Introduction

The increase in organic food production and consumption is a distinct environmental-economic trend worldwide1,2. Organic food production systems depend on ecological processes, biodiversity, and nutrient cycles and aim to sustain the health of soils, ecosystems, and people3. The dynamics of organic food demand vary among countries and regions of the world, depending on economic4, environmental5,6,7 and social circumstances8. In 2021, 3.7 million organic producers were reported in 191 countries, organic agricultural land had expanded to 76 million hectares, and global sales of organic food and drink reached almost 125 billion euros9. With 48.6 billion euros, the United States continued to be the world’s leading market, followed by Germany (15.9 billion euros) and France (12.7 billion euros)9. Swiss consumers spent the most on organic food (425 euros per capita on average), and Denmark continued to have the highest organic market share, with 13 percent of its total food market9.

Organic food production has been regulated at European Union (EU) level since 1991. The EU requirements for organic food are set by regulation (EC) No 834/2007, specifying the principles of organic food production. The latest organic regulation (EU) 2018/848 including more organic foods than the previous regulations was published in June 2018 to ensure more control on environmental and economic impacts of organic food and applied from 1 January 2022.

To assess to what extent food and agricultural production systems affect the environment, a proper assessment method evaluating resource depletion issues and pollutant emissions is needed. The method most widely used to assess agricultural systems’ environmental impact is life cycle assessment (LCA)10,11. LCA is an approach that assesses the environmental impacts and resource use through a product’s life cycle12. This assessment considers flows of materials and energy and results in aggregated impact indicators for resource consumption and pollutant emissions11. Results from LCAs quantify negative impacts of food production systems, which can be used in stakeholder communication and policymaking for identifying sustainable food and agricultural production systems13,14.

However, current LCA studies on organic food face several challenges to estimate environmental impacts and tend to favor intensive agriculture and often disregard multifunctionality of agriculture11. This may be due to a lack of comprehensive operational indicators for some environmental aspects (e.g. biodiversity loss and pesticide effects) or often ignoring certain flows (e.g. soil organic carbon changes) or inconsistent modeling of indirect effects (e.g. indirect land use change)11. Besides, few databases include environmental data of organic food products. Therefore, disposing of data on the environmental impacts of organic food is important for almost all parts of society: policy makers, farmers, agribusinesses, public procurers, the media, and consumers. A review of studies on environmental impacts of organic food may provide estimations of their environmental profiles to guide environmentally friendly food choices.

Several reviews have compared organic and conventional agricultural systems that either considered a few environmental indicators or lacked statistical strength due to the consideration of a small number of papers15,16,17. Many studies showed that food products differ largely in their environmental impacts18,19,20,21,22. Some studies considered only a single environmental impact19,21 or a specific food type, such as animal-based food18,20. A study compared different organic and conventional food items for five environmental indicators, but by reporting the impacts per nutritional value and not per mass or area unit7. A review study on about 20 LCA studies considering both area and output functional units, assessed costs and benefits of organic agriculture across multiple production, environmental, producer, and consumer dimensions23. Another review on 34 comparative LCA studies14, focused on efficiency of LCA to compare environmental impacts of organic and conventional agricultural products.

However, so far, there is no systematic review on a large number of LCA studies conducted on organic food solely or both conventional and organic food, that simultaneously addresses results of LCA on organic food considering key environmental indicators in the field of agriculture per mass and area functional units.

The main aim of this review focusing on environmental LCA of organic food was 1) to identify to which extent the LCA studies cover food categories, environmental impacts, and functional units in different geographical regions, 2) to assess and compare environmental impacts among organic foods and to compare their impacts with conventional foods and 3) to evaluate impacts of LCA methodological choices on differences between environmental impact of organic and conventional products.

We conducted a quantitative review of 100 LCA studies on organic food, assessing eight key impact categories from cradle-to-farm gate. Our findings reveal that organic food has lower environmental impacts per area unit and higher land use and similar climate impact per mass unit compared to conventional. Additionally, organic farming exhibited lower potential for biodiversity loss and ecotoxicity impact. The choice of functional units influences results, underscoring the importance of considering multiple units in evaluating organic food’s environmental impact.

Materials and methods

Selection of LCA studies

To review relevant LCA studies, a selection of studies published before July 2020 was made via ISI Web of Science for 1991–2020, with the key words ‘Organic’ and ‘Life cycle assessment’. The setup of search criteria yielded a total of 2177 publications based on published scientific studies with detailed information about published items and sum of the citations in each year as presented in Fig.1.

Fig. 1: Publication trends and citations over time.
figure 1

Number of publications (a) and sum of times cited (b) per year for publications resulting from the review set up in Web of Science on 22/06/2020.Each bar represents a specific year, showcasing the publication output and citation count over time.

Full size image

Following PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-Analyses)24, selection of relevant studies was performed by including only those studies that conducted LCA of organic food or comparative LCA study, reported environmental impact values (midpoint characterization) for organic food, used real-world experimental sites, and did not scenario-model environmental impacts of organic foods. This yielded a master bibliography of 100 studies including 25 studies on organic food and 75 comparative studies including 94 peer reviewed scientific articles, three scientific reports, two conference papers and one master’s thesis. The reviewed studies are listed in Table 1, which shows the food products, type of the study, focus of the study, country of the authors and the country on which the study focused.

Table 1 Reviewed LCA studies on organic products (O) and LCA studies comparing organic and conventional products (C)
Full size table

Focus areas for review

To review the LCA studies the scheme presented in Fig. 2 was followed. As described in the first section of the Results of this study, the review was guided by an overview of geographical origin and temporal scope of the LCA studies. Moreover methodological choices including system boundary, functional unit and environmental impact were considered under the overview of the studies. Next, the environmental impacts of organic food and the comparison between and within different organic and conventional food categories were presented. The final step is the discussion of the challenges, opportunities and perspectives of the food LCA studies. Methodological choices considered in this review included “system boundary”, “functional unit” and “environmental impact category”. A “system boundary” specifies which processes and activities related to a product life cycle are considered, and which are excluded25,26. A food LCA study should include inputs to the farm and activities at the farm and may include activities that take place after the product has left the farm26,27. The data analysis of this review includes studies on cradle-to-farm gate, i.e. not including activities after the product has left the farm. The so-called “functional unit” in LCA studies is a quantitative measure of the function that is delivered by the system11,26. Functional units for food products can be based on the quantity (mass-based) or/and the land occupied (area-based). They can also reflect the quality or nutritional values of food from their calories, protein content and/or micronutrients. Environmental impacts of food products via LCA are quantified using a set of indicators called “impact categories”12. Eight impact categories were considered in this review including global warming potential (GWP100), acidification potential (AP), eutrophication potential (EP), eco-toxicity potential (ETP), biodiversity impacts (BI), energy use (ENU), water use (WU) and land use (LU). GWP is reported in carbon dioxide equivalents and includes greenhouse gas emissions (GHG) mainly in the form of N2O, CH4 and CO228. AP is an estimation of the potential increase in acidity of an ecosystem, expressed in SO2 equivalents and includes acidification potential from SO2, N2O, NOx NH3 and NO, among others. Eutrophication occurs due to the enrichment of terrestrial and aquatic ecosystems with nutrients often characterized in PO4³– or and NO3– equivalents resulted from \({{{{{{\rm{PO}}}}}}}_{4}^{3-}\), NO, NO2, NH3 and \({{{{{{\rm{NH}}}}}}}_{4}^{+}\) among others. ETP in an LCA context includes fate, exposure, and effects of eco-toxic substances on different species in soil and water29 that were aggregated in a single parameter (toxic equivalence factor) and often characterized in either Comparative Toxic Unit for eco-toxicity (CTUe) or kg 1.4 Dichlorobenzene equivalent (DB- eq.). BI is included as a mid-point impact category in agricultural LCA studies, it essentially considers the effect of farm scale activities on species diversity of plants and animals and their vulnerability30. ENU refers to the depletion of energetic resources and represents a source of GHG emissions from human activities and generally includes, but is not limited to fertilizer production, infrastructure construction and machinery use. WU refers to the quantity of blue water consumption or withdrawals. LU refers to use of land as a resource being temporarily used for cultivation of crops and feeding and housing of animals.

Fig. 2: Components of the review approach.
figure 2

The review approach includes an analysis of geographical and temporal scopes, methodological choices, environmental impacts of organic food, and discussions on challenges and opportunities in food LCA studies.

Full size image

Analysis of selected studies

Data extraction

The selected papers were grouped according to the products that were assessed and each study was analyzed according to the stages of the review approach (see Supplementary Data 1, a tabular overview). Data on the impact categories for each food product were collected considering the following criteria:

  • The study had to include either environmental impacts of organic food or a comparison of organic and conventional food. If studies were conducted on conventional agricultural systems and their results were compared with the results of other studies on organic systems, they were not considered for review.

  • The study had to report midpoint characterization results on impact categories.

  • The study that considered ENU as an impact category had to report midpoint characterization results on energy consumed for agricultural production.

Further, for a study reporting impact categories per mass unit and providing product yield of organic and conventional systems per hectare at farm gate (plant products), the impact values were calculated per unit area for data analysis.

Data analysis

To assess the environmental impacts of organic food, the statistical distributions (mean, standard deviation, and ranges) of the impacts of organic products were calculated using both studies on organic food and comparative studies (see Supplementary Information). To assess how organic and conventional products differed in their environmental impacts, products were aggregated into groups of similar types defined as food categories. Further, impacts of agricultural systems of each category were compared using an analytical approach adapted from a study17. First, the response ratio of environmental impacts for each item within each publication, denoted as RR, was calculated using the following formula:

$${RR}=\left(\frac{{impact\; of\; organic\; system}}{{impact\; of\; conventional\; system}}-1\right)$$

Thus, negative values show lower impacts and positive values indicate higher impacts of organic compared to conventional. Next, median values of the response ratios for each impact were calculated. Results of each impact were not weighted due to the small sample size per food category, therefore all cases contributed equally to the results. A Shapiro–Wilk test was used to test the normality of data related to each impact. Because not all impact ratios were normally distributed, a Wilcoxon Signed Rank test was used to determine whether the median impact ratios differed significantly from zero. Statistical analyses were performed using the R package stats from version 3.6.1 of the R statistical computing environment31.

Results

Overview of the reviewed studies

Table 2 provides an overview of the number of LCA studies and considered product categories for this review divided into studies focusing on either only organic or organic versus conventional production in different periods for publication date (2000-2006, 2007–2013, and 2014–2020) and geographic regions of the world (Asia, America, Europe, and Oceania), shows only the data extracted from the reviewed articles and not the calculated data. There was no study reported from Africa and it was therefore not included in the study regions. The studies vary in considering different impacts, system boundaries and functional units. Studies on both animal and plant products are considered. Animal products include milk, pig, cattle, lamb, seafood, chicken and eggs and plant products include vegetables, grain and cereals, fruits, nuts, and aromatic beverages (e.g. tea). Studies on “alcoholic beverages”, “breads and pasta”, “tomato paste, diced and dried tomatoes” that report their results considering both farm gate and post farm gate were considered under the fruits, grain and cereals and vegetables, respectively.

Table 2 Number of LCA studies, study regions, publication years, considered products, system boundaries, functional units and impact categories
Full size table

Spatial and temporal scope

Table 2 shows that most studies were from Europe and North America. In detail, most of the studies on organic food were from Southern Europe (35%) and for comparative studies from Northern Europe (44%). These results therefore are representative of industrialized agricultural systems and comparisons among the present studies will show differences between environmental impacts of products. It also shows that most of the organic and comparative studies included more plant products than animal products. There was in general greater focus on grain and cereals, fruits and vegetables for plant products and milk and cattle for animal products than on other products. Table 2 also shows that most of the studies (72% of organic and 49% of comparative) were recent (2014-2020). The years between 2007 and 2013 covered 20% of organic and 37% of comparative studies. The 2000-2006 period had few organic and comparative studies.

Methodological choices

System boundary

System boundary in Table 2 was reported for all publications including either cradle-to-farm gate or cradle-to-post-farm gate, but for evaluation of LCA studies, only estimations available on cradle-to-farm gate were considered. Based on Table 2, 64% and 80% of studies on organic food and comparative studies, respectively used cradle-to-farm gate system boundary.

Functional unit

Based on Table 2 more studies used mass-based than area-based functional unit (Table 2). This included 68% of studies on organic products and 65% of comparative studies. 24% and 30% of the studies used both mass- and area-based functional units for organic and comparative LCA studies, respectively (Table 2). Three comparative studies on fruits32,33,34 considered only area-based functional unit and two organic food studies on grain and cereals used protein-based functional units35,36.

Environmental impact

All LCA studies on organic food and 96% of comparative studies evaluated impacts of products and their agricultural systems on GWP for a 100-year time horizon. However, they considered slightly different characterization factors for CH4 and N2O from 2000 until now, which could not be modified to obtain GWP of all reported products with same characterization factors (Table 2). Few of these studies considered soil carbon sequestration (SOC) (Table 2). Of the organic LCA studies only two papers37,38 and of comparative studies only six papers39,40,41,42,43,44 included SOC in their assessment. Some other studies considered emissions from dLUC45,46,47,48,49,50,51,52.

Of 50 studies that considered AP, 11 and 39 were related to organic and comparative studies, respectively and most of these studies were on plant products (69% of organic studies and 55% of comparative studies) (Table 2).

The publications reviewed here covered different ways of reporting EP, because current LCA characterization models either use a single or combined impact category for terrestrial, marine, and freshwater environments53. Most recent studies report their results in separate EP categories33,51,54,55,56,57,58. EP in Table 2 was reported for all publications, without considering the methods that were used but for evaluation of LCA studies and their relevant data analysis only estimations on NO3- equivalents were considered. 48% of the LCA studies on organic food and 54% of the comparative LCA studies considered EP (Table 2).

Of the 100 studies, only 26 analyzed the water toxicity related impacts and 86% of them concerned plant products (Tables 2 and 3). Four studies focused on eco-toxicity potential of animal products44,58,59,60 and use of pesticide or active substances were reported in six studies61,62,63,64,65,66.

Table 3 LCA studies that included eco-toxicity potential an impact category in their analysis
Full size table

The complexity of biodiversity in the broadest sense of the Rio Convention cannot be totally measured and, even for agroecosystems, a single impact category that reflects impacts on a wide range of organisms due to farming operations is not likely to be devised30,67. For this reason, few studies on biodiversity impacts of agricultural production systems were available (Table 4). Of the four LCA studies that included BI of food products, two studies47,52 focused on organic milk and rice production systems, respectively, and the two others reported on both organic and conventional milk production systems (Tables 2, 4).

Table 4 LCA studies that included biodiversity as an impact category in their analysis
Full size table

Of the 13 LCA studies that reported WU (Table 5), 85 % were related to comparative studies mainly on plant products (73%) and 15% were just on plant-based organic food (Table 2).

Table 5 LCA studies that included water use as an impact category in their analysis
Full size table

Of the 35 comparative LCA studies that reported ENU, the highest number of analyses for energy use was for grain and cereals (33%) (Table 2). LU in reviewed studies was not giving information on soil quality aspects and similar to ENU, was more frequent in comparative studies (33%) than in studies on organic products (20%) (Table 2).

Evaluation of LCA studies

Supplementary Table 1 and 2 of Supplementary Information show an overview of the reported environmental impacts of organic food from cradle-to-farm gate per mass and area-unit, respectively, extracted from both LCA studies on organic food and comparative LCA studies (See Supplementary Information for detailed results on the comparison of environmental impacts among organic food per mass and area units). Figures 3 and 4 show an overview of the observed variation of food environmental profiles per mass and area units from cradle-to-farm gate and their associated response ratio, respectively, in different agricultural systems extracted from only comparative LCA studies as well as the data per unit area calculated based on the values per mass unit and yield per hectare. As most LCA studies have been conducted with a strong focus on GWP considering mass-based functional unit, a more detailed comparison of the carbon footprints of organic and conventional food in kg of product is presented in Fig. 5.

Fig. 3: Comparison of cradle-to-farm gate impacts of organic and conventional agricultural products.
figure 3

Comparison of cradle-to-farm gate impacts of agricultural products from organic and conventional systems, measured both per mass (kg) and per area (ha) units. The impact categories include GWP: global warming potential, AP: acidification potential, EP: eutrophication potential, ENU: energy use, and LU: land use. Food categories are BE: cattle meat (beef), LA: lamb, PI: pig, EG: eggs, CH: chicken, SE: seafood, DA: milk, GR: grain and cereals, VE: vegetables, FR: fruits, NU: nuts. The results of studies in each plot are sorted by the average of the organic food impacts. The environmental impact of beef, lamb, pig and chicken are not separated for live and carcass weight, for the aim here is a presentation of the differences between environmental impact of organic and conventional systems. A base-10 log scale is used for the Y-axis of all the graphs.

Full size image
Fig. 4: Response ratios for impacts of organic and conventional products.
figure 4

Response ratios for impacts of organic and conventional products expressed per mass (a: for GWP, AP, EP, ENU and LU, b: GWP, c: AP, d: EP, e: ENU, f: LU) and area (g: for GWP, AP, EP, ENU and LU, h: GWP, i: AP, j: EP, k: ENU) units. Impact categories are GWP: global warming potential, AP: acidification potential, EP: eutrophication potential, ENU: energy use, LU: land use. Response ratios for nuts (two values) are not shown in B and H sections of this figure. Box plot boundaries represent 25 and 75 percentiles, thick line within the box represents the median and whiskers represent 10 and 90 percentiles. Dashed line represents the zero-response ratio where the impacts of organic and conventional products are the same. Positive values indicate higher impacts from organic products and negative values indicate lower impacts from organic products. Numbers within each plot show the number of cases for each impact category or product category. ns= not significantly different from zero (Wilcoxon Signed Rank Test P > 0.05), **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image
Fig. 5: Carbon footprint comparison of organic and conventional foods.
figure 5

Carbon footprint for organic and conventional food (kg CO2 equivalent per kg) based on comparative studies of the same product. (a) includes data for all food items, i.e., ranging from plant-based foods like fruits and vegetables at the low end of the scale to animal foods such as pork and beef at the high end. (b) zoomed in on the lower end of the scale on plant-based food products and milk. The black line marks where the carbon footprint of organic and conventional foods is the same. Points below the line indicate a larger carbon footprint for organic food compared to conventional, while points above the line indicate a higher carbon footprint for conventional food compared to organic.

Full size image

Environmental impacts of organic and conventional production systems

Global warming potential (GWP) per mass unit

Cattle and lamb showed the highest average GWP for both production systems, followed by pig, eggs, nuts, seafood, chicken, milk, grain and cereals, fruits, and vegetables (Figs. 3, 5a, b). When assessing per produced mass unit, organic and conventional did not significantly differ in their GWP\(\left(p=0.0924\,,n=125\right)\). The median response ratio for GWP was −0.057 (Fig. 4a) indicating that organic products on average had 6% lower GWP per mass unit than conventional. The median response ratio was also close to 0 for the different food categories (Fig. 4b), except for nuts that had higher impacts for organic and seafood and lamb that had lower impacts. This was however based on very few studies (n = 2 for nuts, n = 2 for seafood, n = 1 for lamb) (Fig. 3).

Global warming potential (GWP) per area unit

Dairy production systems had the highest GWP per area unit for both production systems (Fig. 3). A significant difference between organic and conventional systems was observed for GWP per area unit \(\left(p=0.0002\,,n=59\right)\) and the median response ratio for GWP was −0.22 (Fig. 4g), i.e. 22% lower impact of organic products per area unit. However, some studies on grain and cereals68,69, fruits70,71 and nuts42 showed lower GWP per ha for conventional systems than organic (24% of the studies) (Figs. 3 and 4h).

Acidification potential (AP) per mass unit

Based on Fig. 3, animal products (except milk) had higher AP for both production systems compared to plant products. The median response ratio for AP was zero and the difference between organic and conventional was not significant \(\left(p=0.7344\,,n=52\right)\) (Fig. 4a). There were differences in the median response ratios between food categories (Fig. 4c). Most studies on organic milk (86%), pig (67%), chicken (60%), grain and cereals (67%) showed lower AP whilst more studies on organic eggs (100%) and vegetables (60%) had higher AP than conventional (Fig. 4c). Manure had the greatest impact on the AP of both chicken66 and eggs production72 because of NH3 emissions73. A study on seafood reported a larger impact of conventional shrimp than organic50, while a study on lamb showed larger AP of organic64.

Acidification potential (AP) per area unit

The difference between organic and conventional systems for AP per area unit was not significant \(\left(p=\,0.1243\,,n=23\right)\,\) (Fig. 4g). Nevertheless, the median response ratio was −0.38 (Fig. 4g). The exceptions were some studies on grain and cereals68,69,74 and on fruits34,54 that showed lower AP per ha for conventional than organic (26% of the studies) (Fig. 4i). The higher AP for organic systems could be explained by model emission factors for NH3 volatilization, which are higher for organic N fertilizers than for mineral N fertilizers75.

Eutrophication potential (EP) per mass unit

Similar to AP, cattle and lamb had the highest EP per mass unit followed by eggs, chicken, pig, grain and cereals, vegetables, milk and fruits (Fig. 3). No significant difference between organic and conventional systems was observed for EP \(\left(p=\,0.2151\,,n=51\right)\), the median response ratio for EP was 0.038 (Fig. 4a). However, 49% of the studies showed lower EP for organic food compared to conventional (Fig. 4a). Larger EP for organic products as compared to conventional system was mainly due to lower yields of both animal and plant products. Most organic milk (82%), pig (67%) and fruits (80%) had lower EP than conventional, whereas organic chicken (80%), eggs (100%), grain and cereals (56%) and vegetables (100%) had higher EP than conventional (Fig. 4d).

Eutrophication potential (EP) per area unit

The highest EP per unit area for both production systems was found in fruit, milk, and pig (Fig. 3). The difference between organic and conventional was significant\(\,\left(p=\,0.0091\,,n=24\right)\) and the median response ratio for EP was −0.47 (Fig. 4g) showing 47% lower EP in organic systems. However, 21% of the studies, including studies on grain and cereals68,69,74 had higher EP for organic than conventional (Fig. 4j). Similar to AP, higher EP for organic systems can be due to higher emission factors for NH3 volatilization of organic N fertilizers compared to mineral N fertilizers. These factors affect NH3 and NOx emissions and indirect N2O emissions, contributing to AP, EP and GWP.

Energy use (ENU) per mass unit

Figure 3 shows ENU per mass unit with the highest average for cattle, lamb and egg followed by pig, chicken, fruits, milk, grain and cereals, and vegetables. Significant differences between organic and conventional products were not found for ENU \(\left(p=0.1008\,,n=63\right)\). Although the variation of ENU was important (from 62% lower to 38% higher ENU in organic products), the median ENU showed 13% lower ENU per mass unit in organic than conventional products (Fig. 4a). Larger ENU in conventional products can be caused by the large amount of energy required for production (and transport) of mineral fertilizers17. Most organic milk (89%), cattle and lamb (100%), and grain and cereals (94%) had lower ENU than conventional, whereas organic pig (67%), chicken (100%), eggs (50%), vegetables (67%) and fruits (63%) had higher ENU than conventional (Fig. 4e).

Energy use (ENU) per area unit

Figure 3 shows a higher average ENU per unit area for fruits and vegetables followed by pig, grain and cereals, and milk. Organic and conventional products differed significantly for ENU per area \(\left(p \, < \, 0.0001\,,n=17\right)\). The median response ratio for ENU was −0.32 (Fig. 4g), showing 32% lower ENU per area unit in organic than conventional products.

Land use (LU) per mass unit

Animal products (except milk) had higher LU per kg for both production systems than plant products (Fig. 3). Production of 1 kg of meat from cattle in both production systems required more land (m2) than the production of 1 kg of meat from lamb, pig or chicken (two, four- and six times more, respectively). Organic and conventional products differed significantly\(\,\left(p \, < \, 0.0001\,,n=46\right)\), organic products required 64% more land than conventional products (Fig. 4a) mainly due to lower yields17,76 Organic milk, meat of pig, cattle, lamb, and chicken, eggs, vegetables, grain and cereals and fruits had 51%, 73%, 83%, 126%, 195%, 80%, 86%, 48% and 108% more LU respectively, than conventional (Fig. 4f).

Eco-toxicity potential (ETP)

Table 3 generally shows lower ETP for organic products than conventional, mainly due to the application of synthetic pesticides in conventional systems. However exceptions were studies on organic fish58, pig59, olive71,77 apple70, grape33, tomato78, bean69, rice79 and wheat bread55. Higher toxicity impacts in some studies can stem from lower yields in organic farming and the use of copper sulfate, that have high characterization factors in some impact assessment methods80.

Biodiversity impacts (BI)

The BI comparing organic and conventional products via LCA were considered in only three studies on milk production (Table 4). A study81 found positive impacts of organic farming in the indicators number of grassland species, grazing cattle, layout of the farm and herd management, but indices in these categories showed a wide range and were partly independent of the farming system. In agreement with that, a biodiversity assessment by a study44 showed that on average the biodiversity loss from organic milk were 33% of the conventional. Furthermore, a study47 compared BI of organic and conventional milk for twelve farms in Denmark, Italy and Germany and showed lowest impact on biodiversity loss for the organic milk, due to the higher share of grassland in their system.

Water use (WU)

Table 5 generally shows lower WU for organic products than conventional with the exceptions of organic chicken66 and olive77. Further, a study82 showed equal amount of water use for both systems.

Discussion and conclusions

Geographical coverage of organic LCA studies

The present review showed that LCA studies are not available for all organic food products from all world regions. Most LCA studies (74%) are based on European production, indicating the need for studies in Africa, South America, Oceania, and Asia where the production conditions are different and organic production is increasing83. Furthermore, organic LCA studies have focused primarily on milk, cereals/grains (in Northern Europe) and fruit (in Southern Europe), showing the lack of studies on environmental impacts of other food categories such as e.g. nuts, seafood, chicken, and vegetables that might be a major part of future diets2. Therefore, with introduction of new dietary choices, there is a need for more studies on LCA of organic foods both in developed and developing countries that would provide further insight and a clearer overview of the environmental profile of organic foods compared to other food systems globally.

Environmental impacts of food products

The environmental impacts of our food system are considerably affected by our dietary choices7,14,17,18,22. In general, meat from ruminants (e.g. cattle and lamb) had the highest impacts per mass unit and other livestock products (e.g. pig, eggs, chicken, seafood, and milk) had intermediate impacts that were higher than those of most plant products (Fig. 3 and Supplementary Table 2). It should be noted that impacts shown in Fig. 3 and Supplementary Table 2 only include GWP, EP, AP, LU and ENU and not e.g. BI and ETP. A study84 showed that, in conventional systems, chicken and pork have larger freshwater ecotoxicity impacts than beef and milk. The present study further showed that organic products generally had lower impacts per unit area compared to conventional products mainly due to lower emissions associated with the inputs used.

The differences between the organic and conventional production systems are mainly based on differences in regulations, where nutrient inputs to organic agriculture are mainly from nitrogen fixation and manure, and conventional agriculture largely depend on synthetic fertilizers and pesticides. However, in both organic and conventional production systems, there may also be differences due to differences in local conditions, intensity levels, type of fertilizer, weed and pest management practices, use of catch crops etc. The trade-off between inputs and yields resulted in almost similar GWP per mass product unit from organic and conventional systems. Manure application85,86 and more diversified crop rotation, often including temporary grassland (ley)87 have the potential to increase soil organic carbon in organic systems showing potential to lower GWP. However, the higher SOC and the lower GWP may be compensated by lower yield in organic systems76. It should be noted that the allocation of emissions caused by the application of manure remains as a general allocation challenge in the LCA of agricultural systems, which has been handled using different approaches. In some cases, all emissions and positive effects (e.g. SOC changes) are allocated to livestock, and in other cases only emissions from stables and storage are allocated to livestock and emissions from application to plant production.

The ENU per area unit in organic system was lower compared to the conventional system mainly due to lower dependency on energy intensive production of synthetic fertilizer and pesticides. The ETP of organic systems generally showed lower impact compared to conventional systems especially when measured per hectare due to lower pesticide inputs and alternative pest and disease measures, but there were also exceptions dependent on the differences in the management of the organic and conventional system and the functional unit (ha or kg). The BI of the organic systems was lower than that of the conventional systems, which is in accordance with other studies88,89,90, but the number of studies was still low. Larger LU for organic systems may result in more natural habitat conversion that increases biodiversity loss and decreases carbon stocks7, but other indirect and rebound effects in the organic food consumption patterns might go in the other direction as discussed in a study11.

Environmental impacts captured by LCA and the need for improvement

LCA has become an established and important tool for assessing environmental sustainability in food and farming systems. It can help to improve production systems, provide a basis for political decision making, deliver consumer information and compare agricultural production systems. In LCA of agricultural systems, environmental impacts are assessed by considering a set of indicators and referenced to agricultural product as the functional unit (mass, area, nutrition value, energy). This reference to food products is an expression of the environmental impact resulting from the production of a certain amount of food product and can be seen as a measure of eco-efficiency91, where the aim is to identify which system can provide the same amount of food product with the lowest environmental impacts. However, LCA faces some challenges when assessing multifunctional agricultural systems such as organic agriculture, where more research is needed. LCA is focused on the supply of products (e.g. crops and animals) by the agricultural system11. In addition, more research is still needed to model potential biodiversity loss, pesticide effects and changes in soil organic carbon in LCA. For this reason, although the use of pesticides affects both toxicity and biodiversity impacts, BI and ETP of food systems were rarely considered in LCA of food products. In this review, only 3% and 21% of LCA studies comparing conventional and organic agriculture considered biodiversity and eco-toxicity impacts, respectively. In a meta-analysis study88 showed 30% higher species richness for organic systems compared to conventional systems. The approach to include biodiversity in LCA suggested by a study92 as recommended by the EU PEF is not suited for comparing impacts on biodiversity of the different systems, as it can only be used for identifying the hotspots within the product systems11. Although a study93 provides characterization factors to differentiate impacts of organic and conventional agriculture on biodiversity, there is still room for improvement of the LCA method to consider the effect of different land management practices on biodiversity in both agricultural systems11. Assessment of eco-toxicity impacts of agricultural systems is also limited by both lack of data14 and the need for several decades of use of a given pesticide to fully understand its health and toxic effects94.

LCA studies on food products to target reduction of negative environmental impacts need to adapt to local geographic and climatic conditions considering different scales. What might seem as an effective mitigation option at the global scale when assessed per kg, might not be an efficient mitigation option at the local scale when assessed per ha. Mass-based functional units work best for global impacts such as GWP, while for impacts that can have large local-scale impacts such as eutrophication, land-based functional units are important for the assessment. For this reason, it is important to use functional units based on both mass and area in LCA studies of agricultural systems. Furthermore, the effects at local scale might be affected by impacts at wider regional scale. Limited attention has been paid to spatially oriented assessments in relation to agri-food LCA studies95,96,97,98,99,100. There is therefore, still considerable scope for progressing in the explicit assessment of spatial variation in LCA using suitable and complementary models (e.g. both land scape and field scale models), databases and technologies such as geographic information systems100,101.

The most widely considered impact on farm scale is GWP including estimations on GHG emissions from different farming activities. Most of the studies analyzing the environmental impact of food products have not included land use change (LUC) for estimation of GHG. Though, some studies on food products have included direct land use change (dLUC) in their analysis but with slightly different approaches. Whether dLUC is included or not as well as the method chosen to quantify LUC can lead to large variations in LUC factors, and thus highly influence the GWP results of different intensities of food production systems11.

So far, soil carbon sequestration has been included in a few studies using different methodologies and time horizons. Land degradation, including processes such as erosion, compaction, salinization, and soil organic carbon loss is another neglected issue in LCA studies11. This is while land degradation is estimated to affect 90% of the soil globally by 2050102. A global indicator of the land degradation could be the percentage change in soil organic carbon103.

Ignoring important impact categories such as biodiversity, eco-toxicological and soil quality impacts and water and resource use in comparative LCA studies is problematic since the overall conclusions will be strongly affected by this choice. There is a need for methodological development regarding biodiversity and toxicity assessment. Further in relation to WU, most of the studies focus on the assessment of off-stream consumptive use (i.e. the part of withdrawn water from a groundwater or surface-water source that is evaporated, transpired, incorporated into products or crops, consumed by humans or livestock, or otherwise not available for immediate use) of blue water while other types of water use are underrepresented104. In addition, since different watercourses perform different functions, more detailed inventories and impact pathways should be considered in assessing water use105.

It is also needed to provide more data on organic food systems for LCA inventories, because lack of data might cause under- or overestimation of environmental impacts and challenges in thorough comparisons between organic and conventional systems14.

Uncertainties in estimates of environmental impacts of agricultural systems and food products may arise from several sources, such as the way current agricultural systems are conceptualized in LCA models, the way environmental impacts are estimated or the setup of the models and input data. Within all these issues there are also multiple interactions that may contribute to uncertainties.

The current review shows that most LCA studies assess the environmental impacts per mass unit and not per area unit. Certainly, LCA results are highly sensitive towards the choice of the functional unit, as observed in this and other review studies comparing organic and conventional systems14. Better performance of organic farming per unit area (e.g. for biodiversity and eco-toxicity) is the main reason for policies favoring a transition to organic farming systems. However, organic agriculture is less productive than conventional agriculture, resulting in similar values per unit product for most environmental impacts. Thus, focusing only on impacts per unit of product may result in decisions in favor of conventional food production systems that may increase negative environmental impacts in the farming region11. Although nutritional quality of organic food may be better7, expressing impacts of single organic and conventional food items per mass unit ignores food quality and the dietary patterns that they are part of. Therefore, the need to develop a more refined functional unit that represents the actual functions of foods is obvious. The use of nutrient-based functional units (e.g. unit protein) has been advocated by several authors106,107. The use of a protein-based functional unit may produce LCA results that more accurately reflect impacts of the actual function of foods than when mass is used as the functional unit. However, providing protein is not the only function of food, and the search for a perfect nutritional functional unit has been a key challenge for LCA of food108. Furthermore, it may significantly increase data requirements107. Assessing the environmental impacts at the diet level could link farm management decisions to diet-level environmental impacts considering an enhanced focus on human nutrition across the entire value chain109. Furthermore, including the entire diet could bridge the gap between diet-level and product-level and provide implementable action plans for both consumers and producers and at the same time take into account that organic consumers tend to lower their intake of animal-based food110,111.