本局引進自動即時預報系統(Auto Nowcaster, ANC)目的是要針對雷暴及邊界層有良好的即時預報，而提升ANC預報的品質有賴於優質的地面觀測資料，自動氣象站的即時觀測資料是最重要的地面資料，而各站資料品質不一。本研究利用2009-2011北臺灣44個自動氣象站資料，分析各站氣象資料的特性，並且做實地的場勘及GPS重新定位，以了解各站確切的地理位置及測站周遭環境，並用攝影資料定量計算出各站受阻擋的比率評估各站資料的代表性，進而篩選出可導入ANC系統的測站。
The Auto Nowcaster (ANC) prediction system has been ported at CWB. This system is able to predict thunderstorm initiation, development, and decay. It also can provide detail prediction of the boundary structure responsible for the storm initiation. However, accurate predictions solely relay on the good quality of surface observations. In this project, the 44 auto weather station (AWOS) data in Northern Taiwan collected between 2009 and 2011 is used to study the short-tem climatology of each station. In order to make the best use of these station data, a site survey is planned. Through the site survey, the GPS location and surrounding environment are documented. Fisheye-view photos will be taken at each side. Using the photogrammetry technique, the blocking percentage can be calculated. This blocking percentage provides a guide line for judging the applicability of station data for the ANC system.
