The Structural Features Of Mesocale Convective Systems In The Kernel In TCs With Rapid Intensification

We have made significant progress during the past20years in track prediction ofNorthwest Pacific tropical cyclone(TC). However, the progress of TC intensity forecastskill is slow, especially in the rapid intensity(RI) change progress. This may be due tothe incomplete understanding of its physical mechanism. Recent studies took the RI processas a combined effect of TC core area, ocean and large-scale environmental field. Andenvironmental field of underlying has a indirectly influence on the TC intensity, by causingimpacts on mesoscale convective systems in TC inner core, especially in dynamic effects.Therefore, it has great value to analysis the impact of mesoscale convective systems inTC circulation on its RI process. In this paper, we use satellite data to analysis the changeof mesoscale convective systems before the RI process, to provide some help to the RIforecasting20-yr (1991-2010) tropical cyclone (TC) intensity in the Western North Pacific(WNP) fromthree forecast centers was analyzed, i.e. China Meteorological Administration(CMA), JapanMeteorological Agency (JMA), and Joint Typhoon Warning Center (JTWC) of the United States.Results show that there is more or less discrepancy in the intensity change of a TC amongdifferent datasets. The maximum discrepancy reaches22hPa/6h (42hPa/6h,33hPa/6h) betweenCMA and JMA (CMA and JTWC, JMA and JTWC). Special attention is placed on the records forabrupt intensity change, which is currently a global puzzle for forecasters. It is foundthat an abrupt intensity change process recorded by one dataset can have an intensity changein another dataset varying from0to≥10hPa/6h with the same sign or with opposite signin some extreme cases. In the total2511cases with rapid intensity changes, only14%caseshave a consensus among all the three datasets and25%cases have an agreement between twoof the three datasets. In spite of such a significant uncertainty, the three datasets agreeon the general statistical characteristics of abrupt intensity changes, including regionaland seasonal distributions, persistency features, and the relationship of abrupt intensity change with TC intensity and moving speed. Notable disagreement is on very strong systems(super typhoons) and very fast moving TCs.Pick up samples who have a consistency in intensity changing in two or three datasetsabove for further analysis. And divided them into four categories, according to theintensity differences and intensity change differences in6hrs. the four categories are:1, RI process samples achieved the strength of TY,2, RI process samples did not achievethe strength of TY,3, non RI process samples achieved the strength of TY,4, non RI processsamples did not achieved the strength of TY. And consider of the deficiencies of the satellitedata, the total sample number is: category1:211, category2:78, category3:229, category4:429.4categories have the same statistics in average scale of TC kernel,between0.5°-1°, with minimum in category2, and maximum in category3. And there existsome extreme cases who have large kernel scale could reach the value of more than2.5°.With the process of strengthen, kernel scale of samples did not achieved the strength ofTY (category2and category4) reduced, as that of samples achieved the strength of TY(category1and category3) changing to the opposite direction. Because of the difference inscale of TC kernel, the kernel scale need to be normalized to a common size. First, calculatethe radial average brightness temperature within600km from the TC center（bt(r)）. Andttehmep ebrraitguhrteena ewstiset dht bieynm p1es1ru1ab ttkurmra（e（ctbitBt） are normalized by subtracting the average brightness）from btng the rad(iru)s: o fBtB=tcbht(arn)g in bgt. The normalized kernel radius(Rt) is cr from negative to positive (rBt)from the physical radius (r) and dividing by te rBt: Rt=r rrBtBt. The value of normalizedkernel radius is defined from0to1. The range of Rtis used to in the following calculation.-60℃was selected as the range of the β-mesocale convective systems(MCS) in TC kernel.The number and area of disturbances in the4categories and the average distance betweenthe center of disturbances and TC (distance of disturbances) are analyzed. The result is:TC before RI process has disturbance numbers, area and distance reduce quicker than TC beforenon-RI process under equal strength. Samples did not achieved TY have more disturbancenumbers, and their disturbance distance change faster as time passed. The area disturbancesin TC reached TY grow quickly, and the amplitude is larger. The disturbance distance ofcategory2shorten quicker than that of category4. And the distance of category3tends tomaintain, while it has a slow Increase in category1.According to slope difference ofdisturbance numbers, area and distance, we could have the4categoriesdistinguished. If the strength of sample is weak, the disturbance area remainedstable or decreasing trend, and the disturbance number and disturbance distance decrease speed is relatively slow, this TC sample is thought to has less prone to rapid strengtheningprocess in the next24hours. Otherwise, if there is an increasing trend in disturbance areaand a rapid decrease trend in distance, the TC is considered in high probability ofenhancing rapidly. It is the same in the samples which is strong enough. There are biggerdisturbance area changes and distance changes before RI process.No matter in which category, the average of normalized radius in the South China Sea(SCS) is bigger than that in the western north pacific (WNP). And the mutual differencesand abnormal samples in every category are less than that in the western north pacific. Therelationship among the four category is just the same as in WNP. And statistical resultsshowed that it has more significant difference between RI process and non-RI process in SCS.The disturbance number, area and distance of category2change quicker than that of category4.And except the number of disturbance, the area and distance change of category3have largeamplitude than of category1.“Saomai” has its RI process after reaching the strength of TY, while the RI processof “Nuri” happened when it is not strong enough to achieve TY.“Saomai” and “Nuri”are typical cases to analysis the differences of the change of β-MCS before TCintensification. They all generate in the WNP in August, and they also have the same path,moving northwest, landing at southeast coastal areas of our country, so they are similarin climate. With the result of case analysis, the relationship between the change of β-MCSand the RI process have great common with the statistics results above. And the furtheranalysis (secondary analysis) of “Saomai” and “Nuri” are in the same changing character.