The results can be sumarised in terms of chronology, construction and crops. In terms of chronology our dating approach reveals that our set of terraces span at least 4,000 years of human history. Our oldest terraces are of Minoan age c. 4,200 years ago (Choiromandres, Eastern Crete) and mid-Bronze age (4,000-3,500 years old) in northern England (Plantation Camp, Cheviot Hills, Brown et al. 2023). We also have terraces that date to the Iron Age (3,000-2,000 years old, Sint Martens Voeren, Belgium) and more terraces that date to the Roman and Medieval periods (Castronovo, Sicily; Soave, Italy). Some sites appear to have had only one major period of us, such as the Choiromandres terraces, but most had long-periods of episodic use up to modern times. Using a combination of OSL, 14C AMS and HyPy 14C AMS we managed to get a chronology for all the sites. However, at one site, Guesswick, Co. Durham UK both OSL and HyPy 14C provided us with multiple dates that were far too old for terrace sediments. This also occurred with ridge and furrow sediments and we believe that it is due to a last ploughing event that was the deepest and brought up material from the base of the soil profiles. Further analysis is underway on this result, and the chronology of all the sites which are being integrated with a Global Terrace database. The chronologies of our terraces have fed directly into the assessment of their role in carbon capture and storage (Zhao et al., 2021, 2022). This work suggests due to a combination of burial and thermal stabilisation, terraces on average increase carbon storage by about 20% in comparison to unterraced slopes (Zhao et al, subm.). Because of the variable performance of the biological proxies for crops and biodiversity, no single method is applicable to all terraces. Most surprisingly the method that worked with most sites was sedaDNA (12 out of 14). From the combination of the pollen phytoliths, charcoal and sedaDNA we can show that a variety of first crops were grown on the terraces. The phytoliths and pollen confirmed that vines were the crop on the Minoan terraces (Choiromandres) and also on some of the terraces in the Soave district (NE Italy). Our data suggests that the Norwegian sites had cereals on them, especially barley, but the latest at Bjornskinn was likely created for potato cultivation in the 18th century CE. The unexpected finding of sedaDNA of cod at sites 200m above sea level accords with records from these ‘fjord farms’ of the application of fish remains as a fertilizer. We also have several examples of undoubted polyculture with the site at Castronovo in Central Sicily recording over 200 taxa including 40 crop plants in the 13th century CE (Brown et al. Subm.). These results uncovered some remarkable historical details, such the variety of crops cultivated in region of Soave in north-eastern Italy, which is famous for its viticultural terraces. While it is commonly assumed that these terraces were built to cultivate grapes for wine production, with Soave wine reputedly the favourite of Emperor Augustus, the sedaDNA evidence suggests they have also been used for different purposes at other periods. In particular the terraces which date from around the turn of the 16th century shows sedaDNA evidence of cannabis (hemp) cultivation. At this time Venice, in order to secure its fleet, had declared a monopoly on hemp production. This illustrates the flexibility of terraces, which opens up the possibility of using them to meet the demand for specialist crops – although maybe not cannabis - as opposed to importing produce from far-flung locations. The unforeseen success of the sedaDNA component of the project is being further investigated, but the data from Castronovo in Sicily suggests that high sedaDNA retention is associated with expansive clays in the soil along with a lack of mixing in a system with high biodiversity.
