The transformation of natural ecosystems to agricultural ecosystems, agricultural intensification, and the cessation of agriculture alter nutrient cycling dynamics. Unfortunately, in ecosystems experiencing rapid rates of erosion, the fingerprint of past land use change has been removed from the landscape. In dryland ecosystems, however, particularly ecosystems transformed by prehistoric farmers and modified little since, soil erosion rates are often sufficiently low to have preserved chemical evidence of these past changes.

Our research aims to understand the long-term ecological consequences of these transformations, across both space and time. We quantified prehistoric agricultural potential across two contrasting field systems on the Hawaiian archipelago using a combination of archaeological and geochemical data. For each field system, we then defined relative changes in agricultural potential: for Maui, we contrasted soil fertility reductions by cultivation with "background" chemical weathering rates; for Hawaii, we indexed soil fertility at different time points during the agricultural intensification process.

Results/Conclusions

On Maui, evidence of intensive prehistoric agriculture starting ~500 years ago is confined to ash-mantled swales. On Hawaii, ~60-km² of ash-mantled, relatively old lava flows was transformed into an extensive field system. On both islands, field systems are restricted to a narrow isohyetal band: rainfall must have been sufficient for sweet potato and dryland taro cultivation but not so excessive that rock- and ash-derived plant nutrients were leached. Modification of these "sweetspots" by farmers on both islands is likely to have increased rootcrop yields. For example, the construction of checkdams across Maui swales trapped sediment and water and likely enabled more intense cultivation. Following any initial increase in agricultural potential, however, cultivation appears to have led to relatively rapid and long-term reductions in agricultural potential, as indexed by soil phosphorus levels. On Maui, losses of soil phosphorus associated with cultivation were ~4 kg P ha^-1 y^-1, nearly 25-fold faster than the rate of loss associated with chemical weathering prior to agriculture. On Hawaii, losses of soil phosphorus appear to have tripled at some point during the 300-year period associated with intensive agriculture, based on an analysis of soil phosphorus levels below a chronological sequence of field walls.

These reductions in agricultural potential may have been abrupt if soil fertility declines followed the imposition of a ritual control system in ~1620 CE--a period of temple construction in Maui and the implied beginning of a phase of agricultural intensification. Conversely, these reductions may have been gradual if they preceded this date.