Knowledge for better food systems

If farm animals only graze pastures and eat by-products – livestock problem solved?

Elin Roos's picture

This post is written by FCRN collaborator Elin RöösElin is a postdoctoral researcher at the Swedish University of Agricultural Sciences working for the Future Agriculture initiative at the same university, which is a strategic multidisciplinary research platform that addresses the sustainable use of natural resources with emphasis on agricultural production and food systems.

About 40% of the cereals and legume grains produced every year are used to feed farm animals.  Many commentators argue that this is highly resource inefficient as around 70% of the human edible energy produced is lost in this process due to metabolic losses in the animals (Papargyropoulou et al., 2014). For example, typically 4 kg of cereals and legumes is needed to produce one kg of edible poultry meat (Röös et al., 2014), although the quantity varies by system.  By contrast raising livestock on by-products that humans can’t or don’t want to eat or waste is often considered to be a resource efficient way of producing protein for human consumption, as is raising livestock on biomass from grasslands unsuitable for the production of human edible foods e.g. natural or semi-natural pastures or other marginal grasslands.  (For a deeper discussion of the concept of efficiency please refer to the FCRN paper on efficiency).

This approach of letting the ecological resource capacity be the constraining factor for livestock production, i.e. to restrict livestock production to feeds not suitable for human consumption has been around for some time and was coined as ‘producing livestock on ecological leftovers’ in a paper by FCRN’s Tara Garnett in 2009 and explored further as one of the scenarios presented in Gut Feelings  (2015).

This concept of restricting livestock production is attractive in several ways. From a food security perspective it has been argued that diverting grains from animal feed to human food would increase available food supply. Further, the concept is compatible with principles of agro-ecology (e.g. adjusting production to context-specific conditions and making use of local resources) and makes use of pastures for food production which a vegan diet would not. It also yields a diet which would contain meat and dairy; products that most people appreciate in their diets. However, despite the popularity of framing sustainable meat production in this way, there are few studies that have looked into how a diet based on this concept would actually look like and what the environmental impacts of such a diet would be. This said, recently Schader et al. (2015) modelled various scenarios of global livestock production in which reliance on food-competing feed crops was progressively reduced.  They found that greenhouse gas emissions and other environmental impacts would be reduced (compared with the reference scenario), and enough food would be supplied although only 11% of protein would come from animal sources, as compared with 38% in the reference scenario.

In a study published late last year my colleagues and I made an attempt to apply this concept to Sweden and see what kind of diet that would give (Röös et al., 2015b).

As a first step we formalised the general ecological leftover (EL) principles a bit further.  These are:

  1. Arable land should primarily be used for the production of plant-based food for humans.
  2. Livestock should be fed biomass not suitable for or wanted by humans.
  3. Grasslands should be used for livestock production if grazing can be justified by reasons other than meat and milk production, e.g. biodiversity conservation, providing a livelihood for vulnerable populations etc.

In principle 3 we introduced a further specification of the concept of ecological leftovers for two reasons.

First, grasslands are only ‘leftover’ if viewed from a human food consumption perspective - there are other legitimate uses for this land, such as bioenergy production or wildlife conservation. In other words, just because there is grassland and it could be used for livestock rearing this does not mean that livestock production is necessarily the preferred use for this land.

However - and this is the second point - if livestock production on grassland does bring additional benefits (i.e. in addition to the production of food) like biodiversity conservation, or supporting the livelihoods or food security of poor and vulnerable people, then this arguably justifies having livestock production on these lands.  Through this additional requirement we therefore recognise the important role that grazing livestock have in some contexts and the important role that grassland could have for purposes other than livestock production.  It would be very interesting to hear the FCRN community’s views on this line of reasoning! (Log in to submit a comment below).

Description: purpose in setting out the general EL-principles above is that it should be possible to use them to develop and define sustainable diets based on the EL-concept in most contexts and regions. However, the principles need to be general enough to be applied very differently depending on the situation e.g. based on the various roles of livestock for biodiversity conservation, food security and livelihoods in the region, on the need to grow winter feed, on legal restrictions and on cultural preferences e.g. when it comes to what is considered an inedible by-product etc. For example, we considered whey as a by-product and used it as feed, but whey could be and is processed into protein powder and other food products intended for direct human consumption.  

In our case, Sweden, we applied the EL-principles in the following way:

  1. Arable land is used to produce all crops for human consumption except those that for climatic reasons cannot be cultivated in Sweden (exotic fruit, coffee, tea and cacao) which are imported. It is also necessary to use some arable land for the production of winter feed for grazing animals and feedstuffs to supplement by-products in the diet of monogastric animals in order to ensure that their diets are nutritionally adequate (by-products contain enough fibre and protein, but not energy).
  2. By-products from food processing and from crop production are fed to animals according to current practices and regulations, e.g. feeding consumer stage food waste to livestock is not allowed.
  3. Semi-natural pastures and meadows are grazed by livestock to support biodiversity preservation; Sweden has a national goal to preserve the current area of semi-natural pastures.  These currently constitute only one percent of the Swedish land area and represent the most threatened eco-system in Sweden.

Following these principles and a requirement that the diet should be nutritionally adequate (fulfil the Nordic Nutrition recommendations) and be culturally acceptable (i.e. resemble current Swedish eating patterns as far as the non livestock components are concerned) we calculated three different diets based on the EL-principles as applied to Sweden.

We also decided that for the diet to be ‘fair’ (i.e. follow the principles of Contraction and Convergence), the use of arable land should not exceed the global per capita availability of arable land of 0.21 ha.  This means that the 2.6 million hectares of available Swedish arable land should provide 13.5 million people with food. In comparison, the Swedish population was 9.8 million in 2015.

In the three different diets, pastures and by-products are utilised differently reflecting different perspectives on how pastures should be ‘best’ managed:

An intensive dairy production system.

This scenario (using only heifers from intensive dairy production to graze the pastures) was based on the viewpoint that using the full genetic potential of the currently available high-producing dairy cow breeds to maximise milk production, and hence reduce impacts per unit of produce, is the most efficient way of using feed biomass.  In other words it is a ‘waste’ to let highly productive dairy cows graze the pastures that contain biomass of varying and sometimes low nutritional quality and that are often located at some distance from the farm, making it difficult to keep high-yielding dairy cows there. As a consequence a lot of cattle (1.8 million – compared to 1.5 million currently) are needed to manage the available pastures (1 % of Swedish land area) as only heifers graze these.    This scenario is a little bit more intensive than the current average Swedish system and can be considered a business as usual scenario except that all feed is produced within the country - at present Sweden imports substantial amounts of soy and other protein feeds.

In this scenario mainly heifers, and not the high-yielding dairy cows (9,300 kg energy corrected milk; ECM per cow and year), are used to graze the pastures. The main feed for the cattle is a grass-clover crop grown on arable land. In addition, all suitable by-products from the production of plant-based foods in Sweden are used as feed for mainly the dairy cows and supplemented with cereals to maximise milk production. All male calves are kept as intact bulls to utilise their faster growth rate (compared with steers) and so reduce their lifetime enteric methane emissions. The bulls are not kept on semi natural grassland due to the potential danger of keeping them far from the farm. This scenario yielded a lot of milk, enough to supply current recommended levels in the diet and to export substantial additional amounts. This dairy production also delivered 150 grams of beef meat per person /week from culled dairy cows and offspring raised for meat. No eggsor poultry meat were produced in this scenario and only 70 grams of pork per person / week, as by-products that could have been used for poultry and pork production were used to boost milk production.

An extensive dairy production system.

This extensive dairy production, in which cows are fed forage only, i.e. some through grazing but most from grass-clover grown on arable land (winter-feed), was motivated by the principle that cattle should be fed only on roughage and not on foodstuffs edible by other animals as cattle possess the unique ability to turn cellulose-rich feed into energy. Based on this view, feeding ruminants grain is just ‘wrong’ even if it raises milk yields.

In this scenario pastures are grazed by dairy cows that are managed less intensively using only grass and no grain or concentrates, producing 5,000 kg ECM per cow annually, and with a longer annual dry period in which they are not milked at all and hence kept on pastures. All male calves are raised as steers and graze the semi natural pastures together with heifers. A small quantity of by-products is fed to young cattle to ensure adequate growth rates but most by-products are used to produce eggs and pork. A total number of 870,000 cattle are sufficient to graze the Swedish pastures. The milk produced is enough to supply one litre of liquid milk per person / week, plus 70g cheese, approximately half the amount compared to the current consumption. When it comes to animal products this system yields: 70 grams of beef, 320 grams of pork, 14 grams of poultry (from culled hens and male chickens from egg laying systems here raised for meat) and three eggs per person /week; here by-products are used to produce pork meat instead of boosting milk production.

  1. Description: pastures using suckler herds.

This scenario was included as it is the least climate impacting way of preserving the pastures due to the need for fewer animals.Suckler herds can be kept in the pastures throughout the year and therefore fewer animals are needed compared to dairy herds, where animals spend considerable time inside. In addition to keeping methane emissions down by keeping fewer animals, less arable land is needed to grow winter feed.

By using pure suckler herds for grazing semi-natural grassland, the number of cattle needed to preserve the Swedish pastures is just slightly over 540,000. No milk is produced in this scenario, but 40 grams of beef per person per week are obtained. All by-products are used to produce eggs and pork: three eggs and 350 grams of pork per person per week.

 In all scenarios the amount of meat in the diet is substantially reduced; from the current 10 servings per week to 2-4 servings per week. To compensate for this reduction legume consumption would have to increase by approximately 200 -700%; from hardly anything in the average Swedish diet today to 2-4 servings per week. The quantity of protein rich foods available in the different diets is summarized in the table below.

Table 1. Amounts of protein-rich foods in the different diets based on the ‘ecological leftover’ approach for Sweden compared with current consumption patterns


Current consumption

Scenario 1: Pastures grazed by heifers from

intensive dairy herds

Scenario 2: Pastures grazed by cows, steers and heifers from extensive dairy herds

Scenario 3: Pastures grazed by cows, steers and heifers from suckler herds

Servings of meat per week

(100 g bone-free per serving)





Servings of seafood per week

(100 g bone-free per serving)





Servings of legumes per week

(60 g dry weight per serving)





Servings of cereals per week

(60 g dry weight per serving)






Eggs per week





Slices of cheese per day

(10 g per slice)





Millilitres of milk per day





Total protein per day (g)

Approx. 100




Total protein % of total energy

Approx. 17%




1This scenario supplies more dairy than these amounts, but consumption was capped to the recommended level.

Greenhouse gas emissions in all dietary scenarios were  in the range 0.4-0.6 tonne of CO2e per person per year, considerably lower than the current level of approximately 1.9 tonne tonnes CO2e (Röös et al., 2015a). Land use stayed well within the ‘fair’ share of 0.21 ha per person/ year in all scenarios which made it possible to use spared land for bioenergy production (eg. biogas from ley) making agriculture self-sufficient in energy. Nitrogen and phosphorous additions were also heavily reduced compared with today and the most threatened ecosystems in Sweden, the semi-natural pastures, were preserved. All in all, these Swedish diets based on the EL-concept for Sweden seem to provide a clear step towards more sustainable eating patterns.

But how relevant is it to design diets based on the EL-concept on a national or regional level? A global estimate, such as that provided  by Schader et al. (2015), gives a notion of the amount of animal products that can be produced on a global scale and that we, the global population, could share among us through free trade. That is meaningful information. Using neo-liberalist theory the most efficient (again see here for more on efficiency) way of organising the food system would be by incentivising free trade to push each region to produce what it can do best (a form of ecological comparative advantage).  By this logic shouldn’t Sweden just stick to producing cereals on the plains in the south and forages for milk and beef production further north, and continue to import protein-rich feed, as well as fruits and vegetables? Or, to take this reasoning one step further, just grow forest and import food. And what about regions or countries with either just arable land or just pastures: should the former be allowed no meat and the latter only meat?  

Despite these obvious limitations with applying the EL-concept on a national/regional basis, our study has gained quite some interest in Sweden. I have been asked to develop a diet for the Nordic countries as well as for smaller regions in Sweden based on the same concept. Why this interest? There may be several explanations at different levels of detail. First, from an overarching perspective, the number of Swedish farmers is diminishing at a rapid rate due to structural changes within the agricultural sector (fewer but larger farms) and financial problems within the livestock sector (difficult to compete with imported products due to higher labour and building costs as well as stricter animal welfare and environmental regulations), and consequently the share of the consumed food that is produced in Sweden has gone down considerably. This development together with the political decision to develop a food strategy for Sweden (meant to be launched in March this year) has spurred a vigorous discussion about self-sufficiency rates; the survival of Swedish farmers; a moral responsibility to supply the Swedish population with domestic foods and not export environmental impacts and drive deforestation by food imports; the conservation of semi-natural pastures for biodiversity reasons and the continued use of arable land for production purposes (climate change could improve crop production conditions in Sweden so some argue that it would be foolish to abandon this land); and the sustainability of the food system in general.  As a more detailed example, whenever the need to reduce meat consumption (Swedes consume an average of some 85 kg of carcass weight meat per year), and especially beef (Sweden is one of the top consumers of beef in the EU), is mentioned the immediate fear raised is the threat to the semi-natural pastures that this could potentially entail due to fewer grazing animals - even though 50 % of beef consumed is imported and a substantial amount of cattle are raised indoors. Hence, there is great interest in understanding more about eating patterns that are nutritionally adequate and could reduce GHG emissions and preserve semi-natural pastures. Our study provides quantitative knowledge of the environmental impacts, the land use and type of foods of a few such diets based on different views on the best use of the pastures. We hope that it can make discussions on the sustainability of the Swedish food system more nuanced and informed.

We very much welcome your views on this blog and the paper.

You can submit a comment below by logging in as a member. Do not hesitate to get in touch if you have forgotten your login details. We encourage you to register as a member (for free) to make use of the other networking benefits on the website.

For more, you can see the original research paper by Röös et al as well as three other papers that explore these issues, but at different scales/in different contexts, and underpinned by somewhat different assumptions; see our summary of Schader et al, our summary of the paper by De Oliviera Silva et al, as well as Call for conservation: abandoned pasture.


Garnett, T. 2009. Livestock-related greenhouse gas emissions: impacts and options for policy makers. Environmental Science & Policy, 12(4), 491-503.

Papargyropoulou, E., Lozano, R., K. Steinberger, J., Wright, N., Ujang, Z.b. 2014. The food waste hierarchy as a framework for the management of food surplus and food waste. Journal of Cleaner Production, 76, 106-115.

Röös, E., Ekelund, L., & Tjärnemo, H. (2014). Communicating the environmental impact of meat production: challenges in the development of a Swedish meat guide. Journal of Cleaner Production, 73(0), 154-164. doi:

Röös, E., Karlsson, H., Witthöft, C., Sundberg, C. 2015a. Evaluating the sustainability of diets–combining environmental and nutritional aspects. Environmental Science & Policy, 47(0), 157-166.

Röös, E., Patel, M., Spångberg, J., Carlsson, G., Rydhmer, L. 2015b. Limiting livestock production to pasture and by-products in a search for sustainable diets. Food Policy. Accepted.

Schader, C., Muller, A., Scialabba, N.E.-H., Hecht, J., Isensee, A., Erb, K.-H., Smith, P., Makkar, H.P.S., Klocke, P., Leiber, F., Schwegler, P., Stolze, M., Niggli, U. 2015. Impacts of feeding less food-competing feedstuffs to livestock on global food system sustainability. Journal of The Royal Society Interface, 12(113).







intolife's picture
Submitted by intolife on

Hi Elin, very interesting to read this.  With your work on developing a diet for the nordic countries, how much of this is related to norway?  these is an ongoing question here with regard to appropriate land-use and meat production (maybe even more polarised than in Sweden due to Norway's limited farmland).  It would be interesting to see these scenarios modelled against the Norwegian system :-)

Will (Oslo)

Will Nicholson


Food Sustainability in the Restaurant Industry