Sulfates present in urban aerosols collected worldwide usually exhibit significant non-zero Δ<sup>33</sup>S signatures (from −0.6 to 0.5 ‰) whose origin still remains unclear. To better address this issue, we recorded the seasonal variations of the multiple sulfur isotope compositions of PM<sub>10</sub> aerosols collected over the year 2013 at five stations within the Montreal Island (Canada), each characterized by distinct types and levels of pollution. The δ<sup>34</sup>S-values (n = 155) vary from 2.0 to 11.3 ‰ (± 0.2 ‰, 2σ), the Δ<sup>33</sup>S-values from −0.080 to 0.341 ‰ (± 0.01 ‰, 2σ) and the Δ<sup>36</sup>S-values from −1.082 to 1.751 ‰ (± 0.2 ‰, 2σ). Our study evidences a seasonality for both the δ<sup>34</sup>S and Δ<sup>33</sup>S, which can be observed either when considering all monitoring stations or, to a lesser degree, when considering them individually. Among them, the monitoring station located at the most western end of the island, upstream of local emissions, yields the lowest mean δ<sup>34</sup>S coupled to the highest mean Δ<sup>33</sup>S-values. The Δ<sup>33</sup>S-values are higher during both summer and winter, and are < 0.1 ‰ during both spring and autumn. As these higher Δ<sup>33</sup>S-values are measured in "upstream" aerosols, we conclude that the mechanism responsible for these highly positive S-MIF also occurs outside and not within the city, at odds with common assumptions. While the origin of such variability in the Δ<sup>33</sup>S-values of urban aerosols (i.e. −0.6 to 0.5 ‰) is still subject to debate, we suggest that oxidation by Criegee radicals and/or photooxidation of atmospheric SO<sub>2</sub> in presence of mineral dust may play a role in generating such large ranges of S-MIF.