|其他摘要||Most rivers in arid and semi-arid areas originate from alpine regions with different water sources. The seasonal frozen ground widely exists in the area, which influences the water infiltration rate due to its preventing rainfall and snowmelt water from infiltration into soil layer. Most meteorological and hydrological stations in Xinjiang are situated in lowland areas. All of these bring difficulties for the study on hydrological process in alpine watershed in Xinjiang. This study chooses the Kaidu river watershed as the study area where the seasonal frozen ground widely is distributed with the rainfall, glacier/snow melt water and ground water as main water sources. The observed soil temperature and moisture data from closed station have been used for analyzing the seasonal frozen-thaw process, and path analysis method has been applied to determine the importance of the seasonal frozen ground on spring peak flow. An existing hydrological model, i.e. SDHydro, was modified to simulate the runoff and its responses to climate change in the Kaidu river watershed. According to the study, the main conclusions are as follows:
(1) The thermal insulation of snow cover prevents the energy transformation between atmosphere and pedosphere, which results in the shallow frozen depth under thick snowpack and thick frozen depth under shallow snowpack. The soil temperature is lower under snow cover, and it will increase quickly with the influence of air temperature as snow disappeared. The increase of soil temperature resulting in snowmelt improves the soil moisture. Hence, soil temperature increases after snow cover disappeared, while soil moisture reach the maximum value before snow cover disappeared, and then decreases.
(2) The results based on path analysis showed that> 0ºC accumulative temperature in spring and precipitation during winter have a significant direct influence on spring peak flow, while <0℃accumulative temperature in winter, representing the freezing status of soil, has the significant indirect influence in the Kaidu river watershed. The determining factors of the seasonal soil frozen depth and the thawing days are different with elevation enhancement after analyzing the influencing factors of the seasonal soil frozen depth and the thawing days in Bayinbuluk, Hejing and Kuerle stations using path analysis method. The determining factor of the seasonal frozen depth changes from <0℃ accumulative temperature and maximum snow depth in low elevation to <0℃ accumulative temperature in middle elevation, and to relative humidity in high elevation. The determining factor of the thawing days changes from wind speed and maximum frozen depth in low elevation to sunshine duration in middle elevation, and to relative humidity in high elevation. The difference depends on the locations and their environment in different elevation.
(3) The SDHydro model which was developed on the basis of the principles of System Dynamics has been modified and used to simulate the streamflow in the Kaidu river watershed. The modification includes the estimation of snowmelt rate by replacing the daily mean temperature with the active >0℃ accumulative temperature, and the water infiltration calculation by expanding the influence of soil physical status on water infiltration from surface soil layer to deeper soil layers. Graphical comparison and statistical analysis on the simulated and observed streamflows indicated that the modified SDHydro model can efficiently simulate the streamflow including the low and high streamflow in the Kaidu river watershed, and the capability of the modified model to reproduce streamflow in low- and normal-flow years was better than that in high-flow years. The simulation tests also demonstrated that the modification of snowmelt estimation has a greater influence on snowmelt runoff than the modification of water infiltration calculation in different soil layers does.
(4) The streamflow in the Kaidu river watershed is more sensitive to precipitation variation than that to temperature variation after analyzing the sensitivity of streamflow to climatic disturbance. Temperature variation has a great influence on intra-annual streamflow distribution. The spring streamflow has an increasing trend and the spring peak flow has a significant increasing trend under future climate change. The starting time of the spring peak flow brings forward significantly under scenario RCP8.5 with the maximum greenhouse gas emission. The response of the occurring time of spring peak flow to temperature variation is not as significant as that of the starting time of spring peak flow. The main reason is that the occurring time of spring peak flow is controlled by spring temperature increasing rate. The starting time of spring peak flow happens from middle and late March to middle and late April, while the occurring time of spring peak flow happens from the beginning of April to middle May.|
张飞云. 开都河流域山区季节性土壤冻融过程对径流的影响[D]. 北京. 中国科学院大学,2016.