CAJ | 학술논문

近年来,临界慢化现象在揭示复杂动力系统是否趋向于临界性突变方面展示了重要潜力.在此基于临界慢化的理念,对含有不同噪声的中国一级气象地理区划的11大区域的气温观测资料进行处理,研究气候突变的前兆信号.首先利用Mann-Kendall (M-K)方法准确的找出各个区域的突变位置,然后计算了表征临界慢化现象的自相关系数以此寻找气候突变前兆信号,并研究了不同噪声对气候突变前兆信号的影响.结果表明,针对不同的信噪比、不同区域的气温资料在气候突变发生前5—10年均出现了自相关系数增大的临界慢化现象,临界慢化现象可能是气候突变发生前的一个前兆信号;噪声对气候突变前兆信号的检测结果影响较小,论证了临界慢化现象检测突变前兆信号的可靠性,为该方法在实际观测资料中的应用提供了实验基础.
근년래,림계만화현상재게시복잡동력계통시부추향우림계성돌변방면전시료중요잠력.재차기우림계만화적이념,대함유불동조성적중국일급기상지리구화적11대구역적기온관측자료진행처리,연구기후돌변적전조신호.수선이용Mann-Kendall (M-K)방법준학적조출각개구역적돌변위치,연후계산료표정림계만화현상적자상관계수이차심조기후돌변전조신호,병연구료불동조성대기후돌변전조신호적영향.결과표명,침대불동적신조비、불동구역적기온자료재기후돌변발생전5—10년균출현료자상관계수증대적림계만화현상,림계만화현상가능시기후돌변발생전적일개전조신호;조성대기후돌변전조신호적검측결과영향교소,론증료림계만화현상검측돌변전조신호적가고성,위해방법재실제관측자료중적응용제공료실험기출.
In recent years, critical slowing down phenomenon has shown great potentials in disclosing whether a complex dynamic tends toward critical cataclysm. Based on the concept of critical slowing down, the observed data of temperature in different regions in China which have different noises are processed in this article to study the precursory signal of abrupt climate change. First, Mann-Kendall(M-K)method is used to find the locations of the abrupt climate change in different regions, then the autocorrelation coefficient which can characterize critical slowing down is calculated;the appearance-time moments of early warning signals of abrupt climate change under the influence of different noises are also stadied. The results show that for different signal-to-noise ratios, the critical slowing down phenomenon has appeared in the data 5-10 years before the abrupt climate change took place, which indicateds that critical slowing down phenomenon is a possible early warning signal for abrupt climate change and the noise has less influence on the test results for early warning signals of abrupt climate change. Accordingly, it demonstrates the reliability of critical slowing down phenomenon to test the precursory signals of abrupt climate change, which provideds an experimental basis for the wide applications of the present method in real observation data.