| 摘要: | 2012年6月11日晚間一道結構完整的梅雨鋒面通過北台灣，劇烈降水造成多處地區發生災情。為了瞭解此個案的降雨特徵，本研究利用中央大學觀測坪的JWD、Parsivel和2DVD，以及北部雨滴譜儀觀測網：翡翠、南港、霞雲站的JWD觀測資料進行此個案的雨滴粒徑分布(DSD)分析比較，並配合雨滴譜儀資料成功蒐集的時間將本研究分成弱降水時期、強降水時期和完整個案時期進行討論。
弱降水時期主要比較JWD、2DVD和Parsivel等三種雨滴譜儀的觀測特性。根據比較結果顯示，三種儀器無論是在DSD或降雨率的表現都具有非常好的一致性，也確認在強降水發生之前，這三種儀器觀測結果都非常相似。強降水時期比較JWD和Parsivel在強降水時期的觀測分析。Parsivel在強降水時中大雨滴容易會有高估的情況發生，是因為Parsivel的儀器限制導致雨滴誤判。比較DSD時序變化和降雨率趨勢，Parisivel的表現較為一致，JWD卻完全沒有抓到最大雨發生的特徵。降雨率方面，JWD和Parsivel觀測的降雨量和10米塔附設傾斗式雨量計的雨量變化趨勢一致，但是雨量上卻存在很大的落差，有可能是雨滴譜儀在強降水情況下的觀測限制所造成。
完整個案時期比較同一時間北部其他JWD測站的觀測資料。根據分析結果顯示，三台外站JWD計算的降雨量和鄰近自動雨量站的觀測結果差異不大。霞雲站在弱降水的情況下小雨滴濃度最高；但是在強降水的情況下最低。翡翠和南港站的DSD變化則非常類似。中央站的JWD觀測到最大的雨滴粒徑比其他站都小，且中大雨滴濃度都比其他站來的高。最後是分別利用五分山S波段雷達的Z-R關係式，以及中央大學C-Pol雷達的K_DP-R關係式進行降水估計。根據比較結果，中央站由於雨滴譜儀的雨量和實際的降水有落差存在，所以利用Z-R關係式估計的降水表現最差。因此在使用雨滴譜儀校驗雷達降水時，雨滴譜儀的觀測表現應該先經過測試才能確保雷達降水校驗的準確性。
In the nighttime of 11th June 2012, a mature Mai-Yu front passed through the northern Taiwan. The extreme rainfall event caused multiple areas flooding. In order to investigate the characteristic of drop size distribution (DSD) accompanied with this heavy rainfall event, we used JWD, Parsivel and 2DVD collocated at NCU, and three JWDs in FeiCui, NanGang and XiaYung to investigate the frontal precipitation. In order to get the complete data, my discussion is organized as follows: Part 1 is focused on the weak precipitation period. Part 2 is focused on the strong precipitation period. Part 3 is the whole period as the front passing through the northen Taiwan.
In the weak precipitation period, we made sure that before the strong precipitation happens, the three type disdrometers (JWD, 2DVD and Parsivel) operate consistently.During strong precipitation period, a significant DSD variation characteristic had been found. Due to the limitation of instrument, Parsivel tended to overestimate the concentration of medium to large drops in the strong rainfall intensity. Comparing the rain drops concentration with the rain rate varies with time, Parsivel showed a good agreement but JWD even did not get the most significant characteristic as the strongest rainfall occurred. The rain rates of JWD and Parsivel varied in the same trend, but compared the rain rates with the rain gauge observation in the 10 m tower at NCU, both of them showed obvious underestimation. We suspected the limitation of instrument made the rain rate underestimated.
We analyzed whole two days’ data of JWD at FeiCui, NanGang and XiaYung on 11th-12th June 2012. The accumulated rainrates of these three stations were similar to the tipping buckets nearby. For the DSD variation, we found the concentration of small raindrops at XiaYung was the most in weak precipitation period but the least in strong precipitation. The DSD variations of FeiCui were similar to NanGang. The largest size of the raindrop detected at NCU was the smallest but the concentration is the highest amont the four JWD stations.
For quantitative precipitation estimation (QPE) of radar, we used the WuFenShan radar data and disdrimeter data to estimate the rain rates based on the Z-R relationship. We also used the C-Pol radar data of NCU to estimate the rain rates based on the K_DP-R relationship. According to the analysis results, the K_DP-R relationship showed a good proformance on the rain rate estimation. The reflactivity of the WuFenShan radar needed to add 3 dBZ for the better rain rate estimation. Besides, the rain rate estimation showed the worst result at the station of NCU. The reason was the rain rate of disdromters underestimated seriously so the performance of the Z-R relationship was bad. To avoid these bad results happening again, we should confirm the disdrometer shows consistency well with tipping buckets before we use the disdromter data. |
