Research On Hydrological Extreme Events And Riverine Ecological Problems Under Changing Environment

As climate change and human activities intensified,the global hydrological cycle and the evolutions of water resources have changed dramatically in the last decades,which have a great impact on the statistical characteristics of hydrological extreme events and the function of water ecosystems,thus causing serious threats to water security in China.On the other hand,the changing environment have also brought new challenges to hydrological researches,some traditional research methods have become invalid as a consequence.How to understand the trends and evolutions of hydro-meteorological events in a changing environment,and how to evaluate their impacts on water resources and water ecology in watersheds,are major issues in theoretical researches as well as practical applications of current scientific research field.This study is aimed to investigate the hydrological extreme events and ecological problems in rivers under changing environment,and the Yichang section,which belongs to the upper Yangtze River basin,is taken as a case study.The dissertation analyzed the spatio-temporal evolutions of hydrological time series,built the non-stationary frequency analysis models for extreme events,explored the variations of correlation structure between hydrological variables and assessed the changes of suitability of riverine habitats under multifactorial influences.The results in this study can provide theoretical basis and technical support to water resource management and ecological civilization construction in the study region.The main conclusions and innovations are summarized as follows:(1)The statistical tests and wavelet analysis methods were applied to explore the nonstationary properties in different streamflow series at the Yichang station,and the results showed clear evidence of the impact of changing environment on its flow regimes.On annual scales,the averaged streamflow series presented significant decreasing trend during the study period,and on seasonal scales,the streamflow decreased in summer and autumn while it increased in winter;as to the extreme events,the low flows presented decreasing trends on all time scales while the high flow series showed increasing trends.The abrupt change points were detected in 1960 s or the early 21 st century in almost all cases.The annual averaged streamflow series at the Yichang station have a primary cycle of 25 years according to the wavelet analysis,and there are seven times of changes between wet and dry in the study period.(2)A non-stationary frequency analysis model was built to investigate the effect of reservoir regulation on frequency characteristics of low flow series,some questions about model selection and residual analysis have also been discussed in this study.The results indicated that stationary models are no longer applicable under changing environment,while the non-stationary models have a better performance according to the residuals.Besides,covariates with explicit physical meanings are more suitable in fitting the hydrological time series than the time covariate and the results are more reasonable in forecasting the future.The low flow guarantee capacity was higher when the design values were calculated by non-stationary models in the study region.(3)A non-stationary standardized streamflow index was proposed in this study to fit the streamflow series,and a modified reservoir index was introduced as covariate to assess the effect of reservoir regulation.The copula method was applied for bivariate modeling of drought duration and severity,in which joint and conditional return periods were considered for drought risk assessment.The results indicated that the non-stationary index performed better than the traditional one and the drought characteristics showed significant differences: the type and parameters of the marginal distributions changed when considering nonstationary properties,and the correlation coefficient between drought characteristics increased in non-stationary models.The joint return period for different indices varies given the same design values,which is a reflection of the non-stationary properties in the monthly streamflow series.The model with modified reservoir index as covariate has smaller conditional return period,implying a worse drought condition compared with the stationary model.(4)A multivariate extension of the Mann-Kendall test was adopted for trend test between flood peak and 7-day flood volume series,and a time varying copula model was constructed for bivariate frequency analysis,with the non-stationary return period and risk of failure applied for flood risk assessment.The results demonstrated that the time-varying copula performs better than the stationary one in terms of AIC values,and the conclusions of the bivariate non-stationary return period and risk of failure were different depending on the design flood event.In the event that both flood peak and volume exceed,the flood risk is smaller with the non-stationary model,which is a joint effect of the time varying marginal distribution and copula function.While in the event that either flood peak or volume exceeds,the effect of non-stationary properties is more complex,the flood risk depends on the initial design standard for the hydraulic project.(5)The Indicator of Hydrological Alteration method was applied to quantify the effect of Three George-Gezhouba cascade reservoirs on downstream hydrological regimes,and a statistical habitat simulation model was proposed to investigate the variation of weighted useable area for Chinese sturgeons under multifactorial influences.The results indicated that the Gezhouba reservoir has little impact on hydrological regimes,while the Three Gorges reservoir caused moderate alteration to downstream flow regimes and it has changed the water temperature structure.During 2007 to 2012,the most suitable spawning time for Chinese sturgeon is in early-and mid-November,while the hydrological conditions in October can no longer meet the requirements for its spawning.Both streamflow and water temperature play an important role in the calculation of weighted useable area for the Chinese sturgeon's habitat,while the water temperature is more significant according to the model's fitting results,the statistical habitat simulation model provided in this study has obvious advantages in habitat evaluation under multifactorial influences.