Doctrine 25 names external protection as one of five layers where scientific value becomes exposed. Practitioners tend to treat that layer as a single question, answered once, for an entire institution: is this asset protected by law, or is it not.
That question is malformed. It assumes an institution holds one kind of exposure. Institutions holding long-term scientific infrastructure typically hold several assets and functions at once, and each one carries its own default authorization threshold before it can be altered or removed. That threshold exists independent of any specific commitment made about it. It is baked into the statutory and administrative architecture underneath the institution, not into whatever promise the institution’s leadership happened to make this year.
This means a single site can sit on several different floors at once, and most stewardship diagnostics never separate them.
The Mistake in Plain Terms #
Ask “is this site protected” and you get one answer for several different things that are not protected the same way, or by the same authority.
The category can rest on statute while the site does not. Congress can authorize a whole class of scientific infrastructure without naming a single site within it. The category is durable. The individual site is not automatically durable just because it belongs to a durable category.
A site’s operating status can rest on administrative designation alone. An agency can create a research site under statutory authority without that site itself ever being named in law. The statute permits the designation. It does not protect the designation from being withdrawn by the same kind of administrative action that created it.
The active research program at a site can rest on something weaker still. A cooperative agreement, an interagency arrangement, or a partnership contract can carry the actual science forward, and that instrument answers to none of the authorities protecting the land or the category.
A prestige designation from a second institution can layer on top of all of it, and that designation belongs to whoever granted it. It does not transfer authority over the underlying site to the granting body, and it does not survive on its own if the underlying designation is withdrawn.
Four instruments, four different issuing authorities, four different floors, all sitting on the same acre of ground.
The Hubbard Brook Case, Layer by Layer #
Hubbard Brook Experimental Forest is not one designation. It is a stack of four.
Layer one, the category. The Forest Service’s authority to create experimental forests traces to the McSweeney-McNary Forest Research Act of May 22, 1928, which authorized a coordinated federal forest research program and the establishment of roughly fourteen regional experiment stations nationally. That Act was repealed and replaced by the Forest and Rangeland Renewable Resources Research Act of 1978, which is the current statutory basis for the experimental forest network. This is the one layer in the stack that Congress built. It authorizes the category. It does not name a single site.
Layer two, the site designation. Hubbard Brook itself was established in 1955 by the USDA Forest Service as a center for hydrologic research. That is an agency action, taken under the statutory authority above, not a congressional act naming Hubbard Brook specifically. No bill created this site. An agency memo did, and nothing in the record suggests a different kind of instrument would be required to un-create it, short of the Goodlander-Pappas bill now pending, which would be the first instrument to require congressional approval before this specific layer could be removed.
Layer three, the research program. The Hubbard Brook Ecosystem Study, the long-term watershed research that produced the acid rain findings and everything built on them since, began in 1963 under a cooperative agreement among F. Herbert Bormann, Gene Likens, Noye Johnson, and Robert Pierce of the Forest Service. That is a contractual instrument between named individuals and institutions. It answers to none of the authorities protecting the land or the site designation above it.
Layer four, the external prestige designation. In 1988, Hubbard Brook was designated a Long-Term Ecological Research site by the National Science Foundation, a different federal agency entirely, with its own funding and its own criteria. This layer adds visibility and a funding stream. It confers no authority over the other three layers, and losing any of them does not require NSF’s involvement or consent.
Why the Stack Matters More Than Any Single Layer #
The naive read of Hubbard Brook says the site is protected because it carries an NSF LTER designation and decades of scientific significance. That read checks layer four and stops.
The operative floor is not the strongest layer in the stack. It is the weakest layer that something depends on. If layer two, the bare agency designation, is withdrawn, the LTER status has nothing left to attach to, the cooperative agreement has no site to operate on, and the category authority in layer one becomes irrelevant to this specific acre of ground. A single administrative action at the second-lowest floor can collapse everything built on top of it, regardless of how durable or prestigious the upper layers appear.
This is the same failure Doctrine 18 names for commitments, applied to structure instead of promises. A Level 5 commitment protecting layer four does nothing if layer two is where the exposure sits. Practitioners who inventory an institution’s protections by listing its most impressive instrument, the statute, the prestige designation, the funding line, are measuring the wrong thing. The question is never “what is the strongest protection this site has.” It is “what is the weakest layer this site’s continued existence actually depends on, and what would it take to remove that layer specifically.”
The Network View: Not Every Layer Generalizes the Same Way #
Hubbard Brook is one site inside a roughly 80-site national network of experimental forests and ranges, established progressively since 1908. The four-layer stack above does not scale uniformly across that network, and assuming it does is its own diagnostic error.
Layer one is shared. The category authority is a single statute covering every site in the network at once. A change here is a network-wide event by definition.
Layer two is replicated, not shared. Every site carries its own instance of the same type of instrument, an individual agency action that created that specific site. Hubbard Brook’s designation and any other site’s designation are separate acts. Withdrawing one has no effect on the others, even though all of them trace back to the same layer-one statute. A network built on a shared foundation can still be dismantled one unshared designation at a time.
Layer three has no network-level form at all. Each site’s active research program depends on its own cooperative agreements, naming its own institutions and researchers. There is nothing to protect at the network level here, because nothing at this layer exists at the network level.
Layer four is present at only a minority of sites. Formal NSF Long-Term Ecological Research status has extended to a small number of experimental forests, not the network as a whole. Most sites in the network have no fourth layer to lose, because they never had one.
This asymmetry is the reason a site-specific protective instrument systematically undercounts the exposure it is trying to fix. Layer one already looks solid across the entire network, so an observer checking only the strongest, most visible layer concludes the network is protected. But layer one was never where the exposure lived. The exposure sits at layer two, and layer two is unprotected precisely because it is replicated rather than shared: each site’s designation stands or falls on its own. A rider or bill written to protect one named facility leaves every other site’s layer-two designation exactly as exposed as it was before, no matter how durable the shared statute above it looks.
Practitioner application at network scale. When a threat is diagnosed as touching one prominent site, check whether the vulnerable layer is shared or merely replicated across the network. If it is replicated, the fix has to name the network, not the facility, because protecting one instance of a replicated layer does nothing for the others. This is a structural conclusion, not a policy preference: it follows directly from where each layer actually sits, independent of which site currently has the public’s attention.
The Cross-Reference to Doctrine 18 #
Doctrine 18 Companion measures the durability of a specific commitment. This companion measures the baseline architecture a function depends on before any commitment is made about it at all. The two are independent axes, and both are required.
A Level 5 statutory commitment can be pointed at the wrong layer. Appropriations language directing staffing funds is genuinely Level 5, durable, hard to reverse, and it does nothing for a site whose exposure sits at the bare designation layer rather than the funding layer. The money is real. It is protecting the wrong floor in the stack.
The Armistice Problem, addressed in the Doctrine 18 companion, adds the second half of this diagnostic. Doctrine 18 asks how durable a given commitment is. This companion asks which layer it is actually pointed at. A commitment can fail either test independently, and celebration should be withheld until both have been checked, not just one. A Level 5 commitment pointed at the correct layer is the only combination that earns the reaction people instinctively give any Level 5 commitment on sight. A Level 2 commitment, however warmly received, is a bullet dodged, not an armistice, regardless of which layer it touches. And a Level 5 commitment pointed at the wrong layer, real, durable, and irrelevant to the actual exposure, deserves the most caution of all, because its evident strength is exactly what makes the misdirection hard to see. Two questions, not one: how durable is this, and is it protecting the floor that is actually weak.
Practitioner Application #
Before accepting that a long-term scientific function is protected, decompose it into the separate instruments it actually depends on: the category authority, the site-specific designation, the operating agreement that runs the science, and any external prestige or funding designation layered on top. Identify who issued each one and what it would take to withdraw it. Then ask which layer is weakest, and whether any Doctrine 18 commitment currently in force is actually pointed at that layer, rather than at whichever layer happens to be most visible.
A site that looks well-protected in aggregate can have nothing at all pointed at the one layer holding the rest of the stack up.
References #
McSweeney-McNary Forest Research Act, 45 Stat. 699 (May 22, 1928), repealed by Forest and Rangeland Renewable Resources Research Act of 1978, Pub. L. 95-307.
Hubbard Brook Ecosystem Study. “About the Forest.” hubbardbrook.org.
USDA Forest Service, Northern Research Station. “Hubbard Brook Experimental Forest.” research.fs.usda.gov.
LTER Network. “Hubbard Brook LTER.” lternet.edu.
Field notes and examples #
Last Updated on July 29, 2026
