Study The Temporal And Spatial Distribution Of Stable Hydrogen And Oxygen Isotopes In Precipitation In Arid Northwest China Based On GCM

The stable hydrogen and oxygen isotopes play an important component in water molecules, and also a vital link in the water cycle process. To compare the observed data（GNIP） with the simulated data（GISS-E（MERRA）, GISS-E（NCEP）, isoGSM（NCEP）, LMDZ（free） LMDZ（ECMWF）, MIROC（free）） to analyze the stable hydrogen and oxygen isotopes (δ¹⁸O and δD) monthly variation, yearly variation, summer and winter variation, interannual variation, spatial distribution, slopes and significance variation, Local meteoric water lin variation, correlation between meteorological factors during 1986-2003 in arid Northwest China. The results as follows:（1） the δ¹⁸O and δD are simulated by the isotope-equip GCM, the monthly variation, yearly variation, summer and winter variation present the markly seasonal variation feature, the value of δ¹⁸O and δD are higher in summer than winter, and the max in June or August, the min in January or December. The δ¹⁸O and δD yearly variation except the variation trend is simulated by LMDZ （free） remarkably, the rest is non-significant.（2） The δ¹⁸O and δD spatial distribution, slopes and significance variation are simulated the remarkably seasonal variation by six isotope-equip GCM. The value of δ¹⁸O and δD are higher in summer than winter, yearly variation non-significant. The δ¹⁸O and δD show enrichment in the west of Xinjiang in summer, the opposite in winter. The reasons lie in the water vapour derived from the westerly wind or Western Siberia. The GISS-E（MERRA） and GISS-E（NCEP） are in good simulation effect.（3） Based on GNIP, the Urumqi, Zhangye and Hotan interannual variation are remarkably, and the tendency is rising in Zhangye, through the 0.01 inspection. Although the data (δ¹⁸O and δD) is discontinuity, the interannual variation is remarkable. Except isoGSM （NCEP） simulate remarkably, the rest is non-significant.（4） Based on GNIP and GCM, the local meteoric water line in Urumqi, Zhangye and Hotan present good results, the best correlation between δ¹⁸O and δD in Hotan, the worst in Urumqi. Except the 5.8856 is simulated by LMDZ （ECMWF）, the rest are larger than 6.7636. The r2 are through the 0.0001 inspection, larger than 0.9. The simulation results by GISS-E （NCEP） are similar with the GNIP.（5） The δ¹⁸O and δD and temperature correlation are good derive from the isotoped-equip GCM, the GISS-E （NCEP） present the better than the rest. Although the most isotoped-equip GCM present the better results, the temperature effect is not simulated by isotoped-equip GCM, especially LMDZ （free） present the opposite trend.