Pedology of archaeological stone-wall bench terraces

Abstract

Agricultural terraces are the most widespread archaeological landforms worldwide. Despite their far-reaching prevalence in many inhabited parts of the world, archaeological soils in terraces (hereafter ‘terrace soils’) are scarcely studied and their nature is poorly defined. This study aims to depict a space–time framework for the development of calcareous soils in archaeological stone-wall bench terraces and suggest modifications to their soil classification and mapping. To do so, we examined soils at the forefront of five terraces in Catalonia (Spain), that had been previously dated with optically stimulated luminescence to periods from Medieval-Late Middle Ages (1200 ± 60 CE) to Modern (1810 ± 15 CE). Local soils surrounding the terraces are Entisols and Inceptisols in the relatively drier areas and Mollisols in the moister areas. Methods applied are field survey, standard physico-chemical analyses, and soil micromorphology. CaCO3 content, P, OM and CEC show irregularly trending values throughout the profiles. Micromorphology shows abundant content of charred components throughout the profiles and slight discontinuous carbonate recrystallization, mostly biogenic. Based on these observations, the main parent material was determined to be local soil that was redeposited by workforce as fill material for terracing. We further interpret that pedogenesis mainly involved structure development under the influence of anthroturbation and bioturbation, inheritance of ex situ aggregates from parent soil, reaggregation with little occurrence of organic matter, and low levels of lessivage and CaCO3 redistribution. We found that pedogenesis in these terraces started in a chaotic manner, considerably faster than pedogenesis in natural soils, primarily influenced by the act of terracing, the maturity of its inherited parent soil material, and relatively large depth. Therefore, early stages of terrace soil development can differ greatly from known soil models, questioning paradigms of pedogenic ‘time zero’ and ‘maturity’ in terrace soil. We suggest four terrace fill settings for terrace material and propose a model for the state of entropy and development of terrace soils. We classify the studied soils as Archaeo-Anthroportic Xerorthents, (Anthrosols (Calcaric, Escalic)), and Archaeo-Anthroportic Typic Haploxerolls (Anthrosols (Calcaric, Escalic, Relocatic, Anthromollic)), both overlying buried Inceptisols (Cambisols). Our primary conclusion is that terrace soils should be separately classified and mapped by treating terracescapes as distinct morpho-stratigraphic units in which age is the same as the age of their geomorphic surface. This study advances the pedological understanding of archaeological stone-wall bench terraces and identifies gaps in the understanding of terrace soils.