Improvements in yield may not ensure food supply longterm.
Modern agriculture has a large carbon footprint and chemical fertilisers are destroying the health of our soils (at the rate of a fertile area of 30 soccer fields/minute). Despite a 700-fold increase in the use of pesticides during the second half of the 20th century, yield remains constant but 20-40% of domesticated plants are still lost to pathogens.
[...]
From a perspective of a longtermist, this case study is an emblematic story of how we often choose to increase yield in the short-term, in exchange for long-term benefits.
Do you have one or two sources you’d recommend that make a strong case that crop yields will plausibly decline on any particular timescale? Particularly if those sources also incorporate factors like potential gains from genetic engineering, potential inefficiencies from increased meat production, etc.?
I see many references to soil health as an ongoing crisis, but I haven’t seen any forecasts about actual outcomes / tipping points that stick in my memory. Our World in Data finds gradually increasing yields since mid-century; the trends seem to have flattened a bit recently, but it’s hard to tell how much of that is the end of easy gains from Green Revolution vs. active negative trends beginning to take effect.
The EA longtermist community is tiny, and it’s easy to imagine even very strong evidence of long-term food supply risk being missed. But even ALLFED, which has a deep focus on related topics, seems to focus almost entirely on risk of crop loss from major disasters rather than ongoing climate change. I’d be interested to know about ALLFED work I’ve missed, or other good resources that people at places like Open Phil should think about.
Regarding sources that argue about risks on crop yields in the future, I have collected some over time. Apologies for not explaining them more thoroughly—they include reports of past/current crop losses as well as predictable ones, of course caused by many different factors:
Future warming increases probability of globally synchronized maize production shocks (2018)
Amplified Rossby waves enhance risk of concurrent heatwaves in major breadbasket regions (2019)
Climate Change Impacts on Global Food Security (2013)
Plant pathogen infection risk tracks global crop yields under climate change (2021)
World agriculture towards 2030/2050: the 2012 revision (2012, UN FAO) The global groundwater crisis (2014)
Hi Aaron, I’m afraid as mentioned I don’t know this area very well so I’m not familar with the literature and cannot point you towards particular papers. I used this fungi case study as an example to point out that risk must be understood as more than the hazard (e.g. drought), but also the vulnerablity (e.g. resilience levels of plants against drought): if resilience has decreased, the trend in yield will not decrease to show the increased risk up until the point where the hazard hits and the vulnerability is exploited.
But I’m happy to point Open Phil towards researchers who do know and think a lot more about this. Tail-risk studies in climate change are generally neglected, so some of our peparedness efforts might have to be based on conversations with experts rather than a literature search.
If there are particular researchers you have in mind, I’d guess that ALLFED would be very interested in talking to them (if they haven’t already). I’d like to share this comment thread with the folks I know there. What are the names you were thinking of?
Carla-nice piece! I direct the Alliance to Feed the Earth in Disasters (ALLFED). We are indeed concerned about tail risks from climate change. We have also considered two different resilient foods for global agricultural shocks related to fungi: mushrooms and fungus grown in bioreactors. Mushrooms have the advantage of the potential to scale extremely rapidly and grow on fiber (e.g. agricultural residues or wood), but they are expensive. Quorn is a fungus that is currently grown on grain, so in a catastrophe, eating the grain would generally make more sense (though Quorn would be better than feeding animals). However, it would be possible to have fungus in a bioreactor grow on fiber, which would be resilient. But unfortunately, neither of these bioreactor options is very economical at this point.
That sounds cool. Happy to see that some of this work is going on and glad to hear that you’re specifically thinking about tail-risk climate change too. Looking at fungi as a food source is obviously only one of the dimensions of use I describe as relevant here, and in ALLFED’s case, cost of the production is surely only one relevant dimension from a longtermist perspective. In general, I’m happy to see that some of your interventions do seem to consider fixing existing vulnerabilities as much as treating the sympoms of a catastrophe. I’ll go through the report you have online (2019 is the most recent one?) to check who you’re already in contact with and whether I can recommend any other experts to you who it might be useful for you to reach out to.
On a seperate note and because it’s not on the Q&A of your website: are you indeed fully funded by EA orgs (BERI, EA Lottery as per report)? I found it surprising that given your admirable attempts to connect with the relevant ecosystem of organisations you would not have funding from other sources. Is this because you didn’t try or because it seems no one except EAs want to grant money for the work you’re trying to do?
Do you have one or two sources you’d recommend that make a strong case that crop yields will plausibly decline on any particular timescale? Particularly if those sources also incorporate factors like potential gains from genetic engineering, potential inefficiencies from increased meat production, etc.?
I see many references to soil health as an ongoing crisis, but I haven’t seen any forecasts about actual outcomes / tipping points that stick in my memory. Our World in Data finds gradually increasing yields since mid-century; the trends seem to have flattened a bit recently, but it’s hard to tell how much of that is the end of easy gains from Green Revolution vs. active negative trends beginning to take effect.
The EA longtermist community is tiny, and it’s easy to imagine even very strong evidence of long-term food supply risk being missed. But even ALLFED, which has a deep focus on related topics, seems to focus almost entirely on risk of crop loss from major disasters rather than ongoing climate change. I’d be interested to know about ALLFED work I’ve missed, or other good resources that people at places like Open Phil should think about.
Regarding sources that argue about risks on crop yields in the future, I have collected some over time. Apologies for not explaining them more thoroughly—they include reports of past/current crop losses as well as predictable ones, of course caused by many different factors:
Future warming increases probability of globally synchronized maize production shocks (2018)
Amplified Rossby waves enhance risk of concurrent heatwaves in major breadbasket regions (2019)
Climate Change Impacts on Global Food Security (2013)
Plant pathogen infection risk tracks global crop yields under climate change (2021)
World agriculture towards 2030/2050: the 2012 revision (2012, UN FAO)
The global groundwater crisis (2014)
Anthropogenic depletion of Iran’s aquifers (2021)
Oxfam Media Briefing 27.4.2017 https://www.oxfam.de/system/files/mb-climate-crisis-east-africa-drought-270417-en.pdf (To the trend in (mostly rain-fed) African agriculture)
Increased future occurrences of the exceptional 2018–2019 Central European drought under global warming (2020)
But this is all a bit here and there, not an expert who can evaluate the validity of this research
Thanks for the links! I appreciate your taking the time to share them, and they all seem relevant to my question.
Hi Aaron, I’m afraid as mentioned I don’t know this area very well so I’m not familar with the literature and cannot point you towards particular papers. I used this fungi case study as an example to point out that risk must be understood as more than the hazard (e.g. drought), but also the vulnerablity (e.g. resilience levels of plants against drought): if resilience has decreased, the trend in yield will not decrease to show the increased risk up until the point where the hazard hits and the vulnerability is exploited.
But I’m happy to point Open Phil towards researchers who do know and think a lot more about this. Tail-risk studies in climate change are generally neglected, so some of our peparedness efforts might have to be based on conversations with experts rather than a literature search.
If there are particular researchers you have in mind, I’d guess that ALLFED would be very interested in talking to them (if they haven’t already). I’d like to share this comment thread with the folks I know there. What are the names you were thinking of?
Carla-nice piece! I direct the Alliance to Feed the Earth in Disasters (ALLFED). We are indeed concerned about tail risks from climate change. We have also considered two different resilient foods for global agricultural shocks related to fungi: mushrooms and fungus grown in bioreactors. Mushrooms have the advantage of the potential to scale extremely rapidly and grow on fiber (e.g. agricultural residues or wood), but they are expensive. Quorn is a fungus that is currently grown on grain, so in a catastrophe, eating the grain would generally make more sense (though Quorn would be better than feeding animals). However, it would be possible to have fungus in a bioreactor grow on fiber, which would be resilient. But unfortunately, neither of these bioreactor options is very economical at this point.
That sounds cool. Happy to see that some of this work is going on and glad to hear that you’re specifically thinking about tail-risk climate change too. Looking at fungi as a food source is obviously only one of the dimensions of use I describe as relevant here, and in ALLFED’s case, cost of the production is surely only one relevant dimension from a longtermist perspective. In general, I’m happy to see that some of your interventions do seem to consider fixing existing vulnerabilities as much as treating the sympoms of a catastrophe. I’ll go through the report you have online (2019 is the most recent one?) to check who you’re already in contact with and whether I can recommend any other experts to you who it might be useful for you to reach out to.
On a seperate note and because it’s not on the Q&A of your website: are you indeed fully funded by EA orgs (BERI, EA Lottery as per report)? I found it surprising that given your admirable attempts to connect with the relevant ecosystem of organisations you would not have funding from other sources. Is this because you didn’t try or because it seems no one except EAs want to grant money for the work you’re trying to do?