Geophysical survey methods have led to high-resolution mapping of subsurface remnants of ancient landscapes at continually expanding spatial scales. Yet, particularly when applied across entire archaeological landscapes, spanning hundreds of hectares, resultant datasets provide little direct information about inhabitation, environments or change over time.Focusing on a 2.5 km2 area around Stonehenge, we show how geophysical soil survey, when combined with targeted sampling and excavation, can enable reliable empirically-grounded identification of complex activity traces. Particular focus lies on anthropogenic dug pits, identification and interpretation of which are vital in European earlier prehistoric archaeology due to their close connection with inhabitation and ceremonial practices. By integrating frequency domain electromagnetic and invasive datasets, and using a semi-automated interpretation scheme, we identified previously unknown concentrations of large pits (with diameters >2.4 m) among several thousand smaller pits and natural features across the Stonehenge landscape. Excavations of a subset of identified features demonstrate that, in this area, our investigative methodology is 66% accurate for identifying large anthropogenic pits. Our results have significant implications for understanding Stonehenge and its landscape setting, revealing elusive forms of Mesolithic to later Bronze Age land use that - even within the world's most intensively researched archaeological landscape - have gone unrecognized until now.These findings underscore both the crucial role of archaeological excavation as an essential basis for reliable interpretation of geophysical data, as well as the perils of inductive visual interpretation of features’ morphologies and their spatial configuration in non-invasive survey data.