4. CONCLUSIONS

In this study, the sources of nitrates in the Huashan watershed were quantitatively analyzed using a combination of nitrogen and oxygen isotopes and the SIAR model, and the uncertainty and sensitivity of the model were evaluated. The conclusions are as follows:
(1) The spatial and temporal variation of nitrate in surface water and groundwater in the Huashan watershed is large, with low concentrations of nitrate in surface water and groundwater in spring and autumn and high concentrations of nitrate in summer and winter; concentrations in surface water are higher in the lower reaches of the watershed, while nitrate concentrations in groundwater are higher in areas near the upper reaches. Surface water and groundwater are both supplied by atmospheric precipitation, but surface water is more affected by evaporation and has a wider range of sources, and the interaction between surface water and groundwater is closely connected.
(2) Using hydrochemical and δ15N、δ18O qualitative analysis, we found that SN is an important source of nitrate in the study area, and in April 2022, groundwater was influenced by MS. Additionally, denitrification occurred in groundwater near the watershed outlet in November 2021, while no denitrification was observed in surface water or groundwater in April 2022.
(3) Using the Stable Isotope Analysis in R (SIAR) model, we have determined that SN contributes on average 30.4% to the nitrate content in the Huashan basin, while AD only contributes 12%. In November 2021, the largest source of nitrate in surface water was from M&S (38.1%), while in groundwater it was from NF (39.8%). Similarly, in April 2022, NF was the largest contributor to nitrate in surface water (38.3%), and M&S in groundwater (29.8%). These results are consistent with the qualitative analysis. SN add NF were found to contribute to 58% and 64% of nitrate in surface water and groundwater, respectively. To optimize watershed management, it is recommended to improve the efficiency of nitrogen fertilizer use through better fertilization and irrigation practices.
(4) The larger uncertainty was associated with SN and NF for both surface water and groundwater, with M&S showing moderate uncertainty. AD showed the least uncertainty. In addition, for δ15N, changing the mean value of M&S had the greatest effect on the results, followed by the sensitivity of SN and NF, and the lowest sensitivity of AD; for δ18O, only the mean value of AD oxygen isotope values had a greater effect on the results. Influence of end-member inputs on nitrogen source apportionment is significant, therefore, it is recommended to prioritize the sampling and establishment of end-member values for pollution sources with higher contributions.