Fig 1.
Environmental Köppen Geiger classes for South Asian subcontinent [20].
Figure created from data under a Creative Commons Public Domain Dedication CC BY 4.0: Beck, H. E. et al. Figshare https://doi.org/10.6084/m9.figshare.6396959 (2018).
Fig 2.
Nested hierarchies of the PAGES LandCover6k land use classification system.
(reproduced from [14]: Fig 1). Reprinted from (14) under a CC BY 4.0 license, with permission from PlosONE, original copyright 2021.
Fig 3.
Comparison of Paleolithic sites (45-20kya), with 12k and 6k time windows.
There are notably fewer sites in both the 6k and 12k time windows than in the pre-20k time period, with particular clusters in 12k and 6k that are discussed below. Created by authors using base maps from Esri. Base map image(s) is the intellectual property of Esri and is used herein under license. Copyright © 2020 Esri and its licensors. All rights reserved. Sources for base maps: Esri, GEBCO, NOAA, National Geographic, DeLorme, HERE, Geonames.org, and other contributors.
Fig 4.
Land use map at 12kya (12,250–11,750 years ago).
There is a notable spread of LU1 HHGF across the majority of the subcontinent, with patches of minimal-extensive and small points of LU2 LLFP. Note the difference in sea-level between 12k and the modern coastline creating different potential land to use (e.g.: Sri Lanka-mainland land bridge). Inset A shows Maldives, which are included at 12k due to the lower sea levels creating a greater coastline. As will be apparent by 6k, when sea levels rise to reach modern levels, such land masses no longer provide enough area to merit their inclusion within the land mass grid. At 12k though they form a large enough landmass due to lower sea levels, but present no evidence for human land use. Map made by authors using ArcGIS and LC6k database.
Fig 5.
Land use at 6kya (6,250–5,750 years ago).
LU1 HHGF remains the dominant land use across the subcontinent, with a slight decrease in minimal-extensive and some increase in LU2 categories. Diversity in the forms of land use therefore seem to increase by 6k, but HGFF remains the main land use across South Asia. It is also important to note the change in sea level to those similar to today, changing the available land from that in 12k to that recognizable today. Map made by authors using ArcGIS and LC6k database.
Fig 6.
A) Excerpt map for Indo-Iranian borderlands and Gujarat, showing the diversity of land use within this north west region of South Asian the 6kya time window (6,750–5,5750 years ago). LU1 agriculture and LU2 herbaceous ground crops can be seen in the alluvial fans in this zoomed in excerpt, as can the broad-based marine/aquatic resources along the Gujarat coast, along with areas of LU1 agriculture. B) Excerpt map for Ganges Plains at 6kya. Areas of LU1 HGFF LU2 LLFP can be seen in this zoomed-in map. Map made by authors using ArcGIS and LC6k database. C) Excerpt map of Sri Lanka at 6kya, showing the prevalence of LU2 broad based and/or marine/aquatic resources along the coastlines. The continued presence of LU1 HGFF LU2 LLFP in the Horton Plains is also highlighted. Maps made by authors using ArcGIS and LC6k database.
Fig 7.
Data coverage and data quality maps.
Top left is data coverage at 12kya, top right is data quality at 12kya. Bottom left is data coverage at 6kya, bottom right is data quality at 6kya. Data coverage has been estimated by putting a raster interpolation on site data for both 12kya and 6kya to create density buffers at 30km to take into account survey work around sites. Data quality was mapped using the 8x8km grid square system of the LC6k database. Map made by authors using ArcGIS and LC6k database.