摘" 要:利用南方地區(qū)15年地面觀測(cè)凍雨數(shù)據(jù),結(jié)合有明顯凍雨形成物理機(jī)制的16種氣象因子,構(gòu)建具有可解釋性的XGBoost凍雨模型,其能夠很好地模擬非訓(xùn)練期間的凍雨個(gè)例和年凍雨日數(shù)的時(shí)間變化。利用1954—2023年的ERA5再分析資料驅(qū)動(dòng)該模型,獲得南方地區(qū)凍雨日數(shù)70年的長(zhǎng)時(shí)間序列。在長(zhǎng)達(dá)70年中,凍雨日數(shù)存在2~3年的顯著周期,且隨著全球溫度的升高,凍雨日數(shù)呈明顯的下降趨勢(shì),其有3個(gè)突變點(diǎn),分別為1958年、1967年和1990年,其中最長(zhǎng)的下降時(shí)段為1990—2023年。該研究為凍雨的氣候研究提供堅(jiān)實(shí)的數(shù)據(jù)支撐。
關(guān)鍵詞:凍雨日數(shù);預(yù)報(bào)模型;可解釋性;機(jī)器學(xué)習(xí);南方地區(qū)
中圖分類(lèi)號(hào):P426.6" " " 文獻(xiàn)標(biāo)志碼:A" " " " " 文章編號(hào):2095-2945(2024)32-0087-04
Abstract: An interpretable XGBoost model of freezing rain was constructed by using surface freezing rain data observed in southern China over the past 15 years, and combining 16 meteorological factors with obvious physical mechanisms of freezing rain formation. The model was able to simulate well the temporal variation of individual freezing rain events and the annual number of freezing rain days during non-training periods. A 70-year long time series of freezing rain days in the southern region was obtained by XGBoost freezing rain model which driven using ERA5 reanalysis data from 1954 to 2023. The freezing rain days showed an obvious decreasing trend with the increase of global temperature, and existed a significant cycle of 2-3 years. There were three mutation points in the long 70-year period, which are 1958, 1967 and 1990, with the longest period of decline from 1990 to 2023. This study provides solid data support for climate change studies of freezing rain.
Keywords: freezing rain days; forecast model; interpretability; machine learning; southern region
凍雨是我國(guó)南方地區(qū)冬季發(fā)生的一種極端天氣現(xiàn)象,其量雖不及雨和雪,但嚴(yán)重的凍雨一旦發(fā)生,對(duì)自然環(huán)境系統(tǒng)和社會(huì)經(jīng)濟(jì)將造成巨大的危害[1-2]。如2008年冬季南方地區(qū)發(fā)生的凍雨事件,覆蓋范圍廣泛、持續(xù)時(shí)間長(zhǎng),導(dǎo)致湖南、貴州、江西等地大面積斷電,交通中斷,農(nóng)作物顆粒無(wú)收,造成超千億元的經(jīng)濟(jì)損失[3]。
目前,凍雨預(yù)報(bào)主要依靠耦合了凍雨參數(shù)化方案的數(shù)值模式,如耦合在WRF模式中的凍雨參數(shù)化方案如Ramer方案、AFWA方案、Thompson方案和RUC方案等,它們根據(jù)凍雨發(fā)生時(shí)氣象要素的統(tǒng)計(jì)閾值和各降水相態(tài)形成的物理機(jī)制所得,但由于凍雨具有高度復(fù)雜性和多變性,目前的凍雨物理及經(jīng)驗(yàn)?zāi)P驮谔幚砩婕皟鲇甑闹T多氣象要素的復(fù)雜性和不確定性時(shí)仍面臨諸多挑戰(zhàn),憑借經(jīng)驗(yàn)統(tǒng)計(jì)得到的氣象要素閾值范圍存在偏差,導(dǎo)致這些凍雨參數(shù)化方案對(duì)凍雨的預(yù)報(bào)評(píng)分相對(duì)較低。……