The new infrared laser spectroscopic techniques enable us to measure the isotopic composition (δ18O and δ2H) of atmospheric water vapor. With the objective of monitoring the isotopic composition of tropical water vapor (West Africa, South America), and to discuss deuterium excess variability (d = δ2H – 8δ18O) with an accuracy similar to measurements arising from isotope-ratio mass spectrometry (IRMS), we have conducted a number of tests and calibrations using a wavelength-scanned cavity ring-down spectroscopy (WS-CRDS) technique. We focus in this paper on four main aspects regarding (1) the tubing material, (2) the humidity calibration of the instrument, (3) the water vapor concentration effects on δ, and (4) the isotopic calibration of the instrument. First, we show that Synflex tubing strongly affects δ2H measurements and thus leads to unusable d values. Second, we show that the mixing ratio as measured by WS-CRDS has to be calibrated versus atmospheric mixing ratio measurements and we also suggest possible non-linear effects over the whole mixing ratio range (~2 to 20 g/kg). Third, we show that significant non-linear effects are induced by water vapor concentration variations on δ measurements, especially for mixing ratios lower than ~5 g/kg. This effect induces a 5 to 10‰ error in deuterium excess and is instrument-dependent. Finally, we show that an isotopic calibration (comparison between measured and true values of isotopic water standards) is needed to avoid errors on deuterium excess that can attain ~10‰. Copyright © 2011 John Wiley & Sons, Ltd.