Virtual water trade at the sub-national scale in South America: the role of Transnational Corporations

The production of agricultural commodities requires a massive use of freshwater resources. However, given the worldwide trade of these commodities, water resources available and used in the production sites are virtually displaced to the importing countries, by means of trading companies, addressed as Transnational Corporations (TNCs). It is therefore pivotal to uncover the role played by companies in the international redistribution of water to enhance water stewardship, a concept related to social, environmental and economic sustainability in water use. Recent research started to emphasise the importance of tracking food commodities flows at the sub-national level, where the environmental and climatic heterogeneities of producing regions are directly related to variations in the unit water footprint (uWF, [m3/ton]) of cultivated crops. This thesis aims to unveil the sub-national spatial and temporal variability of unit water footprints associated with major Transnational Corporations. Specifically, the analysis is focused on traded cocoa, coffee, corn, cotton and soy, from the producing countries of Argentina, Bolivia, Brazil, Colombia, Ivory Coast and Paraguay to major importing countries, through trading companies controlling at least 80% of each analysed market. Evapotranspiration data (ET, [mm]) are estimated by means of the agro-ecological model waterCROP at the cell-grid level (5 arcmin spatial resolution) and they are coupled with sub-national trade data from the Trase database, which provides data on single trade flows ([tonnes]) from the production sites to the final importing country, through exporter and importer companies. Results on the traded virtual water volumes, and related uWFs, demonstrate the complexity of assessing how each company effectively relates to the water resource. Whenever neglecting the spatial heterogeneity of producing sites, upscaled average values lead to inaccurate assessments of resource use. Indeed, the sub-national level of the analysis enables us to uncover how the unit water footprint of a given trader, fixed the commodity, varies according to the producing country and the final importer. For instance, in 2017 Cargill soybean export had a unitary water requirement greater in Brazil than in Paraguay (1552 m3/ton and 1338 m3/ton, respectively), conversely to what found for corn (616 m3/ton and 831 m3/ton, respectively). Always in 2017, Cargill soybean exports from Brazil to China and to Germany showed relevant differences (1562 m3/ton and 1719 m3/ton, being respectively the largest and smallest importers in terms of tonnes). The heterogeneity in volumes and their dependence on the specific geography of production at the fine-scale indicate the importance of providing traders and importing countries with detailed information about the uWF associated with trade flows. Changes in demand can influence how traders source food commodities at the local level, leading to a more conscious use of water resources.