Exploring physical processes controlling the stable isotopic composition of water, including details such as water vapor source, atmospheric circulation, and cloud microphysical processes, is useful for understanding the water cycle. The water stable isotopes (¹H₂O, ¹H²D¹⁶O, and ¹H₂¹⁸O) differ in molecular symmetry and weight. These differences in physical properties lead to a change in the stable isotope composition of water, due to fractionation during phase changes. When water vapor condenses and forms liquid or solid particles, it becomes depleted in ²D and ¹⁸O, because heavy isotopes condense preferentially to light ones.

In this study, we modified the NCAR WRF model to understand the role of different factors in the fractionation of the stable isotopes of water. The experimental stable isotope thermal equilibrium data were converted into isotope saturation vapor pressure, which was then used in the two-stream Maxwellian kinetic equation for calculating the condensation and evaporation or deposition and sublimation of HDO, parallel to that for H₂O. Mass conservation was also considered explicitly for the collection processes as well as during freezing and melting. A frontal system event was selected to reveal the complexity of isotope fractionation. The simulated results showed fairly good agreement with water vapor and rainwater stable isotope measurements and suggested that the decreases in water vapor δD before the front arrived in Taiwan were due to an air mass of continental origin. When the front passed during the early morning of 12 June, both the water vapor sources and the cloud microphysical processes contributed to 50 ‰ decrease in water vapor δD, which returned to background levels after the front had passed. Additional sensitivity experiments showed that the thermal equilibrium assumption commonly used in earlier studies might significantly overestimate the decrease of mean δD by about 11 ‰, while the maximum difference can be more than 20 ‰, during the precipitation event. Cloud microphysical processes, including ice-phase processes, have substantial effects on isotopic fractionation, especially on the vertical redistribution of isotopes. Furthermore, the sensitivity tests suggest that the initial vertical profile and the land–sea contrast in surface sources are quite important in simulating atmospheric stable isotopic composition and should be estimated from observations such as satellite data, without which the underestimation in the decrease of water vapor δD could reach about 34 ‰ and 28 ‰, respectively.

In summary, this study suggests that a better understanding of the relationship between water stable isotope variation and hydrological cycle can be achieved with a combination of multiplatform observations and detailed cloud model simulations.