Non-linear Regression Analysis as a tool to obtain Ground level water vapor isotopic (δ18O) composition

Abstract

Oxygen and hydrogen isotopic composition is a powerful tool to trace origin, movement and mixing of atmospheric water vapor and study contemporary hydrometeorological processes affecting spatio-temporal distribution of water in different hydro-climatic regions. Isotopic composition is an important input parameter for isotope mass balance and isotope enabled atmospheric circulation models. Although laser-based absorption spectrometers can conveniently measure the isotopic composition of water vapor it is not cheap enough to be deployed at large number of locations which is necessary to understand spatial variation in vapor flux. Therefore, a novel, simple and cost-effective method for collecting ambient water vapor for estimating its isotopic composition is developed. This method involves liquid condensation of ambient water vapor on an ice-cooled (0 °C) metallic surface under the supersaturated environment which involves kinetic fractionation due to diffusion of isotopic water molecules through supersaturated boundary layer at the metallic surface (R. D. Deshpande, Maurya, Kumar, Sarkar, & Gupta, 2013). The true isotopic composition of ambient water vapor is estimated from measured isotopic composition of liquid condensate after correcting for the kinetic fractionation using a non-linear regression model. For this, isotopic composition of liquid condensate is compared with the actual vapor collected by complete cryogenic trapping at -78°C. The oxygen isotopic composition of ambient water vapor estimated from liquid condensation is accurate within ±1.8‰ at Ahmedabad compared to the vapor sampled by complete cryogenic trapping. This simple method can be employed for isotope tagging of water vapor at large number of locations in remote areas with minimal resources, particularly in developing countries.