Research On The Roles Of Precipitation Sources In Summer In Eastern China Using Stable Isotopes

Ocean-continent interaction,as a significant part of climate teleconnection systems,impacts global water cycles and regulates regional extreme hydrological events.Eastern China is a typical region where monsoon controls summer precipitation,the Asian Monsoon and ENSO phenomenon would influence the moisture transport to Eastern China,and cause heavy precipitation and severe flood,which adversely affects the development of social economy and safety of the people.Studying the role of oceanic moisture from the Pacific Ocean to summer precipitation in eastern China is important to understand and forecast the formation of rainstorm.Previous studies about the contribution of OM to precipitation in eastern China mainly have focused on analytical models and numerical models,while physical water vapor tracers method based on stable isotope has seldom been applied.This mainly results from the limitation of traditional isotopic methods and scarcity of isotopic observations.This work tries to solve the two problems and analyze the spatio-temporal variation of the role of oceanic moisture from the Pacific Ocean in summer precipitation in eastern China using ¹⁸O and ³H.The main work and conclusions are drawn below:（1）We extend the traditional end member mixing analysis to quantify the process of moisture transport with stable isotope.The improved method makes it possible to evaluate the contribution of evaporation,transpiration and advection to regional precipitation in the study area with a single tracer.In this study,this improved method is applied to analyze long-term mean precipitation sources for eastern China.Results show that in the entire study area,OM predominates summer precipitation,with ET playing a supplemental role.The contributions of OM to precipitation decrease from June to August in Central China and Northern China.Spatially the OM contributions decrease along the transport route from south to northeast.In order to assess the uncertainty of the estimates,the Statistical evaluation index and Gaussian first-order approximation are employed.The 95%confidence intervals of the estimated contributions range from±0.07 to±0.38,with the standard errors between 0.03 and 0.17 for all months and sub-regions.The relative uncertainty contributions of each source to the total uncertainty follow the order of precipitating vapor>advection>ET.The large variances ofδ¹⁸O in precipitable water vapor account for the most uncertainty of the proportional estimates.（2）We build time-and-space-explicit isoscape database.First,spatial and temporal correlations between observations and simulations are investigated at a regional scale.Based on the high correlation between simulations and observations,we compare performances of six isotope-equipped GCMs（i GCMs）simulations and two bias correction methods（BCMs）.Results show that model simulations generally underestimate the precipitation oxygen isotope.The two BCMs are able to reduce biases in the simulated values.The linear scaling（LS）method presents a similar performance with the distribution translation（DT）method in projecting meanδ¹⁸O_p.In addition,uncertainties for the precipitation oxygen isoscape（POI）are estimated by statistical evaluation index and a variance decomposition method.The POI for eastern China is generated by averaging the results from the six i GCM simulations corrected by the two BCMs.In this study,the i GCMs seem to contribute more uncertainty to meanδ¹⁸O_p estimates than the two BCMs,followed by the interaction between i GCMs and BCMs,with the least from BCMs.The POI built by a multi-i GCM and multi-BCM approach adequately preserves the spatiotemporal pattern of isotopic measurements.（3）We analyze the influence of ENSO on precipitation sources in the study area.First,we apply the precipitation oxygen isotope database and LMWL（Local Meteoric Water Line）to extend observed precipitation hydrogen isotope in space and time.Based on the hydrogen and oxygen isotopes,three-end-member-mixing model is employed to quantify the long-term averaged precipitation sources and compared with the results of the modified two-component mixing model.Results show that the contribution of advection,evaporation and transpiration to precipitation calculated by the three-end-member mixing model is close to that by the modified two-end-member mixing model,which illustrates that the three-end-member mixing model could simulate the spatio-temporal characteristic of regional water cycles.We then analyze the ENSO cycle process and moisture transport in different phases using the three-end-member model to investigate the spatio-temporal characteristic and its mechanism.Results show that different intensity of ENSO events present similar spatio-temporal patterns of precipitation source contribution,which is above 90%in Central China,and about 80%in North China.In the developing phase of El Ni（?）o,the advection sources is higher than in the decaying phase,and vice versa.Lastly,standard error,95%confidence intervals and classical error theory are applied to evaluate the reliability and uncertainty.Results show that the standard errors of calculated proportions of different precipitation sources are about 0.03,and the 95%confidence intervals of the estimated contributions are small,which ensures the reliability of the results.From the perspective of different precipitation sources,the contribution of uncertainty of precipitation vapor is higher than90%,the contribution of evaporation vapor is about 10%,and the contribution of the uncertainty of advected moisture and transpirated moisture are less than 1%.