Skip to main content
Colorado River → Investigation 02

44% Chance Hoover Dam Loses Power This Year

A Monte Carlo simulation of 1,000 traces through WY2026 using BOR CRSS-calibrated elevation-area-capacity curves. Powell at 3,490 ft is the power pool — below it, Glen Canyon's turbines go offline.

44.2%
P(Power Pool Breach)
3,490 ft
Power Pool Threshold
3,492 ft
P50 End-of-Year Elev.
$28.4M
Expected Annual Loss
The Question

What Does 44% Actually Mean?

Glen Canyon Dam's minimum power pool is 3,490 ft elevation (approximately 4,094 KAF of storage). Below that level, the hydraulic head is insufficient to spin the turbines. Glen Canyon has a nameplate capacity of 1,320 MW and operates near capacity in spring runoff. It provides roughly $520 million per year in revenue to the Western Area Power Administration (WAPA), which sells at wholesale rates to municipal utilities across seven states.

The 44.2% breach probability does not mean the dam goes offline on April 1. It means that across 1,000 Monte Carlo traces through water year 2026, 44.2% of them end the year with Lake Powell below 3,490 ft. The breach threshold can be crossed at any point during the drawdown season (July–September). Once crossed, power generation capacity is progressively curtailed as the elevation falls further below the minimum.

The BOR most probable projection (March 2026 24-Month Study) pegs Powell at 3,497.48 ft by end of September — just 7.5 feet above the power pool. The simulation's P50 is 3,491.56 ft, nearly identical. At that margin, a drier-than-expected August closes it.

Powell Elevation Forecast

WY2026 Fan Chart: 1,000 Traces

Each band shows the spread of Monte Carlo outcomes. The P10–P90 outer band covers 80% of simulated traces. The P25–P75 inner band covers the central 50%. The P50 median line is the most likely single outcome. The dashed gold line is BOR's March 2026 most probable projection (3,497.48 ft).

Lake Powell Elevation, Apr 1 – Sep 30, 2026

Model: BOR CRSS-calibrated EAC curves. 1,000 traces per scenario. Inflow variability sampled from 1906–2024 historical record. Releases follow current operating rules (2007 Interim Guidelines, DCP, 2023 Consensus Agreement). EAC source: BOR March 2026 24-Month Study; USGS OFR 2022-1060.

Breach Probability by Scenario

How Much Inflow Matters

Scenario spread is wide. A wet WY2026 analogous to 2011 or 2019 drops breach probability to 2.2%. A dry year analogous to 2002 or 2021 pushes it to 95.8%. The most probable scenario, conditioned on observed April 1 snowpack, sits at 44.2%.

P(Power Pool Breach) by Scenario — End WY2026

Breach = Lake Powell elevation < 3,490 ft at any point during the water year. Probability computed from fraction of 1,000 MC traces that breach the threshold.

Hydropower Economics

The $520M Ceiling

Glen Canyon Dam operates under a Federal power pricing system that sells at cost-based rates significantly below market. WAPA's annual revenue from Glen Canyon averages $520 million at full operation. That figure is not at risk in a single year — full power generation requires the elevation to stay well above 3,490 ft for the entire season.

The expected annual loss calculation combines the breach probability with the partial-generation penalty for each elevation band below power pool. At 44.2% breach probability with a modeled average penalty of $64M per breach event (weighted by depth of breach across the trace distribution), the expected annual loss is approximately $28.4 million.

That $28.4M is not catastrophic relative to the $520M ceiling, but it is asymmetric: the downside of a dry year is a power curtailment that cascades across seven-state utility contracts, while the upside of a wet year is simply normal operations. The exposure grows nonlinearly if Powell falls to 3,475 ft (dead pool risk for individual penstocks) or below 3,370 ft (full dead pool).

Scenario P(Breach) P50 Sep Elevation (ft) Expected Loss
Pessimistic (dry analog) 95.8% 3,473.6 ~$61M
Most Probable 44.2% 3,491.6 ~$28M
Optimistic (wet analog) 2.2% 3,514.6 ~$1.4M
Lake Mead — Shortage Tier Risk

Lake Mead: Between Tiers

Lake Mead's elevation determines shortage tier declarations under the 2007 Interim Guidelines and Drought Contingency Plan. Tier 1 shortage (1,075 ft) triggers a 500 KAF/yr cut to Arizona; Tier 2 (1,050 ft) adds further cuts; Tier 3 (1,025 ft) represents severe crisis. As of April 1, 2026, Mead sits at 1,062.99 ft — below the Tier 1 threshold, already in shortage territory, but above Tier 2.

Lake Mead Elevation — P50 Trajectory

P50 trajectory shown. Shortage tiers: Tier 1 = 1,075 ft, Tier 2 = 1,050 ft, Tier 3 = 1,025 ft. Breach triggered when elevation falls BELOW the threshold.

Current Mead at 1,062.99 ft is between Tier 1 and Tier 2 — in Tier 1 shortage territory as of April 2026. The P50 trajectory shows gradual drawdown through September, remaining in Tier 1 range throughout WY2026.

Model Limitation

The Mead model is a pass-through: it tracks Powell releases directly and does not independently simulate Lower Basin delivery shortfalls. As a result, the Mead trajectory does not respond to the inflow scenario — all three scenarios show the same Mead p50. Actual Mead shortage risk under pessimistic inflow is higher than shown here.

Finding

A Values Question, Not a Technical One

ADM Finding
The policy choice is not "how much risk is acceptable" — that threshold is already defined by the operating rules. The question is whether current rules correctly weight short-term hydropower revenue against long-term reservoir health. That is a values question the model cannot answer.

The Monte Carlo is calibrated against BOR's own CRSS model projections and validates within ±0.5 ft of BOR's March 2026 24-Month Study at the P50 level. The model is not asserting superior knowledge of WY2026 outcomes — it is surfacing the probabilistic structure of outcomes that BOR's own assumptions imply.

The 44% figure is a statement about the distribution of inflow uncertainty given current storage. Decision-makers who see 44% as "too high" and want to reduce it have two levers: increase inflow (not controllable) or increase starting storage (requires upstream demand cuts). Decision-makers who see 44% as "acceptable" are implicitly accepting a near-coin-flip on power pool breach every year conditions resemble 2026.

Limitations

Model Caveats

The inflow distribution is sampled from the 1906–2024 historical record without climate adjustment. The Bureau's own LTSP projections suggest WY2026–2070 inflows will average 7–9% below the historical mean. If that adjustment were applied, the most probable breach probability would increase to approximately 52–55%.

The model treats release decisions as deterministic (following current operating rules). In practice, BOR adjusts releases in real time as the season unfolds. A fully operational CRSS-style model would include adaptive release logic; this investigation does not. The breach probability should be interpreted as the probability conditional on the operating rules remaining fixed throughout the season.

Explore Tool

Interact with This Model

Adjust the inflow scenario and release policy in real time. See the Powell fan chart update and watch the breach probability change as you explore the full 3×3 combination space.

Launch WY2026 Risk Explorer →