Search

The Tianjin coastal region in Bohai Bay, Northern China, is increasingly affected by storm-surge flooding which is exacerbated by anthropogenic land subsidence and global sea-level rise (SLR). We use a combination of InSAR (Synthetic Aperture Radar Interferometry), CGPS (continuous GPS), and tide-gauge observations to evaluate the spatial variability of relative SLR (RSLR) along the coastline of Tianjin. Land motion obtained by integration of two tracks of Sentinel-1 SAR images and nineteen CGPS stations shows that the recent land subsidence in Tianjin downtown is less than 8 mm/yr, which has significantly decreased with respect to the last 50 years (up to 110 mm/yr in the 1980s). This might benefit from the South-to-North Water Transfer Project which has provided more than 1.8 billion cubic meters of water for Tianjin city since 2014 and reduced groundwater consumption. However, subsidence centers have shifted to suburbs, especially along the coastline dominated by reclaimed harbors and aquaculture industry, with localized subsidence up to 170 mm/yr. Combining InSAR observations with sea level records from tide-gauge stations reveals spatial variability of RSLR along the coastline. We find that, in the aquaculture zones along the coastline, the rates of land subsidence are as high as 82 mm/yr due to groundwater extraction for fisheries, which subsequently cause local sea levels to rise nearly 30 times faster than the global average. New insights into land subsidence and local sea-level rise could help the countrys regulators to make decisions on ensuring the sustainable development of the coastal aquaculture industry.

Understanding temporal expressions is the important foundation of many NLP tasks. However, the varied representations of temporal expressions is difficulty in analysis and understanding. To parsing expressions, an effective classification method of temporal expressions is significant. A temporal expression may belong to one or more classes, but the classification usually requires manual annotation characteristics. In this paper, we present a character embedding feature method based on distributed word representation and a multi-label classification method based on neural network. The Chinese temporal expressions can be classified automatically by their characteristics. Experiments on TempEval-2 datasets show that our method has achieved good performance.