This is the threshold that decides whether a foot of drawdown is harmless or permanent. The critical head is simply the water level corresponding to the historic-maximum effective stress. Stay above it and compaction is elastic; fall below it — even briefly — and you buy permanent subsidence.
The preconsolidation stress σ′pc is the greatest effective stress the sediment skeleton has ever experienced. Because effective stress is set by head (page 02), there is an equivalent water level — the critical head hc — at which σ′ = σ′pc:
The behavioral switch is a simple comparison:
Critical head is not fixed. Whenever heads reach a new low, that becomes the new historic maximum stress — so hc moves down to the new low and stays there.
This is why drought years matter so much: each new low both causes permanent compaction and resets the threshold lower for next time.
| Total subsidence at end | — ft |
| Permanent (inelastic) portion | — ft |
| Recoverable (elastic) portion | — ft |
| Final critical head (vs. start) | — ft |
The margin is how far heads can fall before crossing the initial critical head. Set it to 0 and every decline is permanent.
The sustainable target is to hold seasonal lows above the critical head. Within that envelope, the basin can be pumped hard seasonally and still rebound with negligible permanent subsidence.
A single new record low — often during drought — converts elastic storage into permanent compaction and lowers hc. The damage is locked in even if the next year is wet.
Critical head gives SGMA a physically defensible basis for a groundwater-level minimum threshold tied directly to the subsidence indicator: protect the surface by protecting the head. DWR's Land Subsidence BMP (2026) builds its threshold guidance on exactly this logic.
σ′pc (and thus hc) is estimated from: (1) the break in slope on a field stress–strain plot of compaction vs. effective stress (Riley's method, page 07); (2) the historic minimum water level in long records; and (3) lab consolidation tests on core (Casagrande construction on the e–log σ′ curve).
Where the modern historic low already exceeds anything before it, the aquifer is being loaded in the virgin range continuously, and hc simply tracks the ongoing decline.
Different interbeds can have different preconsolidation stresses depending on their loading history (e.g., past erosion, desiccation, or earlier pumping). Critical head is therefore a property of a specific interval, not a single number for the whole column. Multiple-depth piezometer–extensometer installations (page 08) resolve this.
Helm's stress-dependent models (1976) and the Holly-site analysis (Sneed & Galloway 2000) formalize how σ′pc varies and evolves.