USING OSL MEASUREMENTS TO DECIPHER SOIL HISTORY IN ARCHAEOLOGICAL TERRACES, JUDEAN HIGHLANDS, ISRAEL Naomi Porat 1 * , Uri Davidovich 4 , Yoav Avni 1 , Gideon Avni 3 , Yuval Gadot 2 1 Geological Survey of Israel, 30 Malkhe Israel St, Jerusalem 9550161, Israel 2 The Department of Archaeology and Ancient Near Eastern Cultures, Tel Aviv University, POB 39040, Tel Aviv 6997801, Israel 3 Israel Antiquities Authority, POB 586, Jerusalem 9100402, Israel 4 Institute of Archaeology, The Hebrew University, Jerusalem 9190501, Israel Received 29 September 2016; Revised 19 February 2017; Accepted 19 February 2017 ABSTRACT Archaeological terraces are a prominent feature of the agricultural sphere in hilly landscape throughout the Mediterranean, and dating of these simply built features is of utmost importance. Excavations and optically stimulated luminescence (OSL) dating of the soil fill of terraces were previously carried out on Mt. Eitan, the Judean Highlands, Israel. Archaeological surveys showed that Mt. Eitan was almost continuously settled from the Middle Bronze Age until modern times. However, OSL dating showed that all extant terraces were constructed 500–200 years ago, during the Ottoman Period, while older ages (Hellenistic to Mamluk periods) are limited to the base of several terraces. Many samples dated to the Ottoman period contain older quartz grains, indicating partial bleaching at the time of terrace construction; these grains might preserve older episodes of terrace building. To test this, we calculated ages for all measured aliquots and used the finite mixture model on the dataset to identify age components not apparent in the OSL age spectrum. Results show that aliquot ages cluster into only a few periods, of which four distinct components fall within the 14th–19th centuries (Christ era). The Hellenistic–Roman and Early Islamic periods are also represented, even in areas where terraces with such ages were not found, while the Persian Iron and Bronze Age periods are not represented at all. The unbleached grains thus preserve episodes of terrace building no longer represented in the landscape. Copyright © 2017 John Wiley & Sons, Ltd. key words: terraces; Israel; ottoman; OSL; soil INTRODUCTION The optically stimulated luminescence (OSL) dating method uses a radiation-sensitive signal in quartz grains that is reset when exposed to sunlight. When grains are re-buried, a new signal accumulates over time due to natural ionizing radia- tion (Wintle 2008); this signal is proportional to the time since burial and to the amount of environmental radiation. While laboratory experiments show that under ideal condi- tions the OSL signal is reset within seconds when exposed to the sun (e.g. Porat et al. 2001), in nature, the process is much slower due to the presence of other opaque minerals such as clays, grain coating by iron oxides or high sediment load in fluvial systems during transport (Olley et al. 2004). Thus, in many samples collected from fluvial environments, the OSL signal in the sediment grains is not fully reset (i.e. poorly bleached); this is evident by the scatter of the mea- sured equivalent dose (D e , the amount of dose received by the sample while buried) on a set of subsamples (aliquots) from a single sample. Typically the D e values of partially bleached sediments will be biased towards higher values, and the D e representative of the time of deposition will be calculated from the lowest D e population which is assumed to be the best bleached, using statistical models such as the minimum age model or finite mixture model (MAM and FMM, respectively; Galbraith & Roberts 2012). The first model isolates the youngest D e population, whereas the lat- ter divides the D e data into statistically significant popula- tions (e.g. Rodnight et al. 2005; Bateman et al. 2010). Finite mixture model has also been used to isolate the most significant young population, and commonly the youngest population comprising more than 10% of all measurements is selected (Rodnight et al. 2006). In these models, the higher D e values are discarded and not taken into account when interpreting the history of the studied sediments. In this paper, we exploit the discarded D e data and the poor bleachability of quartz under some depositional processes to obtain information on past history of quartz grains in ar- chaeological soils. Our main objective is to recognize epi- sodes of agricultural terracing for dry farming which have taken place in the past but are not revealed by OSL dating of currently existing terraces. Terracing of mountain slopes for artificial creation of arable plots was a major technological innovation that has led to the complete alteration of the natural landscape (e.g. Bevan & Conolly 2011). A cornerstone for all terrace studies is the ability to directly date the time of construction and use of terrace walls and fills, a notoriously difficult task [see a comprehensive and critical review of methods used for dating terraces in Davidovich et al. (2012)]. Extensive research has been carried out by our group since 2009 on the history of terracing of the Judean Highlands, Israel, *Correspondence to: N. Porat, Geological Survey of Israel, 30 Malkhe Israel St., Jerusalem 9550161, Israel. Phone: +972-506235388. Fax: +972- 25314332. E-mail: naomi.porat@gsi.gov.il Copyright © 2017 John Wiley & Sons, Ltd. land degradation & development Land Degrad. Develop. (2017) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/ldr.2729