Thin Plate Spline (TPS) is the leading interpolation technique used to generate climate surfaces. The extent of our analysis is from 49° to 32° latitude and 124.7° to 112.9° longitude, which includes multiple states in the Western United States. A comprehensive analysis was done to determine which covariates and polynomial function degree are best suited for each climatic variable (precipitation, minimum temperature, maximum temperature, and maximum temperature). Using TPS we ran a ten-fold cross-validation using elevation (DEM), slope, aspect, solar potential, radio detection and ranging (radar), and two different Normalized Difference Vegetation Index (NDVI) products derived from AVHRR and MODIS as covariates. We also tested a range of possible degrees for polynomial function to determine the best suitable one. We generated two monthly datasets: (1) an average of 1950-2000 and (2) a yearly average ranging from 1950 to 2000 at a spatial resolution of 1km2, which we name ClimSurf. Leading covariate candidates were: precipitation polynomial function degree of 2 with radar being the covariate, maximum temperature polynomial function degree of 2.4, mean and minimum temperature polynomial function degree of 1.8, for all three temperatures DEM was the covariate. A comparison to other products such as PRISM and WorldClim showed strong agreement across large geographic areas but ClimSurf varied at high elevation regions such as in the Sierra Nevada Mountains; a region that has great hydrological, cultural, and ecological importance. Our findings suggest when looking at the uncertainty base for an ecoregion or DEM classes, radar and DEM are the best covariates to generate climate surfaces. ClimSurf is the first climate surface to be produced at this spatial (1km2) and temporal (monthly) resolution and is available for download. Ecological niche models have been an important tool in trying to understand species distributions, yet there has been a lack of consideration for uncertainty, especially for environmental surfaces. The objective of this study was to determine how uncertainty in environmental surfaces influenced predictions from ecological niche models for four alpine mammal species in the Sierra Nevada mountain range. We produced Maximum Entropy models from hundreds of unique current climate surfaces where uncertainty was generated with Monte Carlo methods. We also generated a future climate dataset using a bias-corrected spatially downscaled algorithm to 1 km2. We found that both datasets had similar known patterns but with higher variation in high elevation localities. Compared to the baseline case with no uncertainty, our ecological niche models tended to overestimate the distribution of the species when uncertainty was factored in. By running a sensitivity analysis on American Pika, we determined that January precipitation had the greatest impact on predicted distributions; while October’s maximum temperature and June’s precipitation had the greatest variability. Our future downscale and the uncertainty surfaces will be available to download for our reader.