Land Use in Eastern Iberia

(c) 2023. C Michael Barton, Arizona State University. See CITATION.cff for full citation of this dataset.

Data and R markdown script supporting analyses described in the paper:

Cegielski, W.H., Snitker, G., Barton, C.M., Bernabeu Aubán, J. Cortell Nicolau, A., Pardo Gordò, S., Bergin, S.M., Diez Castillo, A., 2023. A multi-method approach with machine learning to evaluating the distribution and intensity of prehistoric land use in eastern Iberia. Quaternary International In press.

Data files include:

(All data in Unicode UTF-8 character set. Quotes around long strings that may contain comma delimter)

• medland_survey2014_2017.csv: lithic surface collections from surveys in 3 study areas.
• training_data_E_Iberia.csv: lithic assemblages from excavation, dated contexts in eastern Iberia. These data were used for training the Random Forest model used for chronological unmixing.
• C14_S&E_Iberia_all.csv: radiocarbon dates from eastern and southern Iberia used for SPD analysis.

Metadata for Data Files

(See paper for more detailed information)

study.area: valley surveyed
zone: sampling stratum
sector: survey block
subsector: survey patch/collection unit
visibility: visibility of ground surface
collection: total number of artifacts collected
time.date: time and date of survey
xcoord: survey patch centerpoint coordinate
ycoord: survey patch centerpoint coordinate
area.sqm: area of survey patch
Counts of each lithic type collected in survey patch

Description of Lithic Types

Type Code	Type Name	Description
undiag.lithics	undiagnostic lithics	undifferentiated lithics, including:  unretouched flakes, chunks/debris, irregularly or minimally retouched flakes, and unprepared flake cores
flake.core	flake cores	prepared flake cores (e.g., levallois and discoidal cores)
MP.tools	Middle Paleolithic retouched artifacts	Large bifaces ("handaxes"), Mousterian points, and side scrapers (all forms)
notch.dent	notches & denticulates	single notches & series of notches (denticulates)
blade.tech	prismatic blade/bladelet technology	unretouched blades/bladelets, prismatic blade cores, & core preparation/rejuvenation flakes
burins	burins	burins (all forms)
end.scraper	end scrapers	end scrapers (all forms)
ret.blade	retouched blades	marginally retouched blades
invret.blade	invasively retouched blades	invasively retouched blades
microburin	microburins	microburins
backed	backed pieces	backed bladelets, backed points, and backed small flakes
trapeze	trapezes	trapeze microliths
triangle	triangles	triangle microliths
truncation	truncations	truncated blades and bladelets
bifacial.pt	bifacial points	bifacial projectile points (all forms)
dent.sickle	denticulated sickle blades	denticulated sickle blades (usually with silica sheen)

GIS Files of Survey, Sampling Protocol, and Interpolation for Occupational Ubiquity and Land Use Intensity

In ESRI shapefile format: polygons of survey patches ("subsectors"), survey blocks ("sectors"), and sampling strata ("zones"). There is only a single sampling stratum (zone) for Hoya de Buñol. Sectors were only defined for Zone 3 in the Cocina-Catadau study area. See paper for more information.

Metadata for GIS Files

study.area: valley surveyed
zone: sampling stratum
sector: survey block
subsector: survey patch/collection unit

These columns are followed in the patches files with counts of each lithic artifact type collected from that patch.
The R script listed below uses a Random Forest model to estimate the probability values of Occupational Ubiquity and Land Use Intensity for each patch with artifacts.

Spatial Interpolation for Occupational Ubiquity and Land Use Intensity

Chronological unmixing created probability values for ubiquity and intensity for each survey patch with artifacts, linked with the coordinates of the center point of each survey patch (see published paper and R scripts included here). Points generated from the center point coordinates were used to create raster maps of ubiquity and intensity using bilinear spline interpolation for each time period within each study area. The interpolation was carried out in GRASS GIS version 7.4.4 with the v.surf.bspline module. This can be done in the GRASS GUI or on the command line. The relevant command is:

v.surf.bspline input=[points file] column=[probabilities] raster_output=[interpolated raster] ew_step=300 ns_step=300

where...

• [points file] is the name of the GIS file of survey patch center points that also has ubiquity and intensity probabilities for each time period.
• [probabilities] is the column of probability values to interpolate (i.e., ubiquity or intensity for each time period).
• [interpolated raster] is the name of the raster file created through interpolation.

After interpolation, each raster was filtered to remove any negative probability values generate by the spline algorithm using the following GRASS raster map calculator command.

r.mapcalc expression="[filtered raster] = ( ( [interpolated raster] >= 0 ) * [interpolated raster] )"

R scripts

• land-use_in_E-Iberia.Rmd: R-Markdown script to carry out visibility analyses, Random Forest analyses to calculate age estimates of assemblages, and SPD (summed probability distribution) analysis of radiocarbon dates from eastern and southern Iberia
• land-use_in_E-Iberia.nb.html: HTML output file of analyses carried out with R-Markdown script