Published: September 17, 2026 | Updated: September 17, 2026
POLICY & REGULATION
On 1 September 2026 Japan’s grid coordinator, OCCTO, published its outlook for grid congestion in FY2031, the state where more power wants to flow than a transmission line can carry. Across ten transmission areas, curtailment of wind caused by that congestion is counted in two: Hokkaido and Tohoku. Tohoku’s local grid alone carries between 552.3 GWh and 1,823.5 GWh of annual curtailed energy, depending on which scenario is used. The document names one of the drivers directly: offshore wind connected ahead of the reinforcement works, on the condition that output is cut back when the network is congested. For anyone financing these projects, that makes it a statement about the energy yield assumption as much as about where the grid is tight.
👉 Japan’s Offshore Wind Policy and Regulatory Framework Explained
→
Execution Reality
→
Bankability Test
Wind curtailment from congestion is counted in Hokkaido and Tohoku. Tokyo, Chubu, Kansai, Chugoku and Kyushu carry a dash on every wind row. Solar curtailment appears in the same tables, which makes the blank wind rows a finding in themselves. A dash still means only that nothing was counted on these assumptions and this equipment list.
Curtailed energy on Tohoku’s local grid rises by 367.9 GWh or 1,168.1 GWh against last year’s assumption. A footnote attributes part of the congestion to offshore wind connected before the bulk-system reinforcement is complete, under an arrangement that allows output to be cut when the network is congested. The connection choice shows up directly in the yield outlook.
The two scenarios differ only in the vintage of thermal fuel cost data. Even so, wind curtailment on Hokkaido’s Hakodate trunk line moves by roughly three times between them, while Tohoku’s Shimokita AB line barely moves, and the direction of deterioration is opposite in the two areas. This data supports a sensitivity range, and little more.
Wind Curtailment Is Counted in Two of Ten Transmission Areas
The outlook comes in two parts: paper 3, which gives area totals for FY2031, and annexes 1 to 8, which list every substation and transmission line expected to be congested, with curtailed energy broken out by generation category.
Reading the wind rows across all eight annexes, only Hokkaido and Tohoku carry numbers. Tokyo’s 56 blocks, Chubu’s 27, Kansai’s 6, Chugoku’s 7 and Kyushu’s 8 show a dash on every wind entry. Shikoku has three entries, the largest at 157 MWh, or 0.056 per cent, identical in both scenarios. Hokuriku and Okinawa report no congested facilities at all.
The blanks carry meaning because the same tables count solar curtailment: 26 blocks in Tokyo, 20 in Chubu, three in Chugoku and six in Kyushu. The wind rows are empty as an outcome of the modelling.
One distinction has to be kept. This outlook covers curtailment caused by network congestion. It is a separate mechanism from the supply-demand curtailment already familiar in Kyushu, where output is cut because generation exceeds demand. The absence of Kyushu wind entries here says nothing about that second mechanism, and reading the two together would distort any comparison between zones.
Tohoku’s Local Grid Reaches 1,823.5 GWh, With Early Connection Named as a Driver
Paper 3 presents FY2031 under two scenarios whose only difference is the vintage of the thermal fuel cost study used: the transition scenario draws on the February 2025 cost review, the current scenario on the May 2015 edition. The physical assumptions are the same.
On Tohoku’s local grid, the network that branches from the bulk system toward demand centres and generators, annual curtailed energy is 552.3 GWh in the transition scenario and 1,823.5 GWh in the current scenario, increases of 367.9 GWh and 1,168.1 GWh against last year’s assumption. National totals are 1,234.9 GWh and 2,255.5 GWh, equivalent to 0.14 and 0.26 per cent of national demand in FY2031. The document’s own summary line attributes the rise to expanding wind deployment in Tohoku.
A footnote then specifies what sits inside that figure: congestion associated with the increase in offshore wind and similar generation connected before the bulk-system reinforcement works are complete, under an early-connection arrangement conditioned on curtailment during congestion.
Peak figures follow the same pattern. Annual maximum curtailed power, defined as the average of the three days with the largest curtailment, reaches 1.316 GW and 2.039 GW in Tohoku across bulk and local systems combined, equal to 9.8 and 15.2 per cent of the area’s H3 demand, the average power across the three highest demand days of the year. Hokkaido records 0.948 GW and 0.859 GW, or 17.4 and 15.8 per cent, a higher share than Tohoku on both scenarios. The document states that in the eastern areas this ratio is expected to exceed 15 per cent under one of the scenarios.
Lenders size debt against downside yield, typically the P90 level that a project is expected to exceed in nine years out of ten, and then test the debt service coverage ratio (DSCR), which shows how many times over the cash available covers repayment in the tightest year. Where a single change of assumption moves curtailment by a factor of three, that spread lands on the P90 input. The published figures are area aggregates rather than project yields, so the usable question at project level is narrower: does the connection agreement rely on early connection, and when does the reinforcement schedule land relative to commercial operation?
The Same Change of Assumption Moves Areas in Opposite Directions
The equipment-level annexes show how unevenly the scenario difference lands.
In Tohoku, wind curtailment on the Shimokita AB line is 197,087 MWh (19.129 per cent, 2,244 hours) and 195,733 MWh (19.032 per cent, 2,242 hours). The two scenarios almost coincide. On the Mogami trunk line, total congested energy falls to roughly a third between scenarios, while wind curtailment stays at 89,008 MWh and 84,318 MWh. What moves there is the thermal and other categories, while wind holds still.
Hokkaido behaves differently. On the Hakodate trunk line between the Futaba and Kita-Oshamambe switching stations, a 187 kV circuit, wind curtailment is 109,608 MWh (6.464 per cent, 907 hours) in the transition scenario against 36,454 MWh (2.155 per cent, 288 hours) in the current one, close to a threefold difference. The direction is also reversed: Tohoku looks worse under the current scenario, Hokkaido under the transition scenario, consistent with the area ratios of 17.4 and 15.8 per cent.
The useful conclusion is about method. When one changed input moves two areas in opposite directions, neither figure can be carried straight into a yield assumption. The annexes also aggregate several generation categories into one column that cannot be decomposed, so the non-wind split stays unstated here, and equipment-level values are not summed because the same curtailed energy can appear against more than one facility.
The Highest Curtailment Rates Sit in Aomori, and They Do Not Move
Among Tohoku’s 66 kV circuits, the highest wind curtailment rates are on the Aomori side: 28.684 per cent over 2,868 hours on the Nakasaki line, 20.494 per cent over 2,245 hours on the Towada-nishi line, and 10.193 per cent on the Sai line. On the Akita side, wind entries are minimal, with the Noshiro-higashi line at 0.268 per cent. As above, an absent entry is a statement about this equipment list, not about project risk.
These 66 kV values are identical across both scenarios, and the same holds for Hokkaido at 110 kV and below. It appears that only the higher voltage classes were recalculated, though the document gives no reason, so the observation stays at that level.
The practical consequence differs by project size. For generation connecting into a local circuit at these voltages, choosing a different scenario changes the curtailment outlook by nothing at all. Large offshore projects connecting into the bulk system sit on the other side of that line, where the outlook swings with the assumptions. A wide swing is a contracting question before it is a modelling question: it has to be clear who carries the variance between the two cases.
Fifteen of Japan’s 21 Offshore Wind Zones Sit in the Two Counted Areas
Placing the designated promotion zones and promising zones over that map gives the following split. Area names refer to transmission territories rather than prefectures.
| Transmission area | Zones | Wind curtailment counted | Zones included |
|---|---|---|---|
| Hokkaido | 5 | Yes | Matsumae, Hiyama, Ishikari, Gan-u and Minami-Shiribeshi, Shimamaki |
| Tohoku | 10 | Yes | Four Akita zones plus Akita City, Aomori (Japan Sea north and south), Yuza, Sakata, Murakami and Tainai |
| Tokyo | 3 | None counted | Choshi, Isumi, Kujukuri (all in Chiba) |
| Kyushu | 3 | None counted | Goto, Saikai Enoshima, Hibikinada |
Fifteen of the 21 zones sit in the two areas where wind curtailment is counted. That overlap is what makes this a market-level question: the constraint tracks the same coastline as the auction pipeline. Which zones this favours is left open here, because area aggregates sit above project yields and connection terms remain unpublished zone by zone.
Grid congestion is not a national discount to be applied evenly. It is a regional condition attached to specific connection agreements.
Most Japanese offshore wind models still carry a single curtailment haircut applied across the country. The FY2031 outlook no longer supports that treatment. Congestion-driven wind curtailment is counted in two areas, and in one of them the document attributes part of the increase to projects that chose to connect before the reinforcement works were finished.
The reverse move is equally unsupported. Area figures cannot be subtracted from a project’s yield, because they aggregate every generator behind the same constraint. What the data does support is a range. Three times on one Hokkaido circuit, unchanged on a Tohoku one: that is the width a sensitivity case should carry, and it is wider than most models assume.
The question that follows is contractual. Early connection brings forward revenue and accepts curtailment in the interim, and the value of that trade depends on the reinforcement schedule for the specific circuit. Developers and lenders looking at Hokkaido and Tohoku zones should be asking for that schedule alongside the wind resource assessment, because the two now sit in the same calculation.
Related DeepWind Articles
- Japan’s 2050 Grid Master Plan Under Review: Offshore Wind Held at 45 GW as Data Center Siting Reshapes Transmission Scenarios
- Cost Realities of Japan’s Offshore Wind: LCOE and IRR Across 12 Promotion Zones
- Japan’s Offshore Wind Promising Zone Projects: Candidate Areas for Upcoming Auctions
- Japan’s Floating Wind Base Ports Cannot Work Alone: MLIT Sets a 2-Year Build Target and a 400 km Towing Radius
DeepWind Weekly tracks Japan’s offshore wind market beyond headline announcements, focusing on execution risk, cost structure, project viability, supply-chain constraints, and policy implications.
Subscribe to receive weekly intelligence on Japan’s offshore wind market.
Why Isn't Floating Wind Bankable in Japan Yet?
The synchronized supply-chain bottleneck, quantified node by node, with a full sensitivity ranking and a bankability ladder of scaling conditions. Includes Simulator Professional (β) access to reproduce every figure on your own assumptions.
See the full analysis →