摘要
利用喀什市1981-2010年0.8、1.6和3.2m逐月平均地温资料,计算各深层地温气候趋势系数和气候倾向率,并对其变化规律进行研究。结果表明:(1)近30a来,喀什市各深层地温的年均值呈明显的阶段性特征,1984-1992年呈极显著下降趋势(P<0.01),1996-2004年呈极显著上升趋势(P<0.01),全分析期各层地温线性变化趋势不显著。季节变化上,各深层平均地温在冬、春季均呈不显著上升趋势,夏季和秋季各层地温均呈下降趋势,其中0.8m土层在夏季(P<0.01)、1.6m土层在秋季(P<0.05)分别达到极显著和显著下降水平。(2)年代际变化上,20世纪80年代各层平均地温均高于90年代和21世纪初10a。(3)气温的变化是影响深层地温变化的主要因素之一,气温与各深层地温之间呈正相关关系;降水对深层地温也有一定影响,深层平均地温较低与降水量的增加有关。(4)Mann-Kendall突变检验表明,喀什市0.8、1.6m深层地温年均值在1985年和2009年发生了突变,3.2m深层地温年均值在1985年和2008年发生了突变。研究结果可为指导喀什农业生产、应对气候变化提供科学依据。
The variation of monthly average soil temperature at 0.8m,1.6m and 3.2m deep layers in Kashi from 1981 to 2010 was investigated by using linear trend analysis and accumulated variance methods.The results showed that annual average soil temperature at deep layers of 0.8m,1.6m and 3.2m decreased from 1984 to 1992 significantly(P〈0.01)and increased significantly from 1996 to 2004(P〈0.01),but the liner trend of soil temperature was not significant during the whole period.Soil temperature at different deep layers increases significantly in winter and spring but decreased significantly in summer and autumn,of which at 0.8m layer in summer(P〈0.01),1.6m layer in autumn(P〈0.05) was significant respectively.From decadal variation,average soil temperature at different deep layer in 1980s was higher than that in 1990s and first 10 years of 21st century.Air temperature variation was one of the main factors affecting deep layer soil temperature and there was positive correlation between them.Precipitation also had a certain impact on the deep soil temperature,and average soil deeper temperature was related to precipitation increasing.Annual average soil temperature at 0.8m and 1.6m depth had abrupt changes in 1985 and 2009,and 3.2m depth was in 1985 and 2008 by the Mann-Kendall test.The results could provide scientific reference to adapt to climate change for Kashi.
出处
《中国农业气象》
CSCD
北大核心
2014年第3期237-242,共6页
Chinese Journal of Agrometeorology
关键词
喀什市
深层地温
气候倾向率
M-K检验
Kashi city
Soil temperature at deep-layer
Climate linear trend
Mann-Kendall test