A precise understanding and prediction of tropical cyclone(TC)genesis remains one of the fundamental objectives for the meteorological community.Monitoring would be much easier if we could anticipate in advance the re...A precise understanding and prediction of tropical cyclone(TC)genesis remains one of the fundamental objectives for the meteorological community.Monitoring would be much easier if we could anticipate in advance the regions where a TC would form.In this study,we considered 8 cases each of developing and non-developing TCs over the North Indian Ocean(NIO).We found that the stream function averaging over a layer(850-500 hPa)can effectively identify the quasi closed circulation(QCC)before the low-pressure area(LPA)formation.Based on this,we designed an algorithm to track the QCC.The day after an LPA the negative stream-function value at the center of QCC gradually increases in all developing cases.Whereas,in non-developing cases,the negative stream function values are comparatively smaller and remain steady.The total precipitable water within the QCC for developing cases gradually increased on the day of the LPA and persisted until the day of depression.A strong QCC can trap and enhance the availability of moisture through vertical moisture flux transport from the surface in developing lows.However,in non-developing lows,a feeble QCC can only trap moisture at the initial stage but fails to sufficiently moisten the mid-levels.We applied machine learning to identify the threshold values for the stream function and total precipitable water to find the potential of the QCC to become a depression.We tested an algorithm for pre and post monsoon seasons during 2020–2022.The algorithm successfully detected many vortices 5–7 days before the formation of a depression,and it identified depressions 3–4 days in advance.As the thresholds are obtained by machine learning method from the training data,this algorithm could be applied to other basins.This advances our knowledge of the TC origin and aids in its early monitoring.展开更多
文摘A precise understanding and prediction of tropical cyclone(TC)genesis remains one of the fundamental objectives for the meteorological community.Monitoring would be much easier if we could anticipate in advance the regions where a TC would form.In this study,we considered 8 cases each of developing and non-developing TCs over the North Indian Ocean(NIO).We found that the stream function averaging over a layer(850-500 hPa)can effectively identify the quasi closed circulation(QCC)before the low-pressure area(LPA)formation.Based on this,we designed an algorithm to track the QCC.The day after an LPA the negative stream-function value at the center of QCC gradually increases in all developing cases.Whereas,in non-developing cases,the negative stream function values are comparatively smaller and remain steady.The total precipitable water within the QCC for developing cases gradually increased on the day of the LPA and persisted until the day of depression.A strong QCC can trap and enhance the availability of moisture through vertical moisture flux transport from the surface in developing lows.However,in non-developing lows,a feeble QCC can only trap moisture at the initial stage but fails to sufficiently moisten the mid-levels.We applied machine learning to identify the threshold values for the stream function and total precipitable water to find the potential of the QCC to become a depression.We tested an algorithm for pre and post monsoon seasons during 2020–2022.The algorithm successfully detected many vortices 5–7 days before the formation of a depression,and it identified depressions 3–4 days in advance.As the thresholds are obtained by machine learning method from the training data,this algorithm could be applied to other basins.This advances our knowledge of the TC origin and aids in its early monitoring.
