| Tropical cyclones(TCs)are among the most devastating natural phenomena.With continued global warming,increasing numbers and magnitudes of strong TCs are expected,and the coastal regions in different countries around the world are more vulnerable to TCs which bring catastrophic damages and economic losses.The rapid intensification(RI)process,being an inevitable part of major TCs,has still challenged the operational TC intensity forecasts.The current studies have not fully understood the temporal and spatial variations of RI events under the background of obvious climate change.As for the specific RI process,the occurrences of RI events usually involve complex atmospheric and oceanic interactions and are modulated by the internal dynamics as well.Meanwhile,benefiting from the continued development of satellite technology,multi-source remote sensing data would certainly improve our understanding of the intensity change process.It is thus crucial to determine the RIrelated characteristics observed on satellite images.The current study has combined TC best-track datasets,remote sensing data,and statistical data in order to statistically investigate the inter-decadal characteristics of RI events as well as the variabilities of the RI embedded environmental conditions and convection features in the Western North Pacific(WNP).Based on the statistical findings,this study has further applied two remote sensing data to characterize the convection features in different intensity change categories,and the convection evolutions during different periods of RI process,which we hope to provide some theoretical basis in understating this extreme event.The main conclusions are as follows:(1)Significant inter-decadal variations of the RI occurrence and spatial distributions were found in the WNP,and climate change had significantly increased the RI-TC occurrence ratio.Based on the regime shift detection and spatial kernel density estimation methods,this study focused on the comparative analysis of one abnormally high period(2000-2016: P3)and one abnormally low period(1973-1999:P2).The results indicated that the RI activities and RI-related TC activities(RITC and NRITC)demonstrated significant differences between these two periods.The RITCs could reach higher intensity through RI process with their genesis locations being farther east and south in the WNP.Compared to P2,P3 demonstrated more favorable RI occurrence conditions,and the RI events distributions were more concentrated in the main development region.The genesis locations of RITCs and NRITCs exhibited significantly westward and northward migrations,while the LMI distributions for these two TC categories exhibited significantly lower density in the east Philippine Islands and the South China Sea area,respectively.The above changes led to a higher LMI distribution density of TCs in the East China Sea area.The averaged locations of NRITC genesis and LMI locations demonstrated significant poleward migrations toward higher latitudes during P3.By comparing the longitudinal distributions of the TC events and RI events in the main development region,the analyses indicated that the inter-decadal variations of RI events were affected by the combined changes of the TC genesis frequency,the TC prevailing tracks along with the fundamental large-scale background that had been significantly changed by the climate change.(2)Based on the Statistical Hurricane Intensity Precipitation Scheme(SHIPS)developmental dataset,the variabilities of environmental conditions and convection features during RI process were statistically analyzed.The distributions of embedded environments show large variabilities among RI events,indicating a weak interrelationship.While the environmental conditions between RI events and slow intensification(SI)events in RITCs show no obvious difference,the SI events in NRITCs experience significantly more unfavorable environmental conditions,which suggest that unfavorable environments would indeed prohibit further TC intensification.Vertical wind shear plays a more important role in deciding the occurrence of RI.Seen from the TC evolution perspective,no environmental conditions demonstrate significant changes during RI onset as well,indicating that RI is more likely triggered by other processes under similar favorable environments.Meanwhile,statistical analyses also demonstrate that unfavorable environmental conditions would significantly delay the timing of the first RI occurrence.Conversely,the distributions of convection features are more consistent and show higher correlations.Larger magnitudes of TC intensification are related to larger convection feature thresholds,and the convection features are more significantly influenced by extremely different thermodynamic conditions.During the RI onset,all convection features exhibit abrupt changes,and statistical analyses imply that these changes are not solely caused by the increase of TC mean intensity,instead they can be used to indicate the RI occurrence.(3)The above analyses indicate that climate change has significantly altered the RI-embedded large-scale environmental conditions,and increase the probability of RI occurrence during the recent decade,however,as for specific RI processes,the external environmental conditions still exhibit large variabilities,and the statistical results suggest that the RI occurrences are more likely controlled by the internal dynamics,such as deep convective process.It is thus crucial to identify these convection features by use of the multi-source remote sensing data.(4)Based on the 14-year CloudSat-Tropical Cyclone overpass dataset,this paper had investigated the vertical features of TCs in different intensity change categories.The contoured frequency by altitude diagrams(CFADs)show similar arc-like contour profiles for all five categories,with two distinct modes of reflectivity distributions separated by the melting layer.RI has the highest frequency expanding through the whole reflectivity range along with the broadest coverage in the upper branch,and the frequency distributions monotonously increase from neutral to SI and RI category.The most striking features of CFADs are the “discontinuity” of contour lines at 6-11 km altitudes resulting from the abrupt increase of particle size.This “discontinuity” feature is the least obvious in RI category,indicating the strongest updrafts enhancing the concentration of precipitation-sized ice hydrometeors.Intensifying TCs has significantly higher ice water content at higher altitudes,especially for the RI process,which can be seen as an indicator of TC rapid intensity change.More large-size ice particles lifted to higher altitudes further indicate the importance of stronger updrafts and deep convection and the associated latent heat release on TC rapid intensity change.Though limited by the sample sizes,the above results in the current study are ultimately decided by the differences of the TC inner-core convection features.(5)By using the IRBT data from geostationary satellites,this paper further investigated the radial and azimuthal distributions of deep convection feature evolutions during different periods of RI process in a timeline-based framework.The results show that the averaged inner-core IRBTs become significantly colder during the RI onset and RI continue periods,which are contributed mainly by the increasing occurrence percentage of the deep convective clouds.In the left of shear quadrant,especially the upshear left quadrant,IRBTs exhibit the most significant cooling trend along with the largest cooling magnitude.The spatial evolutions of IRBTs further reveal a cyclonic migration of deep convections from downshear to upshear quadrant,which has finally led to a significant increase of inner-core symmetry at RI onset.By comparing five inner-core convection features of RI events with TC events that have the same current intensity,significant differences between the means of these two categories are observed,and the largest changing magnitudes appear just around RI onset,which further indicates that the rapid cooling of IRBTs along with the rapid increase of the occurrence percentages of deep convective clouds are indeed indicative of RI,rather than resulting from the increase of averaged TC intensity. |
