CHARACTERIZING MULTITEMPORAL ALPINE SNOWMELT PATTERNS FOR ECOLOGICAL INFERENCES

Snowmelt patterns and the persistence of snowpatches into the growing season can profoundly affect the distribution of alpine vegetation. Th is study applies Markovian transition probability matrices to the prob lem of characterizing classified multitemporal satellite snow-cover da ta. Transition probability matrices are used to form hypotheses regard ing the effects of snow persistence and ablation patterns on the alpin e treeline using an integrated geographic information system. Four clo ud-free Landsat MSS (Multispectral Scanner) scenes of a portion of Gla cier National Park, Montana were processed to characterize periods of the 1987 snowmelt season. Stratification of the rugged landscape by el evation and slope aspect, achieved through the processing of 1:24,000 base-scale Digital Elevation Models (DEMs) and integrated with the sat ellite characterizations of snow conditions, demonstrated the dynamics of snow-cover conditions as a consequence of topographic position and antecedent snow conditions. Analysis of transition matrices by topogr aphic position and watersheds highlighted areas of significantly late snowmelt, which holds implications for ecological investigations of al pine treeline by considering snow both as a stressor and protector of vegetation, depending upon its spatial pattern and temporal persistenc e.