Recent studies have shown that cirque-floor elevations are of limited use in the determination of paleosnowline altitudes because of ambiguities in their relationship with glacier equilibrium lines, subjective methodologies, and so forth. In particular, cirque-floor elevations will be somewhat higher than snowlines of valley glaciers sourced from the cirques. These studies also suggest that paleosnowlines are better estimated on the basis of reconstructed glaciers. However, cirque-floor elevations, while higher, may define gradients in snowline. In order to examine this possibility, an analysis of cirque-floor elevations was performed in the southern Sawatch Range.

A first-order trend surface of the lowest cirque-floor elevations rises steeply to the southeast, but the trend appears not to be statistically significant. The dominant component of the trend is its east-west component, having a gradient of ~11.5 m km^-1 to the east. This is in marked contrast to the regional trend of late Pleistocene equilibrium-line altitudes (ELAs) determined for reconstructed paleoglaciers that shows a rise of ~3.7 m km^-1 to the east. The non-parallelism of the trends suggests that either the gradient in cirque-floor elevations during “cirque glaciation” was different than during “full glacial” conditions, or the gradient thus defined is meaningless. Approximations of modern snowline gradient suggest that the latter is correct.

A simple degree-day model was used to determine the temperature and precipitation changes required to simulate late Pleistocene ELAs in a small portion of the study area. ELAs here are ~3415 m and require mean summer temperatures to be ~7 °C cooler than present. Required temperature changes are relatively insensitive to slight changes in precipitation.