Widely distributed species are experiencing a continual pattern of range shifts due to anthropogenic expansion and climate change, forcing these species into novel environments and out of critical habitat. The ability to estimate current and forecasted states of demographic parameters of species distributed along a gradient of environments is becoming increasingly important in a time of large-scale environmental change. Consulting models that provide temporally relevant estimates of population dynamics based on the latest realizations of environmental conditions can allow for informed, quick and decisive conservation and management actions. Modelling the drivers of demography across a wide range of environmental conditions will provide a more comprehensive understanding of how species will respond to novel environments. In this study we provide an example of relating seasonal-environmental variables to survival in a widely distributed ungulate species. We used a mule deer (Odocoileus hemionus) survival dataset collected in Utah with seven sites distributed across the multiple ecoregions of the state, allowing for the elucidation of relationships across a variety of environmental conditions. Multivariate analyses predicting survival of young and adult females were performed using geographic location, elevation, and seasonal satellite-derived primary productivity data and weather variables. We developed frameworks for estimating past and current states of survival and predicting short-term (sub-year) forecasts of survival. Furthermore, we investigated adaptive modelling techniques for increasing the certainty of the forecasted predictions of survival. We found that increased winter precipitation had a negative effect on survival across the state. Survival was lower in the northern region of the state and in higher elevations. Furthermore, measures of summer primary productivity had a positive relationship with survival. Lastly, our adaptive modelling demonstration shows that uncertainty of forecasted survival predictions can be reduced with the addition of data. This study provides a framework for developing models that will provide invaluable information to managers in a time of large-scale environmental change.