Sunlit clouds over a mountain horizon

Weather-market research

A clearer view of
the day ahead.

Heliomark Research develops temperature forecasts for weather-market decisions. We study how the day may unfold, where the daily high could land, and how those probabilities change as new weather data arrives.

Explore the research
Temperature paths & daily-high probabilitiesWeather data · Forecast models

The forecasting question

How hot could it get?
How likely is it?

For a weather-market decision, the likely daily high is only part of the picture. The range of possible highs, the chance of each outcome, and the evidence that could change that view all matter.

We develop forecasts that describe those possibilities. Our work brings together observations, forecast guidance and machine learning to estimate how temperatures may change through the day and what that means for the daily high.

The models are currently under evaluation.

Our research

From observations
to market questions.

Connect the course of a day, the probability of its high, and the market outcome that depends on it.

01 /

Temperature pathsHow the day unfolds

Estimate how temperatures may change, whether the daily high has already passed, and how new observations change the range of possible outcomes.

02 /

Daily-high probabilitiesA view of each possible outcome

Study the likelihood of different daily highs and map those forecasts to the temperature ranges used in weather markets.

03 /

Forecasts in contextThe market and its rules

Compare forecast probabilities with market prices and account for the location, measurement source and settlement rules that define each outcome.

Research approach

How we test
a forecast.

We evaluate forecasts on later days the model has not seen. The aim is to understand when probabilities are useful, where they fall short, and how new evidence should change them.

  • 01Match the forecast to the market’s temperature target.
  • 02Use weather data available at forecast time.
  • 03Compare with observed highs and established forecasts.
  • 04Study errors across different weather conditions.
How a forecast can evolveConceptual
Conceptual temperature pathsConceptual illustration of possible temperature paths after the latest observation.ObservedNowPossible outcomes
Conceptual illustration of possible temperature paths from a shared observation.
Weather data · Forecast models · EvaluationBuilt around the daily-high question.
Layers of sunlit clouds and atmospheric mist

As the day unfolds

New observations.
New expectations.

Cloud cover, a clearing sky or a stronger breeze can change the temperature outlook. We study when new observations should shift a forecast and what that could mean for a weather-market decision.

  • 01How much time and sunlight remain for further warming?
  • 02Has the likely high moved into a different market range?
  • 03Does the new evidence justify changing the forecast?

Research in practice

The work behind
the forecast.

We develop proprietary software for our own research and trading. The work includes weather-data pipelines, trained temperature models and retrospective evaluation.

Discuss the research
  • Weather data

    Bring together observations and forecasts while keeping track of where each measurement came from and when it became available.

  • Temperature models

    Develop models of the temperature path and daily high, then study their performance across new dates and weather conditions.

  • Market interpretation

    Connect temperature probabilities to specific market outcomes and examine how pricing and settlement rules affect a decision.

Research enquiries

Bring your
research question.

We welcome questions from people following weather markets and potential research partners.

Get in touch to discuss temperature forecasting, weather data, model evaluation or a potential research collaboration.

hello@heliomarkresearch.com