Start a new topic

Modeling food and agriculture in En-ROADS

I'm looking forward to using En-ROADS in a variety of scenarios and organizations. Many of my of the NGOs and companies I work with are focused on food and agriculture, so I'd like to be able to better understand the land use and methane components of the model.

Have you developed a scenario using the methane and land use parameters? How did you account for changes in ruminant enteric methane, projected deforestation rates, or other impacts of meat reduction?

The Climate Interactive blog post is a good start, but it would be great to stay connected and share more detailed information with my fellow facilitators:

Isaac - Great question. 

We're actively working on building the food system into En-ROADS more directly.  The blog post you linked to is our effort to connect the current version of En-ROADS to the food system and nature-based solutions conversations.

In the future, hopefully by the end of 2020, we'll have more specific sliders like 'diets', 'agroforestry', and other policies being discussed.

What policies are interested examining?  Why those policies?  What potential impacts are you seeing in your research or from your NGOs/companies?

Hi Travis,

Diet is the big one - many of the NGOs I work with (HSUS' Forward Food program, Center for Biological Diversity's Take Extinction Off Your Plate campaign, and others) are interested in science-based communications tools and workshops to show the climate impact of meat reduction. I think En-ROADS could play a helpful role there.

But I also work with groups on new technologies like cell-based meat (a.k.a. clean meat or cultivated meat), where animal meat is grown through cell culture instead of by farming whole animals. This in particular would make an interesting case for En-ROADS, because it would dramatically shift the role that land and energy play in food production. Cell cultures could be grown more efficiently from more land-efficient calorie sources or perennial crops with higher carbon storage potential. Energy demand might increase, but with a shift toward electricity that could be more easily decarbonized. (I'm holding a workshop on these topics in San Francisco later this month:

I'm excited to work with anyone else who is interested in modeling some of these scenarios.

So we're definitely going to take a crack at adding diets.

The life-cycle studies I've seen about cell-based meats and other non-meat production appear to be all over the place, on water use, land use, and energy use.  Do you want to post what you think the best peer- reviewed articles are here?  could be a good resource for others too.


Also, Charles Jones (CI modeler) weighed in on some of your other questions about the food system blog post, saying:

"Energy demand is not broken down to that detail in En-ROADS, but it is flexible enough to try different things. For the numbers in the Table 1 you cite, I used the energy savings estimates in an FAO Report ( ) and identified the setting on the Transport Efficiency slider that resulted in the same amount of energy consumption reduction. I picked transport efficiency because that is the sider that closest approximates a demand reduction like changes in the food system would. If you wanted to approximate your own numbers, look at a graph that shows a value you think should change - I'd recommend "Final Energy Consumption" for "energy used in the food system". Then adjust the demand sector - some of transport and some of buildings and industry - until you see the reduction you'd expect. Note the graphs are in exajoules = 10^9 gigajoules so you would have to convert the units. But this will give you a way to test mental models of climate impact of whatever energy reductions you envision."

Travis, thanks to you & Charles for that helpful modeling tip.

Regarding cell-based meat, you're right - there aren't many environmental assessments (LCAs) out there at the moment, and they come to dramatically different conclusions on energy and resource use. That reflects their speculative nature. Without any operating facilities even at the pilot scale, it's been a challenge for academics to estimate the likely impact of the technology. I and others have written some reviews to help interpret the existing studies. Here are some of the key resources:

  1. Tuomisto et al. 2014 LCA:
  2. Mattick et al. 2015 LCA:
  3. Technical comparison of the leading LCAs:
  4. Fact sheet on environmental impacts for a non-expert audience:
I'll be revisiting this topic over the next few weeks, so I'll have more input by April. I'm also aware of at least one ongoing LCA project that should be released later this year. I'm hopeful that one will answer many of our basic questions about the structure of the industry.

If anyone on your team would like some guidance, I'm happy to have a phone call to talk about the major issues in land and energy use for cell-based meat.

I ran the virtual climate solutions workshop last week with about a dozen members of the Food and Climate Alliance ( En-ROADS went over very well with the group, many of us see it as a valuable tool for discussing the importance of food and diet in combination with more conventional climate solutions.

I developed two scenarios in preparation for the workshop: reducing meat consumption 50% by 2040, and replacing 50% of animal meat with plant-based meat by 2050. Travis, I wonder if you or any of the CI modeling team would be able to comment on my approach. I'd like to know whether the way I went about this fits with how the simulator works 'behind the scenes.'  One of my concerns has to do with how En-ROADS treats continued changes in energy systems over time; does the early time frame of dietary change by 2040 unrealistically affect projected energy use through 2100?

50by40 scenario:

One of the main food-related climate campaigns is '50by40' - targetting a 50% reduction in meat production by 2040. To build the scenario, I started with livestock emissions data from the FAO 2013 report Tackling Climate Change through Livestock. That report breaks down livestock-related emissions by gas (CO2, CH4, N2O) and the CO2 emissions by source (land use change, feed production energy, and energy for transportation). 

To estimate 'business as usual' emissions from livestock, I used projected increases in livestock from the FAO 2018 report, Future of Food and Agriculture to 2050. Assuming emissions-per-livestock-unit stay the same, we can get projected livestock emissions in 2040 by multiplying current emissions by the proportional increase in the number of animals. (This is probably an okay assumption for feed and methane emissions, but problematic for land use change emissions. I'd need to know more about deforestation in En-ROADS to address that issue.) Total annual emissions from livestock in 2040 were:

4.46 Gigatons CO2e from enteric methane

2.88 Gigatons CO2e from N2O

1.33 Gigatons CO2e from feed-related energy emissions

0.48 Gigatons CO2e from transport emissions


To estimate the emissions from plant crops grown to replace animal meats in the 50by40 scenario, I used the ratio of life-cycle emissions from high-protein plant foods to animal meats (on a per-unit-protein basis) from the Poore & Nemecek 2018 meta-analysis of LCA studies. (This could be done more precisely by accounting for the projected proportion of animal and plant proteins in the 2040 diets, but I expect that within-group variation will not be influential here.) I made the blanket assumption that emissions from plant crops were divided among fertilizer N2O (50%), fertilizer production and other industrial energy emissions (25%) and tractor/transport emissions (25%). Total annual emissions from replacement plant crops in 2040 were:

0.25 Gigatons CO2e from N2O

0.12 Gigatons CO2e from energy emissions

0.12 Gigatons CO2e from transport emissions


Total emissions from the ‘50by40’ scenario were simply 50% of current livestock emissions plus the replacement plant crop emissions. I used the energy intensity value from En-ROADS for 2005 (the year of the FAO livestock emissions data) to convert energy and transport emissions to energy values (exajoules/yr).


Finally, I addressed deforestation and afforestation. Based on several studies that have been published over the past few years, I assumed at least 50% of current deforestation is due to livestock farming. Given that even a moderate reduction in global meat demand will drastically decrease the need for additional land clearing for pasture, I assumed that the ‘50by40’ scenario would halve deforestation emissions by 2040. Based on an estimate of the cropland needed to produce additional plant foods from the EAT-Lancet report (~37% of the cropland needed for animal foods) and an assumption that about 50% of unutilized cropland could be reforested, I estimated that roughly 88 million hectares of afforestation would be possible under this scenario.


Altogether, this produces the following estimates for En-ROADS Simulator variables in the year 2040:

174.8 Exajoules/yr, Transport Energy (Fuel) Use

412.7 Exajoules/yr, Industry Energy Use

1.93 GT CO2/yr, Deforestation

16.85 GT CO2/yr, Methane & Other Gases

6%, Available Land for Afforestation

20 years, Time to Secure Land for Afforestation


Resulting slider values:

Transport Energy Use: +0.4 to 0.9%/yr

Building and Industry Energy Efficiency: +0.7 to 1.9%/yr

Deforestation: -3.2 to -3.2%/yr

Agricultural and Waste emissions (CH4 & N20): -28 to -28%

Percent available land for afforestation: +6% to 6%

Time to secure land for afforestation: 20 years

Afforestation planting time: -20 to 10 years

Login to post a comment

Didn’t find what you were looking for? Please submit a ticket to our support team.