1 Simulating the long-term effects of agropastoral landuse decisions: A computational modeling approach to the Prepottery/Pottery Neolithic transition in northern Jordan Isaac I.T. Ullah 2 Problem Domain The Prepottery/Pottery Neolithic Transition in the Southern Levant ● Late PPNB/C (c. 9250 – 8500 B.P.) ● ● ● High levels of settlement centralization, with dense habitation in a few large agglomerated towns, each containing up to3000 people Highly standardized blade-based stone tool technology, advanced knowledge of plaster-making, multistory dwellings with many rooms, large statuary, and spectacular art Late Neolithic (c. 8500 – 7000 B.P.) ● ● ● Generally much less spectacular than the PPNB/C Widely dispersed in small hamlets of only about 20 people each, with fewer larger settlements of a few hundred people Stone tools made from non-standardized flakes, very little art, simple one-room houses, pottery invented, but most pots undecorated coarse-wares 3 YBP Period Name 4 Potential Motivators for PPN/LN Transition 1) Human-Caused Environmental Degradation ● ● Depletion of soil fertility due to intensive farming, increased erosion due to overgrazing and woodgathering for plaster-making Perhaps in conjunction with climate change 2) Increased Social Stress of Life in the First Large Villages ● ● Larger populations, you don't really know everyone, increased occurrence of social friction, few social institutions exist to deal with these stresses Perhaps also in conjunction with environmental degradation, but emphasis on social motivators 3) Conscious Reformulation of Subsistence Behavior to Mitigate Risk ● ● Dispersal spreads risk over many ecotones, spreads access to resources, increases chances of success No specific social or environmental motivators required, but these could be factors 5 6 Project Area and Background Neolithic Sites in Wadi Ziqlab Figure 1 7 Timeline of the Neolithic in Wadi Ziqlab 8 Tell Rakkan I PPNB/C Village ● Limited excavations ● 50-150 people ● Wheat/Barley ● Goats/Sheep ● 9 Tabaqat al Buma LN Hamlet ● Fully excavated ● 5-10 people ● Wheat/Barley ● Goats/Sheep ● 10 11 Simulation Experiment Protocol Basic Research Design ● ● ● ● ● Simulate agropastoral landuse around Tell Rakkan I for the 700 year period of the PPNB/C Several models, parameterized to represent potential Neolithic agropastoral subsistence systems. Systematically vary a small number of the most important components of potential agropastoral subsistence systems between models Keep all other variables static between models Models to serve as “Hypothesis Generators” with which to reexamine the archaeological record 12 Modeling 6 Potential Neolithic Subsistence Systems Agropastoral ratio: Ovicaprids per person: Herd stocking rate: Farming fertility decline: Agropastoral ratio: Ovicaprids per person: Herd stocking rate: Farming fertility decline: 1) Good Pastoralists 2) Good Agropastoralists 3) Good Agriculturalists 20/80 50/50 80/20 26 17 7 ~0.15 animals/ha ~0.15 animals/ha ~0.15 animals/ha 1.00% 1.00% 1.00% 4) Greedy Pastoralists 5) Greedy Agropastoralists 6) Greedy Agriculturalists 20/80 50/50 80/20 26 17 7 ~0.3 animals/ha ~0.3 animals/ha ~0.3 animals/ha 2.00% 2.00% 2.00% 13 Agropastoral Economic Data 14 15 Paleoenvironmental Reconstruction Reconstructing Neolithic Topography “Clipping out” younger areas 16 Reconstructing Neolithic Topography Interpolation 17 18 Reconstructed Topography Modern Topography Tell Rakkan I 19 20 Reconstructed Soil Depths Deep Soils Shallow Soils 21 Reconstructing Neolithic Climate Precipitation 9 27 8 26 January 7 -94 00 July -93 00 -92 00 -91 00 -90 00 -89 00 -88 00 -87 00 -86 00 -85 00 Temperature 11.5 25 -84 00 105 11 100 10.5 95 10 90 9.5 85 9 Number of Storms -94 00 -93 00 Precip per storm -92 00 -91 00 -90 00 -89 00 Year BP -88 00 -87 00 -86 00 -85 00 80 -84 00 Number of Storms per Year Average Precipitation per Storm (mm) 28 July, Degrees Celsius January, Degrees Celsius 10 Climax vegetation – PPNB/C period Grasslands Shrubs Maquis Forests 22 23 24 Results Population Over Time 25 Variation in population over 700 years Figure 9 0.5 Coeficient of Variation 0.4 0.3 0.2 0.1 0 Good Pastoralists Greedy Pastoralists Good Agropastoralists Greedy Agropastoralists Good Agriculturalists Greedy Agriculturalists 26 Variation in population after first population peakFigure 9 0.4 Coeficient of Variation 0.3 0.2 0.1 0 Good Pastoralists Greedy Pastoralists Good Agropastoralists Greedy Agropastoralists Good Agriculturalists Greedy Agriculturalists 27 Pastoralists Agropastoralists Greedy Good Figure 10 Grasslands Shrub s Maquis Forest s Agriculturalists 28 Pastoralists Agropastoralists Greedy Good Figure 10 Grasslands Shrub s Maquis Forest s Agriculturalists 29 Extent of different landcover types within a 2-hour walking-cost catchment after 700 years Grasslands Shrubs Maquis Young Woodland Mature Woodland 30 Extent of different landcover types within a 2-hour walking-cost catchment after 700 years Grasslands Shrubs Maquis Young Woodland Mature Woodland 31 32 Percent of farmable land within a 2 hour walking-cost catchment with reduced fertility 40% Percent of catchment 30% 20% 10% 0% Good Pastoralists Greedy Pastoralists Good Agropastoralists Greedy Agropastoralists Good Agriculturalists Greedy Agriculturalists Cumulative Erosion/Deposition Agropastoralists Agriculturalists Greedy Good Pastoralists Net Erosion 0 Net Deposition 33 Cumulative Erosion/Deposition Agropastoralists Agriculturalists Greedy Pastoralists Net Erosion 0 Net Deposition 34 Cumulative Human Contribution to Erosion and Deposition Over Time 35 Human contribution to erosion within a 2 hour walking-cost catchment 3 2.5 Cubic km 2 1.5 1 0.5 0 Good Pastoralists Greedy Pastoralists Good Agropastoralists Greedy Agropastoralists Good Agriculturalists Greedy Agriculturalists 36 0.05 Human contribution to deposition within a 2 hour walking-cost catchment Cubic km 0.03 0 -0.03 -0.05 Good Pastoralists Greedy Pastoralists Good Agropastoralists Greedy Agropastoralists Good Agriculturalists Greedy Agriculturalists 37 38 General Implications for PPN/LN Transition 1) Severe environmental degradation occurred in several of the scenarios, but only lead to a catastrophic population reduction in one (“greedy” agropastoralists). 2) Environmental degradation related both to degree of reliance on agriculture and the degree of “greediness”. 3) Population variability increases with both degree of reliance on pastoralism and degree of “greediness”. 4) Very large populations (600-1000 people) are only possible with high levels of agricultural dependence. 5) Equally mixed subsistence behavior, and being “good”, is the only way to achieve both very high stability and low environmental impact. The PPN/LN Transition in Wadi Ziqlab 1) Archaeological evidence suggests the Neolithic subsistence system in Wadi Ziqlab more similar to the “agropastoralists” or “pastoralists” of these simulations. 2) These simulations experienced high population variability ue to variability in the subistence base. 3) Thus, Banning's “Reduction of Variability” hypothesis seems to be the most likely motivator for the PPN/LN transition in the Wadi Ziqlab region. 4) Environmental degradation could have been a factor, however, but not necessarily. 5) Further simulations will examine if the switch to the dispersed LN settlement pattern would have mitigated the effects of resource variability on these agropastoral subsistence systems. 39 40 Thank you! I would like to thank Michael Barton, Ted Banning, Geoff Clark, Ramon Arrowsmith, Sean Bergin, Alexandra Miller, Mari Soto-Berelov, Seji Kadowaki, Kevin Gibbs, Lisa Maher, Danielle Macdonald, Helena Mitasova, Kelin Whipple, Gabriel Popescu, Christopher Roberts, Claudine Gravel-Miguel, Julien RielSalvatore, and Leah Abriani for invaluable assistance and facilitation of this work. Contact: isaac.ullah@asu.edu This research was made possible through National Science Foundation Grant BCS0410269 and various grants from the Social Sciences Research Council of Canada.