Significance: This paper illustrates the potential of machine learning-based classification of multisensor, multitemporal satellite data for the remote detection and mapping of archaeological mounded settlements in arid environments. Our research integrates multitemporal synthetic-aperture radar and multispectral bands to produce a highly accurate probability field of mound signatures. The results largely expand the known concentration of Indus settlements in the Cholistan Desert in Pakistan ( ca . 3300 to 1500 BC), with the detection of hundreds of new sites deeper in the desert than previously suspected including several large-sized (>30 ha) urban centers. These distribution patterns have major implications regarding the influence of climate change and desertification in the collapse of the largest of the Old-World Bronze Age civilizations.

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.

A range of data sources are now used to support the process of archaeological prospection, including remote sensed imagery, spy satellite photographs and aerial photographs. This paper advocates the value and importance of a hitherto under-utilised historical mapping resource—the Survey of India 1” to 1-mile map series, which was based on surveys started in the mid–late nineteenth century, and published progressively from the early twentieth century AD. These maps present a systematic documentation of the topography of the British dominions in the South Asian Subcontinent. Incidentally, they also documented the locations, the height and area of thousands of elevated mounds that were visible in the landscape at the time that the surveys were carried out, but have typically since been either damaged or destroyed by the expansion of irrigation agriculture and urbanism. Subsequent reanalysis has revealed that many of these mounds were actually the remains of ancient settlements. The digitisation and analysis of these historic maps thus creates a unique opportunity for gaining insight into the landscape archaeology of South Asia. This paper reviews the context within which these historical maps were created, presents a method for georeferencing them, and reviews the symbology that was used to represent elevated mound features that have the potential to be archaeological sites. This paper should be read in conjunction with the paper by Arnau Garcia et al. in the same issue of Geosciences, which implements a research programme combining historical maps and a range of remote sensing approaches to reconstruct historical landscape dynamics in the Indus River Basin.