过去500年华南地区冷暖变化记录及其对冬季温度变化的定量指示意义

PHENOLOGICAL COLD/WARM EVENTS RECORDED IN HISTORICAL DOCUMENTS AND QUANTITATIVE PROXIES FOR WINTER TEMPERATURE IN SOUTHERN CHINA DURING THE PAST 500 YEARS

摘要整编了历史文献中记载的1400年以来华南地区霜冻灾害及冰冻灾害南界、降雪南界及日数和初／终霜冻灾害日期记录；并结合1951年以来器测气温序列，分析了各类记录与华南地区11～2月平均气温的相关性，揭示其对温度变化的定量指示意义。结果表明：华南地区的霜冻灾害南界、冰冻灾害南界、降雪南界和降雪日数可较好地指示该区域的温度变化；且24。N以南地区（包括广东、广西、台湾中南部和海南）的最早与最晚霜冻灾害日期对华南地区温度变化亦具有指示作用。其中霜冻灾害南界、冰冻灾害南界、降雪南界向南推进1个纬度，华南11～2月平均气温分别约低0．23℃、0．41℃和0．40℃；南岭两侧50km范围内的降雪日数每多（少）i0天，华南11～2月平均气温约低（高）0．22℃。受原始记载缺失及距今越近、记录数量越多等的影响，1400～1500年间有记录的年份只有7年，而1501～1950年间有记录的年份则有309年。1951年以来，冰冻灾害和降雪南界从未到海南，霜冻灾害南界到海南岛的年份有8年；而1501～1950年间，冰冻灾害南界虽也从未到达海南岛，但降雪南界有5年、霜冻灾害南界有22年到达海南岛，且整个区域的初（终）霜冻灾害平均日期也早（晚）于1951年以后。这为下一步重建这一地区更长时段的年分辨率温度序列提供厂重要依据。

Six kinds of phenological cold/warm events are collected from historical documents in Southern China since 1400A.D., including southern limits of frost disaster, freezing disaster and snow, as well as snowfall days and the first and last frost disaster dates. Instrumental temperature observations since 1951 are used to analyze the correlation between each kind of pheuological event and mean temperature from November to February. The results show that southern limits of frost disaster, freezing disaster and snow, along with snowfall days are effective quantitative proxies of temperature, whose correlation coefficients with winter temperature are 0.524, 0.689, 0.553 and 0.512 separately, all of which pass the significant level of et = 0.01. Liner regressions indicate that the winter temperature decrease by 0.23~C , 0.41~C , 0.40~C or 0.22~C separately when the southern limit of frost disaster, freezing disaster or snow is extended southward per degree of latitude or the snowfall days in 50-kin radius buffer zones around Nanling Mountains increases by 10 days. Though the first and last frost disaster dates for the whole region are not significantly correlated with winter temperature, such correlation coefficients are 0.214 and -0.345 in subarea where the latitude is lower than 24~N, the latter pass the significant level of ct=0.1, which means they can partly explain the temperature variation. The freezing disaster and snow never occur in Hainan Island after 1951 but there are 5 years with snowfall phenomenon from 1501 to 1950. As for frost disaster, there happens 8 years and 22 years in Hainan Island after 1951 and from 1501- 1950. Moreover, the average first and last frost disaster dates for the whole region are earlier and later in ancient time than that in present. The amount of original records increases with time process according to statistics, for example, there are only 7 years that have records from 1400 to 1500 while the number is 309 years from 1501 to 1950. The southern limit or disaster dates extracted from historical documents may not be the southernmost or the earliest/latest one that truly happened because of the shortage of records. These results provide an important foundation for reconstructing longer temperature series in higher resolution.