METHODOLOGY

ilika Geospatial: Methodology for Delivering Actionable Probabilistic Weather Forecasts

This document provides an overview of how ilika Geospatial delivers high-impact probabilistic weather forecasts to its clients. Our methodology transforms authoritative global weather models into easy-to-use forecasts that quantify the likelihood of extreme conditions like heat, wind, rainfall, or compound risks such as wildfire potential. These forecasts help our clients in sectors including commodities, energy, agriculture, and insurance logistics make smarter, risk-aware decisions.

The AIFS model provides:

We use this ensemble data to calculate the likelihood that specific events will occur for example, the chance that temperature will exceed 35°C or that wind speeds will surpass operational thresholds.

ilika's forecasts are built on leading global weather models, including ensemble systems that simulate multiple possible weather outcomes. From these, we calculate the probability that a specific condition will occur for example, the chance that temperature will exceed 35°C or that winds will be strong enough to disrupt operations.

Custom Event Definitions

We work closely with clients to define the weather events that are most relevant to their operations, risks, or regulatory frameworks. These events are expressed as threshold-based conditions applied to one or more forecast variables derived from the ensemble model output.

Examples:

• Simple threshold event:

  A forecasted 2 m temperature greater than 35°C
  E = { T₂ > 35°C }

• Wind speed exceedance:

  Where wind speed is computed from the zonal (U) and meridional (V) 10 m wind components:
  E = { sqrt(U₁₀² + V₁₀²) > 10 m/s }

• Compound event:

  High temperature combined with strong winds, relevant for fire weather risk:
  E = { T₂ > 35°C AND sqrt(U₁₀² + V₁₀²) > 10 m/s }

Clients can build these custom events using a wide variety of weather variables available from the ECMWF AIFS forecast. In addition to commonly used variables like temperature and wind, the AIFS model also offers access to less conventional but highly valuable parameters such as cloud cover, snowfall, and total precipitation. Please contact your ilika representative for a complete list of weather variables.

These events are evaluated across all ensemble members to derive the probability of occurrence. Our platform supports customized definitions by:

• Adjusting thresholds for different climate zones

• Combining multiple variables (e.g., temperature + humidity)

• Applying filters by geography or time window (e.g., daytime only)

By allowing this degree of customization, ilika ensures that forecast products are tuned to the unique operational contexts of each client.

Forecast Delivery and Integration

ilika translates complex ensemble forecasts into clear, actionable probability estimates tailored to user-defined events. We quantify the likelihood of those events occurring based on the number of ensemble members that support them. This probabilistic view helps clients plan for a range of outcomes and understand the confidence behind each signal.

Forecasts are delivered as:

We work directly with clients to deliver the data in the format that best fits their internal systems, ensuring that outputs are ready for ingestion into bespoke workflows with minimal friction.

Why It Matters

ilika's probabilistic approach helps you:

• Quantify uncertainty in operational planning

• Focus attention where risk is rising

• Build trust in forecasts by understanding their confidence levels

Our goal is to make complex weather models accessible, relevant, and useful to decision-makers so you can act with confidence when it matters most.