The Newberry twin crater lakes, East Lake and Paulina Lake, are separated by a narrow ridge but are very different in composition. East Lake is a terminal lake with a geothermal gas input. The larger and deeper Paulina Lake is fed by winter snow melt and bottom hot springs, but has an overflow, Paulina Creek. We made a stable isotope study of these two lakes with their hot springs and local meteoric waters in the area for the years 2011, 2012, 2014 and 2015. Crater lake waters plot on a common evaporation line that starts at the local meteoric water value. Most of the hot spring waters have close to meteoric values or are mixtures with lake waters, plotting further up the evaporation line. East Lake is isotopically much more evolved than Paulina Lake confirm its ‘closed lake’ status. We developed a water budget model, deriving the mean atmospheric humidity from the slope of the evaporation line. We calculated the isotopic values of atmospheric moisture from local air temperatures using an empirical model. The isotopic composition of the evaporate was calculated from standard equations applying both equilibrium and kinetic isotope fractionations. Evaporation and precipitation rates for these lakes were from literature values, and watershed areas with infiltration rates were approximated. The models were run for each lake in daily time-steps starting with local meteoric water. The runs start in May with mean precipitation and evaporation and then run for 22 weeks. In November the lakes freeze over and evaporation and meteoric water input stop until the lakes thaw again in May. We invoked a whole lake mixing step with cumulated winter precipitation input (snow melt), and then ran again in daily steps for the next 22 weeks. We ran the models for about 40 years until isotopic steady state was reached. During the dry summer of 2011, East Lake reached its isotopic steady state values in August-September, to fall back to lower values during the wetter summer of 2012. The hydrothermal input at the bottom of Paulina Lake can be solved for assuming water steady state, and these estimates are tested with chemical mass balance models of dissolved components. The water balance of East Lake appeared close to water steady state, suggesting that no major seepage terms are needed.