Automated detection of archaeological mounds using machine-learning classification of multisensor and multitemporal satellite data.

This paper presents an innovative multisensor, multitemporal machine-learning approach using remote sensing big data for the detection of archaeological mounds in Cholistan (Pakistan). The Cholistan Desert presents one of the largest concentrations of Indus Civilization sites (from ca 3300 to 1500 BC). Cholistan has figured prominently in theories about changes in water availability, the rise and decline of the Indus Civilization, and the transformation of fertile monsoonal alluvial plains into an extremely arid margin. This paper implements a multisensor, multitemporal machine-learning approach for the remote detection of archaeological mounds. A classifier algorithm that employs a large-scale collection of synthetic-aperture radar and multispectral images has been implemented in Google Earth Engine, resulting in an accurate probability map for mound-like signatures across an area that covers ca 36,000 km2 The results show that the area presents many more archaeological mounds than previously recorded, extending south and east into the desert, which has major implications for understanding the archaeological significance of the region. The detection of small (<5 ha) to large mounds (>30 ha) suggests that there were continuous shifts in settlement location. These shifts are likely to reflect responses to a dynamic and changing hydrological network and the influence of the progressive northward advance of the desert in a long-term process that culminated in the abandonment of much of the settled area during the Late Harappan period.

Carbon and nitrogen isotopic variability in foxtail millet (Setaria italica) with watering regime.

RATIONALE: Carbonised plant remains are analysed for reconstruction of past climates and agricultural regimes. Several recent studies have used C4 plants to address related questions, and correlations between modern C4 plant δ13 C values and rainfall have been found. The millets were important food crops in prehistoric Eurasia, yet little is known about causes of isotopic variation within millet species. Previous research has shown there to be significant isotopic variation between millet accessions. Here we compare isotope ratios from plants grown under different watering regimes. This allows for a consideration of whether or not Setaria italica is a good proxy for environmental reconstruction. METHODS: We compare stable isotope ratios of Setaria italica plants grown in a controlled environment chamber with different watering regimes. We compare the carbon isotope ratios of leaves and grains, and the nitrogen isotope ratios of grains, from 12 accessions of Setaria italica. RESULTS: We find significant isotopic variability between watering regimes. Carbon isotope ratios are positively correlated with water availability, and on average vary by 1.9‰ and 1.7‰ for leaves and grains, respectively. Grain nitrogen isotope ratios also vary with watering regime; however, the highest isotope ratios are found with the 130-mL watering regime. CONCLUSIONS: The carbon isotope ratios of Setaria italica are strongly correlated with water availability. However, the correlation is the opposite to that seen in studies of C3 plants. The difference in isotopic ratio due to watering regime is comparable with that seen between different accessions; thus distinguishing between changing varieties of Setaria italica and changing climate is problematic. In terms of grain nitrogen isotope ratios, the highest δ15 N values were not associated with the lowest watering regime. Again, δ15 N variation is comparable with that which would be expected from an aridity effect or a manuring effect, and thus distinguishing between these factors is probably problematic.

Curriculum learning-based strategy for low-density archaeological mound detection from historical maps in India and Pakistan.

This paper presents two algorithms for the large-scale automatic detection and instance segmentation of potential archaeological mounds on historical maps. Historical maps present a unique source of information for the reconstruction of ancient landscapes. The last 100 years have seen unprecedented landscape modifications with the introduction and large-scale implementation of mechanised agriculture, channel-based irrigation schemes, and urban expansion to name but a few. Historical maps offer a window onto disappearing landscapes where many historical and archaeological elements that no longer exist today are depicted. The algorithms focus on the detection and shape extraction of mound features with high probability of being archaeological settlements, mounds being one of the most commonly documented archaeological features to be found in the Survey of India historical map series, although not necessarily recognised as such at the time of surveying. Mound features with high archaeological potential are most commonly depicted through hachures or contour-equivalent form-lines, therefore, an algorithm has been designed to detect each of those features. Our proposed approach addresses two of the most common issues in archaeological automated survey, the low-density of archaeological features to be detected, and the small amount of training data available. It has been applied to all types of maps available of the historic 1″ to 1-mile series, thus increasing the complexity of the detection. Moreover, the inclusion of synthetic data, along with a Curriculum Learning strategy, has allowed the algorithm to better understand what the mound features look like. Likewise, a series of filters based on topographic setting, form, and size have been applied to improve the accuracy of the models. The resulting algorithms have a recall value of 52.61% and a precision of 82.31% for the hachure mounds, and a recall value of 70.80% and a precision of 70.29% for the form-line mounds, which allowed the detection of nearly 6000 mound features over an area of 470,500 km2, the largest such approach to have ever been applied. If we restrict our focus to the maps most similar to those used in the algorithm training, we reach recall values greater than 60% and precision values greater than 90%. This approach has shown the potential to implement an adaptive algorithm that allows, after a small amount of retraining with data detected from a new map, a better general mound feature detection in the same map.

Were there royal herds? Understanding herd management and mobility using isotopic characterizations of cattle tooth enamel from Early Dynastic Ur.

During the third millennium BC, Mesopotamia (the land between the Tigris and Euphrates Rivers, in modern Iraq-Syria), was dominated by the world's earliest cities and states, which were ruled by powerful elites. Ur, in present-day southern Iraq, was one of the largest and most important of these cities, and irrigation-based agriculture and large herds of domesticated animals were the twin mainstays of the economy and diet. Texts suggest that the societies of the Mesopotamian city-states were extremely hierarchical and underpinned by institutionalised and heavily-managed farming systems. Prevailing narratives suggest that the animal management strategies within these farming systems in the third millennium BC were homogenous. There have been few systematic science-based studies of human and animal diets, mobility, or other forms of human-animal interaction in Mesopotamia, but such approaches can inform understanding of past economies, including animal management, social hierarchies, diet and migration. Oxygen, carbon and strontium isotopic analysis of animal tooth enamel from both royal and private/non-royal burial contexts at Early Dynastic Ur (2900-2350 BC) indicate that a variety of herd management strategies and habitats were exploited. These data also suggest that there is no correlation between animal-management practices and the cattle found in royal or private/non-royal burial contexts. The results demonstrate considerable divergence between agro-pastoral models promoted by the state and the realities of day-to-day management practices. The data from Ur suggest that the animals exploited different plant and water sources, and that animals reared in similar ways ended up in different depositional contexts.

Potential of deep learning segmentation for the extraction of archaeological features from historical map series.

Historical maps present a unique depiction of past landscapes, providing evidence for a wide range of information such as settlement distribution, past land use, natural resources, transport networks, toponymy and other natural and cultural data within an explicitly spatial context. Maps produced before the expansion of large-scale mechanized agriculture reflect a landscape that is lost today. Of particular interest to us is the great quantity of archaeologically relevant information that these maps recorded, both deliberately and incidentally. Despite the importance of the information they contain, researchers have only recently begun to automatically digitize and extract data from such maps as coherent information, rather than manually examine a raster image. However, these new approaches have focused on specific types of information that cannot be used directly for archaeological or heritage purposes. This paper provides a proof of concept of the application of deep learning techniques to extract archaeological information from historical maps in an automated manner. Early twentieth century colonial map series have been chosen, as they provide enough time depth to avoid many recent large-scale landscape modifications and cover very large areas (comprising several countries). The use of common symbology and conventions enhance the applicability of the method. The results show deep learning to be an efficient tool for the recovery of georeferenced, archaeologically relevant information that is represented as conventional signs, line-drawings and text in historical maps. The method can provide excellent results when an adequate training dataset has been gathered and is therefore at its best when applied to the large map series that can supply such information. The deep learning approaches described here open up the possibility to map sites and features across entire map series much more quickly and coherently than other available methods, opening up the potential to reconstruct archaeological landscapes at continental scales.

Lipid residues in pottery from the Indus Civilisation in northwest India.

This paper presents novel insights into the archaeology of food in ancient South Asia by using lipid residue analysis to investigate what kinds of foodstuffs were used in ceramic vessels by populations of the Indus Civilisation in northwest India. It examines how vessels were used in urban and rural Indus settlements during the Mature Harappan period (c.2600/2500-1900 BC), the relationship between vessels and the products within them, and identifies whether changes in vessel use occurred from the Mature Harappan to Late Harappan periods, particularly during climatic instability after 4.2 ka BP (c.2100 BC). Despite low lipid concentrations, which highlight challenges with conducting residue analysis in arid, seasonally-wet and alkaline environments, 71% of the vessels yielded appreciable quantities of lipid. Lipid profiles revealed the use of animal fats in vessels, and contradictory to faunal evidence, a dominance of non-ruminant fats, with limited evidence of dairy processing. The absence of local modern reference fats makes this dataset challenging to interpret, and it is possible that plant products or mixtures of plant and animal products have led to ambiguous fatty acid-specific isotopic values. At the same time, it appears that urban and rural populations processed similar types of products in vessels, with limited evidence for change in vessel use from the urban to the post-urban period. This study is a systematic investigation into pot lipid residues from multiple sites, demonstrating the potential of the method for examining ancient Indus foodways and the need for the development of further research in ancient organic residues in South Asia.

Tracking the Near Eastern origins and European dispersal of the western house mouse.

The house mouse (Mus musculus) represents the extreme of globalization of invasive mammals. However, the timing and basis of its origin and early phases of dispersal remain poorly documented. To track its synanthropisation and subsequent invasive spread during the develoment of complex human societies, we analyzed 829 Mus specimens from 43 archaeological contexts in Southwestern Asia and Southeastern Europe, between 40,000 and 3,000 cal. BP, combining geometric morphometrics numerical taxonomy, ancient mitochondrial DNA and direct radiocarbon dating. We found that large late hunter-gatherer sedentary settlements in the Levant, c. 14,500 cal. BP, promoted the commensal behaviour of the house mouse, which probably led the commensal pathway to cat domestication. House mouse invasive spread was then fostered through the emergence of agriculture throughout the Near East 12,000 years ago. Stowaway transport of house mice to Cyprus can be inferred as early as 10,800 years ago. However, the house mouse invasion of Europe did not happen until the development of proto urbanism and exchange networks - 6,500 years ago in Eastern Europe and 4000 years ago in Southern Europe - which in turn may have driven the first human mediated dispersal of cats in Europe.

'We are inheritors of a rural civilisation': rural complexity and the ceramic economy in the Indus Civilisation in northwest India.

What role do villages play in 'an urban civilisation'? Although it is likely that most of the populations of South Asia's ancient Indus Civilisation would not have lived in cities, it is not clear what their rural way of life would have encompassed. Using ceramic assemblages excavated from Indus-period villages in northwest India, alongside ethnographic records on village organization and rural craft production, this paper argues that Indus villages were characterized by rural complexity. This comprised a range of activities, including craft production, as well as short- and long-distance socio-economic links. Drawing on historical narratives, we show how South Asian villages have been essentialized and presented as either ideal or conservative extremes. We argue for the importance of a better understanding of the breadth and nuances of the rural sphere, and for a greater research focus on village life in the Indus context.

Revisiting Settlement Contemporaneity and Exploring Stability and Instability: Case Studies from the Indus Civilization.

"Map overestimation," or "the contemporaneity problem," derives from the assumption that settlements identified during surface surveys were occupied throughout individual periods. Inductive and simulation analysis have been used to ascertain the degree of contemporaneity in surface survey data sets, as variation in settlement location is critical for understanding population density and demography, which inform social, economic and political interpretations. This paper revisits the inductive approach to interrogating survey data developed by W. M. Sumner and the simulation model approach developed by R. E. Dewar to explore the survey data from two regions within South Asia's Indus civilization. This analysis demonstrates the strengths and weaknesses of these approaches. It also highlights the variability in settlement systems in different areas within the Indus civilization and shows that consideration of stability and instability within settlement systems is an important factor when considering dynamics of resilience and sustainability.

20,000 years of societal vulnerability and adaptation to climate change in southwest Asia.

The Fertile Crescent, its hilly flanks and surrounding drylands has been a critical region for studying how climate has influenced societal change, and this review focuses on the region over the last 20,000 years. The complex social, economic, and environmental landscapes in the region today are not new phenomena and understanding their interactions requires a nuanced, multidisciplinary understanding of the past. This review builds on a history of collaboration between the social and natural palaeoscience disciplines. We provide a multidisciplinary, multiscalar perspective on the relevance of past climate, environmental, and archaeological research in assessing present day vulnerabilities and risks for the populations of southwest Asia. We discuss the complexity of palaeoclimatic data interpretation, particularly in relation to hydrology, and provide an overview of key time periods of palaeoclimatic interest. We discuss the critical role that vegetation plays in the human-climate-environment nexus and discuss the implications of the available palaeoclimate and archaeological data, and their interpretation, for palaeonarratives of the region, both climatically and socially. We also provide an overview of how modelling can improve our understanding of past climate impacts and associated change in risk to societies. We conclude by looking to future work, and identify themes of "scale" and "seasonality" as still requiring further focus. We suggest that by appreciating a given locale's place in the regional hydroscape, be it an archaeological site or palaeoenvironmental archive, more robust links to climate can be made where appropriate and interpretations drawn will demand the resolution of factors acting across multiple scales. This article is categorized under:Human Water > Water as Imagined and RepresentedScience of Water > Water and Environmental ChangeWater and Life > Nature of Freshwater Ecosystems.

Multi-scale relief model (MSRM): a new algorithm for the visualization of subtle topographic change of variable size in digital elevation models.

Morphological analysis of landforms has traditionally relied on the interpretation of imagery. Although imagery provides a natural view of an area of interest (AOI) images are largely hindered by the environmental conditions at the time of image acquisition, the quality of the image and, mainly, the lack of topographical information, which is an essential factor for a correct understanding of the AOI's geomorphology. More recently digital surface models (DSMs) have been incorporated into the analytical toolbox of geomorphologists. These are usually high-resolution models derived from digital photogrammetric processes or LiDAR data. However, these are restricted to relatively small areas and are expensive or complex to acquire, which limits widespread implementation. In this paper, we present the multi-scale relief model (MSRM), which is a new algorithm for the visual interpretation of landforms using DSMs. The significance of this new method lies in its capacity to extract landform morphology from both high- and low-resolution DSMs independently of the shape or scale of the landform under study. This method thus provides important advantages compared to previous approaches as it: (1) allows the use of worldwide medium resolution models, such as SRTM, ASTER GDEM, ALOS, and TanDEM-X; (2) offers an alternative to traditional photograph interpretation that does not rely on the quality of the imagery employed nor on the environmental conditions and time of its acquisition; and (3) can be easily implemented for large areas using traditional GIS/RS software. The algorithm is tested in the Sutlej-Yamuna interfluve, which is a very large low-relief alluvial plain in northwest India where 10 000 km of palaeoriver channels have been mapped using MSRM. The code, written in Google Earth Engine's implementation of JavaScript, is provided as Supporting Information for its use in any other AOI without particular technical knowledge or access to topographical data. © 2017 The Authors. Earth Surface Processes and Landforms published by John Wiley & Sons Ltd.

Landscapes of Urbanization and De-Urbanization: A Large-Scale Approach to Investigating the Indus Civilization's Settlement Distributions in Northwest India.

Survey data play a fundamental role in studies of social complexity. Integrating the results from multiple projects into large-scale analyses encourages the reconsideration of existing interpretations. This approach is essential to understanding changes in the Indus Civilization's settlement distributions (ca. 2600-1600 b.c.), which shift from numerous small-scale settlements and a small number of larger urban centers to a de-nucleated pattern of settlement. This paper examines the interpretation that northwest India's settlement density increased as Indus cities declined by developing an integrated site location database and using this pilot database to conduct large-scale geographical information systems (GIS) analyses. It finds that settlement density in northwestern India may have increased in particular areas after ca. 1900 b.c., and that the resulting landscape of de-urbanization may have emerged at the expense of other processes. Investigating the Indus Civilization's landscapes has the potential to reveal broader dynamics of social complexity across extensive and varied environments.

Barley heads east: Genetic analyses reveal routes of spread through diverse Eurasian landscapes.

One of the world's most important crops, barley, was domesticated in the Near East around 11,000 years ago. Barley is a highly resilient crop, able to grown in varied and marginal environments, such as in regions of high altitude and latitude. Archaeobotanical evidence shows that barley had spread throughout Eurasia by 2,000 BC. To further elucidate the routes by which barley cultivation was spread through Eurasia, simple sequence repeat (SSR) analysis was used to determine genetic diversity and population structure in three extant barley taxa: domesticated barley (Hordeum vulgare L. subsp. vulgare), wild barley (H. vulgare subsp. spontaneum) and a six-rowed brittle rachis form (H. vulgare subsp. vulgare f. agriocrithon (Åberg) Bowd.). Analysis of data using the Bayesian clustering algorithm InStruct suggests a model with three ancestral genepools, which captures a major split in the data, with substantial additional resolution provided under a model with eight genepools. Our results indicate that H. vulgare subsp. vulgare f. agriocrithon accessions and Tibetan Plateau H. vulgare subsp. spontaneum are closely related to the H. vulgare subsp. vulgare in their vicinity, and are therefore likely to be feral derivatives of H. vulgare subsp. vulgare. Under the eight genepool model, cultivated barley is split into six ancestral genepools, each of which has a distinct distribution through Eurasia, along with distinct morphological features and flowering time phenotypes. The distribution of these genepools and their phenotypic characteristics is discussed together with archaeological evidence for the spread of barley eastwards across Eurasia.

Author Correction: Intensified summer monsoon and the urbanization of Indus Civilization in northwest India.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has not been fixed in the paper.

Intensified summer monsoon and the urbanization of Indus Civilization in northwest India.

Today the desert margins of northwest India are dry and unable to support large populations, but were densely occupied by the populations of the Indus Civilization during the middle to late Holocene. The hydroclimatic conditions under which Indus urbanization took place, which was marked by a period of expanded settlement into the Thar Desert margins, remains poorly understood. We measured the isotopic values (δ18O and δD) of gypsum hydration water in paleolake Karsandi sediments in northern Rajasthan to infer past changes in lake hydrology, which is sensitive to changing amounts of precipitation and evaporation. Our record reveals that relatively wet conditions prevailed at the northern edge of Rajasthan from ~5.1 ± 0.2 ka BP, during the beginning of the agricultural-based Early Harappan phase of the Indus Civilization. Monsoon rainfall intensified further between 5.0 and 4.4 ka BP, during the period when Indus urban centres developed in the western Thar Desert margin and on the plains of Haryana to its north. Drier conditions set in sometime after 4.4 ka BP, and by ~3.9 ka BP an eastward shift of populations had occurred. Our findings provide evidence that climate change was associated with both the expansion and contraction of Indus urbanism along the desert margin in northwest India.

Correction: Journey to the east: Diverse routes and variable flowering times for wheat and barley en route to prehistoric China.

[This corrects the article DOI: 10.1371/journal.pone.0187405.].

Exploring Indus crop processing: combining phytolith and macrobotanical analyses to consider the organisation of agriculture in northwest India c. 3200-1500 bc.

This paper presents a preliminary study combining macrobotanical and phytolith analyses to explore crop processing at archaeological sites in Haryana and Rajasthan, northwest India. Current understanding of the agricultural strategies in use by populations associated with South Asia's Indus Civilisation (3200-1900 bc) has been derived from a small number of systematic macrobotanical studies focusing on a small number of sites, with little use of multi-proxy analysis. In this study both phytolith and macrobotanical analyses are used to explore the organisation of crop processing at five small Indus settlements with a view to understanding the impact of urban development and decline on village agriculture. The differing preservation potential of the two proxies has allowed for greater insights into the different stages of processing represented at these sites: with macrobotanical remains allowing for more species-level specific analysis, though due to poor chaff presentation the early stages of processing were missed; however these early stages of processing were evident in the less highly resolved but better preserved phytolith remains. The combined analyses suggests that crop processing aims and organisation differed according to the season of cereal growth, contrary to current models of Indus Civilisation labour organisation that suggest change over time. The study shows that the agricultural strategies of these frequently overlooked smaller sites question the simplistic models that have traditionally been assumed for the time period, and that both multi-proxy analysis and rural settlements are deserving of further exploration.

Journey to the east: Diverse routes and variable flowering times for wheat and barley en route to prehistoric China.

Today, farmers in many regions of eastern Asia sow their barley grains in the spring and harvest them in the autumn of the same year (spring barley). However, when it was first domesticated in southwest Asia, barley was grown between the autumn and subsequent spring (winter barley), to complete their life cycles before the summer drought. The question of when the eastern barley shifted from the original winter habit to flexible growing schedules is of significance in terms of understanding its spread. This article investigates when barley cultivation dispersed from southwest Asia to regions of eastern Asia and how the eastern spring barley evolved in this context. We report 70 new radiocarbon measurements obtained directly from barley grains recovered from archaeological sites in eastern Eurasia. Our results indicate that the eastern dispersals of wheat and barley were distinct in both space and time. We infer that barley had been cultivated in a range of markedly contrasting environments by the second millennium BC. In this context, we consider the distribution of known haplotypes of a flowering-time gene in barley, Ppd-H1, and infer that the distributions of those haplotypes may reflect the early dispersal of barley. These patterns of dispersal resonate with the second and first millennia BC textual records documenting sowing and harvesting times for barley in central/eastern China.