Biodiversity footprints of 151 popular dishes from around the world

biodiversity footprint associated with a crop while dataset (ii) was used to calculate the biodi- versity footprint related to pasture. Agricultural crops distributions were based on 175 ca. year 2000’s global crop-specific maps [31]. The distribution of biodiversity footprint values for a given crop or pasture and country across the raster cells of the country where the crop or pasture was distributed was presented through percentiles. We used the 50th percentile data for each ingredient. In addition, outer fences were calculated by multiplying the interquartile range by three to identify extreme outli- ers in the datasets. The three crops identified as outliers were excluded and replaced with other crops within the fences when applicable (S1 Table). To determine the biodiversity footprint of each dish, the biodiversity footprints of its ingre- dients were summed up. To calculate the biodiversity footprint of each crop ingredient in turn, we used: ðBiodiversity footprint of a crop ingredient ¼ ðWc � FcÞ; ð1Þ where Wc and Fc represent the fresh weight of a given crop c ingredient (in Mg) in the recipe of a dish and the biodiversity footprint of c (in number of species Mg-1, number of threatened species Mg-1 or range rarity Mg-1), respectively. Ingredients were matched to the closest corre- sponding crops or pasture in the biodiversity footprint dataset (S2 Table). In the case of processed ingredients, the biodiversity footprint of the raw ingredients (calcu- lated using Eq (1)) was multiplied by a corresponding conversion factor (S3 Table) to obtain

the biodiversity footprint before this was used as an ingredient in the calculation of the dish’s biodiversity footprint. The conversion factor was the fraction of ingredient which can be extracted from its raw product. Calorical standardization To ensure that the biodiversity footprint calculations of all dishes were comparable, calories were kept at a constant level for standardization. For all dishes, calories were standardised to be 825 kcal per dish, assuming daily calorical needs of 2000 kcal and a breakfast of 350 calories with the remaining two meals at 825 kcal each. This was done through proportionally chang- ing the amount of the ingredients based on the recipe while keeping their relative proportions constant. The calorical information for each ingredient was taken from the US Department of Agriculture [32]. For recipes where quantity instead of the ingredient’s weight was stated, the weight was assumed as that included in the dataset [32]. In addition, when measurements of the ingredients were not readily convertible to the weights presented by US Department of Agriculture, conversion of the ingredients’ weights were done manually following guidelines, e.g. when fruits, eggs or bread slices were specified in units, they were converted to weight using the tables in the guidelines [33]. Dishes selection and recipes In the selection of dishes from Cable News Network or Taste Atlas (S4 Table contains the links to the dishes for each country and their sources), dishes were excluded if not all of their ingre-

dients were represented in the biodiversity footprint dataset with the exception of salt that was not assessed for land use impacts. Countries for which Taste Atlas was the source of dishes were searched using the “dishes” tab followed by the “Selection” tab. The first 120 dishes were considered for countries with more than 100 entries. 151 dishes were eventually selected. The ingredients for each dish were estimated based on recipes in RecipeDB, a structured recipe database for global dishes, when available [34] or using alternative sources (S5 Table contains the recipes of each dish and its source). Processed ingredients such as tomato sauce used in the PLOS ONE Biodiversity footprint of popular dishes PLOS ONE | https://doi.org/10.1371/journal.pone.0296492 February 21, 2024 4 / 15