In school environments in the Sonoran Desert, dust contains anomalously high manganese (Mn) of geogenic origin, similar to those reported at industrial sites. Non-carcinogenic risk assessment for children exposed to Mn was estimated for two sampling seasons (pre-monsoon, post-monsoon). We determined the total Mn content in school dust samples by portable X-ray fluorescence (PXRF), and we used geochemical indices for Mn-source apportionment. We identified dust mineralogy by X-ray diffraction (XRD). We obtained the oral and lung bioaccessibility by in vitro tests. The results showed a much higher total Mn content (1725.7 mg.kg(-1)) in the <20 mu m school dust size fraction in the pre-monsoon season compared to the average value post-monsoon (582.9 mg.kg(-1)). Moreover, in pre-monsoon season, the estimated hazard quotient (HQ) and hazard index (HI) for Mn were higher than 1, representing a potential health risk. Mn's oral and lung bioaccessibility also differed significantly between the sampling seasons, but the behavior was opposite to total Mn-content. The maximum gastric/intestinal bioaccessibility for Mn ranged from 10.6/1.7 pre-monsoon to 28.5/9.0% post-monsoon. In the simulated lung fluid, the maximum bioaccessibility of Mn ranged from 15.6% pre-monsoon to 64.1% postmonsoon. This variability was probably due to differences in Mn mineralogy (todorokite in pre-monsoon dust; an unknown phase in post-monsoon dust). Our results show that it is necessary to integrate seasonal sampling into risk assessment in arid zones worldwide, where dust resuspension processes commonly occur.