The conservation and management of biodiversity requires knowledge of the geographical distribution of species and the environmental changes that threaten species. Distributional modeling has been used to predict species ranges from limited occurrence data and to model the effects of environmental change. This study develops distributional models for a rare, alpine butterfly that is endemic to the Sierra Nevada, California and an important candidate for conservation. We compare two separate modeling techniques (BIOCLIM and MAXENT) to predict distributions for the butterfly based on 1950–2000 averaged and future (double CO₂) bioclimatic variables. Additionally, distributional changes over six 10 year means in the last century are examined.  MAXENT and BIOCLIM perform similarly under current climate scenarios, but differ under a future scenario.  Modeling over different periods of the 20^th century shows reduced alpine habitat within the Sierra Nevada, this is supported by a future climate model.  These distributional shifts may affect the stability and persistence of alpine species.