---
id: "VOLT-HOME-WP-080"
title: "Are clean charging windows stable across seasons?"
slug: "are-clean-charging-windows-stable-across-seasons"
description: "A seasonal comparison of mean operational low-carbon generation share across paired zone-day observations."
published: "2026-08-30"
cluster: "Carbon-aware household flexibility"
status: "measured"
evidence: "/research-data/home-papers/are-clean-charging-windows-stable-across-seasons.json"
figure: "/research-media/home-papers/are-clean-charging-windows-stable-across-seasons.webp"
figure_alt: "Chart for Are clean charging windows stable across seasons?: seasonal range of mean low-carbon generation share, shown as winter, spring, summer, autumn."
source_ids:
  - "iea-electricity-2026"
  - "acer-retail-2025"
  - "iea-demand-flexibility"
  - "ec-retail-flexibility-2026"
  - "volt-licensing"
peer_reviewed: false
---

## Abstract

Fixed charging clocks assume that environmentally preferable periods recur at similar times throughout the year. The frozen evidence for this paper does not directly estimate charging windows by clock hour. Instead, it groups paired zone-day low-carbon generation shares into winter, spring, summer, and autumn, computes each season’s mean, and reports the range between the highest and lowest seasonal means. Across 44,437 observations, that range is 0.08701431732866471 as a fraction.

The measured seasonal difference supports the limited conclusion that one undifferentiated clean-charging rule should not be assumed to generalize across the year. It does not identify which hours are preferable, how stable those hours are within a season, or whether a device can use them. Low-carbon generation share is an operational proxy, not lifecycle marginal emissions. It does not fully allocate imports or identify the marginal generator responding to charging. This observational result therefore measures neither caused nor avoided household emissions and should not be presented as a causal environmental benefit.

## Plain-language answer

The seasonal averages are not identical. The gap between the highest and lowest mean low-carbon generation share across the four seasons is 0.08701431732866471 in the frozen panel. That is enough to reject the convenience assumption that the same broad generation-mix conditions prevail all year.

But the study does not actually locate a charging window. It does not report an hour-of-day ranking, duration, EV availability constraint, or schedule. Seasonal average share can vary even if the best charging hour remains unchanged, and the best hour can move even if seasonal averages are similar.

The practical answer is therefore: do not assume stability from a fixed clock alone. Recompute schedules from current interval data. Also remember that the environmental measure is an operational generation proxy, not lifecycle marginal emissions or measured household emissions.

## Research question

The registered question asks whether clean charging windows are stable across seasons. The implemented estimand is the range of seasonal mean low-carbon generation share. Each paired zone-day is assigned to winter, spring, summer, or autumn by calendar month, and the largest seasonal mean is compared with the smallest.

This addresses seasonal level variation, not window stability in a strict sense. A “window” normally has a start, end, local-time basis, and feasibility condition. None is part of the primary metric. The statistic can motivate a season-aware scheduling study but cannot identify or compare actual windows.

The question is observational. It does not ask what would happen if charging moved, and it does not estimate a treatment effect. The result is a descriptive range across calendar groups.

## Data and provenance

The public evidence JSON in the frontmatter is the canonical paper artifact. The registered source contracts are `generation_mix`, `day_ahead_prices`, `capture_stats`, `res_accuracy`, and `zone_temp_weighted`. The primary calculation uses paired records with finite low-carbon generation share and the date component needed for seasonal assignment.

The evidence metadata records daily prices from 2021-01-01 through 2026-08-29, detailed intervals from 2025-10-01 through 2026-08-29, and a long-history range from 2015-01-01 through 2026-08-29. The frozen publication cutoff is 2026-08-30T00:00:00Z. Extraction used a read-only transaction and a statement timeout of 180 seconds.

The snapshot SHA-256 is `f77e3ae328f93916e53b1bab7516e1d0ac740a0dbf424cd2b73c81fee2559318`. Analysis code is bound to `57c57de79cdab2b5b6d6c54c485cb5162598c5ba0b0bfe995da40d75e6c52ba9`, protocol to `adb36bf6b447af9f96339249b8becaefc20422499cca1977242866347a97bd4b`, the paper registry to `7bcb91d7476d0a69fe9fa75a5c7782f8117e0153f82f9112b7e1d307d3943717`, and the source registry to `07949550ac443ff673fda5c0209b99f137544f3ffecf6775f109bb9d09663bd6`.

There are no household charging sessions, bills, device constraints, or measured household emissions in the source contract. The reference list contains the exact registered contextual links.

## Method

The analysis begins with paired zone-day rows containing finite low-carbon generation share. It obtains the month from each row’s date and maps the row into a calendar season: winter, spring, summer, or autumn. It calculates the arithmetic mean share within each group. The primary outcome is the maximum of those seasonal means minus the minimum.

The primary unit is a fraction. The reported 0.08701431732866471 is a range between group averages, equivalent to 8.701431732866471 percentage points, not a share for any individual season. The figure data do provide the four seasonal mean percentages, so they can be reported directly without inferring them from bar geometry.

Each available paired row contributes equally. There is no weighting by household population, load, generation, or number of flexible devices. The seasons are calendar categories and do not encode local meteorological onset. Pooling across zones can combine different generation systems and daylight or weather patterns.

The environmental field is low-carbon operational generation share. It is not lifecycle marginal emissions, does not fully allocate imports, and cannot identify the emissions caused by marginal charging. The method does not optimize a device or estimate a bill.

## Results

The seasonal range of mean low-carbon generation share is 0.08701431732866471 across 44,437 observations. The frozen interpretation states: “Seasonal variation means one fixed clean-charging clock cannot generalize across the year.”

The evidence supports the first half of that interpretation directly: seasonal means differ by the reported range. The statement about a fixed charging clock is a practical caution rather than a measured comparison of clock rules. No fixed schedule is run, and no season-specific window performance is reported.

The figure labels are “winter,” “spring,” “summer,” and “autumn.” Their mean low-carbon-share values are 54.6070074863634%, 63.308439219229875%, 61.13876251268832%, and 57.080234152399825%, rendered as 54.6, 63.3, 61.1, and 57.1. Spring is highest and winter lowest, yielding the primary 0.08701431732866471 fractional range.

The result does not reveal within-season variability and does not compare local hours. It also does not quantify emissions. Low-carbon share remains an operational proxy rather than lifecycle marginal emissions, and no measured household-emissions outcome exists. The evidence reports no bootstrap interval and names the interval method “not reported for this estimand.”

## Robustness and placebo checks

The evidence states that within-family Holm control applies to inferential claims and that this output is descriptive without an unadjusted significance claim. It reports no secondary results or paper-specific placebo. We therefore make no claim that the range is statistically significant or stable to another seasonal definition.

Robustness checks could estimate each zone separately, weight by load, use meteorological seasons, compare years, and quantify within-season dispersion. A direct window study would rank local-time intervals within each season and test out-of-sample stability under declared device availability. A placebo could randomize month labels while preserving zone composition.

Those analyses are not in the frozen evidence. The available integrity checks are the read-only extraction, cutoff, and cryptographic lineage. They make the calendar-season range reproducible but do not transform it into an interval-window analysis.

## Limitations

The outcome is a seasonal level range, not a charging-window stability score. It does not compare clock hours, window overlap, schedule regret, or feasible charging energy. The title is broader than the measured metric.

Pooling may confound season with zone coverage or changing data availability. Equal zone-day observations are not equivalent to population or electricity-consumption weights. The paper does not report annual trends or individual seasonal means in its JSON primary result.

Calendar seasons are coarse. Conditions can vary substantially within a season, and local climates or generation systems may not align with the same month grouping. Daily averages also hide intraday structure.

The low-carbon measure is an operational generation proxy, not lifecycle marginal emissions. Imports are not fully allocated and marginal response is not modeled. No customer tariff or load profile is included, so this is not a household bill study. It reports no measured household emissions, makes no causal claim, and is not trading advice.

Stability also has more than one meaning. A window may keep the same local clock time but change sharply in relative quality, or its clock time may move while its average quality remains similar. A range of seasonal means cannot distinguish these cases. It also cannot show whether a schedule selected from historical seasonal averages remains valid when current weather, demand, or generation availability differs.

## Practical implication

A controller should refresh environmental and price inputs rather than assume a year-round clock is valid. If it offers a recurring schedule, the schedule should be re-evaluated as the season and current forecast change. The result justifies that conservative design posture even though it does not specify the best interval.

User-facing language should distinguish “higher operational low-carbon share” from “lower emissions caused by charging.” The former is what the proxy can describe. The latter requires an appropriate marginal or lifecycle method. Low-carbon share is not lifecycle marginal emissions.

Actual household use also requires availability, energy, power, and tariff constraints. A season-aware generation signal alone cannot estimate bill outcomes or guarantee feasible charging. This study should inform data refresh and transparency, not a fixed environmental claim.

## Reproducibility

Open the evidence JSON for `VOLT-HOME-WP-080` and verify the slug, measured status, primary metric, source contracts, cutoff, and all five provenance hashes. Assign each paired finite-share zone-day to the registered calendar season, calculate the four group means, and subtract the smallest from the largest.

The frozen output is 0.08701431732866471 as a fraction across 44,437 observations. A local-hour window analysis, meteorological season definition, zone weighting, year-specific model, or marginal-emissions signal is a new analysis and should not overwrite this result.

The public evidence and figure are aggregate artifacts. Use and redistribution of underlying records remain subject to the exact registered licensing reference.

## Disclosure

Analysis and drafting were model-assisted. This working paper is not peer reviewed. It is not a charging-window optimization, not a household bill study, makes no causal claim, and is not trading advice. Volt has no live traders or live capital. It reports no measured household emissions. Low-carbon generation share is an operational proxy, not lifecycle marginal emissions.

## References

- [International Energy Agency — Electricity 2026](https://www.iea.org/reports/electricity-2026)
- [ACER and CEER — Rewarding flexibility: How retail contract choice can help unlock consumer flexibility](https://www.ceer.eu/wp-content/uploads/2025/11/ACER-CEER-2025-Retail-monitoring.pdf)
- [International Energy Agency — Scaling Up Demand Flexibility](https://www.iea.org/reports/scaling-up-demand-flexibility)
- [European Commission — Communication from the Commission on the Citizens Energy Package](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52026DC0115)
- [Voltcast — Voltcast data licensing and redistribution](https://github.com/ossedk/voltcast/blob/main/docs/voltcast/LICENSING.md)