# Import reliance, winter solar inputs and forced shortages

WattInfo, 25 September 2026.

## Import access

The 2016/17 model winter has 4,368 hourly intervals, October–March in Europe/Zurich. One GW delivered continuously supplies 4.368 TWh. The four matched steps remove one GW at a time (2→1 and 1→0), separately with planned maintenance and a 90-day Leibstadt outage. Subtract the higher-ceiling winter gap from the lower-ceiling gap. The extra requirement ranges from 3.847 to 4.037 TWh. The public chart shares one energy scale with the continuous-delivery illustration.

This large marginal effect shows that the proposed package relies substantially on import access. It is not an estimate of gross or net winter imports: domestic generation and water use can change between cases. The saved generation, heating and hydro assumptions remain uncalibrated. Alternative-schedule ranges for winter imports and net imports are the next checks before those volumes are promoted.

## Winter solar

Sum potential solar GW in winter hourly intervals and divide GWh by GWp to obtain kWh/kWp. The 2035 programme uses 26 GWp across thirteen weather years (2011/12–2023/24). The 2050 programme uses 53 GWp in three selected years (2011/12, 2016/17, 2021/22). Their ranges are about 266–343 and 307–343 kWh/kWp respectively. The three 2050 values equal the corresponding 2035 specific yields; capacities differ.

The separate seasonal illustration uses ZWB and AWS roof/facade portfolios from the published winter-PV study, Table 3, printed page 21: 257.8 and 296.9 kWh/kWp. These are modelled portfolios using 2004–2018 weather. The 2050 programme credits solar with higher winter output per kWp in its selected years. This is a difference between input assumptions, not proof of better sites or technology. The thirteen-year 2035 range overlaps the published portfolio averages. Neither comparison corrects the model's annual calibration or establishes an equally reliable supply plan.

## Older 8.8 GW illustration: a physical hourly lower bound

For each hour, minimum forced shortage = max(0, bus demand − fixed generation − natural-hydro turbine power − pumped-storage turbine power − import ceiling). This optimistic limit assumes all turbine power is available, regardless of stocks. The case has no imports, batteries, demand shifting or additional generator. All terms use the saved case inputs.

The builder checks the complete 8,784-hour ledger: generation sums, electrical balance, natural and pumped stock equations including the cyclic boundary, store and turbine limits, and equality of each saved shortage to the bound. It also checks the displayed 168-hour week against the same dated rows. Tolerance is 0.00002 GW for hourly comparisons and 0.00002 GWh for the annual sum. The equality error is zero in this calculation; the feasible ledger attains the lower bound at every hour. Therefore the archived optimum's shortages cannot be moved to other hours while retaining its minimum annual total. This proof applies to this case only, not to the 2035/2050 programme cases.

The result reflects the model's hypothetical capacities and demand. It does not establish real Swiss outage probabilities, national grid adequacy or a reserve purchase. Annual exports still depend on an assumed trade rule.

## Display definitions

The catalogue keeps prior sources and claims unchanged. A dated claimPresentation map supplies clearer public labels and specific limitations for the earlier context claims. It may change only label and limitation; evidence IDs, values, units, periods, methods and sources remain those of the archived record. The public source page and number popovers use the same labels. In older formula strings, “maximum one-hour shortfall” means the largest hourly power gap, not an assertion that it lasts only one hour.

## Reproduction

Run scripts/build_website_programme_followups.py --check. The evidence file records input hashes, per-case solar sums, all four import subtractions and the physical-bound check. The website validator independently recomputes these checks from the saved inputs. No new dispatch, legal verification, market-price forecast or technology-specific curtailment result is added.