| Flooding is a major issue related to global sustainable development,and the analysis of nonstationarity and attribution of the evolution of flood extremes in basins under changing environment are frontier areas of research in hydrological science.Due to the influence of climate change and human activities,rainfall-runoff relationships in some basins show nonstationary characteristics,which may reduce the simulation and prediction capabilities of hydrological models under changing environment.Therefore,it is necessary to investigate the influence of nonstationarity in rainfall-runoff on the flood simulation of hydrological model.Climate change and human activities may also cause nonstationarity in the flood extreme sequences of river basins.To address this issue,it is crucial to explore how to improve the frequency analysis methods of the nonstationarity of flood extremes,quantify the contribution of climate change and human activities to the nonstationarity of flood extremes,and reveal the influence of nonstationarity on the trend of flood extremes from the perspective of the formation mechanism of floods in basins.The Dongting Lake basin is one of the most important tributaries of the Yangtze River basin.The hydrological processes in the Dongting Lake basin have undergone significant changes under changing environment.A comprehensive understanding of the evolution and causes of flood extremes in the basin can contribute to a deeper understanding of hydrological responses in the Dongting Lake basin under changing environment.To this end,three research areas were conducted in the Dongting Lake basin:(1)Based on daily rainfall data from 134 meteorological(rainfall)stations and daily runoff data from16 hydrological stations,the temporal and spatial variations and nonstationarity characteristics of rainfall extremes and flood extremes in the Dongting Lake basin were analyzed.(2)The applicability of the lumped hydrological model in simulating floods and soil moisture in the Dongting Lake basin was investigated,and the nonstationarity of rainfall-runoff in the Dongting Lake basin and its impact on flood simulation by hydrological models were analyzed.(3)Based on the mechanism of flood formation,the main influencing factors of flood extremes in the Dongting Lake basin were analyzed.The GAMLSS model was constructed with physical influencing factors as explanatory variables to quantify the contribution of climate change and human activities to the nonstationarity of flood extremes in the basin and to reveal the impact of the nonstationarity on the trend of flood extremes.The research results show that:(1)The overall trend of rainfall extremes in the Dongting Lake basin is increasing,with significant spatial differences.107 of 134meteorological(rainfall)stations show an increasing trend in rainfall extremes,of which32 stations show a significant increasing trend,and the point rainfall extremes show an overall gradual increase from southeast to northwest;14 of 16 basins show an increasing trend in annual maximum 3d area rainfall and the annual maximum 7d area rainfall,and the annual maximum 15 d area rainfall in 12 basins showed an increasing trend.The areas with significant increasing trends of point and area rainfall extremes are mainly concentrated in the Mishui basin of Xiangjiang River,the lower reaches of Zishui River and the middle and upper reaches of Yuanjiang River,with striking spatial differences.The sequences of rainfall extremes with significant trends are strongly correlated with the global temperature and Nino index(Nino3.4).(2)The flood extremes in the Dongting Lake basin show visibly trend changes and will maintain the current trend changes in the future.11 of the 16 hydrological stations show decreasing trend in flood peak sequences,and 6 stations such as Leiyang show clearly decreasing trend;among the remaining 5 stations with increasing trend,only Taoi station shows evidently increasing trend.12 of the 16 hydrological stations show decreasing trend in flood volume sequences,and 6 stations such as Chenglingji show significant decreasing trend;while the remaining 4 stations show nonsignificant increasing trend.The stations with significant trends of flood extremes are mainly concentrated in the basin of Xiangjiang River and the middle and upper reaches of Yuanjiang River.The flood peak sequences of 14 of 16 stations will keep the current trend in the future,and so it is for the flood volume sequences of 14 stations.The flood extremes in the basin will basically maintain the existing trend.(3)The TVGM,GR4 J and MISDc-2L hydrological models demonstrate strong applicability in flood and soil moisture simulation in the Dongting Lake basin.The TVGM,GR4 J,and MISDc-2L hydrological models have strong applicability in flood simulation in the Dongting Lake basin.The GR4 J and MISDc-2L models can simulate the soil moisture variation process in the basin well.(4)The rainfall-runoff relationship in the Dongting Lake basin show nonstationarity and has a strong influence on the flood simulation.For basins with nonstationary rainfall-runoff relationship,the model needs a longer rate period,and the longer the time interval,the greater the degree of accuracy degradation,and the difference of rainfall and similarity of rainfall process have a greater impact on flood simulation;when simulating future flood extremes,the closeness of the calibration period to the rainfall of the validation period is more important than the similarity of the rainfall sequences.(5)The change of LUCC in Dongting Lake Basin is small,and soil moisture,peak rainfall and reservoir engineering are the main factors affecting the evolution of flood extremes.The main reason for the significant decrease of flood extremes is the effect of reservoir storage.The stations with the trend of flood extremes increasing significantly all have the characteristics of less large reservoirs and more extreme rainfall.(6)The flood extremes in the Dongting Lake basin show nonstationarity,and the GAMLSS model with physical explanatory variables,such as antecedent influencing factors,has a more reasonable fitting effect.10 of the 16 hydrological stations show nonstationarity in flood peak sequences,and 9 stations show nonstationarity in flood volume sequences.The GAMLSS model with time as the explanatory variable has a good fitting effect on the peak-volume sequences of 16 hydrological stations,but the simulation performance of nonstationary stations with significant trends needs to be improved.After the introduction of physical explanatory variables,such as antecedent influencing factors,the model’s fitting effect is obviously more reasonable.Combining PI-PW model can better quantify the relative contribution of climate change and human activities to the nonstationarity of flood extremes;combining with time-varying function and copula function can reveal the influence of nonstationarity on the trend of flood extremes.In summary,this dissertation systematically reveals the evolution of flood extremes,nonstationarity characteristics and causes in the Dongting Lake basin,quantitatively evaluates the influence of non-smoothness of rainfall runoff on flood simulation by hydrological model,proposes a method for analyzing the frequency of nonstationarity flood extremes that can take into account the influence of pre-soil moisture,realizes the quantitative evaluation of the relative contribution of climate change and human activities to the nonstationary flood extremes,and reveals the influence of nonstationarity on the change of flood extremes.The research results are of great theoretical significance and application value. |
PDF Full Text Request