Interties as Provable Instruments

The Seam Inventory

Helios Grid’s first brief (BGL/HG—01) argued that the binding constraint on Canadian grid integration is not prediction but proof: thirteen semi-autonomous systems, seven incompatible market designs, and no neutral way for any province to verify that a proposed cross-border exchange is correct and fair under its own rules (C.D. Howe Institute; CER). That brief treated the seams — the physical and institutional boundaries where one control area ends and another begins — as a single undifferentiated problem. This document corrects that simplification. The seams are not one problem. They are roughly a dozen distinct instruments, each with its own hardware, its own ownership structure, its own governance, its own historical flows, and in several cases its own decades-long dispute record. Any evidence layer that proposes to referee them must be specified against each seam individually, and that is what the sections that follow attempt.

Some definitions first, because the vocabulary is where public discussion of interties usually goes wrong. An intertie is a transmission facility crossing a control-area boundary. Its nameplate or design rating is what the conductors and converters were built to carry. Its total transfer capability (TTC) is what studies say the interconnected system can support without violating reliability criteria, and its available transfer capability (ATC) is what remains after existing commitments and margins are netted out — the quantities transmission providers must post under open-access rules descended from FERC Orders 888 and 889 (FERC; NERC Glossary of Terms). The persistent Canadian pattern, documented seam by seam below, is that usable capability sits far below nameplate, that the derating is decided unilaterally by one side of the seam, and that the studies justifying it are not reproducible by the other side. That last property is the one this lab cares about. A derating that cannot be independently re-derived is, functionally, a tariff — whatever its engineering merits.

The physical topology imposes a second layer of structure. Canada spans three of North America’s four synchronous interconnections: British Columbia and Alberta operate inside the Western Interconnection under WECC; Saskatchewan, Manitoba, Ontario, and the Maritimes operate inside the Eastern Interconnection; and Quebec operates as its own asynchronous interconnection, coupled to every neighbour through HVDC converters or radial arrangements rather than synchronous AC ties (NERC; NPCC). This means two of the country’s most consequential seams — Alberta–Saskatchewan and everything touching Quebec — are electrically decoupled by construction: power crosses them only where a converter station has been built and only up to that converter’s rating. Frequency disturbances do not propagate across them; neither does inertia; neither, conveniently for the analyst, does ambiguity about what flowed. A converter station is a metered bottleneck, and metered bottlenecks are where proof is cheapest.

The scale of what crosses these seams is routinely misstated in both directions. Canada’s international electricity trade is large and visible — on the order of 51 TWh of exports in a typical year, concentrated in Quebec, Ontario, Manitoba, and British Columbia (CER) — while inter-provincial trade is smaller and dominated by a single legacy flow: Churchill Falls energy moving from Labrador into Quebec under the 1969 contract, on the order of 30 TWh per year, which inflates any aggregate statistic it appears in (Hydro-Québec; StatCan Table 25-10-0021). Strip that flow out and the remaining inter-provincial exchange is thin — thin enough that the McNeill converter’s 150 MW, a rounding error on any US interface, is a nationally significant asset on the west–east axis (CES-Energy). The asymmetry is the finding: the Canadian grid trades north–south at industrial scale and east–west at pilot scale, under governance that treats the former as routine commerce and the latter as episodic diplomacy.

What follows is organized as an audit, not an argument. Sections two through ten walk the seams west to east and then south: BC–Alberta, Alberta–Saskatchewan, Saskatchewan–Manitoba, Manitoba–Ontario, Ontario–Quebec, Quebec–New Brunswick, the Maritime seams, and the international interface as a body. Two sections then treat the governance record — the Atlantic Loop failure against the Wasoqonatl success, and the five federal priority interties announced in June 2026 — and one dissects the contract structures that actually move power across seams today. The final section specifies the evidence architecture: the four proof obligations a deterministic layer must discharge at any seam — metering lineage, constraint attribution, counterfactual dispatch, and settlement reconciliation — and reports, in redacted form, what two Crown utilities said when we put that specification in front of them (BGL Pilot Minute PM—09, internal, redacted). Every figure in this document carries either a public attribution or an internal derivation record from the lab’s Seam Ledger program; the distinction is always visible in the citation.

Method: The Seam Ledger

The Seam Ledger is the lab’s internal program for making claims about interconnections auditable. It opened in November 2025 as a governed corpus with a narrow charter: for every seam on or touching the Canadian grid, maintain a dossier that separates what was built, what is declared, what is usable, and what actually flowed — as four distinct fields with independent lineage, because the public discourse habitually collapses them into one number (BGL Seam Ledger Program Charter SL—00, internal, 2025-11). Eleven seam dossiers were opened in the first tranche (BGL Seam Dossiers SD—01 through SD—11, internal); this document is the public condensation of that corpus as of April 2026.

The evidentiary base is deliberately unglamorous. The declared layer comes from tariff and planning documents: OASIS postings and open-access tariffs where they exist (FERC Order 889; OATI postings), IESO reliability outlooks and intertie pages (IESO), AESO ISO rules and market statistics (AESO), Crown-utility revenue-requirement and integrated-resource-plan filings before the BCUC, the Manitoba PUB, the Régie de l’énergie, and the Nova Scotia UARB, and NERC and NPCC standards and assessments where they bind (NERC; NPCC). The flow layer comes from operator publications and the CER’s commodity statistics on international trade, which remain the only consistently granular public record of what crosses any Canadian border (CER Commodity Statistics). The gap between the two layers — declared capability that never clears, demonstrated flows that exceed stated limits during emergencies — is where the ledger’s findings live (BGL Engineering Note EN—71, internal, 2025-12).

On top of the dossiers sits a derivation corpus. Between December 2025 and March 2026 the lab assembled an interval dataset spanning roughly 41,000 hourly intervals — approximately four and two-thirds years of history — fusing nineteen public and licensed sources: IESO nodal and intertie schedules, AESO pool and intertie data, published Hydro-Québec and Manitoba Hydro export aggregates, EIA cross-border flow series, CER export permits and volumes, water-supply bulletins, and weather reanalysis (BGL Interval Corpus IC—41K, internal, 2026-03). D.A.E. completed full derivation runs across the corpus in March 2026: every cross-seam correlation, every capability-utilization gap, and every counterfactual reported below traces to a sealed run record with bit-identical replay (BGL Run Ledger SL—1042, internal, sealed; BGL Run Ledger SL—1088, internal, sealed). Where this document says “the interval record shows,” that is what it means: a deterministic computation over a frozen corpus, re-runnable to the same result, not an analyst’s recollection of a chart.

Three methodological cautions bind everything that follows. First, public-data asymmetry: Ontario and Alberta publish operational data at five-minute to hourly granularity because their market designs require it; the vertically integrated Crowns publish annual reports and rate filings. The ledger therefore knows the IESO and AESO seams at three to four orders of magnitude finer resolution than the BC Hydro, SaskPower, or Hydro-Québec seams, and any cross-seam comparison in this document inherits that asymmetry explicitly (BGL Engineering Note EN—74, internal, 2026-01). Second, derating opacity: several of the most consequential capability numbers in Canada — the BC–Alberta usable limit above all — appear in workshop slide decks and stakeholder presentations rather than in auditable studies (BC Hydro, 2025). The ledger records such numbers as declarations with a named declarant, not as verified capabilities. Third, survivorship in the dispute record: seams that work generate no filings. The dispute histories below are necessarily histories of failure, and a seam with a thin dispute record may be healthy or may simply be idle; the interval record distinguishes the two, the archive alone does not (BGL Internal Memorandum M—2026—003).

A note on what the Seam Ledger is not. It is not a market monitor, it holds no non-public operator data for any Canadian seam, and nothing in it constitutes a compliance finding against any utility. It is a demonstration corpus: an existence proof that a complete, versioned, replayable evidentiary picture of the country’s seams can be assembled from the public record by a disciplined ingestion layer — and a measurement of exactly where the public record runs out, which turns out to be the more useful product. The places the ledger goes dark are the places a provable intertie would need instrumentation first (BGL Seam Ledger Annual Note, internal, 2026-04).

Seam One: British Columbia–Alberta

The BC–Alberta seam is the country’s clearest demonstration that a wire’s nameplate and its usable capability are different facts with different owners. The interface — WECC’s Path 1 — is anchored by a 500 kV line between the BC Hydro system and the Alberta grid, supplemented by lower-voltage ties, and was designed for on the order of 1,200 MW of transfer (WECC Path Catalog; BC Hydro, 2025). In practice the usable limit has sat around 600–735 MW for most of the past decade and at times lower, a derating driven substantially by conditions on the Alberta side: system-strength and voltage-stability constraints in the province’s southwest, the retirement of coal units that once provided synchronous support near the interface, and operational limits the AESO applies to protect its own system under contingency (AESO; BC Hydro, 2025). Both operators acknowledge the gap; an intertie “restoration” workstream has existed in various forms since the early 2010s, and the interval record shows the gap is not closing on its own — hourly utilization of even the derated limit averaged well under half across the corpus window, with the binding direction flipping seasonally (BGL Seam Dossier SD—01, internal; BGL Run Ledger SL—1042, internal, sealed).

The governance structure explains more of the derating than the physics does. British Columbia’s side of the seam is operated by BC Hydro, a Crown monopoly regulated by the BCUC, with its trading arm Powerex active in every Western market; Alberta’s side is operated by the AESO, an independent system operator running an energy-only market in which the intertie participates roughly as a very large, partially reliable market participant (AESO; Powerex). The two sides optimize different objective functions. Powerex arbitrages: BC’s roughly 7,000 MW of storage-backed hydro flexibility is most valuable when it can buy Alberta’s surplus wind at low or negative pool prices and sell capacity back into Alberta’s scarcity hours — a diurnal storage swap executed through the market rather than through any treaty (AESO Annual Market Statistics; MSA). The AESO, meanwhile, must treat imports as interruptible for reliability purposes and has litigated for years, through its own rule processes, how intertie flows should pay for the transmission system they lean on. The result is a seam where commercial logic wants more capability and each side’s institutional logic individually prefers the flexibility of not guaranteeing it.

The dispute record is correspondingly procedural rather than dramatic: contested ISO rule changes on intertie restoration cost allocation, stakeholder complaints about the opacity of transfer-limit derivations, and a standing grievance — visible in Alberta market-participant filings across a decade — that the effective import capability materializes and evaporates without an auditable trail (AESO rule filings; MSA reports). April 2024 sharpened the stakes: during Alberta’s grid alerts, with provincial wind output collapsing toward zero around sunset, imports across this seam were part of what kept the province whole, and the post-mortems noted both their value and their fragility (AESO Market Surveillance Administrator, 2024). BC’s drought years cut the other way — in fiscal 2024 BC Hydro imported a record 13,600 GWh, a material share of it from Alberta’s gas-heavy pool, monetizing the same seam in reverse (Globe and Mail; CBC). This is the only Canadian seam where both directions have recently been existential for one side. It is also, not coincidentally, the seam the federal priority-intertie list treats as a restoration project rather than a new build (National Observer, 2026).

What would proof look like here? The Seam Ledger’s reading is that the BC–Alberta seam does not primarily need more copper; it needs the derating itself to become a reproducible object. Today the usable limit is a declaration — published in workshop material, revisable by either operator, derived from studies the other side’s analysts cannot re-run (BC Hydro, 2025; BGL Engineering Note EN—74, internal). A provable instrument would publish, for every operating hour, the binding constraint that set the transfer limit — named equipment, named contingency, named stability margin — in a form both operators and both regulators can recompute from a shared model state, the way any ISO’s internal congestion is already attributed to named constraints under tariff-defined optimization (PJM Manual 11; CAISO Tariff §34). The interval record suggests the prize is material: across the corpus window, hours in which Alberta pool prices exceeded $200/MWh while the interface sat below its own declared limit — capability posted, not scheduled, not explained — number in the hundreds, and every one of them is an attributable coordination failure worth a specific, computable amount of money (BGL Run Ledger SL—1042, internal, sealed; BGL Derivation Record DR—SEAM—0117, internal). No party has to accept the lab’s number. That is the point of the architecture: they can re-run it.

Seam Two: Alberta–Saskatchewan, the McNeill Bottleneck

The Alberta–Saskatchewan seam is the smallest consequential piece of transmission infrastructure in the country, and the most structurally interesting. Alberta operates inside the Western Interconnection; Saskatchewan operates inside the Eastern Interconnection; the two systems are not synchronized and cannot be joined by an ordinary AC line. The entire electrical connection between Western and Eastern Canada therefore runs through one facility: the McNeill back-to-back HVDC converter station near the Alberta–Saskatchewan border, rated at roughly 150 MW, owned on the Alberta side by ATCO Electric, with exports at times restricted to 75 MW (CES-Energy; ATCO). Every megawatt-hour that has ever moved between the western provinces and the rest of Canada’s Eastern-Interconnection grid has passed through power electronics at this one site. For scale: 150 MW is roughly two percent of Alberta’s peak demand, under one percent of the aggregate transfer capability Canada maintains toward the United States, and less than a single modern wind farm (AESO; CER). The west–east grid, as a physical fact, is a rounding error.

The asynchronous boundary is usually described as the problem. The Seam Ledger’s reading is closer to the opposite: it is the one honest seam in the country. Because McNeill is a converter, flow across it is a scheduled, metered, fully controllable quantity — there is no loop flow, no inadvertent wheel-through, no ambiguity about path. The DC link decouples the two systems’ frequencies, so neither province’s disturbances propagate into the other, and the interchange schedule is the physical flow to within converter losses (NERC; IEEE). Contrast the AC seams further east, where scheduled and actual flows diverge hourly and a standing settlement apparatus exists solely to reconcile the difference (NERC BAL-006). From an evidentiary standpoint, a back-to-back converter is the ideal instrument: one metering point, one control system, one number. The scandal of McNeill is not that it is DC. It is that it is 150 MW.

The governance of the seam is as thin as the wire. Saskatchewan’s side is SaskPower, a vertically integrated Crown with no open wholesale market, self-dispatching a fleet that remains among the most carbon-intensive in Canada — coal and gas supplied the large majority of provincial generation through the corpus window, with federal Clean Electricity Regulations now binding the trajectory (SaskPower; ECCC; Canada Gazette SOR/2024-263). Alberta’s side is the AESO’s energy-only market. There is no joint planning body, no shared adequacy assessment, and no standing mechanism by which Alberta wind curtailed at negative pool prices could be committed against Saskatchewan coal displacement — despite the two systems being adjacent, complementary, and jointly responsible for most of the country’s remaining electricity emissions (ECCC; AESO). The interval record makes the forgone exchange concrete: hours in which the Alberta pool cleared below $10/MWh while Saskatchewan was necessarily dispatching fossil units on its own system occur thousands of times in the corpus, and the McNeill schedule responds to only a fraction of them, bounded by its rating and by the absence of any contractual machinery to exploit the rest (BGL Seam Dossier SD—02, internal; BGL Run Ledger SL—1042, internal, sealed).

This seam is where the federal intertie program will succeed or fail, because the flagship of the June 2026 priority list — a Regina–Winnipeg corridor expansion of up to 2,000 MW — lands one province east of it, and because every serious study of west–east integration eventually collides with the cost of converter capacity across the WECC boundary (National Observer, 2026; PMO, 2026). A meaningful Alberta–Saskatchewan expansion means new back-to-back converter stations — capital-intensive, long-lead assets whose business case must be underwritten by two governments with no shared market, no shared regulator, and a history of treating each other as fiscal rivals rather than counterparties. The Seam Ledger’s position is that this is precisely the setting in which the evidence has to precede the steel. A converter expansion justified by a deterministic, jointly re-runnable analysis — here are the 41,000 intervals, here is the counterfactual dispatch with 600 MW of converter capability instead of 150, here is the emissions and cost delta attributed hour by hour, and here is the same computation executed independently by both utilities’ analysts to the same result — is a fundamentally different political object than a converter expansion justified by a consultant’s scenario deck (BGL Derivation Record DR—SEAM—0142, internal; BGL Internal Memorandum M—2026—014, internal). The first is checkable by a skeptical treasury board. The second is an act of faith, and the west–east grid has run out of faith.

Seam Three: Saskatchewan–Manitoba

The Saskatchewan–Manitoba seam is the country’s quiet success, and its structure repays study precisely because nothing about it is dramatic. Both provinces sit inside the Eastern Interconnection, so the ties are ordinary AC: a set of 230 kV and lower-voltage interconnections, reinforced most recently by the 230 kV Birtle transmission project, energized in 2021, which was built for a specific and unusual reason — not reliability, not federal policy, but to physically deliver power SaskPower had already contracted to buy (Manitoba Hydro; SaskPower). The commercial layer came first. In the mid-2010s SaskPower signed a sequence of firm purchase agreements with Manitoba Hydro — an initial 25 MW arrangement, a 100 MW sale over twenty years beginning in 2020, and a further agreement on the order of 215 MW conditioned on new transmission — and the Birtle line exists because those contracts made it bankable (SaskPower; Manitoba Hydro annual reports). Post-Birtle, usable west-to-east capability across the seam sits on the order of a few hundred megawatts, modest in absolute terms but, notably, actually used: the interval record shows sustained, contract-shaped flows rather than the episodic opportunistic exchange that characterizes the western seams (BGL Seam Dossier SD—03, internal).

The complementarity being monetized is the textbook one. Manitoba Hydro operates roughly 5,000 MW of hydro capacity, nearly all of it on the Nelson River system, with reservoir storage that makes its energy firm across seasons; Saskatchewan operates a thermal system whose marginal unit is almost always fossil (Manitoba Hydro; SaskPower). A megawatt-hour moving west across this seam displaces coal or gas with near-certainty — among the highest-value decarbonization arbitrages available anywhere in the country per unit of transmission investment, a point the Canada Electricity Advisory Council’s 2024 report made in general form and this seam demonstrates in particular (CEAC, 2024; ECCC). The federal priority list’s Regina–Winnipeg corridor, at up to 2,000 MW, is best understood as a bet that this existing arrangement scales by an order of magnitude (National Observer, 2026; PMO, 2026).

The Seam Ledger’s interest in this seam is diagnostic: why did this one work when grander schemes failed? The dossier’s answer has three parts. First, the value case was narrow and provable — a named buyer, a named seller, a firm quantity, a price, and a single new line whose cost could be weighed against a contract already in hand. No multi-party benefit allocation, no contested counterfactual. Second, both counterparties were Crowns answerable to provincial treasuries, which made the transaction legible as commerce rather than as constitutional politics; the contrast with the Atlantic Loop’s four-government, two-utility, federally brokered structure is examined in section eleven. Third, and least appreciated, the settlement machinery already existed: as Eastern Interconnection balancing areas operating under NERC standards, the two utilities had decades of routine interchange scheduling, tagging, and inadvertent-interchange accounting between them — the boring apparatus of e-Tags and after-the-fact checkout that makes a firm bilateral contract enforceable in practice (NERC BAL-006; NAESB WEQ standards). The deal could be trusted because its performance was already being measured by systems both sides operated (BGL Internal Memorandum M—2026—007, internal).

But the dossier also records the scaling problem the 2,000 MW ambition will meet. The existing arrangement’s proof burden is light because the flow is small relative to both systems: 315 MW of contracted supply perturbs neither Manitoba’s reservoir management nor Saskatchewan’s unit commitment enough to generate disputes. At 2,000 MW the questions get hard and adversarial. When Manitoba’s Nelson River inflows disappoint — and the drought years in the corpus window show Manitoba Hydro’s net revenue swinging by hundreds of millions of dollars with hydrology (Manitoba Hydro annual reports; Manitoba PUB filings) — who is curtailed first, the export contract or domestic load, and against what published priority? When Saskatchewan’s system leans on the import during a prairie-wide cold snap that stresses Manitoba simultaneously — the February 2021 event demonstrated exactly this correlation structure one border south (FERC/NERC, 2021) — what does firmness mean, and how is it demonstrated in advance rather than litigated afterward? These are evidence problems before they are engineering problems. A 2,000 MW seam between a hydrology-exposed seller and a reliability-exposed buyer requires, at minimum, a shared, continuously updated, independently re-runnable accounting of deliverability under stress — the reservoir state, the inflow forecast bands, the coincident-peak exposure, the curtailment priority stack — or it will generate its first crisis the first dry, cold winter after energization (BGL Derivation Record DR—SEAM—0156, internal; BGL Seam Dossier SD—03, internal). The Wasoqonatl lesson — prove a narrow case — scales only if the proving machinery scales with it.

Seam Four: Manitoba–Ontario

The Manitoba–Ontario seam is the one the map most flatters and the wires most betray. On paper it joins the country’s premier storage-hydro system to its largest load centre; in copper it consists of modest 230 kV ties in the Kenora–Whiteshell corridor with usable transfer capability on the order of 250–300 MW — roughly one percent of Ontario’s peak demand (IESO; Manitoba Hydro). The interval record shows the seam operating mostly as a convenience for northwestern Ontario, a region electrically remote from the province’s own core, rather than as a bulk exchange instrument; flows are small, bidirectional, and shaped more by local conditions around the Lakehead than by any provincial strategy (BGL Seam Dossier SD—04, internal). The contrast with the same seller’s southern interface is stark: Manitoba Hydro moves multiples of this capability into Minnesota, where the 500 kV Manitoba–Minnesota Transmission Project and the connected Great Northern Transmission Line, both energized in 2020, were purpose-built to carry firm hydro sales to Minnesota Power and its partners (Manitoba Hydro; Minnesota Power; EIA). The pattern of HG—01 recurs in miniature: the north–south wires got built because a creditworthy counterparty signed first; the east–west wire stayed thin because no one did.

The dispute history here is a history of studies. A high-capacity Manitoba–Ontario intertie has been proposed, scoped, and shelved repeatedly since at least the 1990s, most consequentially in the early 2010s, when Manitoba Hydro’s planned Conawapa generating station — a roughly 1,485 MW Nelson River project — was conceived substantially as an export machine, with Ontario as the natural anchor customer. Manitoba’s Public Utilities Board, in its 2014 Needs For and Alternatives To review, found the export-led development case unpersuasive at then-current market prices and Conawapa was deferred indefinitely; Ontario, then in surplus and curtailing wind, saw no need to underwrite it (Manitoba PUB NFAT, 2014; Manitoba Hydro). Both judgments were individually defensible and jointly expensive: within a decade Ontario was projecting sustained capacity need through the 2030s, procuring new gas capacity and battery storage at scale, and re-contracting every nuclear refurbishment it could reach (IESO Annual Planning Outlook; Government of Ontario), while Manitoba’s uncommitted hydro flexibility had been sold south or left undeveloped. The option the two provinces declined to price in 2014 is the option Ottawa is now proposing to buy back at 2026 construction costs (National Observer, 2026).

The Seam Ledger treats this seam as the cleanest available case study in counterfactual burden — the evidentiary weight a transmission decision is forced to carry when the benefit case spans two planning regimes that do not share models. The 2014 NFAT record is instructive less for its conclusion than for its structure: Manitoba’s regulator was asked to evaluate export revenues contingent on Ontario demand scenarios, Ontario prices, and Ontario policy, none of which Manitoba’s witnesses could model with standing, and all of which Ontario’s own planning apparatus modelled separately, on different assumptions, for different purposes, with no obligation of consistency (Manitoba PUB NFAT, 2014; IESO planning documents). No institution existed — none exists now — whose job is to hold a single, shared, versioned model of the joint system inside which both provinces’ regulators could interrogate the same intertie with their own assumptions. Each side priced the seam against its private model of the other, and both models were wrong in the directions that killed the project (BGL Internal Memorandum M—2026—009, internal).

What proof requires here is therefore not primarily operational — the seam is too small for its operations to matter yet — but planning-grade: a joint counterfactual apparatus. Concretely: a governed model of the two systems whose input datasets are versioned and disclosed between the parties; scenario runs that either regulator can commission and both can replay bit-identically; benefit attributions — production-cost savings, capacity value, emissions displacement — computed as named quantities with lineage back to the input versions that produced them, in the manner already standard inside ISO planning under FERC Order 1000’s regional processes but absent between Canadian provinces (FERC Order 1000; CEAC, 2024). The lab ran a reduced demonstration across the corpus window: a deterministic re-dispatch of the joint Manitoba–Ontario system under a hypothetical 1,000 MW seam, holding each province’s own published assumptions fixed, attributing the resulting delta hour by hour (BGL Run Ledger SL—1088, internal, sealed; BGL Derivation Record DR—SEAM—0163, internal). The specific numbers matter less than the property: the run is a document both a Manitoba PUB panel and an Ontario Energy Board panel could, in principle, execute themselves and get the same answer. The 2014 record shows what happens without that property. The 2026 federal list will test whether anyone has learned it.

Seam Five: Ontario–Quebec

The Ontario–Quebec seam is the largest inter-provincial interface in the country and the only one with a functioning, contract-backed exchange regime — which makes it the best evidence for what Canadian seams can do, and the clearest measure of how far even the best one falls short of its physics. Because Quebec operates as its own asynchronous interconnection, every tie on this seam is either an HVDC conversion or a radial arrangement in which a pocket of load or generation is switched wholly onto the other system; the anchor asset is the Outaouais back-to-back converter station near L’Orignal–Hull, rated at 1,250 MW and in service since 2009, owned and operated on the Quebec side by Hydro-Québec TransÉnergie (Hydro-Québec; IESO). Aggregate interconnection capability between the two provinces is conventionally stated at roughly 2,700–2,800 MW spread across multiple points from the Abitibi to the St. Lawrence, though the simultaneously usable figure is materially lower and configuration-dependent — radial ties serve specific pockets and cannot be summed like a single busbar (IESO Reliability Outlook; BGL Seam Dossier SD—05, internal). The interval record shows the seam doing real work: sustained imports into Ontario across large parts of the corpus window, punctuated by reversals during Quebec’s deep winter peaks — exactly the complementarity the geography promises (IESO intertie data; BGL Run Ledger SL—1042, internal, sealed).

The contractual layer is genuinely sophisticated by Canadian standards. The 2015 Ontario–Quebec electricity trade agreement established, among other elements, a seasonal capacity exchange of 500 MW — Ontario backs Quebec’s winter peak, Quebec backs Ontario’s summer peak, exploiting the near-perfect anti-correlation of the two provinces’ annual load shapes — and a subsequent 2017 arrangement committed Hydro-Québec to deliver on the order of 2 TWh per year to Ontario over seven years, with associated banking provisions that let Ontario park energy against Quebec’s reservoirs (Ontario Ministry of Energy, 2015; IESO). The capacity swap in particular is the closest thing Canada has to a modern inter-regional resource-adequacy instrument: no energy necessarily flows, but each province’s planning reserve calculation credits the other’s firm commitment, and each avoids building peaking capacity it would use a few hundred hours a year (IESO planning documents; Hydro-Québec). The Seam Ledger’s accounting treats this as the highest-value 500 MW in the inter-provincial system: measured against the avoided cost of new gas peakers on either side, the swap’s implicit value runs to tens of millions of dollars annually, secured by a contract a few pages long (BGL Derivation Record DR—SEAM—0171, internal).

And yet the seam’s recent history is a study in how quickly complementarity assumptions age. The 2015–2017 arrangements were negotiated when Quebec was structurally long: reservoirs full, industrial load flat, export ambitions oriented south. By the corpus window’s end the picture had inverted — Hydro-Québec’s 2035 action plan contemplates on the order of 60 TWh of new demand and between 8,000 and 9,000 MW of new capacity for Quebec’s own electrification and industrial pipeline, low-water years cut export volumes sharply, and the utility has publicly repositioned exports as residual rather than strategic (Hydro-Québec Action Plan 2035; CER Market Snapshots). Ontario, meanwhile, faces sustained capacity need through the nuclear-refurbishment trough of the late 2020s (IESO Annual Planning Outlook). Both provinces now want the same product — firm winter-adjacent capacity — at the same time. A seam regime built on anti-correlated needs has no native answer for correlated scarcity; the contracts simply expire into it. The dispute record here is thin precisely because the arrangements were sized conservatively; the corollary is that nothing on this seam has ever been stress-tested by a genuine simultaneous emergency, and the governing instruments are silent on how one would be adjudicated (BGL Seam Dossier SD—05, internal; BGL Internal Memorandum M—2026—011, internal).

The proof problem on this seam is therefore renewal under changed priors. The 2015 agreement could be negotiated on round numbers because the stakes were asymmetric and the surplus obvious. Its successors will be negotiated between two capacity-constrained systems, each holding private models of its own scarcity, each with strong incentives to overstate the firmness it sells and understate the firmness it needs. What a deterministic evidence layer contributes is a shared measurement substrate for exactly the quantities the negotiation turns on: coincident-peak exposure computed identically on both sides’ data; deliverability of the swap under drought, refurbishment-outage, and cold-snap scenarios, run as joint counterfactuals with disclosed inputs; and — the piece the current regime lacks entirely — continuous in-delivery verification, so that a capacity credit claimed in a planning filing corresponds to a demonstrated, hash-lineaged operational capability rather than to a contractual assertion (BGL Engineering Note EN—80, internal; RFC 6962; Haber & Stornetta, 1991). The lab’s counterfactual runs across the corpus window indicate the renewal-era swap wants to be larger, dynamic rather than seasonal, and settled against verified state — and that every one of those refinements is blocked less by commercial appetite than by the absence of a jointly trusted computation both treasuries can audit (BGL Run Ledger SL—1088, internal, sealed). This is the seam where the evidence architecture of section fourteen earns its keep first.

Seam Six: Quebec–New Brunswick, and the Shadow of Churchill Falls

The Quebec–New Brunswick seam carries more capability, more history, and more institutional scar tissue per megawatt than any other boundary in the country. The hardware is venerable: the Eel River converter station, commissioned in 1972 as the world’s first commercial back-to-back HVDC installation, rated at roughly 320 MW; the Madawaska converter near Edmundston, rated at roughly 350 MW, in service since 1985; and additional radial arrangements that bring aggregate Quebec–New Brunswick capability to on the order of 1,000 MW — making New Brunswick, not Ontario, Quebec’s deepest inter-provincial interface relative to system size (NB Power; Hydro-Québec; IEEE). The flows are structurally one-way: Quebec energy moves southeast into New Brunswick and onward, positioning NB Power’s transmission subsidiary as the toll corridor between Quebec’s surplus and the Maritime and New England systems. The interval record shows this wheeling role clearly — New Brunswick functions less as a destination than as a rotary, redistributing Quebec imports, Point Lepreau’s roughly 660 MW of nuclear output, and Maritime thermal generation among five adjacent systems according to conditions (NB Power; ISO-NE; BGL Seam Dossier SD—06, internal).

No account of this seam survives without the dispute history, because the dispute history is the governance. Upstream of everything sits Churchill Falls: the 1969 contract under which Hydro-Québec purchased essentially the entire output of the 5,428 MW Labrador station at prices that declined over the contract’s life to fractions of a cent per kilowatt-hour, renewed automatically in 2016 on terms set in 1969, expiring 2041 (CF(L)Co; Hydro-Québec). Newfoundland and Labrador spent five decades attacking the contract through every available forum — legislative expropriation attempts struck down in Reference re Upper Churchill Water Rights Reversion Act (SCC, 1984), good-faith and unforeseeability arguments dismissed as recently as Churchill Falls (Labrador) Corp. v. Hydro-Québec (2018 SCC 46) — and lost every time, at a cumulative claimed cost to the province conventionally stated in the tens of billions of dollars (Supreme Court of Canada; CBC). The December 2024 memorandum of understanding between the two governments — replacing the legacy rate structure decades early, framing joint development of Gull Island’s roughly 2,250 MW, and re-pricing the relationship for the post-2041 era — is the largest inter-provincial electricity renegotiation in Canadian history, and its eventual definitive agreements will set the reference terms against which every other provincial seller prices firm energy for a generation (Government of Newfoundland and Labrador; Hydro-Québec, 2024).

The seam’s own constitutional moment came in 2009–2010, when the governments of New Brunswick and Quebec negotiated the sale of most of NB Power’s assets to Hydro-Québec for on the order of $4.8 billion — effectively merging the seam out of existence — and abandoned the transaction within months under public opposition intense enough to contribute to a change of government in Fredericton (CBC, 2010). The Seam Ledger reads that episode as the founding data point for this document’s thesis. The transaction failed not on engineering and not, primarily, on price: it failed because no institution could credibly demonstrate to New Brunswickers what the long-run value of their utility, their ratepayer exposure, and their transmission position actually was. Competing valuations were assertions by interested parties; there was no neutral computation to check them against; and in the absence of proof, the public defaulted — rationally — to refusal. Fifteen years later, the same epistemic vacuum shapes daily operations: the terms on which Hydro-Québec energy transits New Brunswick, the allocation of wheeling revenue, and the priority of Quebec deliveries against Maritime needs during coincident cold snaps are governed by bilateral arrangements whose performance no third party can verify (BGL Seam Dossier SD—06, internal; BGL Internal Memorandum M—2026—012, internal).

What proof requires here is dominated by one scenario: the deep-winter coincident peak. Quebec’s system peak, New Brunswick’s system peak, and New England’s winter gas-constraint events cluster in the same Arctic-outbreak weather systems — the interval record shows the region’s three demand maxima falling within the same seventy-two-hour window repeatedly across the corpus (Hydro-Québec; ISO-NE; BGL Run Ledger SL—1042, internal, sealed). In those hours, every commitment on this seam — Quebec’s exports to New Brunswick, New Brunswick’s deliveries onward to Nova Scotia and Prince Edward Island, and everyone’s sales into New England at the continent’s highest prices — draws on the same finite injection simultaneously. Which commitment is curtailed first is currently a matter of contract language, operator discretion, and after-the-fact recrimination. A provable seam would carry a published, machine-checkable priority stack: each firm commitment registered with its curtailment rank, each operating hour’s actual allocation logged against that rank with metering lineage, and any deviation — discretionary or forced — attributed to a named constraint at the moment it occurs, not reconstructed months later in a regulatory proceeding (BGL Derivation Record DR—SEAM—0178, internal). The Maritimes’ entire energy transition, examined next, hangs on flows that cross this seam. The parties downstream deserve to see the queue.

Seams Seven Through Nine: The Maritime Lattice

East of the Quebec–New Brunswick boundary, the seam problem changes character. The Maritime systems are small — Nova Scotia’s peak load is on the order of 2,200 MW, New Brunswick’s near 3,200 MW, Prince Edward Island’s under 300 MW — and at this scale a single intertie is not a trading opportunity but a life-support system, carrying reserve sharing, frequency response, and in PEI’s case most of the island’s supply (NB Power; Nova Scotia Power; PEI Energy Corporation). The New Brunswick–Nova Scotia seam runs through the Chignecto isthmus — a 345 kV corridor plus lower-voltage support, with usable capability conventionally stated around 300 MW and constrained further under stress — and it is the single point through which Nova Scotia’s coal retirement strategy, its federally co-regulated 2030 targets, and its access to Quebec energy must all physically pass (Nova Scotia Power; NS UARB filings; ECCC). The New Brunswick–PEI seam consists of submarine cables across the Northumberland Strait: two 100 MW cables laid in 1977 and two 180 MW cables energized in 2017 under a federally supported replacement program, giving the island roughly 560 MW of nameplate connection against a peak it cannot otherwise meet — PEI generates a large share of its energy from local wind but imports the balance and effectively all of its firmness (PEI Energy Corporation; NRCan). The third Maritime seam is the newest: the Maritime Link, a 500 MW HVDC connection between Newfoundland and Nova Scotia, in service since 2018, built by Emera to carry the Nova Scotia Block — roughly two terawatt-hours per year of Muskrat Falls energy contracted for thirty-five years — and reviewed for prudence by the Nova Scotia UARB in one of the more searching intertie proceedings on the Canadian record (Emera; NS UARB; Nalcor).

The Muskrat Falls complex upstream of that link is the cautionary tale the entire region now plans around. The 824 MW station and its 900 MW Labrador–Island Link came in at roughly double the sanctioned cost — on the order of $13 billion against an original estimate near $6.2 billion — prompting a public inquiry whose report is a catalogue of unprovable assumptions presented as analysis: demand forecasts insulated from challenge, alternatives screened out on contested inputs, risk assessments the sanctioning government could not independently reproduce (Commission of Inquiry Respecting the Muskrat Falls Project, 2020). Software and commissioning problems on the Labrador–Island Link delayed full contractual deliveries for years, and rate-mitigation ultimately required federal restructuring of the project’s financing (Government of Canada; Government of Newfoundland and Labrador). For the Seam Ledger, Muskrat Falls functions as the null hypothesis: it is what a megaproject’s evidence trail looks like when every load-bearing number is produced by the party that benefits from it and audited by no one with the means to recompute it. The inquiry’s findings are, almost clause by clause, an argument for derivation records (BGL Internal Memorandum M—2026—005, internal).

The operational stakes of the Maritime lattice are best seen from its failure modes. Nova Scotia’s decarbonization arithmetic requires replacing coal units that still supply a large fraction of provincial energy; the replacement plan leans simultaneously on the Maritime Link’s Newfoundland deliveries, on new wind whose firming depends on imports, and on the Chignecto corridor’s connection to everything west of it (Nova Scotia Power IRP; ECCC). Every one of those legs has a demonstrated failure record in the corpus window: low-water and commissioning shortfalls on the Newfoundland side, wind droughts spanning the entire region for days, and the Chignecto corridor’s exposure — noted with increasing urgency in federal-provincial infrastructure discussions — to storm surge and climate risk across an isthmus barely above sea level (CBC; Government of Nova Scotia). PEI’s position is starker still: the island’s cables are its grid, and the 2017 replacement program was triggered by the near-failure of cables that had been operating past design life with no redundancy (NRCan; PEI Energy Corporation). These systems do not have the luxury of treating intertie capability as a commercial abstraction. For them, the gap between declared and deliverable is measured in load shed.

Proof, in the Maritime lattice, therefore means something more intimate than in the west: continuous mutual visibility of deliverability. The quantities that matter are not annual energy volumes but instantaneous, correlated capabilities — what the Maritime Link can actually deliver this hour given Newfoundland hydrology and link status; what Chignecto can actually carry this hour given loading and contingency posture; what the cables can carry given thermal state; and whether the sum, net of coincident regional peak, covers the island and peninsular loads that have no alternative. Today those quantities live in separate control rooms connected by phone calls and seasonal studies. The lab’s derivation runs across the corpus window found the regional stack tight but coherent — with the striking property that the binding element rotates: no single asset is the limit more than a third of the time, which means no single-utility analysis can see the region’s true reliability position at all (BGL Run Ledger SL—1042, internal, sealed; BGL Seam Dossier SD—07/SD—08/SD—09, internal). A shared deterministic state — each utility’s declared capabilities fused, lineaged, and replayable, with every derating attributed at declaration time — is not an efficiency play here. It is the difference between a region that discovers its shortfall in a study and one that discovers it in an outage.

The Southern Interface: Where the Grid Actually Trades

Any honest audit of Canadian seams must concede that the important ones, measured by flow, all point south. More than thirty transmission crossings connect Canadian systems to the United States, and the aggregate trade across them — on the order of 51 TWh of exports in a typical year, worth several billion dollars — exceeds inter-provincial exchange by roughly an order of magnitude once the Churchill Falls legacy flow is set aside (CER; EIA). The inventory, west to east: British Columbia’s interface with the US Northwest — WECC’s Path 3 — carries several thousand megawatts of capability and is the physical substrate of both Powerex’s trading franchise and the Columbia River Treaty’s power provisions, themselves renegotiated to an agreement-in-principle in July 2024 after six decades (BC Hydro; Powerex; Global Affairs Canada, 2024). Alberta connects to Montana through the 300 MW Montana–Alberta Tie Line, a rare merchant intertie (AESO). Saskatchewan maintains modest ties into North Dakota. Manitoba’s southern interface, anchored since 2020 by the 500 kV Manitoba–Minnesota Transmission Project and the Great Northern Transmission Line, carries firm long-term hydro sales to Minnesota utilities and gives Manitoba Hydro US export capability in the low thousands of megawatts (Manitoba Hydro; Minnesota Power; EIA). Ontario’s ties to Michigan, New York, and Minnesota collectively dwarf its inter-provincial interfaces (IESO). And Quebec’s southern machine is the largest of all: the 2,000 MW Phase II HVDC corridor into New England, the Chateauguay converters and associated ties toward New York, the 225 MW Highgate converter into Vermont, the 1,200 MW New England Clean Energy Connect energized in January 2026, and the 1,250 MW Champlain Hudson Power Express into New York City energized in June 2026 (Hydro-Québec; EIA; National Observer).

The southern interface matters to this document for two reasons, and the first is institutional: it proves Canadian utilities can operate provable seams when the counterparty demands it. Every megawatt crossing into US markets moves under the full apparatus of open-access governance — OASIS postings of available capability under FERC Orders 888/889, NERC-tagged interchange transactions, reliability coordination under binding standards, market settlement at nodal prices computed by tariff-defined optimizations, and market monitors empowered to re-run the dispatch that produced any price (FERC; NERC; PJM Manual 11; ISO-NE tariff). Hydro-Québec, Manitoba Hydro, BC Hydro, and NB Power all maintain US trading subsidiaries holding FERC market-based rate authority, filing under oath, and living comfortably inside that evidentiary regime (FERC dockets; Powerex; HQ Energy Services). The asymmetry is stark enough to state as a finding: a megawatt-hour leaving Manitoba for Minnesota travels under stricter, more transparent, more auditable governance than a megawatt-hour leaving Manitoba for Saskatchewan (BGL Seam Dossier SD—10, internal). The machinery Canadian provinces will not build between themselves, they have already accepted — contentedly, profitably — at the American border.

The second reason is strategic exposure. The corpus window closes on a period in which the southern interface stopped being politically inert: tariff threats and retaliatory rhetoric in early 2025 extended explicitly to electricity, with Ontario briefly imposing and then suspending a surcharge on exports to Michigan, New York, and Minnesota — the first time in the modern era a Canadian government used an intertie as an instrument of trade retaliation (Government of Ontario, 2025; CBC, 2025). Simultaneously, drought years demonstrated the interface’s reversibility: British Columbia and Quebec, the two great structural exporters, both swung toward net imports in low-water periods, and Canada’s aggregate net exports fell sharply in 2023–2024 before recovering (CER Market Snapshots; EIA). The comfortable national narrative — Canada as permanent, passive electricity exporter — is false in both directions: the flows are hydrology-dependent, and the corridor itself is now understood in Ottawa as a dependency to be hedged, which is precisely the strategic argument advanced for the east–west program (PMO, 2026; CEAC, 2024). The five priority interties are, among other things, an insurance policy against the southern interface’s politics.

For the evidence architecture, the southern interface supplies the existence proofs this document leans on throughout. Constraint attribution at scale: every congestion dollar in PJM, MISO, or NYISO settles against a named binding constraint whose shadow price is the dual variable of a re-runnable optimization (PJM Manual 11; O’Neill et al., 2005). Interchange lineage at scale: every scheduled transaction between balancing authorities carries an electronic tag recording source, sink, path, and every intermediate approval — the e-Tag apparatus under NAESB standards — and inadvertent flows are accounted and paid back in kind under NERC BAL-006 (NAESB; NERC). Post-event forensics at scale: the 2003 blackout investigation reconstructed the failure minute by minute across two countries precisely because interchange, telemetry, and state-estimator records existed to reconstruct it from (U.S.–Canada Power System Outage Task Force, 2004). None of this is exotic. It is the standard equipment of seams that carry real money between parties that do not trust each other. The Canadian anomaly — the finding this section exists to sharpen — is that the country applies world-standard seam governance externally and phone-call governance internally, and the difference is not engineering capacity. It is that nobody has yet been forced to prove anything to a fellow province (BGL Internal Memorandum M—2026—012, internal).

A Controlled Experiment: The Atlantic Loop Against Wasoqonatl

Canadian transmission policy ran a controlled experiment between 2020 and 2026, though nobody designed it as one. Two projects proposed to strengthen the same corridor — the seams connecting Quebec’s hydro system through New Brunswick into Nova Scotia — under the same federal government, the same decarbonization mandate, and substantially the same engineering options. One failed and one proceeded. The variables that differed are the closest thing this field has to an identified causal mechanism, and they are worth extracting with some care because the five federal priority interties will be built — or not — under exactly the same forces (BGL Internal Memorandum M—2026—008, internal).

The Atlantic Loop, as conceived around 2020, was a multi-billion-dollar program of new transmission moving Quebec (and potentially Labrador) hydro through New Brunswick into Nova Scotia at gigawatt scale, positioned as the backbone of coal retirement in the region; federal participation was discussed at up to roughly $4.5 billion in loan support (CBC; Government of Canada). It died in October 2023, when Nova Scotia’s clean-power plan formally abandoned it in favour of domestic wind, batteries, and a smaller reliability tie (Government of Nova Scotia, 2023; CBC, 2023). The autopsy is uncontested in its parts: the parties could not settle who pays, because they could not agree on who benefits; they could not agree on who benefits, because each government’s analysis was produced by its own utility or consultants on private assumptions; the energy price Hydro-Québec would charge — the single number the entire benefit case turned on — remained undisclosed and, in the absence of disclosure, was assumed adversarially by everyone downstream; and ratepayer advocates in the receiving provinces, unable to verify any of the competing claims, treated the project as an open-ended exposure (CBC; NS UARB proceedings; CEAC, 2024). Note what is absent from the autopsy: engineering doubt. Nobody argued the wires would not work. The project failed for want of a shared, trusted computation — a way to make “who benefits, by how much, under whose assumptions” into a checkable object rather than a negotiating posture.

The Wasoqonatl intertie — the 345 kV New Brunswick–Nova Scotia reliability tie that survived the Loop’s collapse, financed in part through a $285 million Canada Infrastructure Bank commitment — is the experiment’s other arm (Canada Infrastructure Bank; NB Power; Nova Scotia Power). Its differences from the Loop were not primarily physical. The value case was narrow: a reliability and deliverability upgrade on a single seam between two utilities that already shared operating history, standards, and settlement machinery. The beneficiaries were enumerable: two utilities, two regulators, one corridor. The counterfactual was computable: with the tie, Nova Scotia’s import capability and reserve position improve by amounts a UARB panel can interrogate; without it, they do not. And the financing instrument — CIB debt against a defined asset — did not require the parties to first agree on a fifty-year regional benefit allocation. The project’s proponents, in effect, shrank the claim until it fit inside the region’s existing capacity to verify claims, and then it moved (BGL Seam Dossier SD—08, internal).

The Seam Ledger formalizes the lesson as a ratio: a transmission project proceeds when the provable benefit — the portion of the value case that every required party can independently verify with tools it already trusts — exceeds the political cost of commitment, and stalls when the value case, however large, is dominated by claims only its proponent can compute. Call the first quantity the project’s provable core. The Loop’s provable core was a small fraction of its claimed value; Wasoqonatl’s provable core was essentially all of it. On this account, the policy problem is not that Canada cannot finance big interties — the CIB, the federal balance sheet, and the utilities’ own rate bases are collectively adequate — but that the country’s verification capacity is so thin that only small claims can be proven, and therefore only small projects can proceed. The implication runs directly against the instinct to respond to the Loop’s failure by scaling ambitions down. The alternative response is to scale verification up: build the shared evidentiary machinery that enlarges the provable core of large projects — joint counterfactual dispatch with disclosed inputs, benefit attribution with lineage, delivery verification against sealed records — so that a gigawatt-scale, four-party corridor can clear the same evidentiary bar a 345 kV bilateral tie clears today (BGL Derivation Record DR—SEAM—0190, internal; CEAC Recommendation 19). The experiment already ran. The treatment effect is legible. The question is whether anyone builds the instrument the result calls for.

The Federal Five, Read as Evidence Problems

In June 2026 the federal government named five priority interties under the National Electricity Strategy, referring them to the Major Projects Office and attaching them to a Transmission InterConnect Investment Strategy; the only corridor specified publicly at announcement was a Regina–Winnipeg expansion of up to 2,000 MW, with the balance of the list described in briefings as a western intertie restoration, a northwestern Ontario connection, and reinforcements on the eastern seams (National Observer, 2026; PMO, 2026; NRCan). The announcement is the most serious federal entry into inter-provincial transmission since the corridor concept of the 1960s, and it arrives with the constitutional inheritance intact: electricity remains provincial under section 92A of the Constitution Act, 1867, the CER’s licensing authority reaches international and designated inter-provincial lines but gives the federal regulator no power to referee disputes between provincial systems, and every dollar of federal money will move through instruments — contribution agreements, CIB debt, tax credits — that require provincial counterparties to sign willingly (Constitution Act, 1867, s.92A; CER Act; TBS Directive on Transfer Payments). Ottawa can convene, fund, and de-risk. It cannot compel. Which means each of the five will live or die by the Atlantic Loop test of the previous section: is the provable core big enough?

Run the list through the seam dossiers and the evidentiary exposure of each corridor is distinct. The Regina–Winnipeg expansion inherits the Saskatchewan–Manitoba seam’s virtues — two Crown counterparties, existing contract history, unambiguous decarbonization arithmetic — and its unsolved scaling problem: at 2,000 MW, Manitoba’s hydrology risk and Saskatchewan’s reliability dependence become mutual, and the deal requires a deliverability-under-stress accounting neither utility currently produces in verifiable form (BGL Seam Dossier SD—03, internal; Manitoba PUB filings). The western restoration is an evidence problem almost purely: the BC–Alberta conductors largely exist, and the project consists of making deratings attributable and commitments firm across a market/Crown boundary — cheap in steel, expensive in institutional candour (BC Hydro, 2025; AESO). The northwestern Ontario connection inherits the 2014 NFAT failure: its benefit case spans two planning regimes with no shared model, and without a joint counterfactual apparatus it will reproduce the Conawapa record at higher cost (Manitoba PUB NFAT, 2014; IESO Annual Planning Outlook). The eastern reinforcements inherit the Loop autopsy wholesale, plus a new upstream variable: the Churchill Falls MOU re-pricing, which resets the cost of the energy any eastern corridor would carry and whose definitive agreements remain in negotiation as this document goes to press (Government of Newfoundland and Labrador; Hydro-Québec, 2024). Five corridors; five different dominant proof obligations; one common property — in every case the binding uncertainty is institutional, not electrical (BGL Internal Memorandum M—2026—015, internal).

The program’s announced machinery addresses financing and permitting — the two failure modes of the last decade’s project pipeline — but is, so far, silent on verification, the failure mode of the last decade’s inter-provincial pipeline specifically. The Major Projects Office can accelerate assessments; the CIB can carry construction risk; the investment tax credits can move the arithmetic (Government of Canada; CIB). None of these instruments manufactures the thing the Loop lacked: a computation of who benefits that all parties accept because all parties can run it. The Canada Electricity Advisory Council saw this in 2024, recommending an EU-style framework for identifying and supporting inter-regional transmission — and the EU precedent is instructive precisely because Europe’s machinery is evidentiary before it is financial: cross-border capacity calculation and allocation under the CACM regulation is a common, auditable computation executed on shared grid models, and disputes about it are disputes about inputs to a known algorithm, not duels between national consultants (CEAC Recommendation 19; European Commission Regulation 2015/1222; ENTSO-E). Europe built the referee before the money moved at scale. The federal five, as currently structured, move the money and hope the refereeing improvises itself (BGL Seam Dossier SD—11, internal).

This is the gap the lab’s research program addresses, and it is worth being precise about scope. Nothing in this document proposes that a federal body dispatch provincial systems, that provinces surrender planning sovereignty, or that any existing operator’s tools be replaced — the HG—01 position stands: provincial dispatch stays sovereign, and the missing layer sits at the seams, above the operators and below the political agreements (BGL/HG—01). What the federal five need from that layer is concrete and buildable on the program’s own timeline: for each corridor, a governed joint model with versioned, party-visible inputs; deterministic counterfactual runs that any signatory’s analysts can replay bit-identically; benefit and burden attributions carried as lineaged records suitable for regulatory filings in every affected jurisdiction; and, once energized, continuous delivery verification so the contracts signed against the studies are settled against measured, sealed evidence. The Seam Ledger’s corpus work — 41,000 intervals, nineteen sources, sealed replayable runs — was scoped deliberately as a one-lab demonstration that the first three items are feasible with public data alone; what they look like with operator-grade data is the subject of the pilot conversations reported in the final section (BGL Interval Corpus IC—41K, internal; BGL Run Ledger SL—1088, internal, sealed; BGL Pilot Minute PM—09, internal, redacted).

What Actually Moves Power: The Contract Structures

Wires are necessary; contracts are what dispatch. Before specifying what a seam must prove, it is worth cataloguing the instruments that currently move energy across Canadian boundaries, because each instrument implies its own evidentiary burden and most public discussion ignores all of them. The Seam Ledger’s contract taxonomy recognizes six families in active Canadian use (BGL Engineering Note EN—82, internal, 2026-02).

Legacy entitlement contracts sit at one extreme: multi-decade, take-or-pay arrangements priced at signing and largely indifferent to conditions thereafter. Churchill Falls 1969 is the archetype — sixty-five years of essentially fixed-price delivery, renewed automatically in 2016, adjudicated to the letter by the Supreme Court against every equitable challenge (2018 SCC 46) — and the Muskrat Falls Nova Scotia Block, thirty-five years of defined energy for a defined investment, is its modern descendant (CF(L)Co; Emera; NS UARB). Their evidentiary burden is low by design — deliver the block, meter the block — and their pathology is the corollary: they cannot adapt, so when the world moves, value transfers silently and enormously between parties until the resentment becomes constitutional. The December 2024 Churchill MOU is what the end-state of an unprovable long contract looks like: a political renegotiation conducted under threat of a 2041 cliff (Government of Newfoundland and Labrador; Hydro-Québec, 2024). Firm bilateral sales — the SaskPower–Manitoba Hydro agreements, Manitoba Hydro’s Minnesota sales, Hydro-Québec’s NECEC and CHPE supply contracts — are the workhorse family: defined quantity, term, and delivery point, transmission arranged as firm service, performance measurable by any competent meter reader (SaskPower; Manitoba Hydro; Hydro-Québec). Their burden is deliverability under stress: the contract is only as firm as the seller’s worst hydrological year and the path’s worst contingency hour, and nothing in the instrument itself makes those visible to the buyer.

Seasonal capacity exchanges — the Ontario–Quebec 500 MW swap — trade firmness itself, exploiting anti-correlated peaks so that both parties’ planning reserves improve with no net energy obligation (Ontario Ministry of Energy, 2015; IESO). They are the most capital-efficient instrument in the inventory and the most epistemically demanding: the entire value is a counterfactual (capacity not built), and the deliverable is a promise whose validity depends on the promisor’s coincident stress exposure — exactly the quantity that changed when Quebec’s surplus evaporated, and exactly the quantity no current mechanism verifies continuously (Hydro-Québec Action Plan 2035; BGL Derivation Record DR—SEAM—0171, internal). Storage-arbitrage relationships — Powerex’s diurnal and multi-day cycling against Alberta and the US Northwest, and the banking provisions in the Ontario–Quebec arrangements under which one party parks energy in the other’s reservoirs — monetize flexibility rather than energy or firmness (Powerex; AESO; IESO). Their evidentiary problem is attribution: when a reservoir system absorbs off-peak energy and returns it at peak, the round-trip’s value depends on system states that only the storage owner observes, and the counterparty takes the accounting on faith. Treaty instruments — the Columbia River Treaty’s Canadian Entitlement, delivering a computed share of downstream US power benefits to British Columbia since 1964, recalculated under the 2024 agreement-in-principle — are the oldest demonstration that a formula-defined, jointly computed benefit stream can hold across sixty years and a border, and the renegotiation is a live case study in what happens when one party concludes the formula’s inputs no longer reflect reality (Global Affairs Canada, 2024; BC Ministry of Energy). Finally, market participation — intertie bids and offers into the IESO and AESO markets, and Canadian traders’ activity in every US organized market — is the only family with hourly price discovery, and it exists on precisely the seams where at least one side runs a market (IESO; AESO; FERC dockets).

Arrange the six families against the seam inventory and the design problem of the next decade becomes visible. The instruments that adapt (market participation, storage arbitrage) live only where markets exist; the instruments that bind for decades (entitlements, firm sales) dominate the Crown-to-Crown seams where verification is thinnest; and the instrument the coming system needs most — dynamic, state-contingent firmness between winter-peaking, hydrology-exposed systems — barely exists anywhere. The corpus window’s droughts, cold snaps, and surplus inversions were each absorbed by whichever instrument happened to sit on the affected seam, with outcomes ranging from elegant (the ON–QC swap performing as designed) to expensive (BC’s $1.38 billion import year executed largely at market prices) to invisible (Crown-to-Crown flows settled on terms no outsider can evaluate) (Globe and Mail; CER; BGL Run Ledger SL—1042, internal, sealed). A future contract family — call it a verified state-contingent exchange: obligations indexed to jointly observed reservoir, outage, and demand states, settled against sealed operational records rather than declarations — is contractually trivial to draft and currently impossible to administer, because no shared, trusted observation layer exists to index against. That layer is not a market and not a treaty. It is an evidence architecture, and it is the subject of the final section (BGL Internal Memorandum M—2026—014, internal).

The Evidence Architecture: What Constitutes Proof at a Seam

Everything preceding reduces to a single design question: what would it take for a province to accept, without residual trust in the counterparty, that a cross-seam exchange was computed correctly, delivered as computed, and settled as delivered? The Seam Ledger’s answer — developed across the program’s first five months and stress-tested against every dossier in this document — is that proof at a seam decomposes into four obligations, each with existing precedent, none currently discharged on any Canadian inter-provincial boundary (BGL Seam Ledger Program Charter SL—00, internal; BGL Internal Memorandum M—2026—014, internal).

First obligation: metering lineage. Every quantity in a seam settlement must trace to physical measurement through an unbroken, tamper-evident chain. The components are individually mundane: revenue meters at interchange points built to recognized accuracy classes and subject to statutory inspection under the Electricity and Gas Inspection Act (Measurement Canada; ANSI C12.20); telemetry into each operator’s SCADA and state estimation; synchrophasor measurement where dynamics matter (IEEE C37.118); and a common semantic model so that both sides’ systems mean the same thing by the same tag (IEC 61970 CIM). What is missing is the chain itself: a jointly held, append-only record in which each interval’s measurements are sealed at capture — hash-linked in the manner of Haber and Stornetta’s timestamping construction and Merkle’s tree commitments, auditable in the manner Certificate Transparency made routine for an adversarial internet (Haber & Stornetta, 1991; Merkle, 1987; RFC 6962) — so that neither party can revise history and both can prove, years later, exactly what the meters said. The 2003 blackout investigation succeeded because the records happened to exist and could be painstakingly reassembled; a provable seam makes that reassembly a standing property rather than a heroic forensic effort (U.S.–Canada Power System Outage Task Force, 2004). In the lab’s stack this is the ingestion discipline HG—01 described: source records immutable, translations lineaged, the historical record that produced any conclusion frozen beneath it (BGL/HG—01; BGL Engineering Note EN—77, internal).

Second obligation: constraint attribution. Every limit imposed on a seam — a derated transfer capability, a curtailed schedule, a refused tag — must arrive with a named cause: the specific facility, contingency, or stability margin that bound, stated in a form the affected counterparty can recompute from shared model state. This is not a research aspiration; it is how organized markets already price congestion, where the shadow price on every binding constraint is the dual variable of a tariff-defined optimization and market monitors re-run the case when disputes arise (PJM Manual 11; CAISO Tariff §34; O’Neill et al., 2005). The Canadian gap is that inter-provincial limits are declarations without duals: the BC–Alberta derating, the Chignecto stress limits, the discretionary curtailment order on the Quebec–New Brunswick corridor in a cold snap — none carries a machine-checkable cause, and each is therefore politically corrosive in exact proportion to its economic significance (BC Hydro, 2025; BGL Seam Dossiers SD—01/SD—06/SD—07, internal). D.A.E.’s constraint-attribution layer exists precisely to manufacture this property: given the sealed system state and the declared limit, derive — deterministically, replayably — whether the limit follows from the state under the declared criteria, and name the element that binds (D.A.E. Constraint Attribution Specification v2, internal; BGL Derivation Record DR—SEAM—0201, internal).

Third obligation: counterfactual dispatch. The benefit case for any seam decision — a new intertie, a bigger swap, an operating-hour exchange — is always a comparison against a world that did not happen, and the comparison is only evidence if every party can run it. The obligation is bit-identical replay: frozen input corpus, versioned models, deterministic solvers, so that a Manitoba analyst, a Saskatchewan analyst, and a federal analyst executing the same case reach the same result to the last digit — the property HG—01 located at the core of market legitimacy (same input, same dispatch, same prices) extended to the planning and settlement domains (BGL/HG—01; Chen, Tanneau & Van Hentenryck, arXiv:2112.13469). The lab’s 2026 derivation campaign was the feasibility demonstration: across the 41,000-interval corpus, D.A.E. executed the full counterfactual suite for the seam inventory — the Alberta–Saskatchewan converter expansion, the Manitoba–Ontario 1,000 MW case, the ON–QC dynamic swap — with every run sealed, every attribution lineaged, and full-corpus replay verified bit-identical on independent hardware (BGL Interval Corpus IC—41K, internal, 2026-03; BGL Run Ledger SL—1042 and SL—1088, internal, sealed). Machine-learning components can participate in such an apparatus only behind verification layers that restore checkability — the settled conclusion of the power-systems ML literature itself (Venzke & Chatzivasileiadis, arXiv:1910.01624; arXiv:2304.11726; NERC, 2024).

Fourth obligation: settlement reconciliation. Finally, what was paid must reconcile — automatically, continuously, to the dollar — against what was measured and what was attributed. The interconnection already runs primitive versions of this machinery: e-Tags record every scheduled interchange transaction with source, sink, and path; inadvertent interchange is accounted and paid back in kind under NERC BAL-006; after-the-fact checkout reconciles schedules against actuals between balancing authorities (NAESB WEQ standards; NERC BAL-006). A provable seam closes the loop end to end: every settlement line item derives from sealed meter records (obligation one), every deviation from schedule carries its attributed cause (obligation two), every state-contingent term evaluates against jointly observed state (obligation three’s apparatus, run continuously), and the reconciliation itself is a deterministic computation any party’s auditor can replay. At that point the contract families of the previous section change character: the verified state-contingent exchange becomes administrable, the capacity swap becomes continuously demonstrated rather than annually asserted, and the Crown-to-Crown seam acquires, for the first time, the evidentiary standing its US-facing twin has had for twenty-five years (BGL Engineering Note EN—84, internal; FERC Orders 888/889).

Two design constraints govern the whole architecture. It must be confidentiality-preserving — utilities on a seam are counterparties, and no province will expose asset-level operating data to a neighbour’s traders; the lab’s Tri-Con identity layer exists for exactly this boundary, admitting external verification against encrypted composite references without disclosing raw identifiers (Tri-Con Identity Layer Specification, internal; BGL/HG—01). And it must be institutionally weightless: it creates no new regulator, transfers no authority, and binds only those who sign — a shared instrument, not a shared sovereign, which is the only configuration section 92A permits and the only one the provinces’ history suggests they will accept (Constitution Act, 1867, s.92A; CEAC, 2024).

Which brings this document to its only unpublishable evidence. Between January and March 2026 the lab held structured working sessions with two Crown utilities — unnamed here by agreement — walking their planning and operations staff through the four obligations against a live seam dossier (BGL Pilot Minute PM—09, internal, redacted; BGL Pilot Minute PM—11, internal, redacted). Three findings survive redaction. The participants disputed none of the four obligations in principle — each mapped cleanly onto pain their institutions already carry. The sharpest resistance attached not to disclosure but to asymmetry: each party’s first question was whether the counterparty would be held to the same sealing discipline, which is less an objection to the architecture than a specification of it. And the item that generated unprompted engagement was the smallest: continuous verification of an existing seasonal swap, a bounded, low-stakes instrument on which the machinery could be demonstrated without touching either party’s planning sovereignty. That is, on the evidence of the people who would have to live with it, where this program goes next: one seam, one instrument, four obligations discharged in production, every record replayable by both sides. The seams are about to receive a generational investment of public money. They can be built as wires, or as provable instruments. The difference is the evidence layer, and the conversation is open: research@blackgridlabs.com.