To collect data, 16 meteorological observation stations, 53 hydrological observation stations, 19 soil and vegetation observation stations, and a 20 acre experimental observation plot in Datan Township were established. For the convenience of data recording, each monitoring point is recorded in abbreviated form. Appendix B Table B1 lists the complete names, abbreviations, and corresponding meteorological parameters of each sampling point, so that readers can match the dataset. A total of 6756 samples were collected, including 1206 precipitation samples, 1101 surface water samples, 161 groundwater standard samples, 3779 soil water samples, and 509 plant water samples.

    The sample will be brought back to the laboratory refrigerator within 10 hours. Collect soil samples at a depth of 100 centimeters. Use a soil drill to drill soil samples into layers of 0-5 cm, 5-10 cm, 10-20 cm, 20-30 cm, 30-40 cm, 30-50 cm, 50-60 cm, 60-70 cm, 70-80 cm, 80-90 cm, and 90-100 cm in the sample plot. Divide the collected soil samples into two parts, one of which is placed in a 50ml glass bottle. The bottle mouth is sealed with a sealing film and transported to an observation station for frozen storage within 10 hours after the sampling date is marked to test stable isotope data. Place another part of the sample in a 50ml aluminum box and test the soil moisture content using a drying method. Plant sample collection: Use sampling scissors to collect the woody stems of plants, peel off the bark, place them in a 50ml glass bottle, seal and freeze for experimental analysis. All collected water samples were analyzed using a liquid water isotope analyzer (dlt100, Los Gatos Research, USA) at the Stable Isotope Laboratory of Northwest Normal University. Each water sample and isotope standard sample were injected continuously for 6 times. To eliminate the memory effect of isotope analysis, the first two injection values were discarded, and the average of the next 4 injections was taken as the final measurement result. To ensure the accuracy of the measurement results, a parallel sample is collected for each sample, and the average of the two measurement results is taken as the final value.

    For meteorological and hydrological data, use Mann Kendall (MK) for testing, eliminate outliers, and fill missing values with interpolation. Use a liquid water isotope analyzer (LWIA) for post analysis to detect raw isotope data. LWIA software can automatically check the fault prompts of the instrument and provide a batch of optional data filtering methods. Through LWIA, it is possible to identify which raw data values of the sample are incorrect and need to be retested to see the reasons for the data errors. At the same time, we will also use SPSS software to check the normality of the obtained isotope data.