Geospatial modeling of pre-contact Hawaiian production systems on Molokaʻi Island, Hawaiian Islands

Abstract

The geographic distribution and relative importance of traditional agricultural systems in Hawaiʻi, based on ethnohistoric and archeological data, is only partially understood. Knowledge of the size and distribution of these systems is critical in estimating island populations, production, and surplus, as well as for assessing societal dynamics and the sustainability of indigenous agricultural systems. We employ geospatial modeling using rainfall, elevation, soil type, streamflow, slope, and substrate age, to create a geographic information system (GIS) model for the distribution of four major production systems on Molokaʻi Island – intensive pondfield irrigation, intensive dryland cropping, extensive colluvial slope agriculture, and fishpond aquaculture. Model results were checked against archaeological data sets for known areas, testing for goodness of fit between model predictions and empirical field evidence. Our model predicts that Molokaʻi could have sustained 9.52 km 2 of irrigated pondfields, 7.98 km 2 of intensive dryland systems, and 18.5 km 2 of colluvial slope agriculture. Fishponds are estimated to have yielded 17.98 metric tons of fish protein per year. The total agricultural production for the island is calculated to be 41,490 metric tons per year (wet weight), with pondfield, dryland, and slope systems yielding 46, 17, and 37 percent of total production respectively. Our results indicate that colluvial slope cultivation, which has been largely ignored in previous studies, was a major contributor to Hawaiian agricultural production, especially on the geologically older islands. Pondfield irrigation and colluvial slope systems had higher caloric efficiency than intensive dryland field systems, providing greater surplus to labor ratios in areas with abundant land suitable to irrigation and slope cultivation. Surplus discrepancies likely influenced population distribution as well as sociopolitical dynamics between districts and among neighboring islands. The model has applicability to the entire Hawaiian archipelago and to other islands in Polynesia, where it may be used to estimate pre-contact agricultural production and carrying capacities, two main factors influencing pre-contact social and political dynamics in Hawaiʻi and other island groups.