<p>High spatial resolution weather forecasts that capture regional-scale dynamics are important for natural hazards prevention, especially for the regions featured with large topographical variety and local climate. While deep convolutional neural networks have made great progress in single image super-resolution (SR) which learns mapping relationship between low- and high- resolution images, limited efforts have been made to explore the potential of downscaling in this way. In the study, three advanced SR deep learning frameworks including Super-Resolution Convolutional Neural Network (SRCNN), Super-Resolution Generative Adversarial Networks (SRGAN) and Enhanced Deep residual networks for Super-Resolution (EDSR) are proposed for downscaling forecasts of daily precipitation in southeast China (100°E -130°E, 15°N -35°N). The SR frameworks are designed to improve the horizontal resolution of daily precipitation forecasts from raw 1/2 degrees (~50km) to 1/4 degrees (~25km) and 1/8 degrees (~12.5km), respectively. For comparison, Bias Correction Spatial Disaggregation (BCSD) as a traditional SD method is also performed under the same framework. The precipitation forecasts used in our work are obtained from different Ensemble Prediction Systems (EPSs) including ECMWF, NCEP and JMA which are provided by TIGGE datasets. A group of metrics have been applied to assess the performance of the three SR models, including Root Mean Square Error (RMSE), Anomaly Correlation Coefficient (ACC) and Equitable Threat Score (ETS). Results show that three SR models can effectively capture the detailed spatial information of local precipitation that is ignored in global NWPs. Among the three SR models, EDSR obtains the optimum results with lower RMSE and higher ACC which shows better downscaling skills. Furthermore, the SR downscaling methods can be extended to the statistical downscaling for other predictors as well.</p>

Previous studies have noted an abrupt decrease in western North Pacific (WNP) tropical cyclone (TC) genesis frequency and a westward shift in genesis location since the late 1990s. The recent application of cluster analysis in TC research shows the effect of detecting the contribution of the Western North Pacific Subtropical High (WNPSH) and the interdecadal Pacific oscillation (IPO) on interdecadal change in WNP TCs. In this work, we also apply a clustering algorithm called pHash + Kmeans to group WNP TCs into three classes based on their genesis environmental conditions. The clustering results show that an abrupt decrease after 1998 is related primarily to a decrease in the dominant class (Class3, located mainly in the southern and eastern WNP), and an increase after 2010 occurs because of a new dominant class (Class1, located mainly in the northwestern WNP), which indicates that the WNP environment suppresses Class3 genesis after 1998 and enhances Class1 genesis after 2010. Three periods (P1: 1979–1997, P2: 1998–2010, and P3: 2011–2020) and three regions (SCS: 100°E-120°E, EQ-30°N; WNP1: 120°E-140°E, EQ-30°N; and WNP2: 140°E-160°W, EQ-30°N) are divided to further confirm the above findings. In P1, high (low) mid-level relative humidity (RH), intense (weak) low-level vorticity, and weak (strong) vertical wind shear (VWS) are distributed in WNP2 (SCS and WNP1), indicating suitable environmental conditions for TC genesis in WNP2 but unsuitable conditions in SCS and WNP1. This situation is the opposite in P2, leading to a decrease in genesis frequency and a westward shift in genesis location. In P3, strong low-pressure vorticity and thermodynamic conditions occur in SCS and WNP1, contributing to an increase in TC genesis frequency.