Life cycle assessment of Italian dairy cheese chain

Dairy sector is growing fast, contributing to important share of global and national economic sector and bringing nutrient components into human diets. However, dairy sector is one of the main contributors to environmental impacts arising from food sector. Cheese sector is strategic to Italian dairy sector and economy, exporting high quality and Protected Designation of Origin (PDO) cheese to international countries. There is urgent need to increase the sustainability of dairy sector, considering the whole dairy chain perspective. The aim of this Ph.D. Project was to assess the environmental impacts associated to the Italian dairy cheese chain. The Project has been conducted in Veneto Region (northeast Italy). Life Cycle Assessment (LCA) methodology have been applied to assess the impacts producing raw milk at farm gate (Manuscript I), Asiago PDO cheese from cradle to dairy plant gate (Manuscript II) and Mozzarella from cradle to grave (Manuscript III). In the Project, the primary data have been collected through direct interview to dairy farmers and visits to each dairy farm, and direct interview and visits to Asiago and Mozzarella cheese plants. Primary data collection regarded 34 dairy farms and an Asiago and Mozzarella cheese dairy plants. Ecoinvent® v3 and Agrifoodprint® v1 databases have been use for secondary data, while data from literature and national inventories have been use to model the post plant phases (distribution, retail, consumption and disposal) in the Manuscript III. Simapro© 8 was the modelling software. The impacts estimated affect human health, ecosystem and resources use. Manuscript I and Manuscript II represent the LCA of Asiago PDO cheese production. Indeed, the 34 dairy farms analyzed are located in the adjacent area of the Asiago plant, creating a unique narrow chain for producing Asiago PDO cheese which is manufactured by the PDO guidelines. In the Manuscript I, the functional unit was one kg of milk, and it was the first LCA on milk production in Veneto Region. The production of purchased feed and on-farm feed (which require land, water, chemical fertilizers and manure, machinery use) and animal emissions (enteric methane, and nitrogen emissions from manure management) were main hotspots for overall impact categories, such as climate change, acidification, eutrophication, water and land use, and energy usage. Minor contribution to final impacts originated from electricity, fuels and detergents use, and waste produced during farm activities (such milking, cleaning). Different allocation approaches were tested beside the biological default allocation, and all approaches modified the final results per kg of milk. The results were similar to those reported in literature. In the Manuscript II, the LCA model represents the production chain to produce one kg of Asiago PDO cheese, ready to sell, at dairy plant gate. The raw milk production represented the main contributor to all impact categories, except for ozone depletion where the cheese-making process was the first driver. Excluding farm phase from the assessment the manufacturing operations resulted the hotspots for overall impacts, except for eutrophication and water depletion mainly caused by wastewater treatment, and land occupation which was occurred due to primary and secondary paper packaging. The main contributor inside the cheese plant were electricity and natural gas usage, and process water, moreover transport of raw milk from farm to cheese plant impacted toxicity and photochemical oxidant formation. Economic allocation was applied and compared to milk solids content allocation, which reduced the final emissions per kg of Asiago than economic allocation. In fact, the milk solids allocation assigned more impact to the co-products (whey and other cheese) than economic allocation. Uncertainty analysis and sensitivity analysis of aging period were included into the study. In the Manuscript III the LCA methodology has been applied to an high industrialized mozzarella plant, the Italian third largest mozzarella plant. The LCA was performed in a cradle to grave perspective, including the post manufacturing phases, as distribution, retail, consumption and end of life phases. The plant used Italian and foreign milk, and distribute the mozzarella to Italy and international countries. The functional unit was one kg of mozzarella consumed. Result confirmed that raw milk production was the main contributor to overall impacts categories, except for ozone depletion where refrigerant used for cooling along the post farm chain were the main hotspot. Manufacturing and packaging were the second most important contributors to final impacts, followed by disposal of wastewater, while minor impacts were associated to distribution, retail and consumption; although relevant contribution was transport of milk and mozzarella, considering the international origin and delivery of the products. Electricity and natural gas usage, together with cardboard packaging for delivery drove the impacts during mozzarella-making process. While the impact arising from post plant phase were mainly determine by energy usage. The normalized results showed ecotoxicity, acidification, eutrophication and climate change as the main impact category contributing to the European impact and these categories are the first scope to apply strategies for reduction. A sensitivity analysis was performed to test different allocation approaches and to analyzed how the final results are influenced by allocation method; finally, a sensitivity analysis was performed determining the difference of impacts among the tradition high moisture mozzarella and the low moisture mozzarella. This analysis highlighted, excluding the difference derived from farm phase, transport is the main cause of larger impact for low moisture mozzarella, because foreign raw milk is generally used for this type of production, and the cooking in oven in the consumption phase, because low moisture mozzarella is largely used as pizza topping, in fact a cooking in an electric oven was assumed