Kortex · portfolio weather concentration · live from the graph

BETAevidence: measured · ERA5 anomaly correlations, 2015–2025

Your wind farms are farther apart than their weather

A portfolio spread across countries can still be one meteorological bet. Kortex traces each asset to its operator and computes how many effective independent weather positions the portfolio holds; using measured multi-year weather correlations: eleven years of ERA5-derived daily wind and irradiance at each asset's cell, de-seasonalised so correlation captures shared weather, not shared summer. Assets 50 km apart share weather; assets 1,000 km apart usually don't — and that is now measured, not assumed.

Technology

Operator group

geographic spread vs measured weather correlation (ERA5 daily anomalies, 2015–2025); assets in the Kortex graph only, coverage disclosed below

The diversification frontier

every operator with ≥5 in-graph assets · x = assets on the map · y = effective independent weather positions · the dashed line is perfect independence · bubble = in-graph MW · amber = ≥40% of MW in one co-moving cluster · click a bubble to inspect

Feed this into a siting screen What timing does to price; the grid's day

Method & receipt. Effective independent positions = 1 / (wᵀRw) with w = capacity shares of the operator's in-graph assets and R built from direct pairwise measured correlations only; transitive chaining is deliberately rejected (chained closure merges continents, not weather systems). Correlations are measured, not assumed: each asset maps to a 1° cell carrying ERA5-derived daily series 2015–2025 (Open-Meteo archive); wind uses daily mean wind speed, solar daily shortwave radiation; series are standardised day-of-year anomalies against a ±7-day smoothed climatology, so correlation measures shared weather rather than shared seasons; Pearson r on ≥1,000 overlapping days; assets in the same cell share its series (r = 1.0, conservative); negative correlations are floored at zero — independence, never extra diversification credit. The largest common-mode cluster uses the portfolio's own measured r ≥ 0.5 pairs. Wind speed is used directly rather than a fabricated fleet power curve — disclosed. This measures historical meteorological co-movement, never firm or deliverable capacity. Coverage: figures describe assets in the Kortex graph; asset counts and MW are shown per operator and no claim of portfolio completeness is made. Source: GET /v2/spatial/weather-portfolio · correlations from measured ERA5 anomaly analysis (basis measured_corr_v1) · sources.