Ancient inhabitants of the Kodiak Archipelago, Alaska clustered in coastal areas as they travelled by kayak, utilizing resources spanning the subtidal and intertidal zones such as sea mammals, fish, and mollusks. Traces of these ancient coastal lifestyles are often ephemeral and lacking visible surficial signatures. As such, archaeological excavations, which are destructive, are required to find these deposits. Archaeological predictive models can be used to create targeted survey strategies and increase the likelihood of locating sites with less disturbance. These approaches rely on a Geographic Information System (GIS) to analyze the landscape and identify suitable areas for investigation, with the most critical feature being paleo-shorelines. Is GIS capable of adequately characterizing those relevant landforms in a highly dynamic project area where geospatial data is relatively coarse and sparse, as is the case for the Kodiak Archipelago? This study uses stratigraphic profiles from Womens Bay, in northern Kodiak Island, and Aiaktalik Island, in the southern portion of the archipelago, to assess the validity of GIS-modelled landforms. Findings suggest that 1/3 arc-second Digital Elevation Models (DEMs) are too coarse to show all relevant topographic features for archaeologists looking for ancestral Alutiiq sites. Despite this shortcoming, the visualization of slope and lumping of landforms into broader depositional/erosional areas can provide helpful boundaries for survey areas that may otherwise be overlooked during field reconnaissance. Although the available DEMs cannot provide the exact location of relict shoreline indicators, such as beach berm complexes, they can be used to identify depositional zones within which beach berm complexes are deposited. This paper provides methodological considerations for landform-based prospection and underpins the importance of unique interpretations based on local geomorphologic histories.