Variation in white-tailed deer (Odocoileus virginianus) mortality during winter affects population growth in cold climates. Across the northern extent of their range, mortality increases with colder temperatures and snow. Few studies have examined the relationships between winter conditions and deer mortality, and no studies have concurrently studied this relationship for different ages of deer across multiple years and landscapes. We used recently developed cause-specific mortality models to evaluate temporal and age-class variation in deer mortality in farmland areas and compared to published results from forest areas in Wisconsin, USA, from 2011-2014. We then used temporally varying snow and temperature covariates to predict mortality trends using telemetry information from 860 deer. Cause-specific mortality in the farmland varied by age and year, similar to results from previous research in the forest. Humanrelated mortality was the leading cause of mortality in the farmland during most years and ranged from 4.3% to 10.3% for juveniles and 3.6% to 9.1% for adults from 2011-2014. Very little predation occurred in the farmland, and this differed from previous research in the forest where predation was the leading cause of mortality. During more severe winters (2013 and 2014), other mortality, usually associated with starvation, was the leading cause of mortality for juveniles in the farmland but not adults. In the forest, we found support for saturating effects of accumulated snow depth days >30.5 cm and accumulated temperature days >0°C on mortality. We also found support for the relationship of mortality with accumulated temperature days >0°C in the farmland but no relationship with snow depth. Deer tolerate sustained cold temperatures, but the timing of winter to spring transition is more important for deer survival in both forested and agricultural areas. In the absence of empirical survival information, managers can use our model to predict annual winter effects on deer survival, which can provide improved inference compared to traditional winter severity indices. Our results suggest changes in predator abundance may have minor influence on overwinter survival compared to winter weather. Based on mortality estimates from previous research, the highest predation rates on juvenile deer in the forest occurred when wolf (Canis lupus) counts were lowest and when wolf abundance was highest, juvenile deer predation rates were lowest.