Mars Colony Viability (Strict Science-Only Baseline)

This repository starts from zero assumptions and enforces a hard rule:

No numeric parameter is used unless it is backed by a verifiable primary source (NASA/ESA technical reports, peer-reviewed journals, or NASA NTRS).

All verified numeric anchors are stored in models/verified_constants.py and documented in VERIFIED_SOURCES.md.

What this repo answers (v0.0.x baseline)

A narrow, falsifiable engineering question:

Given colony size N0, initial stores of O2 and water, local closure fractions, and discrete resupply windows: when does the colony collapse due to O2 or water depletion?

Radiation is tracked only as dose bookkeeping using published cruise vs surface dose-equivalent rates. This repo does not convert dose to clinical outcomes.

Negative-result posture (important)

This project is designed to surface constraints. If a scenario collapses, that is a scientifically useful result (it reveals required closure/resupply conditions).

What this repo explicitly does NOT do (yet)

• Food/calories model (requires verified agricultural yields and energy budgets).
• Fertility / multi-generational physiology at 0.38g (no verified human dataset).
• Long-duration ISRU reliability curves (requires demonstrated multi-year plant datasets).
• Terraforming claims.

See docs/NON_CLAIMS.md.

Quickstart

Install

`�ash python -m venv .venv

Windows PowerShell:

. .\.venv\Scripts\Activate.ps1 python -m pip install --upgrade pip pip install -r requirements.txt �ash python main.py --help Because parent directories may contain unrelated pytest configs, always run: �ash python -m pytest -c pytest.ini

• docs/methodology.md ? equations, units, collapse criterion (no speculative numbers)
• docs/NON_CLAIMS.md ? explicit scope limits and non-claims
• paper/manuscript.md ? manuscript scaffold (Markdown -> LaTeX/PDF later)

Citation

See CITATION.cff.

• If/when a Zenodo DOI exists, it will be added to CITATION.cff and pinned in a new release.

DOI

Zenodo DOI: 10.5281/zenodo.18517605

Published Results (Reproducible)

The following scenarios were executed using the public codebase. All results below are generated directly from model execution.

See:

• 
  esults/published_metrics.csv\
• 
  esults/published_metrics.md\

Key observable metric:

• \days_elapsed: simulated survival time until system collapse.

Published Results (Reproducible)

Source of truth: results/summary.csv (generated by running the public model).

Reproduce (PowerShell): python -m scripts.run_scenarios

Artifacts:

• results/summary.csv
• results/summary.md
• results/figures/*.png (if generated)

Summary table (from results/summary.csv)

scenario_id	N0	o2_storage_days	water_storage_days	o2_local_fraction	water_local_fraction	water_recovery_fraction	launch_window_days	missed_window_probability	collapsed	collapse_day	dose_msv
S1_baseline	12	365	365	0.97	0.98	0.98	780	0.2	False	N/A	N/A
S2_higher_closure_buffers	12	730	730	0.995	0.995	0.98	780	0.1	False	N/A	N/A
S3_scale_stress	50	365	365	0.98	0.985	0.98	780	0.2	False	N/A	N/A

Figures

• 
• 

Falsifiable outcome statements

• S1_baseline: N0=12, O2_local_fraction=0.97, Water_local_fraction=0.98, launch_window_days=780, missed_window_probability=0.2 -> collapsed=False at day=N/A; dose_msv=N/A
• S2_higher_closure_buffers: N0=12, O2_local_fraction=0.995, Water_local_fraction=0.995, launch_window_days=780, missed_window_probability=0.1 -> collapsed=False at day=N/A; dose_msv=N/A
• S3_scale_stress: N0=50, O2_local_fraction=0.98, Water_local_fraction=0.985, launch_window_days=780, missed_window_probability=0.2 -> collapsed=False at day=N/A; dose_msv=N/A

Notes:

• If collapse_day is N/A, the run ended without emitting a collapse day for that scenario (model output), not a manual edit.

Cite this work

If you use this repository in research, please cite the Zenodo record.

• Concept DOI (recommended): https://doi.org/10.5281/zenodo.18481530
  (always resolves to the latest version)
• Version DOI: https://doi.org/10.5281/zenodo.18481531
  (fixed snapshot for strict reproducibility)