No abstract

61The Tibetan Plateau (TP), known as the "sensible heat pump" and the 62 "atmospheric water tower", modifies monsoon circulations and regional energy and 63 water cycles over Asia (Wu and Zhang 1998; Zhao and Chen 2001a; Wu et al. 2007; 64 Xu et al. 2008b; Zhou et al. 2009). Strong ascent over the TP may transport lower-65 tropospheric water vapor and anthropogenic pollutants into the upper troposphere-66 lower stratosphere (UT-LS), which exerts an influence on the local ozone valley 67 (Zhou et al. 1995; Liu et al. 2003; Bian et al. 2011) and the aerosol-layer 68 enhancements near the tropopause (Tobo et al. 2007; Vernier et al. 2015). The TP also In the 1990s, a longer-term field experiment was conducted over the TP with the 84 support of the Japanese Experiment on Asian Monsoon (JEXAM). It estimated the 2008a; Zhang et al. 2012; Chen et al. 2011 Chen et al. , 2013. It found diurnal variations of et al. 2013; Hu et al. 2014; Zheng et al. 2014 Zheng et al. , 2015a Zheng et al. , b, c, 2016 Guo et al. 2015; 161 Zhuo et al. 2016; Wan et al. 2017). These problems may also cause large uncertainties 162 in reanalysis datasets and satellite products (such as air temperature, soil moisture, 163 surface heat fluxes, and radiation) over the TP (Li et al. 2012; Wang et al. 2012; Zhu 164 et al. 2012; Su et al. 2013; Zeng et al. 2016). 165To promote Tibetan meteorological research, the Third Tibetan Plateau 166Atmospheric Scientific Experiment (TIPEX-III), to continue for eight to ten years, OBJECTIVES. 173The field observational objective of TIPEX-III is to constitute a 3-D observation 174 system of the land surface, PBL, troposphere, and lower stratosphere over the TP. 175This system integrates ground-, air-, and space-based platforms based on the 176 meteorological operational networks, the TIPEX-III network, the existing NIOST (Fig. 1a). Consistent with the operational observations of the 265 CMA, at each site the measurement system measures soil water content ( Fig. 1a). The regional network consists of 33 sites over 270 Naqu (Fig. 1c), which began operating in August 2015, and 17 sites over Shiquanhe This network consists of six additional sites at Bange, Namucuo, Anduo, Nierong, 280Jiali, and Biru, and contributes to integrated research on the high-resolution land-281 surface and PBL processes over the central TP and their effects on mesoscale systems. 282These observations have been conducted at Shiquanhe, Namucuo, Naqu, Anduo, Gongshan (98.67°E, 27.75°N) station on the southeastern slope of the TP (Fig. 1b), a 300 key area for gauging water-vapor transports from the Indian Ocean to East Asia. (Fig. 1b). A primary goal of these observations is to explore the cloud (Fig. 1b). A follow-up field campaign using ground-based radars Tuotuohe, Mangya, Golmud, and Xining meteorological stations (Fig. 1b). Using PRELIMINARY ACHIEVEMENTS OF TIPEX-III. 328The implementation of TIPEX-III has enhanced the monitoring capability for the 380(1) Cloud diurnal variation and warm rain process. 3...

No abstract

Abstract:The performance of Day-1 Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement (GPM) mission (IMERG) and its predecessor, the Tropical Rainfall Measuring Mission (TRMM) Multisatellite Precipitation Analysis 3B42 Version 7 (3B42V7), was cross-evaluated using data from the best-available hourly gauge network over the Tibetan Plateau (TP). Analyses of three-hourly rainfall estimates in the warm season of 2014 reveal that IMERG shows appreciably better correlations and lower errors than 3B42V7, though with very similar spatial patterns for all assessment indicators. IMERG also appears to detect light rainfall better than 3B42V7. However, IMERG shows slightly lower POD than 3B42V7 for elevations above 4200 m. Both IMERG and 3B42V7 successfully capture the northward dynamic life cycle of the Indian monsoon reasonably well over the TP. In particular, the relatively light rain from early and end Indian monsoon moisture surge events often fails to be captured by the sparsely-distributed gauges. In spite of limited snowfall field observations, IMERG shows the potential of detecting solid precipitation, which cannot be retrieved from the 3B42V7 products.

We propose a large-margin Gaussian Mixture (L-GM) loss for deep neural networks in classification tasks. Different from the softmax cross-entropy loss, our proposal is established on the assumption that the deep features of the training set follow a Gaussian Mixture distribution. By involving a classification margin and a likelihood regularization, the L-GM loss facilitates both a high classification performance and an accurate modeling of the training feature distribution. As such, the L-GM loss is superior to the softmax loss and its major variants in the sense that besides classification, it can be readily used to distinguish abnormal inputs, such as the adversarial examples, based on their features' likelihood to the training feature distribution. Extensive experiments on various recognition benchmarks like MNIST, CIFAR, ImageNet and LFW, as well as on adversarial examples demonstrate the effectiveness of our proposal.