Towards an environmentally sustainable and healthy Atlantic dietary pattern: Life cycle carbon footprint and nutritional quality

Towards an environmentally sustainable and healthy Atlantic dietary pattern: Life cycle carbon footprint and nutritional quality Xavier Esteve-Llorens, Carmela Darriba, María Teresa Moreira, Gumersindo Feijoo, Sara González- García Accepted Mansucript How to cite: Esteve-Llorens, X., Darriba, C., Moreira, M., Feijoo, G., & González-García, S. (2019). Towards an environmentally sustainable and healthy Atlantic dietary pattern: Life cycle carbon footprint and nutritional quality. Science Of The Total Environment, 646, 704- 715. doi: 10.1016/j.scitotenv.2018.07.264 Copyright information: © 2018 Elsevier Ltd. This manuscript version is made available under the CC-BY-NC-ND 4.0 license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Towards an environmentally sustainable and healthy Atlantic dietary pattern: life cycle 1 carbon footprint and nutritional quality 2 Xavier Esteve-Llorens, Carmela Darriba, Maria Teresa Moreira, Gumersindo Feijoo and Sara 3 González-García 4 Department of Chemical Engineering, School of Engineering, Universidade de Santiago de 5 Compostela. 15782- Santiago de Compostela, Spain. 6 * Corresponding author: E-mail address: xavier.esteve.[email protected] 7 8 Abstract 9 Production and consumption of food has a significant effect on climate change. The effect of 10 different consumption habits on the environment should not be under estimated, as there are 11 different studies that mention the environmental impact associated with different foods, 12 especially those of animal origin. The analysis of the Atlantic Diet (AD), as the most common 13 dietary pattern in Northwestern Spain, serves as an example of a diet with a high consumption 14 of local, fresh and seasonal products, home cooking and low-processed foods. The evaluation 15 was carried out by quantifying the carbon footprint following the Life Cycle Analysis 16 methodology and identifying its nutritional quality according to the value of the Nutrient-rich 17 Dietary index (NRD9.3.). According to the main results, the consumption of livestock products 18 and shellfish is responsible for most GHG emissions (70% of the total). The basic ingredients of 19 the AD, such as vegetables and legumes, make a relatively minor contribution (with an impact 20 of 30% of the total) to the total carbon footprint of 3.01 kg CO2eq·person-1·day-1. As regards 21 nutritional quality, AD has a high nutritional score (474), mainly due to the low intake of sodium, 22 added sugars and saturated fats (nutrients to be limited in healthy diets). In general, both the 23 carbon footprint and the nutritional index score are consistent with those of other studies on the 24 Mediterranean diet, which has been recognised as beneficial. Therefore, it can be concluded 25 that the AD may be recommended from a nutritional and environmental point of view, mainly 26 due to the high intake of fish and vegetables. The communication of this valuable environmental 27 and nutritional information to consumers should be taken into account when considering 28 strategic actions for the adoption of healthy and sustainable dietary patterns. 29 30 Keywords: Atlantic diet; Daily diet; LCA; Meals; Nutritional value; Sustainable diet 31 1. Introduction 32 Nutrition is a basic human need and access to adequate nutrition depends on numerous 33 social, political and economic factors (Heller et al., 2013). Balanced and complete nutrition 34 affects human health and well-being. The effects of nutritional patterns on overweight, 35 cardiovascular disease and other diet-related health problems are widely known (Coelho et al., 36 2016; Risku-Norja, 2011). The selection of one type of food versus another entails direct 37 consequences in the supply chain, as well as environmental, economic and social impacts 38 associated with the production process (Cencic and Chingwaru, 2010). In particular, food chains 39 that support diets are linked to environmental issues such as greenhouse gas (GHG) emissions, 40 embedded energy consumption and land use (Irz et al., 2016; Castañé and Antón, 2017). 41 Therefore, environmental pressures on food systems are relevant to public health agendas 42 (Sáez-Almendros et al., 2013). 43 Food production ranges from agricultural and farming activities to manufacturing, 44 refrigeration, retailing, storage, cooking and final disposal of waste (Garnett, 2011; Sáez-45 Almendros et al., 2013). According to Garnett (2011) and Irz et al. (2016), 15-30% of total GHG 46 emissions in developed countries are derived from food production, distribution and 47 consumption, and agriculture is responsible for 70-80% of water consumption (Heller et al., 48 2013). In this regard, researchers are evaluating the sustainability of food production and eating 49 patterns (Baroni et al., 2007; Donati et al., 2016). According to these studies, 50 lactoovovegetarian or plant-based diets are more environmentally sustainable than those 51 containing resource-intensive products (e.g., meat-rich diets) (Baroni et al., 2007; Risku-Norja, 52 2011). Of special interest is the development of methodologies to analyse the environmental 53 impact of a product or food system with the most objective approach. (Aleksandrowicz et al., 54 2016; Duchin, 2005; van de Kamp et al., 2018; Van Kernebeek et al., 2014). The environmental 55 footprints of some diets (e.g., omnivorous, vegetarian, vegan, omega-3 fatty acids enriched) 56 have been quantified according to the Life Cycle Assessment – LCA methodology (Pimentel 57 and Pimentel, 2003; Coelho et al., 2016). In this sense, numerous studies can be found in the 58 literature in which the relationship between European diets, nutritional quality and environmental 59 aspects are evaluated in detail (Aleksandrowicz et al., 2016; Duchin, 2005; van de Kamp et al., 60 2018; Van Kernebeek et al., 2014). 61 Several studies can be found in the literature where the food trends of Swedish 62 homemade menus were analysed proposing dietary guidelines, with special attention to organic 63 food (Carlsson-Kanyama et al., 2003; Carlsson-Kanyama and Faist, 2000; Carlsson-Kanyama 64 and Linden, 2001). In these studies, the energy needs throughout the life cycle of Swedish diets 65 were estimated with the aim of planning home-cooked meals that were nutritionally rich but 66 included products that were less energy-intensive. 67 Jungbluth et al. (2000) proposed a simplified LCA approach to assess consumption 68 patterns in Switzerland with the aim of identifying environmental-friendly decisions. The authors 69 propose different actions to reduce environmental impacts. Therefore, it was proposed to 70 reduce meat consumption and demand for airborne products as well as to promote the 71 consumption of organic products. 72 The effect of Dutch consumption patterns on agricultural land needs was also assessed 73 in detail (Gerbens-Leenes et al., 2002; Gerbens-Leenes and Nonhebel, 2002). According to 74 these studies, large differences in land requirements for food production were identified not only 75 in the Netherlands, but also for other European countries. A hypothetical analysis of a wheat-76 based diet would require up to six times less soil compared to a meat-rich diet. 77 Finally, Van Kernebeek et al. (2014) addressed the question of whether plant-based 78 diets have lower environmental impact than those with a significant contribution of food with 79 animal origin, but taking into consideration nutritional quality. To this end, a review of more than 80 fifty peer-reviewed studies was conducted. The results showed that diets with higher 81 percentages of food products of animal origin could be associated with higher GHG emissions 82 and land use requirements, but these results were variable depending on the functional unit 83 considered. Special mention was made of the need to assess the overall nutritional quality of a 84 diet and the recommended levels of protein intake as particularly relevant elements to be taken 85 into account when comparing dietary patterns. 86 The most recent literature contains numerous references on the selection of the most 87 appropriate functional units for the calculation of the environmental impacts of food consumption 88 patterns. The energy content (daily calories), protein or fat content of diets are some of the most 89 common examples (Pimentel and Pimentel, 2003; Heller et al., 2013). Moreover, profiles per 90 serving or per potential economic value are also available (Heller et al., 2013; Van Der Werf and 91 Salou, 2015). However, nutritional value must be the crucial element of reference for the 92 definition of a healthy diet. Nutrient-based recommendations should be specifically considered 93 in the search for and promotion of a sustainable dietary pattern that meets these values 94 (Smedman et al., 2010; Heller et al., 2013). 95 The high consumption of fruits, vegetables and whole grains in the diet is closely related 96 to the reduction of the risk of developing chronic diseases such as cancer and cardiovascular 97 diseases, which are the main causes of death in industrialized countries (Cencic and 98 Chingwaru, 2010). Based on this type of food, healthier and more fruit and vegetable-rich diets 99 have been identified in southern countries. In contrast, northern countries have diets rich in 100 animal fats and food products of animal origin. It is interesting to identify different social contexts 101 and cultural values in relation to food (Nordström et al., 2013). While food is an individual issue 102 in northern countries, society in central and southern Europe associates food with the social 103 dimension of sharing a meal (Cencic and Chingwaru, 2010). 104 Spain is one of the European countries with the lowest mortality rates for ischaemic 105 heart disease. Within the country, regional differences have been identified in this regard. In 106 fact, variations have been reported to be up to 40% lower than the average in northern cities 107 (Medrano et al., 2012). The traditional Atlantic diet is a common dietary pattern in northern 108 Portugal and Galicia (northwest Spain), culturally and climatically similar areas and has been 109 associated with a lower likelihood of myocardial infarction and good metabolic health (Calvo-110 Malvar et al., 2016; Atlantic Diet Foundation1). The Atlantic diet is characterised by an abundant 111 consumption of plant-based products, as well as local and fresh products (seasonal food) with 112 reduced cooking time. The consumption of meat (mainly beef and pork) and eggs is reasonable 113 and olive oil is considered as the main source of fat for cooking and seasoning (Calvo-Malvar et 114 al., 2016; Charro et al., 2006). Recently, it has been rated as a world reference for a healthy diet 115 (Vaz Velho et al., 2016). The Atlantic diet differs from the Mediterranean - the most popular in 116 southern Spain, in terms of increased consumption of fish, red meat, pork, milk, potatoes, fruit, 117 vegetables and olive oil (Guallar-Castillón et al., 2013), which implies significant changes in 118 nutrients and functional components. However, both of them can be taken as examples of 119 healthy diet (Tojo and Leis, 2009; Sáez-Almendros et al., 2013). 120 1 https://www.fundaciondietatlantica.com/eng/index.php (accessed July, 2018) This study has a twofold objective: to quantify the carbon footprint of the Atlantic diet 121 through a LCA approach associated with the production of the different foods that make up this 122 diet, while identifying its nutritional quality. The recommended Galician dietary pattern and the 123 corresponding intake data have been taken into account. The main causes of GHG emissions 124 will be highlighted to identify options for improvement. 125 126 2. Materials and methods 127 2.1. Weekly menu based on the Atlantic diet 128 The concept of the Atlantic diet dates back to the traditional menus of Galician gastronomy. 129 With the social awareness of a healthy diet, the benefits of this dietary pattern are reflected in a 130 recent study (Vaz Velho et al., 2016) (see Figure 1). It is characterised by i) a high intake of 131 seasonal foods, vegetables, fruits, potatoes, bread and cereals, chestnut, whole nuts, legumes 132 and honey, fish, molluscs and crustaceans; ii) a moderate consumption of milk, cheese, meat 133 (beef and pork), eggs and iii) cooking methods based on boiling, stewing, grilling and roasting. 134 An abundant intake of complex sugars, fibre, polyunsaturated fatty acids, vitamins, minerals and 135 functional components is therefore guaranteed (Tojo and Leis, 2009; Vaz Velho et al., 2016). 136 137 138 Figure 1. Atlantic diet pyramid. The base and the top of the pyramid include the foodstuffs 139 that must be daily consumed or occasionally consumed, respectively. 140 141 Although studies can be found in the literature that consider individual meals, daily or 142 annual diets (Van Kernebeek et al., 2014), a weekly diet has been considered for analysis, as it 143 may facilitate comparison with other types of dietary patterns. Following the recommendation of 144 Tojo and Leis (2009), a weekly diet has been designed– displayed in Table 1 – consisting of 145 seven daily menus (daily diets) divided in five meals (breakfast, mid-morning snack, lunch, 146 afternoon snack and dinner) has been designed, as similar as possible to the recommended 147 Galician eating habits (Xunta de Galicia, 2013). This weekly diet is based on 2,100 kcal and 148 corresponds to the energy needs of an active Spanish adult woman (regular physical activity) 149 according to FAO (2014). 150 151 Table 1. Atlantic diet based weekly menus designed for this study taking into account the 152 recommended servings of the different food groups. The daily diets have been adjusted to a 153 recommended energy intake of 2,100 kcal. 154 155 Daily diet Monday Tuesday Wednesday Thursday Friday Saturday Sunday Breakfast Glass of milk and bread with tomato Glass of milk. cereals. nectarine Glass of milk. Wholemeal bread. orange Glass of milk. wholemeal cereals. nectarine Glass of milk. wholemeal bread. orange Curd with honey. wholemeal bread Glass of milk. bread. peach Mid-morning Snack Orange Yogurt Yogurt Infusion Apple Kiwi Pear Lunch Steamed cockles. vegetable cannelloni. nectarine. bread Vegetables and fruits salad. black rice with cuttlefish. bread Carrot salad. Galician style conger. two figs. bread Mackerel with potatoes and roasted peppers. watermelon. bread Octopus with potatoes. salad. curd with honey. bread Cod croquettes. vegetable stew. tangerine. bread Padron peppers. roast beef with potatoes Afternoon snack Banana Melon Banana Peach Banana Yogurt Nectarine Dinner Grilled beef steak with rice. bread Grilled pork steak. steamed Brussels sprouts. pear. bread Pumpkin cream. pear. bread Pasta salad. fresh cheese. bread Grilled beef steak with rice and steamed vegetables Chicken steak with pasta and mushrooms. grapes. bread Scrambled eggs with mushrooms and pasta Total energy intake 2,124 kcal 2,101 kcal 2,295 kcal 2,140 kcal 2,051 kcal 2,189 kcal 2,097 kcal 156 Recommended servings of different food groups have been considered for evaluation. Table 2 157 shows the frequency servings for the Atlantic dietary pattern. Although the specific composition 158 of the diet changes with age and sex, this level of uncertainty can be assumed for the 159 estimation of the carbon footprint. 160 161 Table 2. Main recommendations of servings (s) frequency for each food group for the Atlantic 162 Diet adapted from Velho et al. (2016). 163 164 Food group Servings frequency Cereals/Grains 6-8 s·day-1 Fruits 3s or more·day-1 Vegetables 2s or more·day-1 Olive oil 3-4s·day-1 Dairy products 3-4s·day-1 Dried fruits 4-6s·week-1 Legumes 2-3s·week-1 Seafood 3-4s·week-1 Meat 3-4s·week-1 Eggs 3-4s·week-1 Sweets Sparingly monthly 165 2.2. Estimation of the Atlantic diet nutrient composite score 166 One of the main objectives of this study is to analyse the nutritional quality of the Atlantic 167 diet to identify whether this dietary pattern meets healthy parameters. It is well known that 168 consumers are advised to look for nutrient-rich foods rather than discretionary calories. Taking 169 into account the main recommendations from Van Kernebeek et al. (2014), the nutrient intake 170 associated with one single meal cannot be used to assess the nutritional quality of a daily diet. 171 Therefore, the nutritional quality has been analysed through daily menus. This perspective will 172 facilitate comparison with alternative dietary patterns. The Nutrient Rich Food (NRF9.3) score 173 (Drewnowski, 2009; Fulgoni et al., 2009) is considered the cornerstone of a dietary guidance 174 approach to healthy eating. However, the Nutrient Rich Diet (NRD9.3) score was considered in 175 Saturday 88.7 39.2 1680 309 11 1009 17 4479 345 18.1 2.3 1.56 Sunday 111.0 36.3 1680 289 92 921 18 5308 393 18.8 1.9 1.46 Daily average 101.2 39.9 1448 321 23. 1084 33 4938 435 19.3 2.0 1.45 332 As shown in Table 4, the average daily diet reports an intake of numerous nutrients to be 333 encouraged (i.e., protein, fibre, potassium and magnesium) higher than the values 334 recommended in Table 3, as well as the average values corresponding to the Atlantic diet 335 reported in the literature (Fundación Española de la Nutrición, 2004). 336 The high protein intake observed is related to the outstanding consumption of fish and 337 moderate consumption of meat (mainly beef and pork). All designed daily diets exceed the 338 recommended daily protein intake value of 50 g (up to 2.5 times). 339 340 341 Figure 3. Total protein and animal based protein ingestion per daily diet designed 342 under the Atlantic dietary patterns (g·day-1). RDV – Recommended Daily Value (g·day-1). AD – 343 Established average daily protein intake under Atlantic dietary pattern (g·day-1). 344 345 Figure 3 represents the daily protein intake for each designed daily diet, together the 346 average dietary value and the recommended daily value suggested by Fundación Española de 347 la Nutrición (2004). Protein intake per person ranges from 85.5 to 123 g·day-1, with the 348 percentage of animal protein in relation to total dietary protein varying between 48% and 70%. 349 0 20 40 60 80 100 120 140 Monday Tuesday Wednesday Thursday Friday Saturday Sunday Daily average g protein·day-1 Total proteins Animal proteins RDV AD In line with Van Kernebeek et al. (2014), protein intake is positively associated with AP%. 350 According to the intrinsic characteristics of the Atlantic diet, protein intake comes mainly from 351 seafood and meat, as well as 26% (on average) from dairy products such as milk, yoghurt and 352 cheese. 353 Fibre intake can be almost double the recommended value, mainly due to the high intake of 354 fruits, vegetables (e.g., potatoes) and cereals (bread). This high intake of seasonable products 355 also leads to a high dose of potassium. As for magnesium, the remarkable consumption of blue 356 fish (e.g., mackerel) and molluscs (e.g., cockles) along with cereals affects the intake ratio. 357 With regard to other nutrients to encourage, such as vitamins A, C and E, as well as 358 calcium and iron, the amount consumed is within the recommended range. It can be associated 359 with the consumption of a nutrient-enriched product, such as carrots (common as side dish) for 360 vitamin A, pepper (the food product with the highest vitamin C content and a common spice 361 ingredient) and vitamin E, molluscs and dairy products for calcium and fish and molluscs for 362 iron. 363 For nutrients to limit (saturated fat, added sugar and sodium), their intake is below the 364 recommended limits. The consumption of olive oil and dairy products such as cheese is 365 associated with consumption of saturated fats (both food groups present a serving frequency of 366 3-4 s·day-1). For added sugar, the intake is around 4% of the maximum recommended value. In 367 designed daily diets, it is associated with the consumption of bread and whole grain cereals. 368 The consumption of bread is a characteristic of the Atlantic diet, being greater than in other 369 types of diets such as the Mediterranean. The outstanding presence of fish (mackerel, 370 cuttlefish,…), bread and meat is mainly responsible for sodium in the diet. Moreover, the 371 Atlantic diet is characterised by a high intake of unsaturated fatty acids, which makes it one of 372 the highest in the world (Fundación Española de la Nutrición, 2004). 373 Just as a remark, potatoes are a basic food ingredient in the Atlantic diet, unlike other 374 dietary patterns such as Mediterranean or even vegan diets. It is considered an important 375 source of complex carbohydrates, fibre, minerals, vitamins and water. 376 Another point to take into account is the notable difference between the intake of nutrients 377 (mainly fibre, vitamin C, vitamin E, iron, potassium, magnesium and added sugar) estimated for 378 the daily diets designed and those reported in the literature for the Atlantic diet (Fundación 379 Española de la Nutrición, 2004). Dietary scenarios depend on individual meals, which are 380 affected by factors such as local conditions, seasonal food, gender and even the economic 381 profile of the family. The relationship between these factors and the nutrients intake could be 382 further explored, but it is beyond the scope of this study. 383 With regard to the NRD9.3 scores for each diet designed (Table 1), scores range from 418 384 (corresponding to the diet proposed for Tuesday) to 525 (corresponding to the diet proposed for 385 Thursday), as shown in Figure 4. These values are in line with others reported in the literature 386 ranging from 260 to 666, corresponding to other different types of dietary patterns (Nordic, 387 Finish, Indian, English, Mediterranean, vegan…) (Collins and Fairchild, 2007; Risku-Norja et al., 388 2008; Pathak et al., 2010; Saxe et al., 2012). 389 390 391 Figure 4. NRD9.3 scores and AP% (ratio of animal based protein and total dietary protein) for 392 each diet that constitutes the designed weekly menu. Minimum (Min) and maximum (Max) 393 NRD9.3 score values found in the literature are also displayed. Numbers on the left-y-axis 394 represent the AP% (in %). Numbers on the right-y-axis represent the NRD9.3 scores. 395 396 The specific characteristics of each particular type of diet are responsible for the wide range 397 of values in the NRD9.3 index. Moreover, this index is also affected by the above-mentioned 398 parameters (nutrients to encourage/limit as well as RDV) since its estimation is directly 399 dependent on nutrient intake. According to the literature (Van Kernebeek et al., 2014), the 400 relationship between the NRD9.3 score and the percentage of animal protein can vary 401 0 100 200 300 400 500 600 700 0 20 40 60 80 100 Monday Tuesday Wednesday Thursday Friday Saturday Sunday Daily average NRD9.3 AP% AP% NRD9.3 score NRD9.3 score_Min NRD9.3 score_Max considerably between studies and there is no a general trend. Risku-Norja et al. (2009) and 402 Gerbens-Leenes and Nonhebel (2002) identified a reduction in the NRD9.3 score with an 403 increase in the ratio of animal protein to total protein consumed. In contrast, other authors 404 (Collins and Fairchild, 2007; Saxe et al., 2012) identified the opposite trend. Thus, this effect 405 has been also analysed in this study considering the different daily diets proposed for analysis 406 along with the average. The results in Figure 4 do not show a clear correlation. Some daily diet 407 scenarios have a downward trend, while others have an upward trend in the NRD9.3 score, with 408 an increase in the ratio of animal protein to total daily protein consumed (AP%). 409 Van Kernebeek et al. (2014) proposed an association between both parameters (NRD9.3 410 and AP%) considering the results reported in the literature and concluded that the NRD9.3 411 score is negatively associated with the protein ratio. With this consideration in mind, Figure 5 412 shows the association between the NRD9.3 score and the AP% for the weekly diet proposed 413 here. In addition, the NRD9.3 scores corresponding to these AP% values have also been 414 estimated considering the correlation proposed by Van Kernebeek et al. (2014). The estimated 415 values are 1.1-1.3 times higher than those calculated for our weekly menu. Variations in nutrient 416 composition and dietary characteristics are responsible for these differences. However, in line 417 with Van Kernebeek et al. (2014), the same behaviour can be observed and the score is 418 negatively associated with the AP%. In this sense, these results can be useful and provide 419 information to both consumers and policy-makers to achieve healthier food choices in the 420 supermarket or advise on the need to prioritise the intake of plant rather animal protein to 421 reduce the intake of products of animal origin, respectively. 422 423 424 Figure 5. Correlation (grey marks) between NRD9.3 scores (y-axis) and AP% (x-axis) for the 425 weekly Atlantic diet designed for analysis. Marks in black have been estimated considering the 426 correlation established by Van Kernebeek et al. (2014) and the AP% values of our daily diets. 427 428 Finally, the nutritional quality of the diets has been completed with the estimation of 429 individual nutrient-rich indexes to report dietary intake in relation to recommended daily values. 430 Table 5 summarizes the corresponding NR scores per daily diet. NR corresponding to protein, 431 fibre and potassium present a value of 100% since their intake exceeds the recommended 432 values. Magnesium intake also implies outstanding NR indexes, equal or very close to 100%. In 433 accordance with the methodology and in order to avoid credits for the overconsumption of 434 nutrients to encourage, nutrient intake is assumed to be equal or greater than the RDV. 435 Conversely, values are not rounded to 100% for nutrients to limit if the recommended daily 436 value is exceeded. Nutrients to limit such as sodium and saturated fats generally present high 437 NR indexes (above 56 and 76% respectively). In contrast, the intake of added sugar reports NR 438 indexes below 5% regardless of the designed diet. These values are much lower than those of 439 other diets such as Mediterranean one (Castañé and Antón, 2017), where NR indexes of 440 between 80% and 136% can be expected. These high values are mainly related to the 441 consumption of products such as yoghurt and jam. 442 y = -0,3661x + 495,58 R² = 0,0039 400 450 500 550 600 40 45 50 55 60 65 70 75 NRD9.3 Estimated NRD9.3 According to Table 5, the Atlantic diet should report really low NR indexes of Vitamin E 443 and C (for Vitamin E below 3% in most proposed daily menus). Improvements in this diet should 444 focus on promoting the intake of ingredients rich on both components, as they are nutrients to 445 encourage. Consumption of citrus products (e.g., orange, mandarin) and nuts (e.g. almonds, 446 hazelnuts) may contribute to increasing the NR-values for Vitamin C and Vitamin E, 447 respectively. 448 449 Table 5. Nutrient Rich (NR) score for each analysed nutrient. Scores have been calculated with 450 regard to the recommended daily value of each nutrient. 451 Boosting nutrients Limiting nutrients Protein Fiber VitA VitC VitE Ca Fe K Mg Saturated fat Added sugar Na Monday 100 100 91.5 32.9 2.1 82.1 100 100 100 121 3.8 77.5 Tuesday 100 100 34.3 19.7 2.3 48.5 65.9 100 100 71.7 4.6 76.3 Wednesday 100 100 39.7 24.3 1.9 63.1 65.5 100 100 99.4 3.6 68.1 Thursday 100 100 87.0 44.9 2.6 65.8 100 100 100 94.1 4.6 76.8 Friday 100 100 100 38.0 2.3 63.6 61.0 100 100 75.9 2.9 56.3 Saturday 100 100 90.8 3º0.0 2.2 57.7 53.2 100 86.3 90.5 4.6 80.0 Sunday 100 100 90.8 28.0 18.1 52.7 55.6 100 98.2 94.0 3.8 74.7 Daily average 100 100 76.3 31.1 4.5 61.9 71.6 100 97.8 93.3 4.1 74.3 452 3.2. Carbon footprint of the Atlantic diet 453 3.2.1. Detailed analysis of carbon footprint for the designed menus 454 The estimation of GHG emissions (i.e., carbon footprint) corresponding to the menus 455 designed following the recommendations of the Atlantic diet represents an absolute value of 456 21.04 kg CO2eq per person and week, i.e., an average of 3.01 kg CO2eq·person-1·day-1. This 457 value is slightly higher (~5%) than that reported in the literature focusing on the assessment of 458 the Mediterranean dietary pattern, the most widespread diet in Spain (Castañé and Antón, 459 2017; Muñoz et al., 2010) and with characteristics similar to those of the Atlantic. The rationale 460 behind that difference is mostly associated with differences on the dietary patterns as well as 461 with the consideration of refrigeration process at households within the system boundaries, 462 which was excluded from analysis by Castañé and Antón (2017) and which adds to 0.23 kg 463 CO2eq·person-1·day-1. Considering the same system boundaries, the AD presents a carbon 464 footprint around 8% lower than the corresponding to the Mediterranean one (2.86 kg 465 CO2eq·person-1·day-1). The shift towards consumption of fish and fresh products (seasonal 466 food) with limited cooking is behind this difference. 467 Through a more detailed assessment of the factors responsible for the carbon footprint of 468 the AD, the production of the different food products is identified as a hot spot followed by 469 household (cooking and refrigeration) and transport activities. Contributions from the production 470 stage account for approximately 78% of total GHG emissions, with the remaining 22% is split 471 between the household stage (92%) and transport activities (8%). Figure 6 displays the carbon 472 footprint per day, as well as the distribution between the stages included in the analysis (food 473 production, household and transport). 474 475 476 Figure 6. Daily carbon footprint (in kg CO2eq·person-1·day-1) taking into account the distribution 477 between food production, transport and household stages. 478 479 Regarding the stage of food production (with an average of 2.31 kg CO2eq·person-1·day-1), 480 it includes all the background activities carried out in the field and on the farm as well as the 481 corresponding industrial processing, if necessary. According to Figure 7, meat and dairy 482 production (livestock-based items) is primarily responsible for GHG emissions at this stage 483 (26% and 30%, respectively). Moreover, both food categories are primarily responsible for 484 variations in the carbon footprint between different daily diets. Looking more closely at the 485 contribution of meat production, red meat accounts for 23%, followed by white meat (1.6% pork 486 and 1.4% chicken, respectively). 487 In contrast, vegetables and fruits are low-carbon food categories (see Table 1 of SM) but 488 consumed in major shares in the Atlantic diet. Therefore, both categories report contributions of 489 8% of total GHG emissions from the food production stage. 490 491 492 Figure 7. Relative distribution of GHGs emission from food production stage between the food 493 groups involved in the designed 7-day menu. 494 495 The remarkable effect on the carbon footprint of livestock products has been highlighted by 496 numerous studies, including those focusing on very different dietary patterns such as Spanish, 497 Peruvian, Western European, American, British and French (Castañé and Antón, 2017; Coelho 498 et al., 2016; Muñoz et al., 2010; Pimentel and Pimentel, 2003; Sáez-Almendros et al., 2013; 499 Scarborough et al., 2014; Vázquez-Rowe et al., 2017). Both products are an important source 500 of protein and energy, and their production involves resource-intensive activities (e.g., fodder 501 8% 8% 0% 9% 28% 27% 15% 2% 3% 0% Fruits Vegetables Legumes Grains Dairy Meat Fish Eggs Olive oil Sweets production and agricultural activities), as well as methane emissions from rumiant enteric 502 fermentation. 503 The seafood category has an outstanding contribution (15% of the total). This contribution is 504 directly related to the remarkable consumption of fish in the Atlantic diet (Vaz Velho et al., 2016) 505 despite reporting moderate rate of GHG emissions per kg of product (see Table 1 of SM). Grain 506 products such as cereals and bread are basic products of the Atlantic food pyramid and their 507 contributing ratio rises to 9%. 508 In terms of GHG emissions from household activities, the total energy required for the 7-day 509 menu is about 33 MJ per week and person split between cooking (60%) and refrigeration (40%). 510 Energy consumption in cooking activities is slightly lower than that of Castañé and Antón 511 (2017), i.e., 30 MJ for the Mediterranean diet. In this sense, the abundance of fresh food 512 products in the Atlantic diet makes complex cooking processes unnecessary, and therefore, 513 implies low energy requirements for cooking. Taking into account the distribution of the carbon 514 footprint among the contributing stages (see Figure 6), there are no significant differences in the 515 average energy consumption for household activities regardless of the designed daily diet, as 516 shown in Figure 6. The consideration of only boiling, frying and baking as the main cooking 517 processes in the analysis -as recommended by the Atlantic dietis also responsible for these 518 negligible differences between the daily menus with regard to the household stage. Boiling and 519 frying (the most common daily cooking methods) report similar energy requirements (~0.75 MJ 520 per meal and person, on average). For baking, it is considerably higher, about 4.1 MJ per meal 521 and person. 522 Finally, the contribution of the transport stage to the global carbon footprint can be 523 considered negligible since it represents less than 2% of the total (on average) with 0.70 kg 524 CO2eq·week-1·person-1. As far as the origin of food is concerned, Galician products have the 525 lowest GHG emissions due to the shorter distribution distances by lorry. Products of foreign 526 origin are distributed by sea freighter and/or lorry. Maritime transportation does not report 527 outstanding contributions to the carbon footprint despite long distances. Once again, road 528 transport is the main contributing factor to the carbon footprint (five times more than maritime 529 transport). 530 531 3.2.2. Comparison with results from literature 532 Numerous studies available in the literature were developed with regard to the 533 environmental assessment of human diets where special attention was paid to the estimation of 534 the carbon footprint (Castañé and Antón, 2017; Coelho et al., 2016; Notarnicola et al., 2017; 535 Pairotti et al., 2015; Pernollet et al., 2017; Röös et al., 2015; Saxe et al., 2012; Scarborough et 536 al., 2014; Vázquez-Rowe et al., 2017). All these studies highlight the limitations of the analysis 537 in the absence of an established methodology and data. The focus on the carbon footprint is 538 based on the availability of data and the awareness of society to avoid anthropogenic GHG 539 emissions to prevent climate change (Rockström et al., 2009). The comparison between our 540 results for the Atlantic diet and those available in the literature for other types of dietary patterns 541 (e.g., Mediterranean, average European, average Spanish, German, Swedish, French, vegan, 542 vegetarian, Nordic, among others) is complex because the results depend on a wide variety of 543 factors and hypotheses. 544 The number of calories that an average person needs on a daily basis depends on several 545 factors, such as minimum and average dietary energy requirements (Vázquez-Rowe et al., 546 2017), level of activity, gender, age, weight, geographical location and cultural aspects (EFSA, 547 2009). Therefore, the range of energy requirements per capita identified in the literature varies 548 from 1,702 kcal·person-1·day-1 in Indian diets (Pathak et al., 2010) to 3,596 kcal·person-1·day-1 549 in Western European countries (Tukker et al., 2011). The daily energy intake recommended by 550 the Panel on Dietetic Products, Nutrition and Allergies (EFSA, 2009) is 2,000 kcal·person-1·day- 551 1.in European countries. It falls in half the range for a moderately active woman (1,625-2,400 552 kcal·person-1·day-1), which is consistent with the values recommended in other countries such 553 as the United States, Australia and New Zealand, as well as by the European food industry 554 (EFSA, 2009). According to experts, this value (2,000 kcal·person-1·day-1) is more consistent 555 with dietary advice for the general population compared to men (2,200-2,300 kcal·person-1·day- 556 1). Therefore, the value set in our study (2,100 kcal·person-1·day-1) could be considered 557 representative for the assessment and coincides with other relevant studies available in the 558 literature (Castañé and Antón, 2017; Collins and Fairchild, 2007; Peters et al., 2007; Sáez-559 Almendros et al., 2013; Scarborough et al., 2014). 560 supplements in intestinal health. Nutrients 2, 611–625. doi:10.3390/nu2060611 711 Coelho, C.R. V, Pernollet, F., van der Werf, H.M.G., 2016. Environmental Life Cycle 712 Assessment of Diets with Improved Omega-3 Fatty Acid Profiles. PLoS One 11, 713 e0160397. doi:10.1371/journal.pone.0160397 714 Collins, A., Fairchild, R., 2007. Sustainable food consumption at a sub-national level: An 715 ecological footprint, nutritional and economic analysis. J. Environ. Policy Plan. 9, 5–30. 716 doi:10.1080/15239080701254875 717 Committee on Climate Change, 2010. The fourth carbon budget: reducing the emissions 718 through 2020. 719 Donati, M., Menozzi, D., Zighetti, C., Rosi, A., Zinetti, A., Scazzina, F., 2016. Towards a 720 sustainable diet combining economic, environmental and nutritional objectives. Appetite 721 106, 48–57. doi:10.1016/j.appet.2016.02.151 722 Drewnowski, A., 2009. Defining nutrient density: Development and validation of the nutrient rich 723 foods index. J. Am. Coll. Nutr. 28, 421S–426S. doi:10.1080/07315724.2009.10718106 724 Duchin, F., 2005. Sustainable Consumption of Food - A Framework for Analyzing Scenarios 725 about Changes in Diets. J. Ind. Ecol. 9, 99–114. doi:10.1162/1088198054084707 726 EFSA, 2009. Review of labelling reference intake values Scientific Opinion of the Panel on 727 Dietetic Products , Nutrition and Allergies on a request from the Commission related to the 728 review of labelling reference intake values for selected nutritional elements 1 Adop. 729 FAO, 2014. Necesidades Nutricionales del Ser Humano, FAO. 730 FAO, 2010. Biodiversity and sustainable diets united against hunger. Rome. 731 Fulgoni, V.L., Keast, D.R., Drewnowski, A., 2009. Development and Validation of the Nutrient-732 Rich Foods Index: A Tool to Measure Nutritional Quality of Foods. J. Nutr. 139, 1549–733 1554. doi:10.3945/jn.108.101360 734 Fundación Española de la Nutrición, 2004. Consumo de alimentos en Galicia. La dieta atlántica. 735 Madrid. 736 Garnett, T., 2011. Where are the best opportunities for reducing greenhouse gas emissions in 737 the food system (including the food chain)? Food Policy 37, 463–466. 738 doi:10.1016/j.foodpol.2012.04.006 739 Gerbens-Leenes, P., Nonhebel, S., Ivens, W., 2002. A method to determine land requirements 740 relating to food consumption patternsA method to determine land requirements relating to 741 food consumption patterns. Agric. Ecosyst. Environ. 90, 47–58. 742 Gerbens-Leenes, P.W., Nonhebel, S., 2002. Consumption patterns and their effects on land 743 required for food 42, 185–199. 744 Goldstein, B., Hansen, S.F., Gjerris, M., Laurent, A., Birkved, M., 2016. Ethical aspects of life 745 cycle assessments of diets. Food Policy 59, 139–151. doi:10.1016/j.foodpol.2016.01.006 746 González-García, S., Castanheira, É.G., Dias, A.C., Arroja, L., 2013a. Environmental life cycle 747 assessment of a dairy product: The yoghurt. Int. J. Life Cycle Assess. 18, 796–811. 748 doi:10.1007/s11367-012-0522-8 749 González-García, S., Hospido, A., Moreira, M.T., Feijoo, G., Arroja, L., 2013b. Environmental 750 life cycle assessment of a galician cheese: San Simon da Costa. J. Clean. Prod. 52, 253–751 262. doi:10.1016/j.jclepro.2013.03.006 752 González, A.D., Frostell, B., Carlsson-Kanyama, A., 2011. Protein efficiency per unit energy and 753 per unit greenhouse gas emissions: Potential contribution of diet choices to climate 754 change mitigation. Food Policy 36, 562–570. doi:10.1016/j.foodpol.2011.07.003 755 Hallström, E., Carlsson-Kanyama, A., Börjesson, P., 2015. Environmental impact of dietary 756 change: A systematic review. J. Clean. Prod. 91, 1–11. doi:10.1016/j.jclepro.2014.12.008 757 Heller, M.C., Keoleian, G.A., Willett, W.C., 2013. Toward a life cycle-based, diet-level framework 758 for food environmental impact and nutritional quality assessment: a critical review. Env. 759 Sci Technol 47, 12632–12647. doi:10.1021/es4025113 760 Hospido, A., Tyedmers, P., 2005. Life cycle environmental impacts of Spanish tuna fisheries. 761 Fish. Res. 76, 174–186. doi:10.1016/j.fishres.2005.05.016 762 Irz, X., Leroy, P., Réquillart, V., Soler, L.G., 2016. Welfare and sustainability effects of dietary 763 recommendations. Ecol. Econ. 130, 139–155. doi:10.1016/j.ecolecon.2016.06.025 764 ISO 14040, 2006. Environmental management — Life cycle assessment — Principles and 765 framework, Iso 14040. Switzerland. doi:10.1136/bmj.332.7550.1107 766 Jungbluth, N., Tietje, O., Scholz, R.W., 2000. Food purchases: Impacts from the consumers’ 767 point of view investigated with a modular LCA. Int. J. Life Cycle Assess. 5, 134–142. 768 doi:10.1007/BF02978609 769 Leiva, F.J., Saenz-Díez, J.C., Martínez, E., Jiménez, E., Blanco, J., 2015. Environmental impact 770 of Agaricus bisporus cultivation process. Eur. J. Agron. 71, 141–148. 771 doi:10.1016/j.eja.2015.09.013 772 Medrano, M.J., Boix, R., Palmera, A., Ramis, R., Galán, I., López-Abente, G., 2012. Towns with 773 extremely low mortality due to ischemic heart disease in Spain. BMC Public Health 12, 774 174. doi:10.1186/1471-2458-12-174 775 Muñoz, I., Milà I Canals, L., Fernández-Alba, A.R., 2010. Life cycle assessment of the average 776 Spanish diet including human excretion. Int. J. Life Cycle Assess. 15, 794–805. 777 doi:10.1007/s11367-010-0188-z 778 Nordström, K., Coff, C., Jönsson, H., Nordenfelt, L., Görman, U., 2013. Food and health: 779 Individual, cultural, or scientific matters? Genes Nutr. 8, 357–363. doi:10.1007/s12263-780 013-0336-8 781 Notarnicola, B., Tassielli, G., Renzulli, P.A., Castellani, V., Sala, S., 2017. Environmental 782 impacts of food consumption in Europe. J. Clean. Prod. 140, 753–765. 783 doi:10.1016/j.jclepro.2016.06.080 784 Noya, I., Villanueva-Rey, P., González-García, S., Fernandez, M.D., Rodriguez, M.R., Moreira, 785 M.T., 2017. Life Cycle Assessment of pig production: A case study in Galicia. J. Clean. 786 Prod. 142, 4327–4338. doi:10.1016/j.jclepro.2016.11.160 787 Pairotti, M.B., Cerutti, A.K., Martini, F., Vesce, E., Padovan, D., Beltramo, R., 2015. Energy 788 consumption and GHG emission of the Mediterranean diet: A systemic assessment using 789 a hybrid LCA-IO method. J. Clean. Prod. 103, 507–516. doi:10.1016/j.jclepro.2013.12.082 790 Pathak, H., Jain, N., Bhatia, A., Patel, J., Aggarwal, P.K., 2010. Carbon footprints of Indian food 791 items. Agric. Ecosyst. Environ. 139, 66–73. doi:10.1016/j.agee.2010.07.002 792 Perignon, M., Masset, G., Ferrari, G., Barré, T., Vieux, F., Maillot, M., Amiot, M.J., Darmon, N., 793 2016. How low can dietary greenhouse gas emissions be reduced without impairing 794 nutritional adequacy, affordability and acceptability of the diet? A modelling study to guide 795 sustainable food choices. Public Health Nutr. 19, 2662–2674. 796 doi:10.1017/S1368980016000653 797 Pernollet, F., Coelho, C.R.V., van der Werf, H.M.G., 2017. Methods to simplify diet and food life 798 cycle inventories: Accuracy versus data-collection resources. J. Clean. Prod. 140, 410–799 420. doi:10.1016/j.jclepro.2016.06.111 800 Peters, C.J., Wilkins, J.L., Fick, G.W., 2007. Testing a complete-diet model for estimating the 801 land resource requirements of food consumption and agricultural carrying capacity: The 802 New York State example. Renew. Agric. Food Syst. 22, 145–153. 803 doi:10.1017/S1742170507001767 804 Pimentel, M., Pimentel, D., 2003. Sustainability of meat-based and plant based diets and the 805 environment. Am. J. Clin. Nutr. 78(suppl), 660–663. doi:10.1177/0956247808089156 806 Risku-Norja, H., 2011. From environmental concerns towards sustainable food provisioning. 807 Material flow and food consumption scenario studies on sustainability of agri-food 808 systems. 809 Rockström, J., Steffen, W., Noone, K., Persson, Å., Chapin, F.S., Lambin, E., Lenton, T.M., 810 Scheffer, M., Folke, C., Schellnhuber, H.J., Nykvist, B., de Wit, C.A., Hughes, T., van der 811 Leeuw, S., Rodhe, H., Sörlin, S., Snyder, P.K., Costanza, R., Svedin, U., Falkenmark, M., 812 Karlberg, L., Corell, R.W., Fabry, V.J., Hansen, J., Walker, B., Liverman, D., Richardson, 813 K., Crutzen, P., Foley, J., 2009. Planetary boundaries: Exploring the safe operating space 814 for humanity. Ecol. Soc. 14. doi:10.5751/ES-03180-140232 815 Röös, E., Karlsson, H., Witthöft, C., Sundberg, C., 2015. Evaluating the sustainability of diets-816 combining environmental and nutritional aspects. Environ. Sci. Policy 47, 157–166. 817 doi:10.1016/j.envsci.2014.12.001 818 Sáez-Almendros, S., Obrador, B., Bach-Faig, A., Serra-Majem, L., 2013. Environmental 819 footprints of Mediterranean versus Western dietary patterns: beyond the health benefits of 820 the Mediterranean diet. Environ. Health 12, 118. doi:10.1186/1476-069X-12-118 821 Saxe, H., Larsen, T.M., Mogensen, L., 2012. The global warming potential of two healthy Nordic 822 diets compared with the average Danish diet. Clim. Change 116, 249–262. 823 doi:10.1007/s10584-012-0495-4 824 Scarborough, P., Appleby, P.N., Mizdrak, A., Briggs, A.D.M., Travis, R.C., Bradbury, K.E., Key, 825 T.J., 2014. Dietary greenhouse gas emissions of meat-eaters, fish-eaters, vegetarians and 826 vegans in the UK. Clim. Change 125, 179–192. doi:10.1007/s10584-014-1169-1 827 Smedman, A., Lindmark-Månsson, H., Drewnowski, A., Edman, A.K.M., 2010. Nutrient density 828 of beverages in relation to climate impact. Food Nutr. Res. 54. doi:10.3402/fnr.v54i0.5170 829 Sonesson, U., Anteson, F., Davis, J., Sjödén, P.-O., 2005. Home transport and wastage: 830 environmentally relevant household activities in the life cycle of food. Ambio 34, 371–375. 831 doi:10.1579/0044-7447-34.4.371 832 Sonesson, U., Janestad, H., Raaholt, B., 2003. Energy for Preparation and Storing of Food - 833 Models for calculation of energy use for cooking and cold storage in households. 834 Tojo, R., Leis, R., 2009. La Dieta Atlántica, el pescado y las algas - Su importancia en el 835 neurodesarrollo y la función cerebral. Chapter: El papel de la Dieta Atlántica como 836 contrapunto saludable a la Dieta Ocidental actual. Universidad de Santiago de 837 Compostela. 838 Tukker, A., Goldbohm, R.A., De Koning, A., Verheijden, M., Kleijn, R., Wolf, O., Pérez-839 Domínguez, I., Rueda-Cantuche, J.M., 2011. Environmental impacts of changes to 840 healthier diets in Europe. Ecol. Econ. 70, 1776–1788. doi:10.1016/j.ecolecon.2011.05.001 841 van de Kamp, M.E., van Dooren, C., Hollander, A., Geurts, M., Brink, E.J., van Rossum, C., 842 Biesbroek, S., de Valk, E., Toxopeus, I.B., Temme, E.H.M., 2018. Healthy diets with 843 reduced environmental impact? - The greenhouse gas emissions of various diets adhering 844 to the Dutch food based dietary guidelines. Food Res. Int. 0–1. 845 doi:10.1016/j.foodres.2017.06.006 846 Van Der Werf, H.M.G., Salou, T., 2015. Economic value as a functional unit for environmental 847 labelling of food and other consumer products. J. Clean. Prod. 94, 394–397. 848 doi:10.1016/j.jclepro.2015.01.077 849 Van Kernebeek, H.R.J., Oosting, S.J., Feskens, E.J.M., Gerber, P.J., De Boer, I.J.M., 2014. The 850 effect of nutritional quality on comparing environmental impacts of human diets. J. Clean. 851 Prod. 73, 88–99. doi:10.1016/j.jclepro.2013.11.028 852 Vaz Velho, M., Pinheiro, R., Sofia, A., 2016. The Atlantic Diet – Origin and features. Int. J. Food 853 Stud. 5, 106–119. doi:10.7455/ijfs/5.1.2016.a10 854 Vázquez-Rowe, I., Larrea-gallegos, G., Villanueva-rey, P., Gilardino, A., 2017. Climate change 855 mitigation opportunities based on carbon footprint estimates of dietary patterns in Peru 1–856 25. doi:10.1371/journal.pone.0188182 857 Xunta de Galicia, 2013. Menús saludables ricos en fibra. Santiago de Compostela. 858 Ziegler, F., Nilsson, P., Walther, B.M.Y., 2003. Life Cycle assessment of frozen cod fillets 859 including fishery-specific environmental impacts. Int. J. Life Cycle Assess. 8, 39–47. 860 doi:10.1007/BF02978747 861 862