Identification and interpretation of archaeological phenomena is typically based on visual cues and the physical presence of "something archaeological," such as a diagnostic artifact, landscape modification, or structural element. Yet many archaeological features, i.e. the discrete archaeological deposits related to past human behavior, lack clear indicators of human activity that provides clues to the feature's origin. At the Cape Krusenstern beach ridge complex, located in northwest Alaska, ambiguous features, that could be natural or anthropogenic (vegetation anomalies), or are of unknown cultural function (indeterminate), comprise 60% of the identified features at the complex. These ambiguous features represent a large gap in our understanding and interpretations of the occupation history of Cape Krusenstern and the Arctic. The goal of this thesis was to identify anthropogenic features and interpret the original human behaviors that contributed to their formation, through soil geochemical analysis. I sought to identify 1) which features are natural and which are anthropogenic; and 2) what behaviors created the cultural features (e.g. occupation of houses or caching of marine versus terrestrial food resources). I used photometric phosphates spot tests and inductively coupled plasma mass spectrometry (ICP-MS) to geochemically characterize bulk sediment samples from ambiguous features. I then used a variety of statistics, including principal component and discriminant function analysis to identify patterning in elemental compositional data. I compared results to geochemical expectations for different types of cultural features based on prior research and my own analysis of cultural and non-cultural control samples. Analysis indicated that a single feature is natural, and the other tested features are anthropogenic features. However, the analysis did not aid in definitely identifying specific human behaviors (i.e. house/occupation versus storage activities) that could have created the ambiguous anthropogenic features. Broadly, food storage features showed slightly greater enrichment levels and less overall variation than house/occupation feature samples. Overall this thesis indicates that there are likely more house (7.9 to 10.2% increase) and food storage features (1.5 to 5.2% increase) present at the Cape Krusenstern beach ridge complex than previously thought. Increasing the number of house and food storage features suggests that the occupation history at the complex is potentially more intense than previously established. These results also suggest that geochemical analysis has potential use for feature identification at a broader landscape scale than previously performed in other archaeological applications of soil geochemistry. Last, this thesis shows there is potential in using previously collected bulk samples to gain in-depth information that can guide future work at the complex.