// Methodology
How the Model Works
Where the data comes from, how the calculations work, what we assume, and where the limits are. We built this because it needed to exist. Here's exactly what's inside it.
// 01 / Data Sources
Where the numbers come from
Every number on this site traces back to one of four types of sources. We don't make up values, and we don't use unpublished data. If a number appears in the model, it came from one of these:
- NASA Technical Reports Published through the NASA Technical Reports Server (NTRS). This includes crew health standards (NASA-STD-3001), life support system specifications, ISS operational data, and human research program findings.
- Peer-Reviewed Research Studies published in journals like Acta Astronautica, Aviation Space and Environmental Medicine, and the Journal of Applied Physiology. Bone loss rates, muscle atrophy data, and psychological timeline models come primarily from here.
- ISS Operational Data Publicly reported numbers from ISS operations: water recycling efficiency, CO₂ levels, crew sleep data, maintenance hours, and life support performance. These are the closest thing to real-world space habitat data that exists.
- Analog Studies Data from environments that simulate aspects of spaceflight: Antarctic winter-over stations, submarine deployments, NASA's HI-SEAS habitat, the Mars-500 isolation study, and Concordia Station. Psychological timeline data comes heavily from these.
When a value is a model parameter rather than a directly measured number, we say so. The glossary marks which terms are framework-specific and which are established science.
// 02 / How It Calculates
What the model actually does
The model tracks twelve systems simultaneously. For any given crew size, mission duration, and set of conditions, it calculates how each system performs and where the first failure is likely to occur.
The Scenario Engine takes your inputs (crew size, duration, habitat type, recycling rates, exercise compliance, privacy levels, leadership stability) and runs them through rate equations for each system. Food depletes based on calories per person per day minus any onboard production. Water depletes based on usage minus recycling efficiency. Air quality depends on scrubber capacity vs CO₂ output. Physical condition depends on exercise compliance vs degradation rates. Psychology depends on crew size, confinement duration, privacy, and routine stability.
The Cascade Simulator starts with a single failure and propagates it through the other eleven systems using coupling rates. If the air system degrades, it affects cognition. Degraded cognition affects decision quality. Poor decisions affect maintenance. Reduced maintenance affects hardware. Each coupling has a rate and a delay, based on the best available data from the sources listed above.
The Mission Generator works in reverse. You give it a goal (crew size, destination, duration) and it solves backwards for the minimum viable configuration: what recycling rates, exercise compliance, and habitat specs would you need to make it work?
All three tools use the same underlying rate equations and the same data sources. They're three different views of the same model.
// 03 / Key Assumptions
What the model assumes
Every model has assumptions baked in. These are ours. Changing any of them would change the outputs.
Oxygen
1.83 lb per person per day
From NASA mass-balance tables. Consistent across all three simulators.
Food
1.3-1.8 kg per person per day
Packaged food mass including packaging. Based on ISS provisioning logistics.
Bone Loss
1-2% per month in the hips, spine, and legs
NASA Human Research Program data. The model treats exercise as a linear mitigation factor. The real relationship is more complex.
Muscle Atrophy
Up to ~10% in 1-2 weeks (specific muscles)
Postural muscles (soleus, paraspinals) without countermeasures. Overall volume decline 5-10% in the first month, even with exercise.
Psychology
Fracture risk peaks Day 60-120
Based on third-quarter phenomenon from Antarctic and submarine analog studies. Varies significantly by crew size and composition.
Water Recycling
93% baseline (ISS operational)
ISS Water Recovery System. Recent demonstrations have reached 98% with the Brine Processor Assembly.
Exercise
2 hours/day prescribed
ISS ARED + T2 protocol. Compliance is a slider in the model because crews always negotiate it down.
Cascade Coupling
System-to-system rates and delays
These are model parameters, not measured constants. They represent our best estimate of how fast a failure in one system affects another.
// 04 / Limitations
What the model does not do
This is important. The model is a planning tool and a teaching tool. It is not a mission design system.
// Honest limitations
The model does not account for individual variation. Real crews are made of specific people with specific medical histories, personalities, and skill sets. The model works with averages.
The cascade coupling rates are estimates. We have good data on individual system degradation rates, but the system-to-system propagation speeds are modeled, not measured. Nobody has run the experiment.
The psychology model is the least certain part of the framework. Human behavior under extreme confinement is studied but not precisely predictable. The Day 60-120 fracture range is a pattern from analog studies, not a law of nature.
The model assumes no resupply. Any mission with resupply capability changes the math significantly.
Radiation modeling is simplified. The model uses career dose estimates from NASA standards but does not model solar particle events, shielding effectiveness, or individual radiosensitivity.
We built this framework because the alternative was no framework at all. The published research exists in hundreds of separate papers across dozens of fields. Nobody had assembled it into a single working model that you could actually run. That's what this is. It's imperfect, it's in beta, and we're improving it as better data becomes available.
// 05 / Evaluate It Yourself
How to check our work
Every claim on this site went through a three-pass fact-check against primary sources. The glossary defines every term we use. The system pages show the math behind each subsystem. The simulators let you change the inputs and see how the outputs respond.
If you find something that looks wrong, it might be. We'd rather know about it than not. The about page has contact information.
The model is a living system. As new ISS data comes in, as analog studies publish results, as the science moves, the model updates. Every number you see reflects the best available data at the time you're reading it.