and P HeimbachldquoLow-frequency SSTand upper-ocean heat content variabilityin the North Atlanticrdquo Journal of Climate vol 27 no 13pp 4996ndash5018 2014

[73] I-F Pun I-I Lin and M-H Lo ldquoRecent increase in hightropical cyclone heat potential area in the Western NorthPacific Oceanrdquo Geophysical Research Letters vol 40 no 17pp 4680ndash4684 2013

[74] P Hall and J D Hart ldquoBootstrap test for difference betweenmeans in nonparametric regressionrdquo Journal of the AmericanStatistical Association vol 85 no 412 pp 1039ndash1049 1990

[75] X Wang and C H Bishop ldquoImprovement of ensemble re-liability with a new dressing kernelrdquo Quarterly Journal of theRoyal Meteorological Society vol 131 no 607 pp 965ndash9862005

[76] T M Hamill J S Whitaker M Fiorino and S G BenjaminldquoGlobal ensemble predictions of 2009rsquos tropical cyclonesinitialized with an ensemble Kalman filterrdquoMonthly WeatherReview vol 139 no 2 pp 668ndash688 2011

[77] Q Li P Xu X Wang et al ldquoAn operational statistical schemefor tropical cyclone induced wind gust forecastsrdquo Weatherand Forecasting vol 31 no 6 pp 1817ndash1832 2016

[78] Y Wang and C-C Wu ldquoCurrent understanding of tropicalcyclone structure and intensity changes - A reviewrdquo Meteo-rology and Atmospheric Physics vol 87 no 4 pp 257ndash2782004

[79] J Xu and Y Wang ldquoDependence of tropical cyclone in-tensification rate on sea surface temperature storm intensityand size in the western north pacificrdquo Weather and Fore-casting vol 33 no 2 pp 523ndash537 2018

[80] M DeMaria J Kaplan and J-J Baik ldquoUpper-level eddyangular momentum fluxes and tropical cyclone intensitychangerdquo Journal of the Atmospheric Sciences vol 50 no 8pp 1133ndash1147 1993

[81] D Hanley J Molinari and D Keyser ldquoA composite study ofthe interactions between tropical cyclones and upper-tro-pospheric troughsrdquoMonthly Weather Review vol 129 no 10pp 2570ndash2584 2001

[82] H Yu and H J Kwon ldquoEffect of TC-trough interaction on theintensity change of two typhoonsrdquo Weather and Forecastingvol 20 no 2 pp 199ndash211 2005

[83] X Chen Y Wang and K Zhao ldquoSynoptic flow patterns andlarge-scale characteristics associated with rapidly intensifyingtropical cyclones in the South China seardquo Monthly WeatherReview vol 143 no 1 pp 64ndash87 2015

[84] W M Gray ldquoGlobal view of the origin of tropical distur-bances and stormsrdquoMonthly Weather Review vol 96 no 10pp 669ndash700 1968

[85] S Shu and L Wu ldquoAnalysis of the influence of Saharan airlayer on tropical cyclone intensity using AIRSaqua datardquoGeophysical Research Letters vol 36 no 9 2009


Hindawiwwwhindawicom Volume 2018

Journal of

ChemistryArchaeaHindawiwwwhindawicom Volume 2018

Marine BiologyJournal of


BiodiversityInternational Journal of


EcologyInternational Journal of


Hindawiwwwhindawicom

Applied ampEnvironmentalSoil Science

Volume 2018

Forestry ResearchInternational Journal of



International Journal of

Geophysics

Environmental and Public Health

Journal of




Microbiology


Public Health Advances in

AgricultureAdvances in


Agronomy




MeteorologyAdvances in

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018Hindawiwwwhindawicom Volume 2018

ChemistryAdvances in

ScienticaHindawiwwwhindawicom Volume 2018


Geological ResearchJournal of

Analytical ChemistryInternational Journal of


Submit your manuscripts atwwwhindawicom

Most of the earlier studies have focused on hurricanes inthe North Atlantic or the East Pacific basins [5 6 8]However this study concentrates on TCs over the NorthwestPacific Basin (NWP) e NWP is controlled by the mon-soon background with about 70 of all TC genesis oc-curring in the monsoon trough [9 10] is is very differentfrom other ocean basins For example in the East Pacific andNorth Atlantic basins the main genesis source is AfricanEasterly Waves [11] Hence the relative importance of theindividual predictors for TC intensification is likely to bedifferent in various basins [6 12] Rios-Berrios and Torn [13]also suggested it is problematic to consider all basins jointlybecause the relative importance of environmental conditionsfor TC intensity change depends on the ocean basin

e present study seeks to address an important but asyet unanswered question ldquoDo significant differences inenvironmental factors including atmospheric conditionsand ocean heat conditions exist between rapidly normallyand slowly intensifying TCs in the NWPrdquo In this study arange of environmental factors is examined includingvertical wind shear (deep-layer upper-mid mid-low andlow-level) net moisture inflow upper-level outflow(strength and direction) and ocean heat conditions (bothSST and ocean heat content (OHC)) Earlier limited studiesof the NWP TC basin either examined different TC cate-gories [14] or focused only on one environmental factor [12]

Specifically the following environmental factors andtheir relationships with RI NI and SI are assessed Envi-ronmental vertical wind shear (VWS) is one of the mostimportant factors influencing TC intensity change and it iswidely accepted that strong VWS generally inhibits TCintensification [5 7 12 15ndash18] VWS also provides prob-ability estimates of RI in dynamical-statistical models[5 6 16 19] and can assist in determining if and when RIwill occur [20ndash22] e deep layer 200ndash850 hPa VWS iscommonly used whereas other studies showed that theVWS at different levels can also impact TC intensity change[12 23ndash25] Hence it is too simplistic to assess the impact ofVWS on TC intensity change by using a VWS calculatedfrom just two levels [26ndash29] In this study the VWS between

different vertical levels are used to analyze the impact ofVWS on the three TC intensification rates RI NI and SI

Environmental moisture is also suggested as a possibleimportant factor for TC intensification Both theoretical andmodeling studies showed that high environmental moisturefavors TC intensification [30 31] whereas dry air intrusioncan cause TC decay [21 32] e environmental relativehumidity (RH) of the lower and middle troposphere ishigher for RI cases than non-RI cases [5ndash7] However otherstudies suggested that moisture might have a negative or nosignificant impact on TC intensification [33ndash35] ereforethe impact of environmental moisture on TC IR requiresfurther more detailed examination In this study the im-pact of environmental moisture on the TC IR is studied byusing the low-level to mid-level net moisture inflow definedas the sum of net moisture inflow from all sides of a boxsurrounding the TC center and extending from the surfaceto 500 hPa is quantity is used to represent the total low-level to mid-level moisture transport from the environmentinto the TC core

Due to the asymmetric structure [36 37] lowRichardson number [38 39] and weaker inertial instability[40ndash43] of the TC upper-level outflow the outflow can easilyinteract with the upper-level larger-scale environment andthe inner-core of a TC thereby playing a mediating rolebetween the environment and the storm core [41] As such itcan affect the secondary circulation and consequently theTC intensity [37 41 43ndash49] e upper-level radial outflowis much stronger [6 37 43] and usually is more concentratedclose to the TC center for intensifying cases than for non-intensifying cases [43] In addition TC intensification isassociated with the multiple upper-level outflow channelsand TCs usually weaken when one of the efficient upper-level outflow channels is cut off [44 50ndash53] Kaplan andDeMaria [5] and Shu et al [14] noted that upper-tropo-spheric flows for RI cases are more easterly than for non-RIcases Rappin et al [41] showed that the ventilation ofoutflow in north and northeast directions minimizes thehurricane energy expenditure leading to larger in-tensification rates Hence the direction of TC outflowpossibly modifies the effect of outflow on TC intensitychange In this study the significant differences in theimpacts of both the strength and direction of outflow areexamined for all three different TC intensification cate-gories As far as the authors are aware this relationshipbetween outflow direction and TC IR has not previouslybeen statistically analyzed

e effect of the ocean heat conditions on TC intensitychange has been addressed for decades [54ndash57] SST iswidely known to affect TC intensification [5 6 58] Nu-merous supporting case studies exist [21 59ndash61] Statisticalanalyses [8 12] showed that TCs typically intensify (weaken)over warmer (colder) ocean eddy regions with higher(lower) SST Some studies have shown that the OHCmeasured from the surface to the depth of the 26degC isothermis important for TC intensity change [59] and that SSTaloneis not always a good measure of the total heat content thataffects TC intensity change [62] is hypothesis is furthersupported by studies showing that the use of OHC can help

0

10

20

30

40

50

60

70

80

90

100

0

5

10

15

20

25ndash7

0ndash6

5ndash6

0ndash5

5ndash5

0ndash4

5ndash4

0ndash3

5ndash3

0ndash2

5ndash2

0ndash1

5ndash1

0 ndash5 0 5 10 15 20 25 30 35 40 45 50 55 60

Cum

ulat

ive f

requ

ency

()

Freq

uenc

y (

)

TC intensification rate (unit kt 12hrndash1)

85

RINI

SI

95

Figure 1 Frequency distribution and cumulative frequency dis-tribution of the 12 hr intensity changes of all TCs in the 1995ndash2015in Northwest Pacific basin (10156 cases)













1113868111386811138681113868 (1)




Quv 1113946500hpa



qe minus1g

qu

qs 1g

qv

qw 1g

qu

qn minus1g

qv

(3)



∯D





OHC ρcp 11139460

h T26∘C( )T(z)dz (5)



















max Num Percentage












45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ


TSSTSTY

STYSuperTY

(a)


TSSTSTY

STYSuperTY

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ

(b)

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E


TSSTSTY

STYSuperTY

140E 150E 160E 170E 180EQ

(c)








0

10

20

30

40


Freq

uenc

y (

)

SINIRI


Freq

uenc

y (

)

0

10

20

30

40

50

60


SINIRI


Freq

uenc

y (

)

0

5

10

15

20

25


SINIRI












Freq

uenc

y (

)

0

10

20

30

40

50

60

70


SI (146 hr)NI (183 hr)RI (188 hr)








5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)






SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy


























1113868111386811138681113868 (1)




Quv 1113946500hpa



qe minus1g

qu

qs 1g

qv

qw 1g

qu

qn minus1g

qv

(3)



∯D





OHC ρcp 11139460

h T26∘C( )T(z)dz (5)



















max Num Percentage












45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ


TSSTSTY

STYSuperTY

(a)


TSSTSTY

STYSuperTY

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ

(b)

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E


TSSTSTY

STYSuperTY

140E 150E 160E 170E 180EQ

(c)








0

10

20

30

40


Freq

uenc

y (

)

SINIRI


Freq

uenc

y (

)

0

10

20

30

40

50

60


SINIRI


Freq

uenc

y (

)

0

5

10

15

20

25


SINIRI












Freq

uenc

y (

)

0

10

20

30

40

50

60

70


SI (146 hr)NI (183 hr)RI (188 hr)








5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)






SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy
















Quv 1113946500hpa



qe minus1g

qu

qs 1g

qv

qw 1g

qu

qn minus1g

qv

(3)



∯D





OHC ρcp 11139460

h T26∘C( )T(z)dz (5)



















max Num Percentage












45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ


TSSTSTY

STYSuperTY

(a)


TSSTSTY

STYSuperTY

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ

(b)

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E


TSSTSTY

STYSuperTY

140E 150E 160E 170E 180EQ

(c)








0

10

20

30

40


Freq

uenc

y (

)

SINIRI


Freq

uenc

y (

)

0

10

20

30

40

50

60


SINIRI


Freq

uenc

y (

)

0

5

10

15

20

25


SINIRI












Freq

uenc

y (

)

0

10

20

30

40

50

60

70


SI (146 hr)NI (183 hr)RI (188 hr)








5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)






SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy



























max Num Percentage












45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ


TSSTSTY

STYSuperTY

(a)


TSSTSTY

STYSuperTY

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ

(b)

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E


TSSTSTY

STYSuperTY

140E 150E 160E 170E 180EQ

(c)








0

10

20

30

40


Freq

uenc

y (

)

SINIRI


Freq

uenc

y (

)

0

10

20

30

40

50

60


SINIRI


Freq

uenc

y (

)

0

5

10

15

20

25


SINIRI












Freq

uenc

y (

)

0

10

20

30

40

50

60

70


SI (146 hr)NI (183 hr)RI (188 hr)








5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)






SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy



















45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ


TSSTSTY

STYSuperTY

(a)


TSSTSTY

STYSuperTY

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E 140E 150E 160E 170E 180EQ

(b)

45N

40N

35N

30N

25N

20N

15N

10N

5N

100E 110E 120E 130E


TSSTSTY

STYSuperTY

140E 150E 160E 170E 180EQ

(c)








0

10

20

30

40


Freq

uenc

y (

)

SINIRI


Freq

uenc

y (

)

0

10

20

30

40

50

60


SINIRI


Freq

uenc

y (

)

0

5

10

15

20

25


SINIRI












Freq

uenc

y (

)

0

10

20

30

40

50

60

70


SI (146 hr)NI (183 hr)RI (188 hr)








5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)






SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy




















0

10

20

30

40


Freq

uenc

y (

)

SINIRI


Freq

uenc

y (

)

0

10

20

30

40

50

60


SINIRI


Freq

uenc

y (

)

0

5

10

15

20

25


SINIRI












Freq

uenc

y (

)

0

10

20

30

40

50

60

70


SI (146 hr)NI (183 hr)RI (188 hr)








5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)






SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy
























Freq

uenc

y (

)

0

10

20

30

40

50

60

70


SI (146 hr)NI (183 hr)RI (188 hr)








5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)






SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy




















5

55

6

65

7

75

8V

WS

200ndash

850

hPa (

m sndash1

)

SINIRI

(a)

SINIRI


4

45

5

55

6

VW

S 20

0ndash50

0 hP

a (m

sndash1)

(b)

SINIRI


3

35

4

VW

S 50

0ndash85

0 hP

a (m

sndash1)

(c)

SINIRI


12

14

16

18

2V

WS

850ndash

1000

hPa

(m sndash1

)

(d)






SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy


















SINIRI

0

5

10

15

20

25

30


Freq

uenc

y (

)


(a)

SINIRI

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


(b)

SINIRI

0

5

10

15

20

25

30

0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 ge8

Freq

uenc

y (

)


(c)

SINIRI

0

10

20

30

40

50

0-1 1-2 2-3 3-4 4-5 ge5

Freq

uenc

y (

)


(d)








0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy




















0

20

40

60

80Pr

obab

ility

()

SINIRI

(a)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(b)


0

20

40

60

80

Prob

abili

ty (

)

SINIRI

(c)


0

20

40

60

80Pr

obab

ility

()

SINIRI

(d)








21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy




















21

22

23

24

25

26Q

uv fl

ux (k

g m

ndash2 sndash1

)

SINIRI

(a)

0

5

10

15

20

25

30

35

40


Freq

uenc

y (

)


SINIRI

(b)


SINIRI


09

1

11

12

Out

flow

stre

ngth

(10ndash5

sndash1)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)

(b)






10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy


















10

20

30

40

50

SW337

NE444

SE29

Outflow direction

Freq

uenc

y (

)

NW913

SINIRI

(a)

0

5

10

15

20

25

30

35


NWSW

NESE

TC in

tens

ifica

tion

rate

(uni

t kt

12h

rndash1)

(b)


SINIRI


284

286

288

29

292

294

SST

(degC)

(a)

SINIRI


40

50

60

70

80

90

OH

C (K

J cm

ndash2)

(b)




5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy
















5 Conclusions



SINIRI

0

5

10

15

20

25

30

35

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

Freq

uenc

y (

)

(a)

SINIRI

0

5

10

15

20

25


Freq

uenc

y (

)(b)


SINIRI

lt26 26 265 27 275 28 285 29 295 ge30SST (degC)

0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(a)

(a)

SINIRI


0

10

20

30

40

50

60

70

80

90

Prob

abili

ty (

)

(b)








Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy




















Data Availability




Acknowledgments





References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy















References




























































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy








































































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy



































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy
















Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy















Documents

TheDependenceofNorthwestPacificTropicalCyclone …downloads.hindawi.com/journals/amete/2019/9456873.pdf · 2019. 12. 12. · ResearchArticle TheDependenceofNorthwestPacificTropicalCyclone