Tackling the Drought — Forecasting Lake Mead & Sizing Water Recycling

2021 HiMCM Problem B · produced end-to-end at the award-winning standard · real USBR data · no API calls
Open the 22-page paper (PDF) View the model code ← HiMCM coaching

Deterministic compliance — all pass

Page countPASS22 pp (≤25)
Summary lengthPASS427 words (band 340–560)
Summary numericPASS27 results
AnonymityPASScontrol number only
ReferencesPASSpresent, in-text cited
AI-Use reportPASSpresent
FiguresPASS17 (target ~14)

Headline results (real data, validated)

  • Volume curve validated to 0.19% vs USBR Table 1
  • Level 2050: 1052 ft (Model 1, R²=0.90, back-test 8.3 ft) vs 976 ft (Model 2, stress)
  • 2050 shortfall 2.49 maf; recycling covers 100% to 2043, 83% by 2050
  • Monte-Carlo mean 82% (p10 59%); full coverage needs recovery ≥0.72
17 figures · 8 tables · coupled pipeline: level → storage → budget → recycling.

Selected figures

Seven detected drought periods
Two forecasts diverge after 2030
Volume curve validated <0.2%
Out-of-sample back-test
Recycling vs the shortfall
Monte-Carlo coverage
Decision flip at ρ=0.72
Key-findings panel