By using the multi-fractal detrended fluctuation analysis method, we analyze the nonlinear property of drought in southwestern China. The results indicate that the occurrence of drought in southwestern China is multi-fractal and long- range correlated, and these properties are indifferent to timescales. A power-law decay distribution well describes the return interval of drought events and the auto-correlation. Furthermore, a drought risk exponent based on the multi-fractal property and the long-range correlation is presented. This risk exponent can give useful information about whether the drought may or may not occur in future, and provide a guidance function for preventing disasters and reducing damage.
In recent years, the phenomenon of a critical slowing down has demonstrated its major potential in discovering whether a complex dynamic system tends to abruptly change at critical points. This research on the Pacific decadal oscillation(PDO) index has been made on the basis of the critical slowing down principle in order to analyze its early warning signal of abrupt change. The chaotic characteristics of the PDO index sequence at different times are determined by using the largest Lyapunov exponent(LLE). The relationship between the regional sea surface temperature(SST) background field and the early warning signal of the PDO abrupt change is further studied through calculating the variance of the SST in the PDO region and the spatial distribution of the autocorrelation coefficient, thereby providing the experimental foundation for the extensive application of the method of the critical slowing down phenomenon. Our results show that the phenomenon of critical slowing down, such as the increase of the variance and autocorrelation coefficient, will continue for six years before the abrupt change of the PDO index. This phenomenon of the critical slowing down can be regarded as one of the early warning signals of an abrupt change. Through calculating the LLE of the PDO index during different times, it is also found that the strongest chaotic characteristics of the system occurred between 1971 and 1975 in the early stages of an abrupt change(1976), and the system was at the stage of a critical slowing down, which proves the reliability of the early warning signal of abrupt change discovered in 1970 from the mechanism. In addition, the variance of the SST,along with the spatial distribution of the autocorrelation coefficient in the corresponding PDO region, also demonstrates the corresponding relationship between the change of the background field of the SST and the change of the PDO.
A strong （weak） East Asian summer monsoon （EASM） is usually concurrent with the tripole pattern of North Atlantic SST anomalies on the interannual timescale during summer, which has positive （negative） SST anomalies in the northwestern North Atlantic and negative （positive） SST anomalies in the subpolar and tropical ocean. The mechanisms responsible for this linkage are diagnosed in the present study. It is shown that a barotropie wave-train pattern occurring over the Atlantic-Eurasia region likely acts as a link between the EASM and the SST tripole during summer. This wave-train pattern is concurrent with geopotential height anomalies over the Ural Mountains, which has a substantial effect on the EASM. Diagnosis based on observations and linear dynamical model results reveals that the mechanism for maintaining the wave-train pattern involves both the anomalous diabatic heating and synoptic eddy-vorticity forcing. Since the North Atlantic SST tripole is closely coupled with the North Atlantic Oscillation （NAO）, the relationships between these two factors and the EASM are also examined. It is found that the connection of the EASM with the summer SST tripole is sensitive to the meridional location of the tripole, which is characterized by large seasonal variations due to the north-south movement of the activity centers of the NAO. The SST tripole that has a strong relationship with the EASM appears to be closely coupled with the NAO in the previous spring rather than in the simultaneous summer.
A regional atmosphere-ocean coupled model, RegCM3-POM, was developed by coupling the regional climate model （RegCM3） with the Princeton Ocean Model （POM）. The performance of RegCM3-POM in simulating a persistent snow storm over southern China and the impact of the Madden Julian oscillation （MJO） on this persistent snow storm were investigated. Compared with the stand-alone RegCM3, the coupled model performed better at reproducing the spatial-temporal evolution and intensity of the precipitation episodes. The power spectral analysis indicated that the coupled model successfully captured the dominant period between 30 and 60 days in the precipitation field, leading to a notable improvement in simulating the magnitude of intraseasonal precipitation variation, and further in enhancing the intensity of the simulated precipitation. These improvements were mainly due to the well-simulated low-frequency oscillation center and its eastward propagation characteristics in each MJO phase by RegCM3-POM, which improved the simulations of MJO-related low-frequency vertical motions, water vapor transport, and the deep inversion layer that can directly influence the precipitation event and that further improved the simulated MJOprecipitation relationship. Analysis of the phase relationship between convection and SST indicated that RegCM3-POM exhibits a near-quadrature relation between the simulated convection and SST anomalies, which was consistent with the observations. However, such a near-quadrature relation was not as significant when the stand-alone RegCM3 was used. This difference indicated that the inherent coupled feedback process between the ocean and atmosphere in RegCM3-POM played an important part in reproducing the features of the MJO that accompanied the snow storm.
