Fieldnote
A watershed crosses boundaries
Overlapping jurisdictions and overlapping models cannot decide what is the same place
A watershed is defined by water, not by law. Cedar Creek drains a valley that two counties divided along a road a century ago, so its upper reaches answer to one water office and its lower reaches to another, while a bioregional council and a conservation district each draw their own lines across it. Every one of those institutions has a legitimate interest in the creek, and every one of them would, left to itself, describe it through its own boundaries. This note is about the two mistakes that overlapping boundaries invite: treating a jurisdiction as the identity of the place, and treating spatial overlap as proof that two descriptions are about the same place.
The first mistake is easy to see once named. If Cedar Creek's identity encoded either county, the creek would have two identities from the start, or one that was wrong for half its length. Document 08 avoids this by keeping jurisdictions as relationships: the watershed's RID, 451.4ZFNMDBX5YHB833N (Cedar Creek), is related to both counties at once, each relationship with its source and effective dates. Even the jurisdiction segment of a registered name is only historical issuance metadata, which must not change when the resource's governance does (§8.2).
Why overlap is not identity
The second mistake is subtler, because spatial overlap feels like evidence. Suppose the county water office's delineation of Cedar Creek and the council's delineation overlap by ninety per cent. Surely they describe the same watershed? Often they do — but overlap cannot tell you that. A delineation of the "Upper Cedar" headwaters overlaps the council's polygon entirely and is still a part, not the whole. A groundwater basin can overlap a surface watershed almost exactly and be a different living system with different behaviour and different stewards. Two fields on either side of a fence may share most of a sampling grid and remain two fields. Geometry is a representation, produced by someone's method at some date; how much two representations overlap measures the similarity of the representations, not the identity of what they represent.
That is the general rule of RSM ID applied to places. A spatial join is a search tool, and a very useful one for discovering candidate equivalences. It is not an identity decision. If the county and the council have each independently registered an identity for what turns out to be the same creek, reconciling them needs evidence and an authorized decision, recorded so that it can be audited and, if necessary, questioned later.
A worked case
In the scenario, both the council's space and the county water office's space describe Cedar Creek and cite the same RID, because both chose to refer to the registered watershed rather than mint their own. Their delineations disagree at the eastern divide. A spring near that divide falls under the upper county's rules on the council's map and under the lower county's on the county's map. Nothing about the creek's identity is in doubt; the governance metadata that policy relies on is. Document 08 says what a resolver must do then: never assume the issuance jurisdiction settles the applicable policies, and apply explicitly defined handling rules — including conservative denial — where location is missing or disputed (§8.6). Monitoring data about the spring is therefore withheld from public requesters until the dispute is resolved, while the watershed itself keeps resolving normally.
What remains open
Two questions stay with governance. The first is what continuity means for a living system whose extent genuinely changes: if a reservoir floods part of a sub-catchment, or a river is diverted, is the watershed still the same referent? Document 08 says that an updated boundary changes the description and spatial references, "not necessarily the underlying RID" (§13.3), which leaves room for cases where it does; the criteria for those cases are not yet written (SPEC-04). The second is how disagreements between authoritative spaces about one place should be mediated — which delineation a permit regime relies on, and who decides — which is a question of authority scope rather than identity (SPEC-06). The habit worth keeping in the meantime is to let the shared RID make disagreements visible, and to let only governed decisions, never overlap, change what is the same.