Optimizing Chloride Tracing Technology To Estimate The Impacts Of Vegetation Change On Potential Groundwater Recharge On The Loess-Covered Region

Potential groundwater recharge refers to the deep drainage of water in the unsaturated zone but has not yet reached the groundwater table.As a crucial part of soil water balance,groundwater recharge plays an irreplaceable role in assessing the evolution of the water cycle in the terrestrial system.Chloride-based tracing technology is widely used to estimate potential groundwater recharge.However,the results of different research scales and/or objects differ greatly because of the method diversity and scarcity of basic data.On the other hand,the“Green for Grain”Project made great conversion from shallow-rooted plants to deep-rooted plants,the effect of vegetation change on potential recharge may be directly and/or indirectly change the soil water balance.Therefore,seeking an effective method for investigating the direct or indirect effects of a long-term vegetation change on potential groundwater recharge has become a key technique to explore how vegetation changes affect the transformation mechanism of soil water balance components.It have irreplaceable significance for achieving the goal of sustainable development of eco-economic-social relationships in the new era.Based on the increasing environmental problems caused by‘Grain to Green’projects on the Loess Plateau,using the mass balance theory and multiple tracing techniques,this study is mainly carried out from two aspects:method optimization of potential recharge estimation（parameter reconstruction and method optimization）and the response mechanism of potential recharge to vegetation change.This study first clarified the groundwater recharge mechanism in the loess regions,and reconstructed the long-term average atmospheric chloride deposition information according to the soil chloride records stored in the deep loess profile.Then,assessed the result differences of recharge estimation by Cl-based approaches to determine the optimal method.The two steps could decrease the uncertainties caused by the data and parameter absence needed for estimating recharge.The spatial variability of potential recharge under shallow-and deep-rooted vegetation and its dominant factors were analyzed by using the selected Cl-based method.According to the interactions between vegetation change and recharge mechanisms,the spatial variability of each water balance component and their responses to vegetation were further explained.The main conclusions are as follows.（1）The use of multiple tracers identified the potential groundwater is mainly recharged in the form of piston flow,supporting that the Cl-tracing method could be used for potential recharge estimation in the unsaturated zone.The single-peak distributions of chloride,nitrate and tritium in the vertical profile highlight the dominance of piston flow in soil water infiltration.The oxygen isotope composition in precipitation is more enriched than that in groundwater,while the chloride concentration in precipitation is lower than that in groundwater,which proved that rainfall infiltration in the unsaturated zone mainly occurs in the form of piston flow.The comparison results of double isotope showed that soil water affected by evaporation contributes more to the potential recharge,so the contribution of piston flow is much greater than that of preferential flow.Results were applicable to other regions with similar deep unsaturated zones in the Loess Plateau.（2）The soil water and chloride contents on regional scale were measured to provide data support for potential groundwater recharge.The distributions of soil water and chloride vary greatly in the horizontal and vertical directions.Mean soil water content and storage increased horizontally along the northwest-southeast direction,while water deficit varied with vegetation types.In the vertical direction,mean soil water content and storage were the highest in the relatively stable layer,followed by that in the rapidly changing layers and the active layers,which are opposite to water deficit.The estimated mean soil water contents under shallow-and deep-rooted vegetation were 12.9±3.2 and 6.4±2.8 g 100g^-1,respectively.Soil chloride concentrations had large spatial variation.Vertically,most soil chloride profiles had peak depths above 6 m and were stabilized below 6 m.（3）The modified inversely transforming chloride storage method（ITCS）and inversely transforming chloride mass balance method（ITCMB）were used to reconstruct atmospheric chloride deposition（ACD）information at different locations and periods to solve the problem of insufficient observed ACD data.According to the tritium age dating method and fertilization history,the modified methods can effectively exclude the fertilizer influence.According to the results of the multimodel inference method and stepwise linear regression method,climate and geographic factors had much greater influences on the ACD reconstruction than vegetation and soil factors in different periods.The distribution of the predicted ACD concentrations in 1963-1993 was similar to the annual average potential evapotranspiration distribution with an average value of 2.36±0.21 mg L^-1,while the distributions of ACD concentration in the stable period（before 1963）were opposite to precipitation distribution with an average value of 1.53±0.68 mg L^-1.The modified methods performed satisfactorily in estimating long-term average ACD fluxes and concentrations.（4）Assessing the uncertainties of multiple chloride-based mass balance methods can confirm the optimal way to estimate potential recharge.The CMB and CWMB methods were the best methods in estimating the potential recharge under the shallow-and deep-rooted vegetation,respectively.The potential recharges estimated by TP(33.6±11.2 mm year^-1)had no significant difference with those estimated by CMB,CS and CP(23.2±17.8～38.0±16.6 mm year^-1)under the shallow-rooted vegetation.After the conversion from shallow-to deep-rooted vegetation,the recharge rates quantified by TP/TP-WB,CMB,CS,CP and CWMB were reduced by7%～80%.The large discrepancy（CV=0.71）in estimated drainages originates from methodological theory such as the time scales of representation,sensitivity to exogenous Cl inputs and root water uptake-induced soil water depletion.（5）On point-scale,the potential groundwater recharge under shallow-rooted plants had obvious spatial variability,and it directly decreased by 50%after the transition to deep-rooted plants.The mean potential recharge of the 110 soil profiles was 22±36 mm year^-1,accounting for 5%±5%of the mean annual precipitation.Low recharge rates(<20 mm year^-1)under shallow-rooted vegetation mainly occurred in the arid region,while high recharge rates(>40 mm year^-1)were mostly distributed in the semi-humid region.Recharge rates under deep-rooted vegetation with mean value of11±37 mm year^-1 were much smaller than those under shallow-rooted vegetation(averaged 39±22 mm year^-1).Plant age and precipitation dominated the change of potential recharge.However,regardless of shallow-or deep-rooted vegetation,climatic factors played the most important roles in affecting recharge.（6）The established spatial distribution of plant age was used for evaluating the vegetation impact on the regional potential recharge.The conversion from shallow-to deep-rooted plant changed the transformation path among the components of the water balance system.The estimated mean age of deep-rooted vegetation in the loess region was 17±7 years.After the conversion from shallow-to deep-rooted vegetation,the change in soil water storage calculated by soil water was-122±31 mm year^-1.The estimated potential recharge under shallow-rooted vegetation was 40±12 mm year^-1.While the recharge rate under deep-rooted vegetation had no recharge in most areas of the loess tableland,with an average value of 0.3±2.9 mm year^-1.The predicted average annual evapotranspiration in the loess region was 452±95 mm year^-1,accounting for 96±10%of the mean annual precipitation.Higher evapotranspiration(>500 mm year^-1)mainly occurred in areas with dense deep-rooted vegetation.While lower evapotranspiration(<300 mm year^-1)mainly occurred in the northwest of the loess area.The conversion from shallow-to deep-rooted vegetation significantly increased the evapotranspiration.Such conversion also reduced both soil moisture and groundwater recharge,those change was significantly related to the plant age,indicating the overwhelming dominance of vegetation-related factors in influencing soil water balance.This study reconstructed the atmospheric chloride deposition information using the mass balance theory and multiple tracing techniques,which fills the parameter history information gap on estimating potential groundwater recharge,and optimize the method of estimating potential groundwater recharge within changing environment by using the limited-natural chloride records,which is of great significance to improve the accuracy of current assessment of groundwater resources.The study clarified the response mechanism of soil water balance components to vegetation,which provides the new reference ideas on how to balance the relationships between vegetation planning and water sustainability or water conservation.