•Evaluation of pastoral livelihood security facing climate change in drylands.•Increasing precipitation variability had negative or positive effects on herd size.•Identification of critical regimes beyond which livelihoods are endangered.•Mobility compensates for climate change but not for increasing income needs.•Appropriate mitigation strategies need to include sufficient resting of pastures. Livestock is the most important source of income for pastoral livelihoods in drylands. Pastoralists have developed flexible resource utilization strategies that enable them to cope with the high spatio-temporal resource variability typical to these areas. However, climate change in the form of decreasing mean annual precipitation accompanied by increasing variability has important consequences for rangeland productivity and thus pastoral livelihood security. Here, we use a spatial simulation model to assess impacts of changing precipitation regimes, and to identify limits of tolerance for these changes beyond which pastoral livelihoods cannot be secured. We also examine strategies to control these limits. Our results indicate that: (i) while reduced mean annual precipitation always had negative effects, increased precipitation variability can have negative, none or even positive effects, depending on the vegetation's recovery potential. (ii) Depending on income requirements there are limits of tolerance to decreases in mean annual precipitation beyond which precipitation regimes overcharge the coping capacity of the pastoral household and threaten its livelihood. (iii) There are certain strategies, in particular “Increasing mobility” and “Diversifying income for coping with income risks from pastoralism”, that allow the limits of tolerance to be shifted to a certain extent. We conclude that it is important to consider climate change and human requirements together to create appropriate climate change mitigation strategies in pastoral systems. Our results also shed new light on the discussion on disequilibrium rangeland systems by identifying mechanisms that can support fluctuating but non-degrading herbivore-vegetation dynamics. The paper finishes with remarks on the broader potential of the presented modelling approach beyond rangelands.