AIMxHIE discussion - Fish Welfare Initiative: Ice slurry machine that can turn pond water into ice slurry (7th April 2026)

[aaronboddy]

Fish Welfare Initiative is an animal welfare charity, working to improve the welfare of farmed fishes as much as possible.

On 7th April, their co-founder Haven King-Nobles will join us to tap into the wisdom of the High Impact Engineers community to discuss various ideas for engineering projects that are currently bottlenecking their work.

The sign-up link is here: https://luma.com/ld1efipl

There are currently 14 ideas they're thinking about:

1. Ice slurry machine that can turn pond water into ice slurry
2. Mobile stunning machine for the Indian context
3. Satellite-enabled farmer water quality app
4. Platform that interprets satellite data as estimated water quality values
5. Static solar-powered aerator
6. Mobile solar-powered aerator
7. Kit for retrofitting existing aerators with solar-panels/batteries
8. Camera-based fish behavior monitoring system
9. Stocking density measuring cameras
10. Simple image-based app for hatcheries to use to estimate the number of baby fishes they have during packing/transport
11. Early warning monitoring system that can notify us when harmful microbiomes start talking over the pond
12. Targeted medicine/antibiotics delivery system for fishes showing early signs of bacterial infections
13. Remote controlled device that would let us take water samples from the middle of the pond
14. Inexpensive handheld and accurate water quality monitoring devices

We're free to explore and discuss any of the ideas with Haven, but to focus our conversation on the day, I asked Haven to select the idea with the highest priority, which is the "ice slurry machine that can turn pond water into ice slurry":

In India, currently the most viable way to reduce suffering at slaughter seems to be chill kill, and so we are seeking to develop it. One of the biggest bottlenecks right now though is making an ice slurry, both because a) it’s hard to make a true ice slurry, and b) the farms we work with “harvest” the fishes on site in rural locations, so it’d be very hard to bring a huge amount of ice slurry in. So the idea for this machine would be to develop something that could pull the pond water, which there’s a lot of, and turn that on the farm into a slurry. The ponds do sometimes have electricity, though if nothing else it could maybe run on diesel. Ideally the machine would also filter the water some so the slurry didn’t turn out green (that’s the color of the pond water)

It would be cool if people could also start the conversation in the Impact Forge, so that we can really hit the ground running on the call itself :)

[aaronboddy]

Thanks to everyone who attended the call on 7th April!

Here are the notes from my AI Notetaker for the benefit of those who weren't able to attend (and as a reminder for those who were!)

Feel free to continue the conversation in this channel.
And we might schedule a follow-up call at some point in the future :)

Overview of Fish Welfare Initiative

• FWI started in 2019 as AIM's first class, currently working primarily in India
• Operating on-the-ground program with approximately 180 farms monitoring water quality with stocking and sea caps
• Unique ability to influence farmer practices and conduct R&D trials due to direct farm relationships
• Main focus areas: improving water quality monitoring and addressing fish slaughter methods
• Current slaughter method involves suffocation over approximately one hour; seeking to implement chill kill (15 minutes) or ideally electrical stunning

Ice Slurry Machine for Chill Kill (Top Priority)

• Major challenge: Chill kill requires ice slurry, which is difficult to manufacture effectively
• Current method uses crushed ice that floats to the top, leaving warmer water at bottom where fish are located
• Ideal solution: machine that can produce ice slurry on-farm using pond water (30°C and very dirty/filthy)
• Salinity consideration: Adding salt would make ice slurry easier to produce but creates welfare concerns for freshwater species; FWI limits salinity to under 5 ppt
• Deep Chill company explored but requires salinity and fresh water solutions unavailable
• Technical considerations discussed:
  • Ice temperature (not just zero degrees) affects cooling rate
  • Surface area of ice crystals critical for faster cooling
  • Ice should be crushed closer to use time to maintain temperature
• Current ice production uses outdated "dinosaur age" ice crusher machines
• Potential alternative: Simple submersion pump to recirculate water and prevent stratification
• Implementation concern: Any added complexity faces farmer resistance (e.g., workers resist wearing gloves despite recommendations)

Continuous Water Quality Monitoring (High Priority)

• New idea from Claude: Continuous pH and temperature monitors to predict ammonia risk
• Rationale: pH and temperature affect unionized ammonium levels; could alert farmers via WhatsApp/SMS to test ammonia with strips
• Previous experience: Used commercial continuous monitors (~$100/month) that measured pH, temperature, and oxygen but not ammonia
• Major challenges with existing monitors:
  • Too expensive for programmatic scale
  • Require cleaning every week
  • Break frequently
  • Biofouling issues including lice eggs being laid on sensors
• Target specifications: $100-$1000 per unit, maintenance required no more than once every 6-12 months
• Welfare relevance hierarchy: Dissolved oxygen (DO) > ammonia > pH/temperature
• DO is most welfare-relevant but appears most difficult and expensive to measure continuously
• Amanda identified potential opportunity: oceanographic floats already measure DO, pH, and nitrate continuously for years ($20,000-$30,000 but potentially much cheaper for freshwater pond application)
• Cost reduction opportunity: Farms highly concentrated in regions (within 10 square kilometers); could use local base stations with solar-powered sensors to reduce uplink costs
• Alternative explored: Camera monitoring ammonia test strip with machine vision, but strips don't work continuously

Farmer Behavior and Implementation Challenges

• Current FWI program: Staff visit farms by motorcycle, take water samples to lab, call farmers with recommendations
• Farmers do respond to FWI recommendations, making this intervention scalable if monitoring can be automated
• Farmers currently use drop method/kits for water quality testing
• Larger farmers send samples to labs, but this represents small percentage
• Farmers typically only test water quality after observing behavioral cues (fish stop feeding, mortality along pond boundary)
• Major barrier: Farmers don't see water quality as sufficient issue until economic impacts (mortality) occur
• FWI distributed 10 API water testing kits; only 4 of 10 still being used
• Reasons for non-use: farm operators (not owners) lack incentive; farms destroyed; difficulty getting accurate measurements
• Cultural resistance: Workers refuse basic practices like wearing gloves despite 10-20 years of traditional methods
• Cost sensitivity: Indian Major Carp (IMC) are low-value species, making any cost addition a significant burden from livelihood perspective

Fish Counting System for Hatcheries

• Problem: Poor counting at hatchery-to-nursery stage leads to overstocking, fragility, resource competition, and high mortality (>50% sometimes)
• Current method: Spoon-based estimation ranging from 50,000 to 1.1 million fish per bag depending on species
• Hatcheries incentivized to not oversell (can remove excess and sell separately)
• Fish transported in transparent plastic balls with oxygenation
• Fish size at this stage: 3.5-4 days old, millimeter-sized
• Proposed solutions explored:
  • Image-based counting via phone app with confidence intervals (preferred for simplicity)
  • Physical counting machine (like Sincere Aqua shrimp counters)
  • Density measurement using suspended particle techniques
  • Ultrasonic or electrical resistance sensors
• Challenges: Fish are fragile; dewatering and weighing would cause significant stress and long-term impacts
• Image-based approach: Could likely achieve 10% accuracy with high-quality training data from precise counting machine
• Stocking density issues occur at multiple life stages:
  • Hatchery to nursery: 40k-50k fish per bag (45-60 days)
  • Nursery to breed out: Fewer fish, 5-6 months
  • Grow out stage: 6-12 months (where FWI currently works)
• Implementation concern: Farmers might resist accurate counting (prefer "free" fish like old scale vs. electric weighing machine analogy)

Satellite Imagery for Water Quality (Likely Discontinued)

• FWI close to ruling out satellite imagery approach for predicting water quality
• Technical approach: Used XGBoost model with all available bands from Sentinel-2 ESA satellite (flyover every 5 days)
• Some signal detected but results not good enough or consistent enough for programmatic investment
• Planet Labs satellites explored but not responding; may have higher frequency but fewer spectral bands
• Challenges: High noise, chaotic systems, many confounding biological factors
• Physical mechanisms may make some water quality parameters invisible to satellite imagery

Other Ideas Mentioned

• Aerator ideas ranked lower priority
• Ammonia sources: fish decomposition, excess feed, sludge at pond bottom
• Sludge management: Shrimp farms remove sludge, but fish farmers view it as food source and avoid removal due to cost
• Ideal intervention characteristics: Simple, visual indicators (e.g., something that turns red when action needed)

Next Steps and Collaboration

• This was the second High Impact Engineers × AIM charities meeting
• Format: Bring AIM charity founders to discuss problems with engineers
• Meeting recorded with AI notes to be posted in Impact Forge Slack
• Engineers encouraged to reach out to Haven via email or Hive Slack with ideas
• FWI can test solutions relatively easily due to farm access and budget for promising ideas
• Context: Only three fish charities in world; FWI's water quality program is potentially best intervention for fish welfare globally