Background/Question/Methods Arctic lakes offer a unique landscape for studying the population dynamics of fish within the context of climate variability given the variability in timing and the short duration of the growing season. Our goals were to quantify empirical vital rates (survival, growth, trend, and recruitment) based on a long time series of mark- recapture data, and to use that information to 1) characterize three “types” of systems as a function of lake morphology and fish life history (e.g., large, deep, char), 2) investigate the drivers of inter-annual variability in survival (including climate change and density dependence), 3) build a stage-based population matrix model to better understand current, and predict future, effects of climate on population trends. Results/Conclusions Two of the systems behaved very differently from the other, apparently due to whether or not the system is open versus closed, and the degree of intra-specific piscivory versus cannibalism. Annual survival rates are relatively high for larger (>240 mm), adult fish (mean 0.74), less so but still high for smaller adult fish (mean 0.54), and all top models all include a climate signal (e.g., DATE ICE OUT). Declining population trends in conjunction with high adult survival rates indicate recruitment and early life-stage survival is low, variable, and limits these populations. We also observed a negative relationship between the survival of large fish and a lagged (1-year) recruitment signal, suggesting density dependence drives population dynamics in some years. And finally, climate appears to drive growth rates and survival rates through different pathways and at different times of year, and with a complex interaction between temperature and food availability.