Essential oils and CO2 extracts of Timur (Zanthoxylum armatum DC syn. Zanthoxylum alatum Hemsl.) from fruits cropped in the Nepalese districts Surkhet, Jajarkot and Salyan from 2012 to 2015 were analysed to explain their typical odour and characterize their volatile profile. Best extraction conditions to ensure the right balance of the terpenic fresh notes compared with the sulphury fruity character were obtained with milled fruits and CO2 extraction. Volatile composition allowed supplying zones discrimination. Among more than 130 volatile constituents, identified by GC–MS and quantified by GC‐FID with the use of predicted relative response factors, only 20 were found to have an odour impact by GC‐O using a hybrid methodology between detection frequency and time intensity methods. Limonene, α‐pinene and myrcene explain the terpenic notes. Linalool, citronellal, 6‐methyl heptanal and octanal bring the fatty and aldehydic tone. The spicy facet is linked to (E)‐methyl cinnamate and (Z)‐ethyl cinnamate. 4‐mercapto‐4‐methylpentan‐2‐one, 3‐mercaptohexanol, 3‐mercaptohexyl acetate and 1‐p‐menthene‐8‐thiol are the key olfactive constituents contributing to the typical tropical and grapefruit odours. Their presence was confirmed using the stable isotope dilution assay and reported for the first time in Timur.

Representation of olfactory data is a key parameter in GC‐O to simplify interpretation. Actual presentation of hybrid results, taking into account both detection frequency and odor intensity, can be improved to be more informative. We propose in this study a new way of representing olfactograms taking into account the three dimensions: detection frequency, odor intensity and odor description using bubble charts. Odor description is represented by colors that were associated to the 25 olfactory families defined in this work using binomial probability statistics without any smelling step.