4. DISCUSSION
4.1 Influential factors of seasonal variation of precipitation isotopic compositions
In the summer monsoon period (May-September), the mostly ocean-originated moisture with high humidity and weak evaporation results in low δ value of the precipitation (Peng et al., 2010; Uemura et al., 2012; Cai et al., 2017). In the winter monsoon period (October-February), there is relatively less precipitation due to the influence of continental dry air masses carried by West Wind Circulation, hence, the precipitation δ value is significantly higher than that in summer (Liu et al., 2009; Peng et al., 2010). However, due to the blockage of the Mountains in western Fujian, the continental air masses tend to enter the East China Sea eastward and then transport southward to Xiamen (Chen et al., 2016), so the wetter moisture (humid air with higher relative humidity) brought from the sea makes the δD and δ18O of winter precipitation in Xiamen lower than that in the same latitude (Guo et al., 2021), showing significant marine air masses characteristics. The period between winter and summer monsoon periods is called seasonal shift period. During this period, the warm and humid air masses from the ocean meet with cold and dry air masses from the land, and then convective activity occurs in the study area, resulting in frontal rainfall for long duration. The isotopic composition of precipitation occurred during this period falls in between summer and winter monsoon periods, with a significant signal of mixed air masses from ocean and land. In addition, the study area is significantly affected by typhoon processes produced in the Western Pacific Ocean from June to November each year, and heavy precipitation often occurs under the influence of typhoons, which show significant lower δD and δ18O than the normal precipitation (Fudeyasu et al., 2008; Xu et al., 2019).
Five environmental effects could affect the variation of atmospheric precipitation isotopes, namely, latitude, temperature, altitude, precipitation amount, and continental (Dansgaard 1964), among which the most significant effects are the temperature and the precipitation amount in the coastal region (Wolf et al., 2020). The lower temperature will result in a lower δ value of precipitation due to the larger isotopic fractionation coefficient at low temperatures, which is generally positive in high latitude region (Merlivat and Jouzel, 1979). The higher precipitation amount will also result in a lower δ value due to the evaporation and isotopic exchange of rain drops with ambient water vapor during the descent (Gedzelman and Lawrence, 1982). In this study, the monthly mean δ18O values of atmospheric precipitation in the Xiamen for 15 months were correlated with the monthly mean temperature and monthly total precipitation amount to analyze the temperature and precipitation amount effects during the study period (Fig. 6). The results show that the δ18O value is significantly negatively correlated with the temperature, showing the characteristic of an inverse temperature effect. This may be due to the influence of monsoon climate in the study area, in that the influence of moisture sources on δ18O values overrides the influence of precipitation fractionation process on δ18O values which leads to an inverse temperature effect (Liu et al., 2009; Xie et al., 2011). In addition, δ18O values were negatively correlated with the amount of precipitation, but the correlation was not significant, and the points with larger deviations were March and April 2019 (seasonal shift period) (Fig. 6), which appears to be due to the influence of longer duration of frontal rainfall compared to previous years in March to May 2019. This ultimately led to a less pronounced precipitation amount effect because the δ18O of precipitation are mostly larger than -4‰ during seasonal shift period (Fig. 5 e).
[Insert Figure 6]
The re-evaporation that occurs during precipitation is an important physical process that affects precipitation isotopes (Salamalikis et al., 2016; Zongxing et al., 2016). Evaporation during the precipitation of raindrops in conditions with small amounts of precipitation and low humidity will cause the meteoric water line slope to decrease (Vodila et al., 2011). Therefore, higher temperature, stronger evaporation and lower atmospheric humidity will result in a smaller slope of the meteoric water line and a consequent change in the intercept (shown in Fig. 5 d). The slope and intercept of the meteoric water line in different seasons indicates differences in the influence of evaporation processes on isotopic composition of atmospheric precipitation (Peng et al., 2007; Vodila et al., 2011). The higher slopes of the meteoric water lines in Xiamen in summer monsoon period indicate a weaker precipitation re-evaporation process, whereas the lower slope in winter monsoon period indicate a stronger precipitation re-evaporation process (Fig. 5 c and d). Although summer temperatures are higher than winter and high temperatures promote re-evaporation, high humidity and abundant precipitation in summer will inhibit re-evaporation during the precipitation process (Guo et al., 2021). On the contrary, the dry climate in winter, with scarce and slow precipitation, facilitates the re-evaporation of rainwater during the precipitation process (Peng et al., 2007). This may be an important reason why the slope of meteoric water line is significantly lower in winter than in summer. The slope of seasonal shift period is between summer and winter monsoon period, indicating a transitional climate condition (Fig. 5 e). In addition, the D-excess of precipitation reflects the isotopic composition of the moisture and contains important information about the source regions, including the state of the evaporation process and the evaporation rate (Kong and Pang, 2016). The evaporation effect of moisture originating from warm and humid marine areas at low latitudes is smaller, and the evaporation process is slower, resulting in a smaller D-excess; whereas the evaporation effect of moisture originated from local or dry and cold inland areas at high latitudes is larger, resulting in a higher D-excess (Frohlich et al., 2002; Vreča et al., 2006; Xie et al., 2011).