Skip to main content
Colorado River → Investigation 01

The Water That Doesn't Exist

The Colorado River has been legally allocated 4.1 MAF more than it physically contains on average. This isn't a drought — it's structural overcommitment baked into every water rights agreement since 1922.

16.5 MAF
Legal Allocations
12.4 MAF
Average Supply
4.1 MAF
Paper Water Gap
2.1 MAF
Unexercised Tribal Gap
The Problem

4.1 MAF of Paper Water

The 1922 Colorado River Compact divided the river's flow between the Upper and Lower Basins based on a measurement period that happened to be one of the wettest decades in the past 500 years. The compacting parties allocated 7.5 MAF per basin — 15 MAF total — plus 1.5 MAF to Mexico under the 1944 treaty. That's 16.5 MAF of legal claims against a river whose long-term average flow at Lee Ferry is roughly 12.4 MAF.

The 4.1 MAF gap is called paper water: water that exists on legal documents but not in the river. For the first decades of the compact, rapid reservoir construction absorbed the discrepancy by drawing down Lake Powell and Lake Mead. Since 2000, that buffer has eroded. The reservoirs have shed roughly 17 MAF of combined storage — a volume equivalent to 1.4 years of average flow.

A further 2.1 MAF of tribal water rights remain quantified but largely unexercised. When tribes build out the infrastructure to deliver this water, structural demand increases by another 17% without any change in supply.

The key distinction: A drought is a temporary deviation below the mean. Structural overcommitment means the mean itself is insufficient. Drought-response tools (conservation, fallowing, shortage sharing) buy time. They don't close a 4.1 MAF structural gap.

Water Budget

Where the Water Goes

The Sankey below shows the average annual water budget for WY2020–2025: supply sources on the left, demand categories on the right. Agriculture consumes nearly half the total, with alfalfa and hay alone accounting for 32%. Municipal and industrial use is large in absolute terms but relatively stable. Reservoir evaporation and riparian ET are structural losses that scale with storage levels.

Colorado River Annual Water Budget — Average WY2020–2025

Sources: Richter et al. 2024 (Science Advances) — crop-specific consumption; BOR Water Accounting Reports — M&I; USGS/DRI — reservoir evaporation; NASA GRACE — groundwater depletion; BOR Natural Flow Dataset — Lee Ferry natural flow.

Structural Trajectory

Supply Flat, Demand Rising

Supply has remained bounded by hydrology since 2000, oscillating around the 12.4 MAF average. Total demand has grown slowly but steadily as municipal populations expand and tribal water rights are incrementally exercised. The legal allocation ceiling of 16.5 MAF has been a ceiling on paper since 1922. The vertical line marks 2026 — the present operating point.

Colorado River Demand vs Supply Trajectory, 2000–2050

Supply values 2000–2025 from BOR Lee Ferry natural flow data. 2026–2050 supply is 20-year moving average (flat assumption). Demand trajectory from BOR water accounting + Reclamation 2026 LTSP projections. Legal allocation is fixed at 1922 compact + 1944 treaty values.

Finding

A Structural Feature, Not a Drought Artifact

ADM Finding
The paper water gap is baked into the legal framework, not induced by climate. Every drought-response measure — conservation, fallowing, shortage sharing — operates within a system that was over-allocated before any of them were needed.

This distinction matters for policy design. Shortage-sharing agreements (like the 2007 Interim Guidelines and the 2023 Consensus Agreement) define who absorbs pain when storage falls below trigger elevations. They are distributive mechanisms, not supply mechanisms. They answer "who goes without" but not "where does the missing 4.1 MAF come from."

Demand reduction at the 2–3 MAF scale — the magnitude needed to close the structural gap — requires either permanent fallowing of agricultural water rights, large-scale transfers, or a renegotiation of the compact itself. All three are politically consequential. The model doesn't resolve which is right. It establishes that one of them is necessary.

Limitations

What This Model Doesn't Capture

The 12.4 MAF average supply figure is a 20th-century mean. Paleo-reconstructions suggest the long-term (500-year) average is closer to 11.5 MAF, which would widen the structural gap to 5.0 MAF. Climate projections through 2050 suggest continued aridification — BOR's most likely scenario in the 2023 LTSP models a 7–9% reduction in natural flow. Neither adjustment is incorporated into this investigation's structural gap estimate; both would make the finding more severe.

The tribal unexercised gap of 2.1 MAF is a rough aggregate; individual tribal settlements vary in their terms, timing, and infrastructure requirements. The trajectory chart models this as a gradual linear ramp, which likely understates the lumpiness of actual tribal buildout.