Identifying plant water sources is helpful to understand the ecohydrological processes at the soil-vegetation-atmosphere interface,and it is extremely significant to the theory and practice,which for the rational planning and utilization of water resources in this area.However,there are certain differences and uncertainties between various methods of plant water source partitioning,which will affect the accuracy of quantified results.As such,this study took the Changwu tableland as the study region,and dynamically collected soil(n=515)and xylem samples(n=110)under apple trees with the age of 18 and 26 years old during the growing seasons in 2020 and 2021(from May to October).The stable and radioactive isotopes(?2H,?17O,?18O and 3H)of soil and xylem samples were determined.Then the factors affecting the accuracy of plant water isotopes were investigated,i.e.,the overall uncertainty and organic pollution interference,and further to discuss the effect on the plant water source partitioning.By analyzing the uncertainty in plant water source partitioning and quantifying the proportion of each component contributing to the overall uncertainty,we selected the optimal plant water source partitioning method.Finally,the selected method was used to identify the apple trees water use source in different years and ages,and then understand their response mechanisms to seasonal drought.The main conclusions are as follows:(1)Organic pollution had a certain impact on the plant water source partitioning.There was a significant difference between the isotopic values before and after correction ofδ2H,δ17O andδ18O(p<0.05),indicating that organic contamination had a significant effect on the isotopic determination of plant water.The effectiveness ranking of the correction method was:correction of methanol and ethanol contamination?correction of methanol contamination>correction of ethanol contamination.In addition,by analyzing the factors affecting the accuracy of plant water isotope measurements,the overall uncertainty of different isotopes was affected by various factors.The accuracy ofδ2H andδ17O determination were mainly interfered by the isotopic true value of high concentration standard samples;and the measurement ofδ18O was more influenced by the isotopic actual value of high concentration standard samples.Moreover,the organic contamination was dominated by methanol-like substances.Specifically,organic contamination had little effect on the quantification of plant water sources when hydroxide stable isotopes were used as tracers,however,when 17O was used to identify plant water sources,organic contamination interfered to some extent.(2)There were uncertainties and differences in the quantification of different plant water source partitioning methods.Specifically,there were significant differences in the quantification results obtained by different tracers and mixing models(p<0.05),however,plant water source quantifications were approximately the same before and after correction for deuterium bias.For the total uncertainty,the contributions were the largest for mixing models(37%),intermediate for tracers(28%),while the deuterium bias correction methods had the minor role in deviating plant water source partitioning.Furthermore,by comparing the effectiveness of different tracers and mixing models,the 2H18O combined with the Mix SIAR model was recommended as the best combination to quantify plant water use sources.(3)Water use strategies of apple trees with different years and ages were roughly the similar,but there were certain differences in the contribution ratios.There were certain differences between the isotopic compositions of precipitation,soil water and xylem water of apple trees with the age of 18 and 26 years in 2020 and 2021.Monthly precipitation and xylem water isotopic compositions of apple trees showed obvious seasonal changes;the?17O and?18O of soil water profiles under apple trees indicated a fierce change in shallow layers,and had a trend of decreasing variation with increasing soil depth.In addition,by using the optimal plant water partitioning method to quantify the apple trees water use source in 2020 and 2021,it can be seen that apple trees mainly depended on 0-3 m soil water in June and July,and gradually increased the proportion of soil water absorbed by apple trees in the soil layer below8 m from August to October.Moreover,the proportion of soil water absorbed by apple trees in the months of 2021 was higher(15%-45%)than in 2020(9%-32%).After analyzing apple trees water use of different ages,we can know that 18-year apple trees mainly utilized the 0-3m soil water in June and July,with the proportions of 59%-77%and 59%-79%,respectively.At the same time,the contributions of soil water in this layer to the 26-year trees were 74%-79%and 63%-72%,respectively.Until the arrival of summer(from August to October),apple trees turned to absorb soil water below 3 m with the ratio of 38%-64%.This study analyzed the uncertainties in identifying plant water source from basic data and partitioning methods,and clarified the dominant factors.Based on this,the optimal method was recommended to explore the plant water source,and further to understand plant water use strategies in arid and semi-arid areas.This study provided valuable insights into reasonable utilization,scientific management and sustainable development of water resources on the Loess Plateau. |