Reconstructions of local resource procurement networks at Cerro Baúl, Peru using multispectral ASTER satellite imagery and geospatial modeling Benjamin Vining Department of Anthropology, Wellesley College, 106 Central Street, Wellesley, MA 02481, United States abstract article info Article history: Received 7 February 2015 Received in revised form 30 April 2015 Accepted 1 May 2015 Available online 12 May 2015 Keywords: Multispectral remote sensing Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Spectral reflectance Lithological mapping Kernel Density Estimation Cerro Baúl Multispectral remote sensing is increasingly common in archaeology, but is largely oriented towards site-detection applications. A high-resolution approach using ASTER data to map environmental and geological resources in the vicinity of Cerro Baúl is outlined here. Spectral mapping evaluates several band ratio and relative band depth image transforms that target diagnostic reflectance/absorption features of earth materials. Available geological resources had a high degree of spatial heterogeneity, which conditioned household strategies for obtaining raw materials. However, some variability cannot be explained by the geological environment alone. Using Minimum Convex Polygons and Kernel Density home-territory estimates, I outline the spatial organization of resource networks that differed from household to household. Differences in household resource networks suggest that socio-economic factors conditioned the use of materials, including locally-available ones. These results demonstrate the utility of multispectral remote sensing to construct palaeoecological models on a scale congruent with archaeological questions. © 2015 Elsevier Ltd. All rights reserved. 1. Introduction Satellite remote sensing is an increasingly significant archaeological research method. Early interest in remote sensing focused on utilizing spectral data, particularly from the Landsat family of sensors, to charac- terize environmental resources in areas of archaeological interest (e.g., Cox, 1992; Custer et al., 1986; Pope and Dahlin, 1989). While this continues to be an application, much use of multispectral information in archaeology has shifted its focus predominantly towards detecting archaeological sites. Such “site detection” approaches represent the majority of more recent archaeological remote sensing programs (see, e.g., Comer and Harrower (2013); Lasaponara and Masini (2012a); Rowlands and Sarris (2007)). Despite initial enthusiasm about the potential of multispectral remote sensing, the focus on site detection has had a number of impli- cations: some of these effects are positive (such as the improvements in detection methods) while others have distracted from the wider application of remote sensing to archaeological problems. The effective spatial resolution of early remote sensing platforms superseded the nominal on-the-ground footprints of many individual archaeological features, making it difficult to resolve sites. This led to some initial frustration with multispectral data. As higher spatial resolution imagery became available, remote sensing archaeology naturally gravitated towards these data. The radiometric resolution of earlier high spatial resolution platforms—their capability to resolve different bandwidths of light as discrete channels—was reduced compared with lower spatial resolution, higher radiometric resolution platforms, however. The amount of electromagnetic radiation a sensor can resolve is a limiting factor. Spectral resolution is defined as the number, position, and spec- tral widths of each band in an imaging sensor, while spatial resolution is the effective ground area imaged by one pixel. Tradeoffs between spatial and spectral resolution arise from problems of resolving limited electromagnetic radiation in higher wavelengths over a decreased Instantaneous Field of View (IFOV). As the IFOV decreases (increasing spatial resolution), the amount of detectable electromagnetic reflec- tance decreases, requiring sensors with narrower IFOV to be sensitive to a broader reflectance bandwidth (reducing radiometric resolution). High spatial resolution sensors (sub-meter to meters) operate in the panchromatic and visible–near infrared spectra, with low spectral resolution (≤5 bands). Conversely, higher spectral resolution satellite platforms offer significantly more spectral data at longer wavelengths (≥5 bands and bands beyond the NIR range), but at the cost of lower spatial resolution (15–90 m). The interest in very high-spatial resolution sensors for site detection applications has meant that the dimension of spectral resolution is relatively underutilized in archaeological remote sensing. With a few recent exceptions (e.g., Menze and Ur, 2012; Lasaponara and Masini, 2012b; Altaweel, 2005), analysis of archaeological effects on spectral reflectance (of earth materials, vegetation, and the features themselves) Journal of Archaeological Science: Reports 2 (2015) 492–506 E-mail address: bvining@bu.edu. http://dx.doi.org/10.1016/j.jasrep.2015.05.001 2352-409X/© 2015 Elsevier Ltd. All rights reserved. Contents lists available at ScienceDirect Journal of Archaeological Science: Reports journal homepage: http://ees.elsevier.com/jasrep