Combining Data from Field Surveys and Archaeological Records to Predict the Distribution of Culturally Important Trees

Aim: Indigenous communities involved in conservation planning require spatial datasets depicting the distribution of culturally important species. However, accessing datasets on the location of these species can be challenging, particularly when the current distribution no longer reflects areas with the full range of suitable growing conditions because of past logging. We test whether using occurrence data from community‐based field surveys and archaeological records in species distribution models can help predict the distribution of monumental western redcedar trees (Thuja plicata)—large, high‐quality trees suitable for cultural purposes such as carving dug‐out canoes, totem poles and traditional houses. This species is critically important to indigenous people of the Pacific Northwest of North America, but trees suitable for traditional carving and building are diminishing in abundance due to logging.Location: Our analysis covers the spatial extent of the traditional territory of the Heiltsuk First Nation, which encompasses a portion of the Great Bear Rainforest in British Columbia, Canada.Methods: We built and compared species distribution models using the machine learning program, Maxent, based on occurrence data from field surveys and archaeological records of culturally modified trees.Results: Our findings highlight similarities and differences between the predictions from these species distribution models. When validating these models against occurrences from an independent dataset, the archaeological record model performs better than the field survey model. These findings may arise because the independent dataset was collected on an unlogged island—an environment that aligns more closely with the historic forest conditions revealed by the archaeological records than the current distribution revealed by the field surveys.Main conclusions: We demonstrate and discuss the utility of using archaeological data in species distribution modelling and conservation planning when the target species is associated with shifting environmental baselines, data limitations and an important cultural resource.