Citation: Hill, C.; Madella, M.; Whitehouse, N.J.; Jiménez-Arteaga, C.; Hammer, E.; Bates, J.; Welton, L.; Biagetti, S.; Hilpert, J.; Morrison, K.D. Per Capita Land Use through Time and Space: A New Database for (Pre)Historic Land-Use Reconstructions. Land 2024, 13, 1144. https://doi.org/10.3390/land13081144 Academic Editor: Yimin Chen Received: 22 May 2024 Revised: 27 June 2024 Accepted: 12 July 2024 Published: 26 July 2024 Copyright: © 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). land Article Per Capita Land Use through Time and Space: A New Database for (Pre)Historic Land-Use Reconstructions Chad Hill 1,* , Marco Madella 2,3,4 , Nicki J. Whitehouse 5 , Carolina Jiménez-Arteaga 2, Emily Hammer 6 , Jennifer Bates 7 , Lynn Welton 8 , Stefano Biagetti 2,3 , Johanna Hilpert 9 and Kathleen D. Morrison 1 1 Department of Anthropology, University of Pennsylvania, Philadelphia, PA 19104, USA; kathy.morrison@sas.upenn.edu 2 CASEs, Department of Humanities, Universitat Pompeu Fabra, 08002 Barcelona, Spain; marco.madella@upf.edu (M.M.); carolina.jimenez@upf.edu (C.J.-A.); stefano.biagetti@upf.edu (S.B.) 3 School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, Johannesburg 2017, South Africa 4 ICREA, 08010 Barcelona, Spain 5 Archaeology, School of Humanities, University of Glasgow, Glasgow G12 8QQ, UK; nicki.whitehouse@glasgow.ac.uk 6 Department of Near Eastern Languages and Civilizations, University of Pennsylvania, Philadelphia, PA 19104, USA; ehammer@sas.upenn.edu 7 Department of Archaeology and Art History, College of Humanities, Seoul National University, Seoul 08826, Republic of Korea; jbates01@snu.ac.kr 8 Department of Near and Middle Eastern Civilizations, University of Toronto, Toronto, ON M5S 1C1, Canada 9 Department of Prehistoric Archaeology, University of Cologne, 50931 Cologne, Germany; johanna.hilpert@uni-koeln.de * Correspondence: chadhill@sas.upenn.edu Abstract: Anthropogenic land cover change (ALCC) models, commonly used for climate modeling, tend to utilize relatively simplistic models of human interaction with the environment. They have historically relied on unsophisticated assumptions about the temporal and spatial variability of the area needed to support one person: per capita land use (PCLU). To help refine ALCC models, we used a range of data sources to build a new database that attempts to bring together PCLU data with significant time depth and a global perspective. This new database can provide new nuance for our understanding of the variability in land use among and between time periods and regions, data that will have wide applicability for continued research into past human land use and present land-use change, and can hopefully help improve existing ALCC models. An example is provided, showing the potential impact of new PCLU data on land-use mapping in the Middle East at 6000 BP. Keywords: land use; land cover; ALCC; PCLU 1. Introduction Humans have been modifying their environments for millennia, becoming, over time, the driving force in changes to the global environment [1–4]. However, contention remains around when human land-use changes started and their impact on the environment through time [4–6]. Among the many impacts of past human land use are changes in vegetation type and cover, with the latter being an important driver of climate change [7]. While we can track land-use changes in the present in a variety of powerful ways [8,9], modelling how humans impacted land cover in the past through changes in land use is more difficult. However, understanding (pre)historic land use and land cover change is crucial for understanding climate change in the present [4]. Historical models of anthropogenic land cover change (ALCC models) are commonly used in climate modeling [3,10–12] and are generally constructed based on published estimates of past population combined with estimates of land-use requirements per person, or per capita land use (PCLU) [11]. While there are ALCC models covering the last several Land 2024, 13, 1144. https://doi.org/10.3390/land13081144 https://www.mdpi.com/journal/land https://doi.org/10.3390/land13081144 https://creativecommons.org/ https://creativecommons.org/licenses/by/4.0/ https://creativecommons.org/licenses/by/4.0/ https://www.mdpi.com/journal/land https://www.mdpi.com https://orcid.org/0000-0002-8397-8105 https://orcid.org/0000-0002-9324-1545 https://orcid.org/0000-0002-7044-6492 https://orcid.org/0000-0002-7774-1988 https://orcid.org/0000-0002-7100-4741 https://orcid.org/0000-0002-4809-3416 https://orcid.org/0000-0003-0936-3070 https://orcid.org/0000-0002-2144-2372 https://doi.org/10.3390/land13081144 https://www.mdpi.com/journal/land https://www.mdpi.com/article/10.3390/land13081144?type=check_update&version=1 Land 2024, 13, 1144 2 of 22 thousand years [1–3], these models have been constrained by uncertainties in population data and by the limited availability of per capita land-use (PCLU) values applicable to different time periods, regions, subsistence strategies, and societies. This paper presents a new global database of PCLU values, spanning several millennia and multiple subsistence strategies, intended to improve the available data for ALCC models while providing an important resource for further research questions on human–environment interactions. This database builds on a wide range of data sources and helps to provide significantly greater nuance to our understanding of the variation of human land use through time and space. ALCC Models and PCLU ALCC models rely on PCLU estimates to reconstruct land-use change through time. They do this by multiplying population estimates by the area supposedly needed to support each person in the population, i.e., a PCLU value. Researchers working on ALCC models acknowledge that PCLU is highly variable [11], and for most parts of the world, it is difficult to reconstruct per capita land use with sufficient resolution earlier than 1960. Therefore, ALCC models have generally relied on relatively simplistic estimates [13]. Until the latest version, for instance, the most widely used ALCC model, the History Database of the Global Environment (HYDE), applied a single hindcast PCLU value for agriculture and a single value for pastoralism before 1960 [2]. The diverse range of competing ALCC models is, at least in part, likely due to the restricted variability in how PCLU values are applied. Thus, to improve accuracy, ALCC models need better models of past human land use. One approach would be improving the quality of PCLU data both regionally and temporally. More data about how and why land use varies across time and space would allow for their more nuanced application in ALCC models and in general land-cover/land-use research. Furthermore, current ALCC models, such as HYDE and KK10 [1–3], rely heavily on data at the national boundary level. This is convenient for tracking modern land-use and land-cover change and for incorporating existing national-level population data sets, but is problematic for modeling land use backwards in time. For instance, land-use data in the HYDE model often show national boundaries persisting through time, well be- fore the earliest existence of those boundaries (Figure 1). But getting rid of the effects of modern national boundaries is difficult, given that much of the data are at this reso- lution. One possible alternative approach might involve connecting per capita land use to climatic zones [14] rather than national boundaries by using either modern climate classification [14–16] or, potentially, models of past climate classifications [17,18], if PCLU values can be reliably linked to climate classes. This will allow for more accurately dis- tributed per capita land-use values in the past and will release the data from the constraints of modern borders (Figure 2). Land 2024, 13, 1144 3 of 22 Land 2024, 13, x FOR PEER REVIEW 3 of 25 Figure 1. Sample of cropland estimates (in km2 per grid cell) from HYDE 3.2 [1] from 8000 BCE to 1500 AD. The border between Israel and Egypt has a real land cover difference in the present along the modern-day border, as visible in the satellite basemap, but this difference incorrectly persists backwards in time through at least 4000 BCE in the HYDE model. Microsoft product screen shot(s) reprinted with permission from Microsoft Corporation, Redmond, WA, USA. Figure 1. Sample of cropland estimates (in km2 per grid cell) from HYDE 3.2 [1] from 8000 BCE to 1500 AD. The border between Israel and Egypt has a real land cover difference in the present along the modern-day border, as visible in the satellite basemap, but this difference incorrectly persists backwards in time through at least 4000 BCE in the HYDE model. Microsoft product screen shot(s) reprinted with permission from Microsoft Corporation, Redmond, WA, USA. Land 2024, 13, 1144 4 of 22 Land 2024, 13, x FOR PEER REVIEW 4 of 25 Figure 2. Köppen–Geiger global climate classification data visualization from [14]. Classes (and vis- ualization) following Beck et al. (2018) [14]: Table 2. The purpose here is to show the variability and distribution of climate classes; see [14] for specific definitions. 2. Materials and Methods 2.1. Data Sources The database presented here (see SI below) incorporates information from a wide variety of sources. It is a comprehensive, but not exhaustive, database built from legacy data obtained from literature searches and published databases. As a starting point, we incorporate PCLU values gathered by Goldewijk et al. [11], drawn from a range of modern and historic PCLU values from a limited number of countries. Their approach assumed that since historic land-use data are extremely limited, they would build a few representa- tive samples of PCLU values for different regions and interpolate those values to neigh- boring countries that could be assumed to have similar land-use trajectories through time. The best source of modern PCLU data, and the source for a significant amount of the PCLU data used by Goldewijk et al. [11] and others, are derived from data published by the Food and Agriculture Organization (FAO) of the United Nations [19,20], which main- tains a global database that includes data on land-use area and population size from 1960 to the present. FAO land-use data are gathered from a variety of sources, including ques- tionnaires filled out by each individual country and submitted to the FAO annually, esti- mates calculated by the FAO, data reported by other intergovernmental agencies, and data reported by official national publications and websites. Though limited to the last 64 years, this database is an excellent global baseline for per capita land-use values tied to national boundaries, as commonly used in ALCC models. Sampled data for cropland hectares per capita and pasture land hectares per capita from FAO data for 1960, 1970, 1980, 1990, 2000, and 2010 are included as a separate table in the database for all available countries. How- ever, the spatial resolution of FAO data is poor, being homogenized to national bounda- ries, and the quality of data reporting may vary by region and through time [21,22]. The primary goal of the current database was to collect PCLU values from before 1960. How- ever, we have also included values from 1960 to the present, supplementing FAO data Figure 2. Köppen–Geiger global climate classification data visualization from [14]. Classes (and visualization) following Beck et al. (2018) [14]: Table 2. The purpose here is to show the variability and distribution of climate classes; see [14] for specific definitions. 2. Materials and Methods 2.1. Data Sources The database presented here (see SI below) incorporates information from a wide variety of sources. It is a comprehensive, but not exhaustive, database built from legacy data obtained from literature searches and published databases. As a starting point, we incorporate PCLU values gathered by Goldewijk et al. [11], drawn from a range of modern and historic PCLU values from a limited number of countries. Their approach assumed that since historic land-use data are extremely limited, they would build a few representative samples of PCLU values for different regions and interpolate those values to neighboring countries that could be assumed to have similar land-use trajectories through time. The best source of modern PCLU data, and the source for a significant amount of the PCLU data used by Goldewijk et al. [11] and others, are derived from data published by the Food and Agriculture Organization (FAO) of the United Nations [19,20], which maintains a global database that includes data on land-use area and population size from 1960 to the present. FAO land-use data are gathered from a variety of sources, including questionnaires filled out by each individual country and submitted to the FAO annually, estimates calculated by the FAO, data reported by other intergovernmental agencies, and data reported by official national publications and websites. Though limited to the last 64 years, this database is an excellent global baseline for per capita land-use values tied to national boundaries, as commonly used in ALCC models. Sampled data for cropland hectares per capita and pasture land hectares per capita from FAO data for 1960, 1970, 1980, 1990, 2000, and 2010 are included as a separate table in the database for all available countries. However, the spatial resolution of FAO data is poor, being homogenized to national boundaries, and the quality of data reporting may vary by region and through time [21,22]. The primary goal of the current database was to collect PCLU values from before 1960. However, we have also included values from 1960 to the present, supplementing FAO data wherever possible to provide more detailed information for specific subnational regions or societies. The database presented here, therefore, moves forward from these existing sources by providing additional PCLU values from a variety of ethnographic, archaeological, and historical sources. Many PCLU values are transcribed or calculated from data collected by early-to-mid-20th century ethnographers, development specialists, and geographers, Land 2024, 13, 1144 5 of 22 e.g., [23–28]. Allan [27], for instance, synthesized population and land-use area values for a diverse range of peoples and subsistence practices drawn from earlier ethnogra- phies to compare strategies among different environments and examine the variability of African food production. Additional data come from projects that calculate land-use values backward in time, based on archaeological and climatic data, for specific regions of the world. The PAGES Rhine LUCIFS project, for instance, focused on the reconstruction and quantification of concrete PCLU values for various prehistoric time slices (early Neolithic, Iron Ages) in the Rhineland [29–32]. One major limitation is a general lack of published data on per capita land use for hunt- ing/gathering/foraging societies. Only a few values were found in older sources [27,33–36], rep- resenting data for just a few countries. Instead, the bulk of the hunter/gatherer/fishing/forager (HGFF) values [see 13] in the database come from Binford’s Constructing Frames of Reference [37]. As part of a wider study of variability in hunter-gatherer subsistence strategies, Binford presents a database of 339 hunter-gatherer groups, with ethnographic and environmental variables for each one of them. These data incorporate the earlier ethnographic data published by Mur- dock [38] and include total population, an estimate of the total area occupied by each group, and a calculated value for density (in persons per 100-square-kilometer unit). These data have been converted to hectares per person and included in the database. 2.2. Data Structure The PCLU database presented here includes data categories for: continent, country, subregion, time period, land-use categories, PCLU value, data source, comments, and cited in. Although subregions are included where possible, most data are only specified to the “country” level. This is not an ideal spatial resolution for these data, and in fact, the variable size of countries makes it somewhat problematic. In many cases, sources identify a more specific subregion, and these can be included in the database, but it is harder to standardize these subregional categories. In some cases, sources identify a particular ethnic group or population that is being recorded, and this is included in the comments section where applicable. However, the primary reason that “country” is used as the organizing spatial category is also because this aligns with the primary ALCC models (HYDE 3.2 and KK10) and will allow this data to be most easily incorporated into those models. Subsistence strategies are recorded, where available, using the global land-use classifi- cation system developed by the LandCover6k project [13,39,40]. This classification system, designed in consultation with global climate modelers, historians, archaeologists, and geog- raphers, provides a scale-independent [41], nested series of land-use categories. At the top level, LU1, land use is divided into general categories meant to enable the broad-scale anal- yses. The applicable values for this database are “hunting/gathering/fishing/foraging”, “agriculture”, and “pastoralism”. Each of these categories can be further subdivided into more fine-grained divisions, LU2 and LU3, that provide more detailed information about land use. For agriculture, for instance, LU2 categories include “herbaceous ground crops”, “swidden/shifting”, “wet cultivation”, and “agroforestry/arboriculture” (see [13] for a complete description). PCLU values are recorded in hectares per capita (ha/cap), the standardized unit for this type of data in ALCC models. In many sources, PCLU is either not directly reported (for instance, “population” might be reported separately from “area under cultivation”) or reported in different units [42]. In such cases, a PCLU value has been calculated to standardize the data according to the database. 2.3. Distribution of Values The spatial and temporal distribution of PCLU values in the database is not random and reflects a few important biases. Although it includes ~1850 separate PCLU values covering nearly all of the 206 modern states, the three major land-use categories, and a temporal range of ~8000 years, there is significant redundancy, and the entries are heavily Land 2024, 13, 1144 6 of 22 skewed toward the present and some particular regions. This is summarized below in Tables 1 and 2 and discussed in the results below. Table 1. Distribution of PCLU database values by region and LU1 category. Continent # of Agriculture Values # Pastoralism Values # of HGFF Values Total North America 186 17 223 426 Latin America 91 58 20 169 Europe (including former USSR) 203 95 9 280 Asia 244 42 24 310 North Africa and Middle East 311 29 0 340 Sub-Saharan Africa 152 41 21 214 Oceania 17 11 60 88 Total 1204 293 357 1854 Table 2. PCLU counts, by date, in the database. 8000– 4000 BCE 4000– 2000 BCE 2000– 1 BCE 0–1000 CE 1000–1800 CE 1800–1900 CE 1900–1950 CE 1950- Present Count 53 42 73 50 85 307 274 835 3. Results The database provides several important take-home messages about PCLU values that can inform ALCC models, presented below. 3.1. Variation among Subsistence Strategies One of the significant things the database highlights is variation in PCLU values for different subsistence strategies based on the LC6k categories. There are some expected patterns here that reflect significant differences between agriculture, pastoralism, and HGFF. At the broadest level, we can see those differences in a density plot of the PCLU values for these three categories (Figure 3) and in spatial distributions as mean values per country (Figure 4). Agriculture and pastoralism have relatively similar ranges, but the PCLU values for agriculture mostly cluster below 1 ha/cap, while higher pastoralism values above 1 ha/cap are much more common. As should be expected, HGFF requires significantly more land per person. Similarly, because HGFF economies are broadly influenced by the local biomass, biodiversity, and natural resources in a given landscape, the area required per person can vary much more than for food production strategies. Land-use requirements are heavily dependent on resource availability and so tend to vary with latitude and other climatic factors such as temperature and precipitation. Additionally, although it was long assumed that hunting and gathering peoples were passively dependent on naturally occurring resource avail- ability, it is increasingly clear that many foraging groups, across a deep span of time, have been capable of both long-term modification and management of ecosystems [43–45] and of adapting foraging practices [46] in order to increase resource availability and decrease land-use requirements per person. Thus, PCLU values for HGFF can span a dramatic range, from as little as 10 ha/cap, overlapping with the high ranges of agriculture and pastoralism, to well over 10,000 ha/cap. Land 2024, 13, 1144 7 of 22Land 2024, 13, x FOR PEER REVIEW 7 of 25 Figure 3. Density plot for all pastoralism, hunter/gatherer/fishing/foraging (HGFF), and agriculture per capita land-use (PCLU) values. Figure 3. Density plot for all pastoralism, hunter/gatherer/fishing/foraging (HGFF), and agriculture per capita land-use (PCLU) values. 3.2. Variation among LU2 Categories The PCLU database includes further land-use classification refinements following the LC6k system [13]. As mentioned above, Land Use 2 (LU2) and Land Use 3 (LU3) refinements allow for finer-grained distinctions to be made among HGFF, pastoral, and agri- cultural groups, though LU2 and LU3 values are not available for all entries in the database. Similar to the LU1 distribution, LU2 PCLU distribution follows relatively expected patterns, especially when only considering values from before 1960. Among agricultural groups, wet cultivation, such as paddy rice farming, has the smallest per capita land use, and swidden/shifting the largest. Among the pastoral groups, anchored pastoralism has the smallest PCLU values, overlapping significantly with swidden/shifting cultivation, and mobile pastoralism the largest (Figure 5, n.b. land-use type definitions follow [13]). 3.3. Geographic Distribution A major goal of this database was to build a comprehensive collection of PCLU values that spans the entire globe at the national level and includes the maximum possible time depth. The inclusion of FAO data ensures there is at least one entry for almost every nation (Figure 6-top), providing blanket global coverage of recent land-use values. However, FAO data are already available and easily accessed (http://www.fao.org/faostat/en/, accessed on 14 June 2024). What archaeologists and ALCC modelers are most interested in is the availability of values before the start of FAO data (1960), which already tracks declines in per capita land use that accompanied the last half-century of global population growth and declining availability of new land to convert to agriculture or pasture [2,47]. While there are ~1800 PCLU values in the database from sources other than the FAO, many of these, too, are recent values. Figure 6 (middle) shows a plot of the PCLU values in the database that predate 1960. While there are nearly 900 PCLU values from before 1960, and these are spread over the last 10,000 years (Table 2), the geographic spread of this data is less complete. There are many countries, representing a significant proportion of the globe, with no coverage before 1960. This includes much of the Sahara, the Arabian Peninsula, Eastern Europe, and Central America. http://www.fao.org/faostat/en/ Land 2024, 13, 1144 8 of 22Land 2024, 13, x FOR PEER REVIEW 8 of 25 Figure 4. Comparison of mean PCLU per country (all dates): agriculture (top), pastoralism (middle), HGFF (bottom). Figure 4. Comparison of mean PCLU per country (all dates): agriculture (top), pastoralism (middle), HGFF (bottom). Land 2024, 13, 1144 9 of 22 Land 2024, 13, x FOR PEER REVIEW 9 of 25 As should be expected, HGFF requires significantly more land per person. Similarly, because HGFF economies are broadly influenced by the local biomass, biodiversity, and natural resources in a given landscape, the area required per person can vary much more than for food production strategies. Land-use requirements are heavily dependent on re- source availability and so tend to vary with latitude and other climatic factors such as temperature and precipitation. Additionally, although it was long assumed that hunting and gathering peoples were passively dependent on naturally occurring resource availa- bility, it is increasingly clear that many foraging groups, across a deep span of time, have been capable of both long-term modification and management of ecosystems [43–45] and of adapting foraging practices [46] in order to increase resource availability and decrease land-use requirements per person. Thus, PCLU values for HGFF can span a dramatic range, from as little as 10 ha/cap, overlapping with the high ranges of agriculture and pastoralism, to well over 10,000 ha/cap. 3.2. Variation among LU2 Categories The PCLU database includes further land-use classification refinements following the LC6k system [13]. As mentioned above, Land Use 2 (LU2) and Land Use 3 (LU3) re- finements allow for finer-grained distinctions to be made among HGFF, pastoral, and ag- ricultural groups, though LU2 and LU3 values are not available for all entries in the data- base. Similar to the LU1 distribution, LU2 PCLU distribution follows relatively expected patterns, especially when only considering values from before 1960. Among agricultural groups, wet cultivation, such as paddy rice farming, has the smallest per capita land use, and swidden/shifting the largest. Among the pastoral groups, anchored pastoralism has the smallest PCLU values, overlapping significantly with swidden/shifting cultivation, and mobile pastoralism the largest (Figure 5, n.b. land-use type definitions follow [13]). Figure 5. Comparison of second-level land-use (LU2) PCLU distributions for all pre-1960 entries. 3.3. Geographic Distribution A major goal of this database was to build a comprehensive collection of PCLU values that spans the entire globe at the national level and includes the maximum possible time depth. The inclusion of FAO data ensures there is at least one entry for almost every nation (Figure 6-top), providing blanket global coverage of recent land-use values. However, FAO data are already available and easily accessed (http://www.fao.org/faostat/en/, Figure 5. Comparison of second-level land-use (LU2) PCLU distributions for all pre-1960 entries. 3.4. Environmental Distribution Understanding the long history of land-use change is critical for tracking changing biomes and even the earth system. Archaeological and paleoecological data have estab- lished that, for the entire Holocene, humans have been modifying the landscape and increasingly affecting the environment while increasing the carrying capacity of the land through changes in technology, the manipulation of ecosystems, the modification of domes- tic plants, and a wide range of productive practices [48,49]. While some of these cultural practices, such as irrigation canals in arid environments, worked to partially transcend nat- ural limitations, in many contexts of the past, environmental conditions might be foreseen to limit the number of people that could be supported by a given unit of land. We should expect a strong, though not deterministic, relationship between environmental conditions and per capita land-use requirements regardless of time period and subsistence strategy. To a significant extent, this is true in the database. For this analysis, we used high-resolution Köppen–Geiger climate maps [14] to com- pare with PCLU data. The Köppen–Geiger [K–G] classification system divides the world into five main classes, based primarily on vegetation, and 30 subtypes, based on precipi- tation and temperature [15]. This classification system has been widely used for climate and climate change research. There are, though, significant problems comparing the PCLU database to K–G classification. First is the issue of resolution. While Beck et al. [14] have published K–G maps at 1 km × 1 km resolution, the PCLU data are primarily divided by national boundaries representing, in many cases, huge areas of land containing many different K–G classes. Large countries, such as China, the United States, and Argentina, may contain as many as 25 different K–G classes as the dominant environmental condition in at least one 1 km2 cell (Figure 6-bottom). However, for a basic comparison, we compared the PCLU database values to the most common K–G value for each country. This is a gross simplification of the climate for each country, though in the majority of cases, the most common K–G class does represent over 50% of the area in each country. Additionally, the K–G data represent a modern snapshot of environmental conditions (1980–2016) and may not match the conditions present when each entry in the database was recorded, especially for records in the distant past. The results of this analysis show, as expected, that PCLU values do vary based on dominant K–G class per country. Temperate regions have smaller and narrower ranges of PCLU values (Figure 7). Land 2024, 13, 1144 10 of 22Land 2024, 13, x FOR PEER REVIEW 11 of 25 Figure 6. Comparison of total PCLU values per country (top), total PCLU values before 1960 per country (middle), and the total number of Köppen–Geiger classes per country (bottom). Figure 6. Comparison of total PCLU values per country (top), total PCLU values before 1960 per country (middle), and the total number of Köppen–Geiger classes per country (bottom). Land 2024, 13, 1144 11 of 22Land 2024, 13, x FOR PEER REVIEW 13 of 25 Figure 7. Density plot for PCLU values based on Köppen–Geiger class. 3.5. Subnational Regions It is clear from the above that for at least some regions, national boundaries are far too broad to be described by a single PCLU value. For these regions, it would be better to have finer-grained subnational boundaries. Where possible, these sorts of additional georeferencing data have been included in the database. However, the availability of this sort of finer-grained detail is limited. The HGFF data published by Binford [37] includes enough information to visualize smaller spatial units for PCLU values for some countries, such as the US and Australia (Figure 8). But in most cases, there are limited subnational data. Figure 7. Density plot for PCLU values based on Köppen–Geiger class. 3.5. Subnational Regions It is clear from the above that for at least some regions, national boundaries are far too broad to be described by a single PCLU value. For these regions, it would be better to have finer-grained subnational boundaries. Where possible, these sorts of additional georeferencing data have been included in the database. However, the availability of this sort of finer-grained detail is limited. The HGFF data published by Binford [37] includes enough information to visualize smaller spatial units for PCLU values for some countries, such as the US and Australia (Figure 8). But in most cases, there are limited subnational data. 3.6. Temporal Variation Tracking PCLU change through time is particularly difficult because of the dearth of historic PCLU data [11,50]. In only a few regions do we have some reasonable continuity of PCLU values into the past. However, by aggregating the global data, we can see obvious trends in PCLU values through time. Figure 9 shows the box plot for all PCLU values divided into arbitrary time periods based loosely on the density of data in the database. Land 2024, 13, 1144 12 of 22Land 2024, 13, x FOR PEER REVIEW 14 of 25 Figure 8. Mean HGFF PCLU value per modern US state and Canadian province. 3.6. Temporal Variation Tracking PCLU change through time is particularly difficult because of the dearth of historic PCLU data [11,50]. In only a few regions do we have some reasonable continuity of PCLU values into the past. However, by aggregating the global data, we can see obvious trends in PCLU values through time. Figure 9 shows the box plot for all PCLU values divided into arbitrary time periods based loosely on the density of data in the database. Figure 8. Mean HGFF PCLU value per modern US state and Canadian province.Land 2024, 13, x FOR PEER REVIEW 15 of 25 Figure 9. Plot of time ranges in PCLU database. Changing land use through time can be difficult to model accurately. We know that, on average, per capita land use declines through time. But these changes can suffer signif- icantly from problems of equifinality. Land use can be affected not only by changing en- vironmental conditions but also by changing social structures and subsistence technolo- gies. The anthropogenic modification of species and soils through time has had a dramatic effect on productivity and thus land-use requirements. This is especially marked for Zea mays, for instance, with domestication characterized by decreasing profligacy but the mas- sively increasing size of individual ears [51]. Similarly, chinampas, created by the Aztecs and still in use in some parts of Mexico, require the careful creation of rich, organic soils from excavated canals built up into floating gardens and have one of the highest levels of productivity of any technological intervention [52]. The strength of the current database lies in its broad range of subsistence types, time periods covered, and global range, rather than the total number of data points over a sig- nificant time span for any one country. In only a few examples does the database include many entries that cover a significant time span for the same country. In the US, for exam- ple, the database includes 200+ entries, and these entries span from 600 CE to the present. However, the vast majority of those dates are from 1800 CE onwards and represent a broad variety of environmental conditions rather than a linear sample of a single area through time. For a few countries, such as China, the database contains more temporally expansive entries. The database contains ca. 80 entries for the modern state boundaries of China. While the specific entries may be similarly spread across a range of environmental conditions, they do provide a more evenly distributed range of dates, from 1 CE to the present, allowing a more complete temporal visualization of PCLU through time (Figure 10. Figure 9. Plot of time ranges in PCLU database. Land 2024, 13, 1144 13 of 22 Changing land use through time can be difficult to model accurately. We know that, on average, per capita land use declines through time. But these changes can suffer sig- nificantly from problems of equifinality. Land use can be affected not only by changing environmental conditions but also by changing social structures and subsistence tech- nologies. The anthropogenic modification of species and soils through time has had a dramatic effect on productivity and thus land-use requirements. This is especially marked for Zea mays, for instance, with domestication characterized by decreasing profligacy but the massively increasing size of individual ears [51]. Similarly, chinampas, created by the Aztecs and still in use in some parts of Mexico, require the careful creation of rich, organic soils from excavated canals built up into floating gardens and have one of the highest levels of productivity of any technological intervention [52]. The strength of the current database lies in its broad range of subsistence types, time periods covered, and global range, rather than the total number of data points over a significant time span for any one country. In only a few examples does the database include many entries that cover a significant time span for the same country. In the US, for example, the database includes 200+ entries, and these entries span from 600 CE to the present. However, the vast majority of those dates are from 1800 CE onwards and represent a broad variety of environmental conditions rather than a linear sample of a single area through time. For a few countries, such as China, the database contains more temporally expansive entries. The database contains ca. 80 entries for the modern state boundaries of China. While the specific entries may be similarly spread across a range of environmental conditions, they do provide a more evenly distributed range of dates, from 1 CE to the present, allowing a more complete temporal visualization of PCLU through time (Figure 10). Land 2024, 13, x FOR PEER REVIEW 16 of 25 Figure 10. PCLU database values vs. time for China. Each dot represents one PCLU entry in the database. While Goldewijk et al. [11] focused on specific individual countries for which there were historic PCLU values that could be used to construct temporal regional PCLU trajec- tories, one goal of our database was to build a more comprehensive aggregation of values from any country or time period. This means that rather than building PCLU trajectories for a specific country that can be interpolated to other neighboring regions, our database can potentially combine all historic values from across a region to include a greater range of such values. For instance, aggregating values from all northern European countries gives a greater breadth of temporal values, spanning 6000 years (Figure 11 n.b., this figure only includes agriculture and pastoralism PCLU values). Figure 10. PCLU database values vs. time for China. Each dot represents one PCLU entry in the database. While Goldewijk et al. [11] focused on specific individual countries for which there were historic PCLU values that could be used to construct temporal regional PCLU trajec- tories, one goal of our database was to build a more comprehensive aggregation of values Land 2024, 13, 1144 14 of 22 from any country or time period. This means that rather than building PCLU trajectories for a specific country that can be interpolated to other neighboring regions, our database can potentially combine all historic values from across a region to include a greater range of such values. For instance, aggregating values from all northern European countries gives a greater breadth of temporal values, spanning 6000 years (Figure 11 n.b., this figure only includes agriculture and pastoralism PCLU values). Land 2024, 13, x FOR PEER REVIEW 17 of 25 Figure 11. PCLU database values vs. time for all of Northern Europe. Each dot represents a PCLU value in the database. One crucial assumption made both here and in HYDE is that the best PCLU estimate for countries and areas for which no data exists is calculated by finding adjacent areas with the most similar ecological and environmental features. While this is undoubtedly the easiest way to fill in blank spots based primarily on proximity, a more nuanced ap- proach might compare the particular features of individual subsistence practices to find the best estimate of local land-use requirements at a particular time and place. In this case, the careful application of global ethnographic data might play an extremely important role. Comptour et al. [53], for instance, suggest that Congo Basin yam hills/raised beds could be used as an analogue for raised fields in pre-Columbian Latin America. Such a time-consuming approach is beyond the scope of this paper but could make an important contribution to the development of more nuanced PCLU models. 4. Discussion 4.1. Population and Land Use The relationship between land use and population is complex. As noted, small pop- ulations practicing extensive forms of land use such as HGFF may use large areas of land at a low intensity, while large, aggregated populations may use the same amount of land but at a higher intensity. While each group uses the same area, the potential consequences for land cover and other human impacts are significantly different. Land “use” (as meas- ured simply by area without reference to production strategy, land-use intensity, or man- agement) is thus of limited utility and should be seen as a starting point rather than the goal of analysis. Land-use intensification refers to strategies that aim to increase productivity while holding land constant, a process associated with a range of factors, including but not lim- ited to population growth [54,55]. The intensification of agricultural production allowed the development of population aggregates such as towns and cities, and it has been doc- umented as a response to other factors, such as commercial opportunities and even ine- quality and resource aggregation, meaning that it is only roughly correlated with popula- tion increase. Figure 11. PCLU database values vs. time for all of Northern Europe. Each dot represents a PCLU value in the database. One crucial assumption made both here and in HYDE is that the best PCLU estimate for countries and areas for which no data exists is calculated by finding adjacent areas with the most similar ecological and environmental features. While this is undoubtedly the easiest way to fill in blank spots based primarily on proximity, a more nuanced approach might compare the particular features of individual subsistence practices to find the best estimate of local land-use requirements at a particular time and place. In this case, the careful application of global ethnographic data might play an extremely important role. Comptour et al. [53], for instance, suggest that Congo Basin yam hills/raised beds could be used as an analogue for raised fields in pre-Columbian Latin America. Such a time- consuming approach is beyond the scope of this paper but could make an important contribution to the development of more nuanced PCLU models. 4. Discussion 4.1. Population and Land Use The relationship between land use and population is complex. As noted, small populations practicing extensive forms of land use such as HGFF may use large areas of land at a low intensity, while large, aggregated populations may use the same amount of land but at a higher intensity. While each group uses the same area, the potential consequences for land cover and other human impacts are significantly different. Land “use” (as measured simply by area without reference to production strategy, land-use intensity, or management) is thus of limited utility and should be seen as a starting point rather than the goal of analysis. Land 2024, 13, 1144 15 of 22 Land-use intensification refers to strategies that aim to increase productivity while holding land constant, a process associated with a range of factors, including but not limited to population growth [54,55]. The intensification of agricultural production allowed the development of population aggregates such as towns and cities, and it has been documented as a response to other factors, such as commercial opportunities and even inequality and resource aggregation, meaning that it is only roughly correlated with population increase. Even if it is not the sole causal factor, on a global scale, over the course of the Holocene, population increases have been associated with decreasing land use per capita. This can be visible within HGFF societies (e.g., the increasing use of low-ranked prey species with demographic pressure [56–58]) as shifts between major subsistence practices (e.g., the shift from HGFF to food production [59,60]) and within food-producing societies (e.g., as changes in technology increase food production efficiency [61]). The continued existence of low-intensity forms of land use provides evidence, however, that this general trend is by no means universal. Further, even societies with high land-use intensity often combine more and less intensive forms of production (for example, irrigated rice farming with extensive grazing and the collecting of wild plants [54]), making the use of a single per capita land-use value problematic even for a small region. Given these difficulties, ALCC modelers have adopted different strategies, either using a single PCLU value for all time periods [1] or assuming a steady process of historical land-use intensification [3], resulting in a reduction of PCLU over time. Neither solution is ideal, and the use of more focused PCLU values appropriate to specific times and places, while presenting limitations, may allow ALCC models to better capture the impacts of historical land-use variability. An important part of this goal for model improvement is to better incorporate land-use shifts beyond those from HGFF to food production. As critical as farming was (and is) as a driver of land-cover change, other forms of subsistence also modified vegetation and carbon cycles [62,63], and aggregating the impacts of these forms of land use is a desideratum for future research. 4.2. Increased Nuance within LU1 Categories ALCC models and the PCLU estimates they currently utilize are primarily designed to capture the most dramatic shifts that affect land cover, i.e., shifts from HGFF to agri- culture and the concomitant deforestation that accompanies this change in many (but not all) places. This focus sets aside the effect of many other changes affecting land use, including changes in agriculture technology such as the development of plowing [61,64–66], the management of soils to increase crop productivity [67–69], changes in wild resource management in HGFF societies [46,57], and shifts in surplus production [70–73], etc. Our current database helps capture some of that variation by providing additional ways to track differences among major subsistence strategy classes (LU1), including more nuanced land-use classification (LU2) and a greater range of time depth. 4.3. Potential Applications The limited variability of the PCLU data utilized in ALCC models has been previously highlighted [11,13] and is one of the key motivating factors for the construction of the current database. This new database brings together a large number of values for the entire globe across a deep range of time, with the potential to provide more nuance to present ALCC models. In fact, improved PCLU data are already helping to refine such models [1,11]. However, as described above, there is a limit to the availability of global PCLU data before 1960. While Klein Goldewijk et al. [11] have introduced a significant improvement in the latest version, with the addition of some 456 PCLU values comprised of historical agriculture and pastoralism estimates, our new database builds on that to produce a much larger dataset for global modeling, with 1854 values across agriculture, pastoralism, and hunting/gathering/fishing/foraging. However, it currently remains Land 2024, 13, 1144 16 of 22 unclear if this expanded dataset is sufficient to significantly affect ALCC models such as HYDE. One way this data could be used would be as a resource from which to construct regional lookup tables (LUTs) for key time slices. Such tables would provide increased variation in ALCC models. This could be done by taking the best data available, either incorporating direct data for each country, subsistence strategy, and time period, where available, by applying appropriate regional data where necessary, or by making informed decisions about how to apply values from other regions and time periods based on envi- ronmental and subsistence strategy similarities as needed. As an example, we constructed a LUT for the Middle East at 6 kya based on the land-use data published as part of the LandCover6k project [13] There are six land-use categories coded for the Middle East at 6k in the published example, and these are shown in Table 3. These include agriculture with and without irrigation, extensive/minimal, hunting/gathering/fishing/foraging (HGFF) with low-level food production, and mobile-regular pastoralism. A single PCLU value has been assigned for each land-use type for this region and time period based on what is available from the PCLU database and our own assessment of which values are appropriate to apply based on date, location, and context. Table 3. PCLU lookup table (LUT) for the Middle East at 6 kya. LC6k Land-Use Category Mean/Single PCLU Range Agriculture, herbaceous ground crops, rainfed 0.5 0.21–0.79 Agriculture, herbaceous ground crops, with water modification 0.48 0.28–0.53 Extensive, minimal 0 0 HGFF, low-level food production 4248 588–13,333 No evidence, underwater 0 0 Pastoralism, mobile-regular 1.5 0.5–3 For rainfed agriculture, the database contains 32 entries for historical reconstructions of PCLU from 6000 kya to 1000 kya. The closest match in the database in time and space is from Wilkinson [74,75], with estimates for rainfed agriculture at various levels of annual rainfall in northern Mesopotamia. These values have been averaged, and their range has been included. However, even more nuances could be included by combining his- toric rainfall estimates with these data to plot different PCLU areas based on regional variation in rainfall. For irrigation, a mean value is derived from the combination of Akker- mans [76] and Wilkinson [74], who provide values for irrigated agriculture in the region for 8000 kya and 4100 kya, respectively. There is no PCLU for areas that were underwater or had minimal human presence. Finally, although there is no PCLU value for HGFF in the Middle East in the past or present, a mean was calculated for the range of entries that come from broadly similar environmental conditions. This includes arid parts of sub-Saharan Africa; arid areas in the Americas, such as Arizona; and areas in Oceania, such as western Australia. 4.4. Impacts on HYDE The LUT example provided here can be plugged into the population data that un- derlies HYDE 3.2 to see how a more nuanced approach to PCLU values impacts this test region. Although HYDE allocates national population data to a finer-resolution grid, here a simplified version is presented using just the national-level data and the “baseline” es- timate (rather than the upper or lower estimates) [1]. Table 4 presents the total land-use area estimates for just two land-use categories—rainfed agricultural land and irrigated Land 2024, 13, 1144 17 of 22 agricultural land—as calculated in HYDE 3.2 and with the PCLU substituted from the LUT presented above. Figure 12 presents this data as a map of the region, comparing land-use estimates for these two types. The ratio of rainfed to irrigated agriculture for the LUT estimates is based on the ratio of these two types of agriculture, calculated from the LC6k estimate for the Middle East for 6k [13]. Table 4. Total land area, in km2, for rainfed and irrigated agriculture, using modern national boundaries, for HYDE and with the proposed 6k PCLU lookup table. Country HYDE 3.2 Rainfed Agriculture HYDE 3.2 Irrigated Agricultural Area Irrigated Agricultural Area Using LUT Rainfed Agricultural Area Using LUT Jordan 2267.859 0 245.09 0 Saudi Arabia 0 0 0 0 Iraq 13,129.88 749.3536 606 858.11 Yemen 0 0 0 301.63 Oman 0 0 0 0 United Arab Emirates 0 0 0 0 Qatar 0 0 0 0 Bahrain 0 0 0 0 Kuwait 0 0 0 0 Syrian Arab Republic 14,090.65 99.83208 935.52 574.19 TOTAL (km2) 29,488.38 849.18 1786.61 1733.93 The population data are constant between these two examples; only PCLU values and the relative percentage of each land-use type per country have changed. But the impact on total land use is relatively large and not unidirectional. At 6k, for the Middle East, HYDE 3.2 relies on a single large value for rainfed agriculture [11], 4 ha/c, de- rived from Butzer [77]. However, more recent estimates for rainfed agriculture are much lower than this, as noted above, so the total area estimates where rainfed agriculture was most significant in modern-day Iraq and Syria are an order of magnitude larger in HYDE 3.2. Although a common concern is that HYDE underestimates land use in earlier periods [78], this shows the potential ways that it may also overestimate land use in some regions. Additionally, the distribution of land is spread more widely across the region based on these models. 4.5. Alternate Approaches Building lookup tables, such as Table 3, provides a good starting point for ways to use the expanded PCLU database presented here to improve past land-use models at various time slices. However, a different approach may still be needed to get at some of the crucial variables in human land use that are not adequately captured by the limited historic PCLU data (such as surplus production, etc.). Instead of relying solely on a combination of estimates of per capita land-use and population models, one alternative approach would incorporate direct evidence for historic land-use and land-cover change. The LandCover6k project is attempting to build such a global database [13,39,40], combining global pollen data for land cover with archaeological data for land use for six key time slices chosen in consultation with climate modelers [79]. This database is being constructed at a resolution that is significantly higher, with a global grid of 8 km × 8 km cells. Global datasets like this can significantly improve models of pre-industrial anthropogenic land-cover change and provide a more realistic building block for earth system model (ESM) simulations [80]. However, such approaches are time consuming, harnessing decades of global archaeological data that are rarely published with this sort of synthesis in mind. So, in the meantime, Land 2024, 13, 1144 18 of 22 improved PCLU data, providing a better model of how land use has varied across time and space, may provide a useful tool for modelers. Land 2024, 13, x FOR PEER REVIEW 20 of 25 Figure 12. (1) HYDE 3.2 baseline total land rainfed agricultural land at 6k; (2) HYDE 3.2 baseline total irrigated agricultural land at 6k; (3) Total rainfed agriculture land area using HYDE 3.2 baseline population data and PCLU from lookup table above; (4) Total irrigated agriculture land area using HYDE 3.2 baseline population data and PCLU estimates above. Table 4. Total land area, in km2, for rainfed and irrigated agriculture, using modern national bound- aries, for HYDE and with the proposed 6k PCLU lookup table. Country HYDE 3.2 Rainfed Agri- culture HYDE 3.2 Irrigated Agri- cultural Area Irrigated Agricul- tural Area Using LUT Rainfed Agricul- tural Area Using LUT Jordan 2267.859 0 245.09 0 Saudi Arabia 0 0 0 0 Iraq 13,129.88 749.3536 606 858.11 Yemen 0 0 0 301.63 Oman 0 0 0 0 United Arab Emirates 0 0 0 0 Qatar 0 0 0 0 Bahrain 0 0 0 0 Kuwait 0 0 0 0 Syrian Arab Republic 14,090.65 99.83208 935.52 574.19 Figure 12. (1) HYDE 3.2 baseline total land rainfed agricultural land at 6k; (2) HYDE 3.2 baseline total irrigated agricultural land at 6k; (3) Total rainfed agriculture land area using HYDE 3.2 baseline population data and PCLU from lookup table above; (4) Total irrigated agriculture land area using HYDE 3.2 baseline population data and PCLU estimates above. 5. Conclusions Understanding the role of humans in land-cover change over the last ten thousand years is a critical part of modeling climate change into the future. Since we know that current ALCC models lack nuance in important details, such as the variability of per capita land use through space and time and across subsistence practices, it is critical that we improve these models with all available resources. The database presented here is an important step towards building finer-grained models of past human impacts. The incorporation Land 2024, 13, 1144 19 of 22 of significant additional sources of PCLU data can only help improve the quality of land- use models, as evidenced by the example provided from the Middle East. Beyond its applicability to climate models, this database should also be of interest to archaeologists looking at how land use, subsistence strategies, technologies, and populations have co- varied and interacted in different ways. Supplementary Materials: The database and an accompanying bibliography are both available here: https://doi.org/10.6084/m9.figshare.25631802. Author Contributions: Conceptualization, C.H., M.M., K.D.M. and N.J.W.; data curation, C.H., C.J.-A. and K.D.M.; writing—original draft preparation, C.H.; writing—review and editing, C.H., J.B., M.M., K.D.M., N.J.W., C.J.-A., E.H., L.W., S.B. and J.H.; visualization, C.H.; supervision, K.D.M. All authors have read and agreed to the published version of the manuscript. Funding: This study was undertaken as part of LandCover6k, a working group of the Past Global Changes (PAGES) project, which in turn received support from the Swiss Academy of Sciences and the Chinese Academy of Sciences. The study was also part of the “HoLa—Holocene Land Use” focus group that received funding from INQUA and of the “Land Use: from Global to Local” project of the Planetary Wellbeing Initiative at University Pompeu Fabra. Support was also provided by “Big data analysis on historical climate and land coverage changes for their prediction during next generation semiconductor manufacturing and the 4th industrial revolution era”, grant number A0342- 20220007, funded by Research Support Fund Industry-University Agreement Samsung Electronics DS, informally called the Korean Working Group (KWG). Data Availability Statement: The entire database presented here, along with the extended bibliogra- phy, are available via the Supplemental Information linked above. Conflicts of Interest: The authors declare no conflicts of interest. References 1. Klein-Goldewijk, K.; Beusen, A.; Doelman, J.; Stehfest, E. Anthropogenic Land Use Estimates for the Holocene—HYDE 3.2. Earth Syst. Sci. Data 2017, 9, 927–953. [CrossRef] 2. Klein Goldewijk, K.; Beusen, A.; Janssen, P. Long-Term Dynamic Modeling of Global Population and Built-up Area in a Spatially Explicit Way: HYDE 3.1. Holocene 2010, 20, 565–573. [CrossRef] 3. Kaplan, J.O.; Krumhardt, K.M.; Ellis, E.C.; Ruddiman, W.F.; Lemmen, C.; Goldewijk, K.K. Holocene Carbon Emissions as a Result of Anthropogenic Land Cover Change. Holocene 2011, 21, 775–791. [CrossRef] 4. Ruddiman, W.F. The Anthropogenic Greenhouse Era Began Thousands of Years Ago. Clim. Chang. 2003, 61, 261–293. [CrossRef] 5. Ruddiman, W.F. The Early Anthropogenic Hypothesis: Challenges and Responses: Early anthropogenic hypothesis. Rev. Geophys. 2007, 45, 207. [CrossRef] 6. Nikulina, A.; Macdonald, K.; Zapolska, A.; Serge, M.A.; Roche, D.M.; Mazier, F.; Davoli, M.; Svenning, J.-C.; van Wees, D.; Pearce, E.A. Hunter-Gatherer Impact on European Interglacial Vegetation: A Modelling Approach. Quat. Sci. Rev. 2024, 324, 108439. [CrossRef] 7. Cui, J.; Piao, S.; Huntingford, C.; Wang, X.; Lian, X.; Chevuturi, A.; Turner, A.G.; Kooperman, G.J. Vegetation Forcing Modulates Global Land Monsoon and Water Resources in a CO2-Enriched Climate. Nat. Commun. 2020, 11, 5184. [CrossRef] 8. Klein Goldewijk, K.; Ramankutty, N. Land Cover Change over the Last Three Centuries Due to Human Activities: The Availability of New Global Data Sets. GeoJournal 2004, 61, 335–344. [CrossRef] 9. Ellis, E.C.; Ramankutty, N. Putting People in the Map: Anthropogenic Biomes of the World. Front. Ecol. Environ. 2008, 6, 439–447. [CrossRef] 10. Hurtt, G.C.; Chini, L.P.; Frolking, S.; Betts, R.A.; Feddema, J.; Fischer, G.; Fisk, J.P.; Hibbard, K.; Houghton, R.A.; Janetos, A.; et al. Harmonization of Land-Use Scenarios for the Period 1500-2100: 600 Years of Global Gridded Annual Land-Use Transitions, Wood Harvest, and Resulting Secondary Lands. Clim. Chang. 2011, 109, 117–161. [CrossRef] 11. Klein Goldewijk, K.; Dekker, S.C.; van Zanden, J.L. Per-Capita Estimations of Long-Term Historical Land Use and the Conse- quences for Global Change Research. J. Land Use Sci. 2017, 12, 313–337. [CrossRef] 12. Pongratz, J.; Reick, C.; Raddatz, T.; Claussen, M. A Reconstruction of Global Agricultural Areas and Land Cover for the Last Millennium. Glob. Biogeochem. Cycles 2008, 22, GB3018. [CrossRef] 13. Morrison, K.D.; Hammer, E.; Boles, O.J.C.; Madella, M.; Whitehouse, N.J.; Gaillard, M.-J.; Bates, J.; Vander Linden, M.; Merlo, S.; Yao, A.; et al. Mapping Past Human Land Use Using Archaeological Data: A New Classification for Global Land Use Synthesis and Data Harmonization. PLoS ONE 2021, 16, e0246662. [CrossRef] 14. Beck, H.E.; Zimmermann, N.E.; McVicar, T.R.; Vergopolan, N.; Berg, A.; Wood, E.F. Present and Future Köppen-Geiger Climate Classification Maps at 1-km Resolution. Sci. Data 2018, 5, 180214. [CrossRef] https://doi.org/10.6084/m9.figshare.25631802 https://doi.org/10.5194/essd-9-927-2017 https://doi.org/10.1177/0959683609356587 https://doi.org/10.1177/0959683610386983 https://doi.org/10.1023/B:CLIM.0000004577.17928.fa https://doi.org/10.1029/2006RG000207 https://doi.org/10.1016/j.quascirev.2023.108439 https://doi.org/10.1038/s41467-020-18992-7 https://doi.org/10.1007/s10708-004-5050-z https://doi.org/10.1890/070062 https://doi.org/10.1007/s10584-011-0153-2 https://doi.org/10.1080/1747423X.2017.1354938 https://doi.org/10.1029/2007GB003153 https://doi.org/10.1371/journal.pone.0246662 https://doi.org/10.1038/sdata.2018.214 Land 2024, 13, 1144 20 of 22 15. Kottek, M.; Grieser, J.; Beck, C.; Rudolf, B.; Rubel, F. World Map of the Köppen-Geiger Climate Classification Updated. Metz 2006, 15, 259–263. [CrossRef] 16. Rubel, F.; Kottek, M. Observed and Projected Climate Shifts 1901–2100 Depicted by World Maps of the Köppen-Geiger Climate Classification. Meteorol. Z. 2010, 19, 135. [CrossRef] 17. Yoo, J.; Rohli, R.V. Global Distribution of Köppen–Geiger Climate Types during the Last Glacial Maximum, Mid-Holocene, and Present. Palaeogeogr. Palaeoclimatol. Palaeoecol. 2016, 446, 326–337. [CrossRef] 18. Willmes, C.; Becker, D.; Brocks, S.; Hütt, C.; Bareth, G. High Resolution Köppen-Geiger Classifications of Paleoclimate Simulations. Trans. GIS 2017, 21, 57–73. [CrossRef] 19. FAO. FAOStat Land Use Domain; FAO: Rome, Italy, 2020. 20. FAO. FAOStat Annual Population; FAO: Rome, Italy, 2019. 21. The World Bank. The World Bank Global Strategy to Improve Agricultural and Rural Statistics; The World Bank: Washington, DC, USA, 2011. 22. Carfagna, E.; Pratesi, M.; Carfagna, A. Methodological Developments for Improving the Reliability and Cost-Effectiveness of Agricultural Statistics in Developing Countries. In Proceedings of the 59th World Statistics Congress of the International Statistical Institute, The Hague, The Netherlands, 25–30 August 2013. 23. Schoolcraft, H.R. Notes on the Iroquois: Or, Contributions to American History, Antiquities, and General Ethnology; E. H. Pease & Co.: Albany, NY, USA, 1847. 24. Simpson, J.H. Palestine. Report on Immigration, Land Settlement and Development [and Appendix Containing Maps]; [Parliament. Papers by command] Cmd. 3686-3687; H.M. Stationery Off.: London, UK, 1930. 25. Stadelman, R. Maize Cultivation in Northwestern Guatemala; Carnegie Institution of Washington Publication; Johnson Reprint Corporation: New York, NY, USA, 1940. 26. Piper, C.V.; Vinall, H.N.; Oakley, R.A.; Carrier, L.; Baker, O.E.; Cotton, J.; Juve, O.A.; Bradshaw, N.P.; Sheets, E.; Marsh, C.D. Our Forage Resources; USDA Yearbook; United States Department of Agriculture: Washington, DC, USA, 1924. 27. Allan, W. The African Husbandman; Oliver and Boyd: Edinburgh, UK, 1965. 28. Humphrey, N. The Relationship of Population to the Land in South Nyeri. In The Kikuyu Lands; Government Printer: Nairobi, Kenya, 1945; pp. 1–17. 29. Dikau, R.; Herget, J.; Hennrich, K. Land Use and Climate Impacts on Fluvial Systems during the Period of Agriculture in the River Rhine Catchment (RhineLUCIFS)—An Introduction. Erdkunde 2005, 177–183. [CrossRef] 30. Lang, A.; Preston, N.; Dickau, R.; Bork, H.-R.; Rudiger, M. Examples from the Rhine catchment. Pages News 2000, 8, 11–13. [CrossRef] 31. Houben, P.; Burggraaff, P.; Hoffmann, T.; Kleefeld, K.; Zimmermann, A.; Dikau, R. Reconstructing Holocene Land-Use Change and Sediment Budgets in the Rhine System. Pages News 2007, 15, 17–18. [CrossRef] 32. Schmidt, I.; Hilpert, J.; Kretschmer, I.; Peters, R.; Broich, M.; Schiesberg, S.; Vogels, O.; Wendt, K.P.; Zimmermann, A.; Maier, A. Approaching Prehistoric Demography: Proxies, Scales and Scope of the Cologne Protocol in European Contexts. Phil. Trans. R. Soc. B 2021, 376, 20190714. [CrossRef] [PubMed] 33. Russell-Smith, J.; Lucas, D.; Gapindi, M.; Kapirigi, N.; Namingum, G.; Giuliani, P.; Chaloupka, G. Aboriginal Resource Utilization and Fire Management Practice in Western Arnherm Land, Monsoonal Northern Australia: Notes for Prehistory, Lessons for the Future. Human Ecol. 1997, 25, 159–195. [CrossRef] 34. Gerbens-Leenes, W.; Nonhebel, S. Food and Land Use. The Influence of Consumption Patterns on the Use of Agricultural Resources. Appetite 2005, 45, 24–31. [CrossRef] [PubMed] 35. Debenham, F. The Bushman’s Way of Life: Freedom and Contentment in Arid Surroundings. Times Br. Colon. Rev. 1954, 13, 7. 36. Tindale, N.B. Aborigines: The White Contact. In Australian Encyclopaedia; Butterworth: Sydney, Australia, 1958. 37. Binford, L.R. Constructing Frames of Reference: An Analytical Method for Archaeological Theory Building Using Hunter-Gatherer and Environmental Data Sets; University of California Press: Berkeley, CA, USA, 2001; ISBN 978-0-520-22393-6. 38. Murdock, G.P. Ethnographic Atlas; University of Pittsburgh: Pittsburg, PA, USA, 1967. 39. Gaillard, M.-J.; Morrison, K.; Whitehouse, N. Past Anthropogenic Land Use and Land Cover Change at the Global Scale for Climate Modelling Studies: PAGES LandCover6k Working Group. Quat. Perspect. 2015, 22, 25–27. 40. Morrison, K.; Hammer, E.; Popova, L.; Madella, M.; Whitehouse, N.; Gaillard, M.-J. Global-Scale Comparisons of Human Land Use: Developing Shared Terminology for Land-Use Practices for Global Change. Pages Mag. 2018, 26, 8–9. [CrossRef] 41. Di Gregorio, A.; Jansen, L.J.M. Land Cover Classification System (LCCS): Classification Concepts and User Manual, for Software Version 1.0; FAO: Rome, Italy, 2001; ISBN 978-92-5-104216-8. 42. Schroeder, S. Maize Productivity in the Eastern Woodlands and Great Plains of North America. Am. Antiq. 1999, 64, 499–516. [CrossRef] 43. Kay, A.U.; Kaplan, J.O. Human Subsistence and Land Use in Sub-Saharan Africa, 1000BC to AD1500: A Review, Quantification, and Classification. Anthropocene 2015, 9, 14–32. [CrossRef] 44. Sheuyange, A.; Oba, G.; Weladji, R.B. Effects of Anthropogenic Fire History on Savanna Vegetation in Northeastern Namibia. J. Environ. Manag. 2005, 75, 189–198. [CrossRef] 45. Gammage, B. The Biggest Estate on Earth: How Aborigines Made Australia; Nachdr., Allen & Unwin: Sydney, Australia, 2012; ISBN 978-1-74331-132-5. https://doi.org/10.1127/0941-2948/2006/0130 https://doi.org/10.1127/0941-2948/2010/0430 https://doi.org/10.1016/j.palaeo.2015.12.010 https://doi.org/10.1111/tgis.12187 https://doi.org/10.3112/erdkunde.2005.03.01 https://doi.org/10.22498/pages.8.3.11 https://doi.org/10.22498/pages.15.1.17 https://doi.org/10.1098/rstb.2019.0714 https://www.ncbi.nlm.nih.gov/pubmed/33250025 https://doi.org/10.1023/A:1021970021670 https://doi.org/10.1016/j.appet.2005.01.011 https://www.ncbi.nlm.nih.gov/pubmed/15950317 https://doi.org/10.22498/pages.26.1.8 https://doi.org/10.2307/2694148 https://doi.org/10.1016/j.ancene.2015.05.001 https://doi.org/10.1016/j.jenvman.2004.11.004 Land 2024, 13, 1144 21 of 22 46. Munro, N.D.; Bar-Oz, G.; Meier, J.S.; Sapir-Hen, L.; Stiner, M.C.; Yeshurun, R. The Emergence of Animal Management in the Southern Levant. Sci. Rep. 2018, 8, 9279. [CrossRef] 47. Ramankutty, N.; Foley, J.A. Estimating Historical Changes in Global Land Cover: Croplands from 1700 to 1992. Glob. Biogeochem. Cycles 1999, 13, 997–1027. [CrossRef] 48. Ellis, E.C.; Beusen, A.H.W.; Goldewijk, K.K. Anthropogenic Biomes: 10,000 BCE to 2015 CE. Land 2020, 9, 129. [CrossRef] 49. Cropper, M.; Griffiths, C. The Interaction of Population Growth and Environmental Quality. Am. Econ. Rev. 1994, 84, 250–254. 50. Weiberg, E.; Hughes, R.E.; Finné, M.; Bonnier, A.; Kaplan, J.O. Mediterranean Land Use Systems from Prehistory to Antiquity: A Case Study from Peloponnese (Greece). J. Land Use Sci. 2019, 14, 1–20. [CrossRef] 51. Wills, D.M.; Whipple, C.J.; Takuno, S.; Kursel, L.E.; Shannon, L.M.; Ross-Ibarra, J.; Doebley, J.F. From Many, One: Genetic Control of Prolificacy during Maize Domestication. PLoS Genet. 2013, 9, e1003604. [CrossRef] [PubMed] 52. Palka, J.W. Ancestral Maya Domesticated Waterscapes, Ecological Aquaculture, and Integrated Subsistence. Anc. Mesoam. 2023, 35, 208–236. [CrossRef] 53. Comptour, M.; Caillon, S.; Rodrigues, L.; McKey, D. Wetland Raised-Field Agriculture and Its Contribution to Sustainability: Eth- noecology of a Present-Day African System and Questions about Pre-Columbian Systems in the American Tropics. Sustainability 2018, 10, 3120. [CrossRef] 54. Morrison, K.D. The Intensification of Production: Archaeological Approaches. J. Archaeol. Method Theory 1994, 1, 111–159. [CrossRef] 55. Boserup, E. The Conditions of Agricultural Growth: The Economics of Agrarian Change under Population Pressure; Allen & Unwin: London, UK, 1965; ISBN 978-0-202-36387-5. 56. Stiner, M.C. Paleolithic Population Growth Pulses Evidenced by Small Animal Exploitation. Science 1999, 283, 190–194. [CrossRef] 57. Stiner, M.C.; Munro, N.D.; Surovell, T.A. The Tortoise and the Hare: Small-Game Use, the Broad-Spectrum Revolution, and Paleolithic Demography. Curr. Anthropol. 2000, 41, 39–79. [CrossRef] [PubMed] 58. Yeshurun, R.; Bar-Oz, G.; Weinstein-Evron, M. Intensification and Sedentism in the Terminal Pleistocene Natufian Sequence of El-Wad Terrace (Israel). J. Hum. Evol. 2014, 70, 16–35. [CrossRef] [PubMed] 59. Kennett, D.J.; Winterhalder, B. (Eds.) Behavioral Ecology and the Transition to Agriculture; Origins of human behavior and culture; University of California Press: Berkeley, CA, USA, 2006; ISBN 978-0-520-24647-8. 60. Redding, R.W. A General Explanation of Subsistence Change: From Hunting and Gathering to Food Production. J. Anthropol. Archaeol. 1988, 7, 56–97. [CrossRef] 61. Sherratt, A. Plough and Pastoralism: Aspects of the Secondary Products Revolution; Cambridge University Press: Cambridge, UK, 1981. 62. Marchant, R.; Richer, S.; Boles, O.; Capitani, C.; Courtney-Mustaphi, C.J.; Lane, P.; Prendergast, M.E.; Stump, D.; De Cort, G.; Kaplan, J.O.; et al. Drivers and Trajectories of Land Cover Change in East Africa: Human and Environmental Interactions from 6000 years Ago to Present. Earth Sci. Rev. 2018, 178, 322–378. [CrossRef] 63. Stephens, L.; Fuller, D.; Boivin, N.; Rick, T.; Gauthier, N.; Kay, A.; Marwick, B.; Armstrong, C.G.; Barton, C.M.; Denham, T.; et al. Archaeological Assessment Reveals Earth’s Early Transformation through Land Use. Science 2019, 365, 897–902. [CrossRef] [PubMed] 64. Greenfield, H.J. The Secondary Products Revolution: The Past, the Present and the Future. World Archaeol. 2010, 42, 29–54. [CrossRef] 65. Hill, A.C. Specialized Pastoralism and Social Stratification—Analysis of the Fauna from Chalcolithic Tel Tsaf, Israel. Ph.D. Thesis, University of Connecticut, Mansfield, CT, USA, 2011. 66. Marciniak, A. The Secondary Products Revolution: Empirical Evidence and Its Current Zooarchaeological Critique. J. World Prehistory 2011, 24, 117–130. [CrossRef] 67. Bishop, R.R.; Gröcke, D.R.; Ralston, I.; Clarke, D.; Lee, D.H.J.; Shepherd, A.; Thomas, A.S.; Rowley-Conwy, P.A.; Church, M.J. Scotland’s First Farmers: New Insights into Early Farming Practices in North-West Europe. Antiquity 2022, 96, 1087–1104. [CrossRef] 68. Scharl, S.; Zerl, T.; Eckmeier, E.; Gerlach, R. Earliest Archeological Evidence of Fertilization in Central Europe. J. Plant Nutr. Soil Sci. 2023, 186, 375–382. [CrossRef] 69. Hart, J.P.; Winchell-Sweeney, S. Resetting Archaeological Interpretations of Precontact Indigenous Agriculture: Maize Isotopic Evidence from Three Ancestral Mohawk Iroquoian Villages. Am. Antiq. 2023, 88, 497–512. [CrossRef] 70. Kuijt, I. What Do We Really Know about Food Storage, Surplus, and Feasting in Preagricultural Communities? Curr. Anthropol. 2009, 50, 641–644. [CrossRef] [PubMed] 71. Bogaard, A. The Archaeology of Food Surplus. World Archaeol. 2017, 49, 1–7. [CrossRef] 72. Frangipane, M. From a Subsistence Economy to the Production of Wealth in Ancient Formative Societies: A Political Economy Perspective. Econ. Polit. 2018, 35, 677–689. [CrossRef] 73. Prats, G.; Antolín, F.; Alonso, N. From the Earliest Farmers to the First Urban Centres: A Socio-Economic Analysis of Underground Storage Practices in North-Eastern Iberia. Antiquity 2020, 94, 653–668. [CrossRef] 74. Wilkinson, T.; Miller, N.; Reichel, C.; Whitcomb, D. On the Margin of the Euphrates: Settlement and Land Use at Tell Es-Sweyhat and in the Upper Lake Assad Area, Syria; Excavations at Tell es-Sweyhat, Syria; Oriental Institute of the University of Chicago: Chicago, IL, USA, 2004; Volume 1. https://doi.org/10.1038/s41598-018-27647-z https://doi.org/10.1029/1999GB900046 https://doi.org/10.3390/land9050129 https://doi.org/10.1080/1747423X.2019.1639836 https://doi.org/10.1371/journal.pgen.1003604 https://www.ncbi.nlm.nih.gov/pubmed/23825971 https://doi.org/10.1017/S0956536122000402 https://doi.org/10.3390/su10093120 https://doi.org/10.1007/BF02231414 https://doi.org/10.1126/science.283.5399.190 https://doi.org/10.1086/300102 https://www.ncbi.nlm.nih.gov/pubmed/10593724 https://doi.org/10.1016/j.jhevol.2014.02.011 https://www.ncbi.nlm.nih.gov/pubmed/24661906 https://doi.org/10.1016/0278-4165(88)90007-4 https://doi.org/10.1016/j.earscirev.2017.12.010 https://doi.org/10.1126/science.aax1192 https://www.ncbi.nlm.nih.gov/pubmed/31467217 https://doi.org/10.1080/00438240903429722 https://doi.org/10.1007/s10963-011-9045-7 https://doi.org/10.15184/aqy.2022.107 https://doi.org/10.1002/jpln.202300150 https://doi.org/10.1017/aaq.2023.44 https://doi.org/10.1086/605082 https://www.ncbi.nlm.nih.gov/pubmed/20642151 https://doi.org/10.1080/00438243.2017.1294105 https://doi.org/10.1007/s40888-018-0133-3 https://doi.org/10.15184/aqy.2019.153 Land 2024, 13, 1144 22 of 22 75. Wilkinson, T.J.; Gibson, M.; Widell, M. Models of Mesopotamian Landscapes: How Small-Scale Processes Contributed to the Growth of Early Civilizations; Archaeopress: Almería, Spain, 2013. 76. Akkermans, P.M. Villages in the Steppe–Later Neolithic Settlement and Subsistence in the Balikh; Berghahn Books: New York, NY, USA, 1993. 77. Butzer, K.W. Early Hydraulic Civilization in Egypt: A Study in Cultural Ecology; Prehistoric archeology and ecology; University of Chicago Press: Chicago, IL, USA, 1976; ISBN 978-0-226-08634-7. 78. Kaplan, J.O.; Krumhardt, K.M.; Gaillard, M.-J.; Sugita, S.; Trondman, A.-K.; Fyfe, R.; Marquer, L.; Mazier, F.; Nielsen, A.B. Constraining the Deforestation History of Europe: Evaluation of Historical Land Use Scenarios with Pollen-Based Land Cover Reconstructions. Land 2017, 6, 91. [CrossRef] 79. Gaillard, M.-J. LandCover6k: Global Anthropogenic Land-Cover Change and Its Role in Past Climate. Pages Mag. 2015, 23, 38–39. [CrossRef] 80. Harrison, S.P.; Gaillard, M.-J.; Stocker, B.D.; Vander Linden, M.; Klein Goldewijk, K.; Boles, O.; Braconnot, P.; Dawson, A.; Fluet-Chouinard, E.; Kaplan, J.O.; et al. Development and Testing Scenarios for Implementing Land Use and Land Cover Changes during the Holocene in Earth System Model Experiments. Geosci. Model Dev. 2020, 13, 805–824. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. https://doi.org/10.3390/land6040091 https://doi.org/10.22498/pages.23.1.38 https://doi.org/10.5194/gmd-13-805-2020 Introduction Materials and Methods Data Sources Data Structure Distribution of Values Results Variation among Subsistence Strategies Variation among LU2 Categories Geographic Distribution Environmental Distribution Subnational Regions Temporal Variation Discussion Population and Land Use Increased Nuance within LU1 Categories Potential Applications Impacts on HYDE Alternate Approaches Conclusions References