Predication of stable isotopes (¹⁸O and ²H) in precipitation of Bangkok metropolitan using artificial neural network

Stable isotopes (¹⁸O and ²H) in precipitation of Bangkok has been sampled since 1968 when GNIP stablished its first station in Thailand. In this study, the role of various local (wind speed, potential evaporation, vapor pressure, air temperature, and precipitation amount) as well as regional parameters (teleconnection indices such as IOD, BEST, NAO, SOI, and QBO) on stable isotopes content in Bangkok precipitation has been investigated. Firstly, simple artificial neural network (ANN) as well as Deep Learning Neural Network (DNN) models have been used to predict stable isotopes content in precipitation. Comparing the simulated and real stable isotopes data shows that both DNN and ANN models can simulate the stable isotopes in precipitation with acceptable accuracy. Secondly, studying the fractional importance of various parameters on stable isotopes content of precipitation demonstrates that among the local parameters (precipitation amount and potential evaporation) and among the regional parameters (BEST index) have the dominant role in controlling the stable isotopes content of precipitation.