Meeting summary: Tech x FAW group discussion 5: CO₂ removal in intensive aquaculture

[rhjiang]

This is a summary of the call associated with this post. Happy for Carlos or anyone else to add to/correct my notes!

The designated discussion topic was CO₂ removal in intensive aquaculture, but we also ended up talking about transportation for the live seafood market, a major welfare issue that Carlos has recently become aware of through his work.

CO₂ removal

• High CO₂ levels are a persistent challenge in RAS. Pond systems don't see this challenge because they struggle to get dissolved oxygen levels sufficiently high.
• Some of the CO₂ is produced by fish, but some is by bacteria.
• Standard practice for removing CO₂ is to run water through a column with air pumped through, where the water releases CO₂ to the air. This method sees some efficiency struggles
  • As water CO₂ levels get lower or air CO₂ levels get higher, transfer to air becomes less efficient
  • Most carbon is stored in the water as bicarbonate, which then reacts to restore CO₂ levels when they lower
• A few levers to pull
  • More efficient air pumps may increase the effectiveness/decrease the price of the standard approach
  • Increasing surface area of the water in the column may increase transfer efficiency
  • Lowering pH of the water or increasing the temperature of the water makes CO₂ transfer to air more efficient. But currently it is costly to do these things
• Hybrid greenhouses + RAS facilities may be a good idea, as the greenhouse plants can absorb the CO₂ released into the air via the water column CO₂ stripper. However, you would need a really big greenhouse. Doable, but you need someone with the money and space who wants to do this
  • Maybe even combine it with a datacenter to use the heat for efficient CO₂ removal from water. But you still need to use energy to cool the water down again after.

Transportation for live seafood market

• The live seafood market is big, but infrastructure is bad/not well-informed.
  • People doing live transportation don't know much about water quality and aren't doing anything to maintain it. Lots of animals die
  • Possibly the reason for this is that, until recently, live seafood was usually sourced more locally. But as the restaurant/tourism sectors boom, they're starting to transport animals from all over
• This should be tractable to change
  • Traders are wealthy, so they should be able to adopt technology
  • There are big profit incentives to increase transport survival rates
  • First step to addressing this would be to test the hypothesis that you can greatly improve survival by getting a trader to let you do a trial of some proposed intervention

[michaelstjules]

(Left this same comment in the other thread)

Seems like there's a substantial risk of this improving the efficiency of aquaculture, e.g. allowing higher stocking densities, and so increasing the number of animals farmed. By default, you might expect any measurable welfare improvements through water quality to be compensated for by higher stocking densities, reaching the same average welfare level as before, and the primary change is just higher stocking densities and more farmed animals.

[JoePaterJoePaterJoePater]

Agree in this case. I was thinking about this at the weekend and vibe coded a python script that calculates the new equilibrium for a welfare improvement that reduces mortality (and hence increases demand through lower costs). I may write a bit more about this another time (are you aware of any existing analysis?), but here's the script if you want to have a play.

liveability-suffering-tradeoff.py

[michaelstjules]

I don't remember any existing analysis off-hand, but Samara Mendez at Rethink Priorities and Jacob Schmeiss would be good people to talk to about estimating these risks, given their expertise in agricultural economics. I've been thinking generally it would be good to have them more involved in impact estimates and risk assessment for welfare tech and other interventions that might reduce mortality rates or increase growth.

That being said, not all of the problem is equilibrium modelling. Some of it is just understanding how specific farmers will respond, given their costs and interests, and possibly actually just checking how they respond in trials (although we might worry that they'll act differently depending on whether or not we check, if they correctly assume we'd prefer they not increase stocking densities or production).

[michaelstjules]

I have some notes on economic modelling for animal welfare here, in case it's useful. It's a bit rough, but probably some useful links there at least.

[JoePaterJoePaterJoePater]

Had a skim, this is great! Shame it never got published. I think second order effects deserve more discussion in general, be that economic/political/technological.

I'll reach out to Samara and Jacob, thanks for the leads.