The AI Journal

An operating plan for a naturally intermittent stream should reserve a season in which artificial releases do not maintain continuous surface flow. Departures should require an explicit ecological justification. This is a modest but consequential reversal of the burden of proof: a manager proposing year-round wetting would have to account for the habitat it removes. The reservation would apply only where evidence establishes recurrent natural drying, and would protect a locally determined range of timing and duration. It would neither impose an annual drying date nor authorise abstraction to manufacture one.

The biological reason is concrete. Sampling terrestrial invertebrates in dry riverbeds and adjacent riparian habitats in Australia and Italy, researchers found that 66 of 320 taxa appeared only in riverbed samples; assemblage differences reflected turnover in taxa rather than merely altered abundances (Steward et al., 2011). Occurrence only in those samples does not establish obligate dependence on dry channels. It does, however, defeat the convenient assumption that maintaining the neighbouring bank automatically conserves everything inhabiting the exposed bed. The dry channel has a conservation claim that a wetted-area target cannot represent.

Simply providing an equivalent area of permanently dry ground would not establish compensation for that loss. In an experiment in six intermittent pools in Queensland, researchers compared invertebrates in leaf packs kept wet, allowed to dry, or kept terrestrial. Assemblages in previously submerged packs remained distinct from those in continuously terrestrial packs after drying (Dell, Alford and Pearson, 2014). This experiment does not identify a required duration of exposure. It establishes why habitat history belongs in the decision: land that has recently been underwater can offer conditions absent from land that stayed dry.

The proposal extends an established principle into an operational constraint. Magnitude, frequency, duration, timing and rate of change are central to the natural flow regime framework (Poff et al., 1997). A reservation for drying would give those dimensions practical force when the flow reaches zero. Its specification should describe permissible supplemental releases during the expected drying season, the reaches allowed to disconnect, and the observations that would trigger reconsideration. Natural rainfall would remain free to interrupt the dry interval.

A universal quota of dry days would be indefensible. An analysis of 25,207 drying events at 894 United States gauges found multiple drying regimes at 77 percent of gauges over the study period (Price et al., 2021). That classification supplies no ecological optimum, but shows why a stream cannot safely be assigned one immutable drying pattern. I propose that each reservation instead draw on relatively unaltered local records and comparable reaches, track exposed habitat as well as discharge, and be evaluated across several wetting cycles. Its success criterion would be persistence of the seasonal assemblages for which it was adopted.

The strongest objection is that this gives historical hydrology priority over living organisms. If an isolated pool now holds the last population of a vulnerable aquatic species, withholding water to preserve a former seasonal pattern could make restoration an instrument of extinction. More generally, neither riverbed surveys nor leaf-pack experiments demonstrate that their terrestrial beneficiaries outweigh aquatic losses from withholding releases. The literature warrants recognising a cost of permanent wetting; it does not settle every trade-off in favour of drying.

That objection requires an exception with teeth. Where evidence indicates that retaining water is necessary to avert irreversible loss, the reservation should yield, and the intervention should be assessed for its effects on surrounding seasonal habitat. But an exception justified by a threatened population differs materially from treating every additional wetted day as an improvement. The proposed rule makes the displacement visible and requires a reason for accepting it. A stream’s capacity to sustain successive communities deserves protection even when one community is temporarily out of sight.