Supporting Information Eriksson et al. 10.1073/pnas.1209494109 SI Text Net primary productivity (NPP) was estimated from mean Climate and Net Primary Productivity Reconstructions. We used the annual temperature and total annual precipitation (interpolated HadCM3 model to build a detailed reconstruction of worldwide over a 1° grid from the predictions by the climate model) using climate for the last 120 ky. HadCM3 is a general circulation model the Miami model shown in the work by Leith (16) and described consisting of a coupled atmospheric, ocean, and sea ice model in the work by Adams et al. (17). This approach takes the min- (1, 2). The resolution of the atmospheric model is 2.5° latitude imum of the temperature- and precipitation-limited rates of NPP by 3.75° longitude by 19 unequally spaced levels in the vertical. as the predicted annual NPP. The two limiting rates are calculated The resolution of the ocean is 1.25° by 1.25°, with 20 unequally independently as simple empirical functions of the appropriate spaced layers in the ocean extending to a depth of 5,200 m. The climate variable. The temperature function describes a sigmoidal model contains a range of parameterizations, including a detailed increase in NPP with temperature, with the strongest increase radiation scheme that can represent the effects of minor trace between 10 °C and 25 °C. The precipitation function describes an gases (3). The land surface scheme includes the representation asymptotic increase in NPP with precipitation, with positive NPP of the freezing and melting of soil moisture, and terrestrial evap- at all nonzero precipitation values. The model does not treat oration includes the dependence of stomatal resistance on tem- seasonality explicitly or the effects of CO2, humidity, light capture, perature, vapor pressure, and CO2 concentration (4). The ocean or plant and soil types, but it does provide a simple estimate of model uses the Gent–McWilliams mixing scheme (5). The sea NPP based on observations; hence, it is a good first-order charac- ice model uses a thermodynamic scheme and contains parame- terization of the major climatic controls on ecosystem productivity. terizations of ice drift and leads (6). In this version of the model, interactive vegetation is not included. Notes on Fossil and Archaeological Evidence for Anatomically Modern Multiple snapshot simulations covering the last 120,000 y have Human Colonization Outside of Africa. Arrival dates for anatomi- been performed with HadCM3. The boundary conditions and setup cally modern humans (AMHs) in different regions around the of the original set of simulations have been previously documented world were used to compare fossil and archaeological evidence in detail in ref. 7. The snapshots were done at intervals of every 1 ky with predictions from the model. We summarize the current state from the preindustrial (PI) to the last glacial maximum (LGM; 21 of knowledge of human arrival in different regions in Fig. S7 and ky B.P.), every 2 ky from LGM to 80 ky B.P., and every 4 ky from Tables S1 and S2. Although accurately dated and taxonomically 80 to 120 ky B.P. Boundary conditions are variable between diagnostic human remains are the only unequivocal evidence of snapshots but constant for each simulation. Orbital parameters are the arrival of AMHs in a region, artifacts in the same sites are taken from the work by Berger and Loutre (8). Atmospheric often used to support modern human settlement preceding the concentrations of CO2 were taken from the works by Petit et al. (9) date of the fossils. Niah Cave artifacts, for example, are dated to at and Loulergue et al. (10), and CH4,andN2O were taken from least 46 ky B.P. (18), whereas human occupation at Lake Mungo European Project for Ice Coring inAntarctica(EPICA)(11).All has well-supported dates going back to 50–46 ky B.P. (19). Al- ice-core data were on the EDC3 timescale (12). though Lake Mungo is the only Australian site with human re- The PI to LGM (21 ky B.P.) snapshots use the ICE-5G ice- mains older than 40 ky B.P., archaeological remains from several sheet reconstructions in the work by Peltier (13) as described in other sites confirm human settlements in different Australian ref. 7. The prescription of ice-sheet evolution from LGM to 120 – – ∼ ky B.P. differs slightly from the original simulations. The pre- regions as old as 44 48 ky B.P. (20 22). Older dates of 60 ky B.P. vious study (7) assumed that, during glaciations, the area of the have been suggested for sites of Nauwalabila and Malakunanja continents covered by ice remains similar to the LGM coverage in Northern Australia, but they have been criticized on the basis and ice thickness rises synchronously with d18O [the SPECMAP of possible bioturbation (23, 24). 18 In the Americas, signs of human occupation in both North and record of d O history in the work by Martinson et al. (14) was ∼ used to constrain the evolution of the volume of land ice]. South America by 14 ky B.P. are suggested by several sites that However, this assumption results in unrealistically large extents do not include fossil remains (Table S2). fi of Northern Hemisphere ice sheets during the initial glaciation. In Eurasia, it is more dif cult to use archaeological evidence to fi To attempt to resolve this problem, another set of snapshot sim- infer the arrival of AMHs in speci c regions because of the ulations was performed, in which the ice extent and height pro- presence of archaic hominin species that could, in principle, be ducing a particular ice volume for PI to LGM were mapped onto the makers of artifacts not directly associated with modern human similar ice volumes for LGM to 120 ky B.P. as given by the remains. SPECMAP dataset [i.e., the total volume of ice prescribed in Recently reported skeletal remains at Zhiren Cave (South the simulations in the work by Singarayer and Valdes (7) is the China) have been suggested as evidence of AMH settlement in same used here, but the evolution of ice extent and height dif- East Asia at least 100 ky B.P. (25). However, the remains exhibit fers]. The ice sheets used in these simulations are closer in areal a mosaic of archaic and derived features, which leaves doubt evolution to those ice sheets derived using ice-sheet models (15). regarding their genuine taxonomic status (26, 27). Other fossil The initial conditions for the snapshot simulations were taken remains of AMHs in China have been associated with dates from the end of each of the simulations in the work by Singarayer older that 50 ky B.P. [for example, the Liujiang remains in and Valdes (7). The simulations were spun up to adjust to the Southern China >68 ky B.P. (28)], but the dates are considered ice-sheet boundary conditions for 470 y. The climate means problematic because of uncertainty in their stratigraphic context described here are averages of years 470–500. (ref. 29, p. 620). 1. Pope VD, Gallani ML, Rowntree PR, Stratton RA (2000) The impact of new physical 2. Gordon C, et al. (2000) The simulation of SST, sea ice extents and ocean heat parametrizations in the Hadley Centre climate model: HadAM3. Clim Dyn 16: transports in a version of the Hadley Centre coupled model without flux adjustments. 123–146. Clim Dyn 16:147–168. Eriksson et al. www.pnas.org/cgi/content/short/1209494109 1of7 3. Edwards JM, Slingo A (1996) Studies with a flexible new radiation code. 1. Choosing 17. Adams B, White A, Lenton TM (2004) An analysis of some diverse approaches to a configuration for a large-scale model. Q J Roy Astron Soc 122:689–719. modelling terrestrial net primary productivity. Ecol Modell 177:353–391. 4. Cox PM, et al. (1999) The impact of new land surface physics on the GCM simulation of 18. Barker G, et al. (2007) The ‘human revolution’ in lowland tropical Southeast Asia: The climate and climate sensitivity. Clim Dyn 15:183–203. antiquity and behavior of anatomically modern humans at Niah Cave (Sarawak, 5. Gent PR, McWilliams JC (1990) Isopycnal mixing in ocean circulation models. J Phys Borneo). J Hum Evol 52:243–261. Oceanogr 20:150–155. 19. Bowler JM, et al. (2003) New ages for human occupation and climatic change at Lake 6. Cattle H, Crossley J (1995) Modeling arctic climate change. Philos Transact A Math Mungo, Australia. Nature 421:837–840. Phys Eng Sci 352:201–213. 20. Fifield LK, et al. (2001) Radiocarbon dating of the human occupation of Australia 7. Singarayer JS, Valdes PJ (2010) High-latitude climate sensitivity to ice-sheet forcing prior to 40 ka BP: Successes and pitfalls. Radiocarbon 43:1139–1145. over the last 120 kyr. Quat Sci Rev 29:43–55. 21. Gillespie R (2002) Dating the first Australians. Radiocarbon 44:455–472. 8. Berger A, Loutre MF (1991) Insolation values for the climate of the last 10 million 22. O’Connell JF, Allen J (2004) Dating the colonization of Sahul (Pleistocene Australia- years. Quat Sci Rev 10:297–317. New Guinea): A review of recent research. J Archaeol Sci 31:835–853. 9. Petit JR, et al. (1999) Climate and atmospheric history of the past 420,000 years from 23. O’Connell JF, Allen J (1998) When did humans first arrive in greater Australia and why the Vostok ice core, Antarctica. Nature 399:429–436. is it important to know? Evol Anthropol 6:132–146. 10. Loulergue L, et al. (2008) Orbital and millennial-scale features of atmospheric CH4 24.