基于统计降尺度模型的博斯腾湖流域未来气温和降水变化趋势分析

SDSM统计降尺度模型是解决空间尺度不匹配问题的有效工具,它使气候变化响应研究得以在区域尺度上展开。本文将SDSM模型应用于博斯腾湖流域分析它的适用性,并对流域未来气温和降雨的变化趋势进行了预测。以日平均气温、最高气温、最低气温和日平均降雨量为预报量,选取合适的NCEP大气环流因子为预报因子,建立预报量与预报因子间的回归关系。利用1961年-1990年、1991年-2001年的实测数据和NCEP大气变量分别对SDSM模型进行率定和验证,效果较好。把HadCM3输出的A2、B2情景下的大气环流变量作为模型输入变量,模拟流域未来3个时期(2020s、2050s和2090s)的气温和降水变化。结果显示,流域未来日平均气温、最高气温和最低气温都呈现明显上升趋势,升高幅度为:日最高气温日平均气温日最低气温,且A2情景下气温增幅略大于B2情景;冬季气温增幅最小,夏季增幅最大;与全国降水的增加趋势不同,博斯腾湖流域未来的年降水量呈现明显减少趋势。分析结果可为博斯腾湖流域开展气候变化的水文响应研究以及气候变化的适应性研究提供科学依据。

Analysis of Trends in Future Temperature and Precipitation in the Bosten Lake Basin Based on a Statistical Downscaling Model

Growing interests in simulating changes in hydrological processes at global and regional scales correspond to increasing floods and draughts at the global scale induced by climate change, which would have a significant impact on social and economic systems, food security and the survival and development of humans. General circulation models (GCMs) perform reasonably well in simulating current climate with respect to annual or seasonal averages at large spatial scales, thereby showing the capability to provide inputs for studies associated with continental or global scale hydrological responses to climate change. However, the models perform poorly at small spatial and time scales relevant to regional impact analyses. The Statistical Downscaling Model (SDSM) is considered a very useful tool to address spatial-matching problems. It enables the study of climate change response to be carried out at local scales. In this paper, the capability of SDSM in the Bosten Lake basin was evaluated and the model was used to analyze the trends in future temperature and precipitation in this watershed. First, mean temperature, maximum temperature, minimum temperature and mean precipitation were selected as the predictands. Appropriate meso-scale atmospheric variables were selected as downscaling predictors. Then, a statistical transfer function between them was constructed by the United States National Centers for Environmental Prediction (NCEP) re-analysis data set and observations at four stations. The first 30 years (1961-1990) of the obtained data series (1961-2001) was used to calibrate the model. The remaining ten years (1991-2001) data series was used to validate the model. Results showed good agreement between the observations and the outputs from the model. On the basis of this, with the outputs of Hadley Centre Coupled Model, version 3 (HadCM3) under A2 and B2 scenarios, respectively, the calibrated SDSM model generated future daily temperature and precipitation series, which were subsequently used to analyze the temporal trends. Results showed that there is an obvious increasing trend for the future mean temperature, maximum temperature and minimum temperature in the Bosten Lake basin at a daily, monthly, seasonal and yearly scale, respectively. The increase range in A2 scenario is larger than that in B2 scenario. The broadest range was found in summer while the lowest is in winter. Different from the average temporal trends of China, the annual precipitation of the watershed in the future would exhibit a decreasing trend for both A2 and B2 scenarios. This study would provide a scientific basis for further studying climate change responses over the Bosten Lake basin.