The Mechanisms Of Groundwater Recharge Originating From Rainfall In Semi-arid Regions

In semi-arid and arid regions,evaporation is high,while rainfall is little.Groundwater is the main source of water supplies,including drinking water and irrigation water.Rainfall is the most important contribution to groundwater recharge.Understanding the mechanism of groundwater recharge clearly,and estimating the recharge accurately thus are critical for sustainable groundwater resources management.However,the infiltration process is affected by many factors,such as the thickness of the vadose zone,soil types,and climatic conditions,to mention but a few.Moreover,recharge varies greatly in both space and time and is difficult to observe directly.Recharge,therefore,remains a poorly understood hydrological process.In this thesis,recharge was explored through numerous field-and modeling approaches.We carried out a field experiment in the Guanzhong Basin,China which is a semi-arid region.The soil water content,water level,percolation,and weather data were measured continuously.We systematically analyzed the process of infiltration originating from rainfall,and the factors influencing the infiltration process,evaluated methods to estimate recharge,as well as quantified the uncertainty of predictive recharge by the ways of numerical modeling and statistical analysis.Several important conclusions can be drawn below:(1)The patterns of rainfall infiltration for groundwater recharge through aeolian sand,coarse sand,and fine sand under different water table conditions were systematically studied.We analyzed the influence of rainfall intensity,duration of rainfall,thickness of the vadose zone,soil type,and initial soil water content on the infiltration process.The extinction depth of aeolian sand is 78 cm,and the observed minimum amount of precipitation that can recharge the groundwater is 10.4 mm.We analyzed the recharge characteristic curves in response to rainfall through fine sand and coarse sand under different water table depths systematically.Impulse response functions,including Poisson distribution and Rayleigh distribution functions,were used to simulate the recharge rates.The results show that the Poisson distribution function can be better used to characterize the recharge replenishment process for the coarse sand.The depth interval coefficient of the Poisson distribution decreases as the thickness of the vadose zone increases.In addition,the depth interval coefficient related to the Poisson distribution function of coarse sand is larger than fine sand.The Rayleigh distribution function has the characteristics of positive skew distribution.It thus can better characterize the groundwater recharge process.The Rayleigh distribution function can be better used to characterize the recharge replenishment process through fine sand.(2)The performances of four commonly used methods(FAO-56 method with the skin evaporation enhancement,maximum entropy production,water table fluctuation method,and Darcy’s law method)estimating bare soil evaporation under different water table depths are systematically evaluated.Based on the evaporation process of bare soil,we proposed a simple method that considers film flow to estimate evaporation during the stage 2 evaporation process.The results show that the extinction depth can be considered as a first-order indication to assess the performances of the FAO-56 method with the skin evaporation enhancement,maximum entropy production,and water table fluctuation method.The maximum entropy production method can obtain the best results independent of water table depths among all methods,but it overestimated evaporation under dry conditions,especially if the water table depths are larger than the extinction depth,and the results will overestimate more.The FAO-56 method with the skin evaporation enhancement underestimated the actual evaporation if the water table depth is less than the extinction depth,and the FAO-56 method with the skin evaporation enhancement overestimated actual evaporation if the water table depth is larger than the extinction depth.The water table fluctuation method only can be used if the water table depth is less than the extinction depth.In principle,Darcy’s law method can be used to estimate bare soil evaporation.However,reliable soil water retention parameters are required.To better reproduce the dynamics of evaporation under dry conditions,we proposed a simple method to estimate bare soil evaporation.The proposed method considers the film flow during the second stage evaporation process.Moreover,the proposed method is formulated in terms of pressure head,relative humidity,as well as soil water content depending on the input data availability.In contrast with actual evaporation from three laboratory experiments,the proposed method can reproduce evaporation dynamics.Compared to the actual evaporation from a lysimeter,the proposed method can reproduce the evaporation dynamics very well.In contrast with the maximum entropy production method,the proposed method can improve the accuracy of the estimation of evaporation by 10.6% on an annual scale.The results have significance for estimating potential recharge accurately.(3)The performances of the specific yield estimation obtained through different methods are evaluated systematically in the context of the water table fluctuation method.The results show that a constant specific yield including the pumping test and the difference between saturated water content and field capacity cannot obtain satisfactory recharge.However,both depth-dependent methods to obtain specific yield can improve the accuracy of estimation of recharge significantly under different water table depth conditions.Care must be taken if the depth-dependent method based on the soil water retention curve is used,as it is highly sensitive to the chosen parameters.Given the uncertainty of specific yield,multiple approaches to estimate specific yield are suggested,if it is possible.(4)We set up a one-dimensional saturated and unsaturated flow model to predict groundwater recharge,quantify the uncertainty of predictive recharge,analyze the parameter identifiability,and assess the worth of observed data.The results show that the prediction of recharge can be highly uncertain,even though a good fit between observed data and simulated data was obtained during the calibrated and validated periods.The uncertainty can be reduced by 64% through calibration.In addition,the maximum uncertainty variance increased through the loss of the observations of soil water content close to the water table and observations of the soil water content from the topsoil and water table.Neglecting observations from soil water content in the middle of the lysimeter has less impact on uncertainty.These analyses allow identifying the most important observations for predicting recharge.Moreover,in terms of the estimation of parameters,the porosity can be calibrated to a great degree of certainty,while the parameters related to evaporation remain uncertain.The results indicate that quantifying the uncertainty is important for predictive recharge to provide reliable information to water managers.In particular,the results have important implications for the design and operation of field experiments.