Dehydration can have negative effects on animal physiological performance, growth, reproduction, and survival, and most animals seek to minimize these effects by reducing water losses or seeking water sources. Much-but not all-of the research on animal water balance comes from dryland ecosystems. However, animals inhabiting mesic regions may also experience desiccating conditions, for example within urban heat islands or during heatwaves and droughts. Here we examined how spatial variation in impervious surface and spatial and temporal variation in microclimate impact water demand behavior of terrestrial arthropods and mollusks in three areas of mesic Northwest Ohio, with analysis of taxa that exhibited the greatest water demand behavior. Water demand behavior was measured as the frequency that individuals were observed at an artificial water source (a moistened pouch), relative to the frequency at a control (a dry pouch). Overall, terrestrial arthropods and mollusks were found about twice as often at the water source than at the control (equivalent to 86 more observations on the wet pouch than on dry at each site, on average), with ants accounting for over 50% of the overall response in urban areas. Daily fluctuations in vapor pressure deficit (VPD) best predicted daily variation in water demand behavior, with increased demand at higher VPD. Mean VPD was generally highest near urbanized areas, but effects of VPD on water demand behavior were generally lower in urbanized areas (possibly related to reductions in overall abundance reducing the potential response). On certain days, VPD was high in natural areas and greenspaces, and this coincided with the highest arthropod water demand behavior observed. Our results suggest that terrestrial arthropod communities do experience periods of water demand within mesic regions, including in greenspaces outside cities, where they appear to respond strongly to short periods of dry conditions-an observation with potential relevance for understanding the effects of climate change.