Influence Of ENSO On Western North Pacific Tropical Cyclone Intensity: A Model Sensitive Study

Using the Tropical Cyclone Year Books of 1949-2002 published by the China Meteorological Press, the annual mean genesis locations and activity areas of tropical cyclones (TCs) in the Western North Pacific (WNP) are computed. TCs'mean genesis locations and activity areas are separated into 3 groups according to neutral, El Nino, and La Nina years. It is found that in El Nino years, most of the TCs generate in the southeast WNP, and in La Nina years, they generate in the northwest WNP. In El Nino years, the territory of TC activity in WNP is wide from west to east, and is narrow from south to north; in La Nina years, it is narrow from west to east, and is wide from south to north. The SVD analysis is also used in this study. Using the SVD analysis, it is found that the TC is stronger and more active in the southeast WNP in El Nino years, and is weaker and less active in La Nina years. Using the SVD left time series as the TC activity index, the relationship between different variables and the WNP TC activity is analyzed. From 2-layer CISK model, the growing ratio of TC intensity is derived. The results show that the bigger the vertical wind shear is, the lower the growing ratio of TC intensity. At different latitude, under the same vertical wind shear, TCs tend to have small growing ratio of intensity at high latitude. The variation of TC intensity under different strength of vertical wind shear is also checked by MM5 model sensitive experiments. There are 42 experiments carried out with the strength of vertical wind shear from 0-30 m s-1 and the intensity of initial bogus from 5-30 m s-1. The main conclusion of the model experiments represents that TCs are weaker under the circulation with vertical wind shear than that without vertical wind shear; strong TCs (intensity> 20 m s-1) are affected less by the vertical wind shear with the strength under 30 m s-1.Model sensitivity experiments are carried out to understand the influence of the atmospheric condition and sea surface temperature (SST) on the intensities of TCs in the WNP. Three model experiments with different cumulus parameterization, planetary boundary layer scheme, bogus, and atmospheric and oceanic circulation are carried out. In each experiment, the specified initial and lateral atmospheric conditions and the SST are averaged for the neutral, El Nino, and La Nina years or the combination of different atmospheric conditions and SST are used. Each set consists of 44 modeled TCs generated at different locations. Experiment I and II use the "mean field", and experiment III uses the "typical field". Mean filed is the simple average of reanalysis data. Typical field is the average of the reanalysis data, which excludes the data when TC is not active. Comparing the results from different experiments, it is found that using Betts-Miller cumulus parameterization, Blacadar planetary boundary layer scheme, and intenseinitial Bogus, the time and locations that model TCs get their maximum intensity are more consist with the observation. Using the typical field, the distribution of TC intensity in the WNP is more similar to the observation than the results of the experiments using mean field.Although different experiments are carried out using different model set up, the same results are found. Model results show that the TCs generated in the southeastern WNP are stronger than those generated elsewhere, consistent with observations. TC intensities are influenced by the SST, vertical wind shear (VWS), low level vorticity (VOR) and relative humidity (RHU); after calculating the maximum potential intensity and growing ratio of the "mean field" and "typical field", similar results are found. Differences in the TC intensities between the El Nino and La Nina conditions are mainly caused by the differences in the atmospheric conditions instead of SST. For TCs with similar genesis locations, differences in their intensities are primarily caused by differences in the VOR. All the model results are also used to regress the TC intensity change using the SST, VWS and VOR. It is found that the regressed TC intensity change could explan 40% of the model TC intensity change. The regressed TC intensity change also shows the same conclusion as the observations.