Fig 1.
The distribution of key Paleolithic sites in Iran differentiated by lithic typology.
The sites are overlain on elevation data (SRTM), progressing from low (green) to high (white) altitudes. The brown speckled polygons mark dune fields. Regions referenced in the text are labelled. Dispersal routes (A-C) previously suggested by Vahdati Nasab et al. [13] are marked as black arrows. The location of archaeological sites discussed in the text is numbered as follow; 1. Kashafrud, 2. Chah-e Jam, 3. Mirak, 4. Soofi Abad, 5. Moghanak and Otchunak, 6. Darband, 7. Shanidar, 8. Bawa Yawan, 9. Qaleh Kurd, 10. Kaldar, 11. Qaleh Gusheh, 12. Ghar-e Boof. Map created by one of our authors (PSB).
Fig 2.
Modern mean annual precipitation for Iran (WorldCLIM).
Map depicts the current limits of westerly and monsoonal moisture overlain on monthly rainfall through the year. Monsoon limits come from quantitative analyses of global monsoons [46, 47]. Higher rainfall is found in the north and west, particularly in the Zagros, Alborz and Kopet Dagh Mountains further south rainfall declines in the Dasht-I Kavir and Lut Deserts. Map created by one of our authors (PSB).
Fig 3.
Mapped Iranian paleohydrology superimposed on an elevation map of Iran (Green low, white high).
The locations of climatic proxy sites discussed in the text are displayed, numbered and, if dated, color differentiated by the Marine Isotope Stage to which they relate (S2 Table); [1] Lake Urmia, [2] Lake Van, [3] Lake Zeribar, [4] Mirabad lake, [5] Qaleh Kurd cave, [6] Zarand Playa, [7] Toshan, [8] Agh Band, [9] Neka-Abelou, [10] Kalat-e Naderi, [11] Pir Ghar cave, [12] Kopet Dagh Mountain fans, [13] Neyshabour fans, [14] Kashmar fans, [15] Nimbluk, [16] Neor Lake, [17] Isfahan, [18] Sefidabeh, [19] Gavkhoni, [20–23] Makran terraces, [24] Minab fans and terraces, [25] Jazmurian Playa, [26] Sabzevar fan, [27] Kashmar (Shesh Taraz) fans and terraces, [28] Anar fan, [29] Bam fan, [30] South Golbaf basin. Map created by one of our authors (PSB).
Fig 4.
Age-depth modelling results, luminescence ages and simplified loess-paleosol stratigraphy from the Neka-Abelou (A) and Toshan (B) sites, southern Caspian Lowlands, northern Iran. Age-depth modelling was conducted at 1 cm depth resolution using the rBacon Bayesian age model [76] under default priors except for ’thick’, which was set to 30. The blue line represents the modelled median age, while black lines represent minimum and maximum 95% confidence ranges. Luminescence ages from A) Neka-Abelou (fading corrected pIR IRSL at 225 C) and B) Toshan (non-fading corrected pIR IRSL at 290 C) are taken from Kehl et al. [72] and Lauer et al. [73] respectively, and shown as red circles with 1 sigma error uncertainties. Simplified loess and soil stratigraphic logs for the sites are taken from Kehl et al. [72] and Lauer et al. [73]. Marine oxygen isotope stage (MIS) boundaries are shown by age [76, 82], with equivalent depths on the age-depth and stratigraphic plots being inferred from intercept with the median of the age depth model. MIS 3 and 5 are shaded grey.
Fig 5.
The relationship between dated archaeological sites in Iran and diverse regional environmental proxies.
A) Dated archaeological sites with Middle (brown) and Upper (blue) Paleolithic affinities. B) Orbital parameters (eccentricity, green; and precession, blue) which influence local insolation receipt (black). C) Pollen record from Lake Urmia, Zeribar and Mirabad lakes from along the northern to central Zagros, showing arboreal to non-arboreal pollen ratio, suggestive of forest/shrubs vegetation during MIS 5 and 1 [63, 67, 93]. D) Speleothem studies from Qaleh Kurd in western and Pir Ghar in the eastern Iranian Plateau, showing wetter conditions during phases of MIS 5 and also the Holocene [44, 78]. E) Alluvial fan and river terrace abandonment [80, 81]. F) Paleosol formation phases, weakly developed loess soils are shown by black diagonal lines [72, 73]. G) Oxygen isotope values from the LR04 global benthic stack [82].
Fig 6.
Hydrologically active catchments during MIS 5 (A) and MIS 3 (B), based upon dated proxies which reflect enhanced humidity.
Dated proxies such as lakes and rivers reflect an accumulation of overland flow from upland areas, where this is the case adjacent catchments containing hydrological networks originating in the same uplands are also displayed as likely humid. Map created by one of our authors (PSB).
Fig 7.
Annual precipitation estimates from paleoclimate models for MIS 5e (~130 ka Krapp et al. [61], and ~130 ka Otto-Bliesner et al. [60]), and MIS 3, overlain with broadly contemporaneous proxy sites.
MIS 3 data is from the Krapp et al. [61] model and displays the mean value during the broad window of 50–40 ka when age depth modelling of northern Iranian loess (Fig 4) suggests phases of enhanced humidity. Map created by one of our authors (PSB).
Fig 8.
Distance from water analyses highlighting the distance from mapped paleohydrology for all areas of the country.
The inset graph shows box plots of the distance from water of LP, MP and UP sites. Map created by one of our authors (PSB).
Fig 9.
The distribution of LP and MP archaeological sites in relation to MIS 5 paleoclimate data.
Sites which are discussed are numbered as follows; 1.Mirak, 2.Qaleh Kurd, 3.Bawa Yawan, 4.Ghar-e Boof, 5.Shanidar. The base map is the Otto-Bliesner et al. (130 ka) mean precipitation climate model [60]. This is overlain by catchments where MIS 5 humidity and active hydrology (hatched) are suggested by our data (see Fig 6). Simple representation of the broad routes through Iran are overlain as arrows. Map created by one of our authors (PSB).
Fig 10.
The distribution of MP and UP sites in relation to MIS 3 data.
The base map is the Krapp et al. (50–40 ka) mean precipitation climate model [61], and is overlain by catchments where MIS 3 humidity and active hydrology (hatched) are suggested by our data (see Fig 6). Map created by one of our authors (PSB).