Since its detection in the aurorae of Jupiter approximately 30 years ago, the H 3 + ion has served as an invaluable probe of giant planet upper atmospheres. However, the vast majority of monitoring of planetary H 3 + radiation has followed from observations that rely on deriving parameters from column-integrated paths through the emitting layer. Here, we investigate the effects of density and temperature gradients along such paths on the measured H 3 + spectrum and its resulting interpretation. In a non-isothermal atmosphere, H 3 + column densities retrieved from such observations are found to represent a lower limit, reduced by 20% or more from the true atmospheric value. Global simulations of Uranus’ ionosphere reveal that measured H 3 + temperature variations are often attributable to well-understood solar zenith angle effects rather than indications of real atmospheric variability. Finally, based on these insights, a preliminary method of deriving vertical temperature structure is demonstrated at Jupiter using model reproductions of electron density and H 3 + measurements. The sheer diversity and uncertainty of conditions in planetary atmospheres prohibits this work from providing blanket quantitative correction factors; nonetheless, we illustrate a few simple ways in which the already formidable utility of H 3 + observations in understanding planetary atmospheres can be enhanced. This article is part of a discussion meeting issue ‘Advances in hydrogen molecular ions: H 3 + , H 5 + and beyond’.