Climate Change In The Qiantang River Basin And Its Influence On Local Runoff

Exceeded greenhouse gas emissions have already caused severe global warming during mid to late 20th century, which have simultaneously affected the hydrological cycle and redistributed the precipitation worldwide. Scientists all over the world focused on the global warming, due to its influence on the environment of human and other living beings, as well as hydrological cycle. Regional climate change investigation can provide significant information for tackling climate change, regional meteorological disaster warning and regulation of regional water resource, which benefits the local industrial manufacture, residents living and economic development.This thesis mainly focused on the climate change and its impact on water resource in Qiantang River basin, which is a vital fresh water supply to Zhejiang Province. This thesis employed various commonly used meteorological and hydrological statistical analysis techniques, including general circulation models, statistical downscaling and hydrological models, to analyze the spatial and temporal changes of climate and the influence on water resource in the Qiantang River basin. The main work and conclusions of this thesis are as follows:(1) The non-parameter trend analysis method, abrupt change analysis and periodicity analysis were applied to investigate the spatial and temporal changes of average and extreme climate change in the Qiantang River basin. The result showed an obvious warming trend, especially the average minimum temperature (Tmin), which was identified with significant increase.1996 was identified as change point for Tmin, after which the warming was greater. The temperature in winter increased most for all four seasons. The extreme low temperatures showed significant upward trend, while the trend of extreme high temperatures was insignificant. The spatial distribution of the trend was different among the extreme temperatures, as well as the seasons. The abrupt change in extreme climate occurred earlier in average climate.(2) The trends in precipitation indices, except summer precipitation, mostly were identified with insignificant trend, which were different form temperature changes. The trends in different ranges of precipitation indicated different directions, which the light precipitation (TP1 and TP2) decreased and the heavy precipitation increased. The PD95/TP9, which represents the extreme precipitation, exhibited opposite trend before/after 1985, and the difference in these two sub-periods was significant. The analysis of consecutive dry/wet days demonstrated the transition to wetness in recent decades. The study region received relatively abundant summer precipitation after 1992. Empirical orthogonal functions (EOF) captured the majority of the variation in the summer precipitation, which was confirmed by the accumulative variance contribution of the first three principal components, as well as the similarity of fluctuation between summer precipitation and the first principal component (PCI).(3) SWAT and HIMS models were applied to simulate the hydrological process in typical basin of Lanjiang, which is a sub-basin in Qiantang River basin.10 sensitive runoff parameters participated in runoff calibration in SWAT, which were screened by applying LH-OAT in SWAT sensitivity analysis module and SWAT-CUP software simultaneously, but the runoff calibration in HIMS was run directly without parameters sensitivity analysis, due to its few simple parameters in the runoff calibration. The simulation result in these two models is reasonable in both daily and monthly runoff simulations. The simulation effects displayed a little difference among stations, and generally the Quzhou station simulated most accurately, and the accuracy in Jinhua was relatively low. The comparison result showed SWAT displayed better hydrological simulation for monthly scale, while HIMS showed better simulation result for daily scale.(4) Statistical downscaling model (EOF-MLR) was established in the Qiantang River basin, which was based on the technique of Empirical orthogonal functions (EOF) and Multiple Linear Regression (MLR). The large-scale predictors, which were derived from NCEP re-analysis data, were processed with EOF analysis to obtain the first leading 10 principal components (PCs). Subsequently, link the large-scale predictors (PCs) to regional predictands (temperatures and precipitation) to get regression equations. Generally, the statistical downscaling models showed high skill to simulate the regional temperatures and precipitation, but the temperature simulation was much better than precipitation. The statistical downscaling models were also applied to different scenarios of HadCM3, GFDL and ECHAM5 to construct local historical temperatures and precipitation, and the results were skillful to simulate the local historical temperatures and precipitation, when compared to the direct output of GCMs simulations. This result simultaneously proved the applicability of the statistical downscaling models to large-scale predictors derived from GCMs.(5) The statistical downscaling models were also applied to A2 and A1B scenarios of the three GCMs to construct local temperatures and precipitation in mid and late 21st century. The predicted local future temperatures and precipitation were compared to those during baseline period (1961-1990). The results showed a significant increase in temperatures in the 21st century, and the warming in late 21st century is greater than that in the mid-21st century. The temperatures downscaled by HadCM3 displayed greater warming than those of the other two GCMs. The temperatures downscaled under A1B scenario would be warmer than those under A2 scenario in mid-21st century, while the simulation of A2 scenario was becoming warmer than that of A1B in late 21st century. The warming of same seasonal temperature indictor ranked the same in the two future periods. There were various differences of spatial distribution of representative month in summer and winter of future temperature changes between different month and different temperatures. For the same representative seasonal monthly temperature, the spatial distributions of warming in the study region were similar in mid and late 21st century. The changes in future precipitation were distinct in wet and dry seasons. Generally, the precipitation in dry season was lower in 21st century than that in baseline while in the wet season was higher than in baseline. The change of seasonal precipitation in late 21st century was greater than that in mid-21st century, especially for the spring and summer precipitation, which significantly increased in late 21st century. The difference of precipitation in January between future and baseline showed similar spatial distribution under different scenarios and two different future periods. The difference in June, however, showed different spatial distribution under different scenarios and future periods, due to the complex influence on the summer precipitation.(6) We combined the results of statistical downscaling of GCMs with specified future climate scenarios to get integrated scenarios, which were applied to the control period of 1970-1989 afterwards. The temperatures and precipitation under integrated scenarios were set as input of the hydrological models to investigate the change of runoff in Lanjiang basin under climate change. The annual runoffs in Jinhua station, which was simulated by HIMS under integrated scenarios, were lower than control, while the other stations were higher than the control by HIMS. The annual runoff simulated by SWAT under integrated scenarios was much greater than that of HIMS. Furthermore, the annual runoff simulation showed some change pattern under different integrated scenarios. For the same hydrological model, there was little difference between the annual runoff changes of the two emission scenarios of 30 series, while the increase of annual runoff was much greater under A2 emission than that under A1B of 70 series. Similarly, the monthly runoff simulated by SWAT was higher than that of HIMS, but the inner-annual variation of monthly runoff was similar between the two models. The monthly runoff under 30series was higher than that of 70 series, but there was little difference between the two emission scenarios for the same series. Furthermore, the remarkable feature is the monthly runoff change in wet seasons was different from that in dry seasons. Compared to baseline monthly runoff, the monthly runoff in wet seasons increased, while the monthly runoff in dry seasons decreased. This indicates different pattern of climate change influence on runoff in the study region.