http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Russell L. Elsberry,William M. Clune,Grant Elliott,Patrick A. Harr 한국기상학회 2009 Asia-Pacific Journal of Atmospheric Sciences Vol.45 No.3
One of the objectives of the combined Tropical Cyclone Structure (TCS08) and THORPEX Pacific Asian Regional Campaign (T-PARC) field experiment during August and September 2008 was to improve understanding and prediction of western North Pacific tropical cyclone formation. One approach during TCS08 was to subjectively monitor the performance of the four global models as to whether any of the more than 50 deep convective regions labeled as TCSxxx were predicted to become a tropical depression. When a 850-mb vorticity maximum associated with the convective region could be followed in all four models, a consensus of the track forecasts to 72 h provided surprisingly good guidance for the pre-tropical cyclone seedlings and tropical depressions. An experienced analyst was able to recognize signatures in each of the models that indicate tropical cyclone formation was likely or unlikely. When all four global model forecasts were in agreement that formation was likely, high confidence could be given to the predicted scenario. This four-model consensus approach to predict tropical depression formation was most effective for the four pre-tropical cyclone seedlings that would later become typhoons, and did predict with less advance warning two seedlings that later became tropical storms. However, the global models did not consistently predict the four pre-tropical cyclone seedlings that only attained tropical depression status. The four-model consensus had a relatively small number of false alarms. Even though a three-model consensus had somewhat more false alarms, in most of the low latitude cases the lack of time continuity and a systematic decrease in consensus time to formation will make these cases relatively easy to recognize as false alarms.