## Solar Irradiance technology ### Introduction OpenWeather’s [Solar Irradiance](https://openweathermap.org/api/solar-irradiance) and [Solar Panel Energy Prediction](https://openweathermap.org/api/solar-panels-and-energy-prediction-2) services provide solar radiation data and estimates of photovoltaic energy generation for projects of any size. They combine established scientific methods with OpenWeather models and data through APIs designed for straightforward integration. Solar irradiance and solar irradiation are closely related, but describe different quantities: **Solar irradiance · W/m²** Solar power received per unit area. The API provides average and maximum irradiance for the requested interval. **Solar irradiation · Wh/m²** Total solar energy received per unit area over a specified period. ### Converting irradiance to irradiation Multiply the average irradiance over an interval by the interval’s duration in hours: ```text Irradiation (Wh/m²) = Average irradiance (W/m²) × Interval duration (h) ``` Example at 400 W/m² average irradiance: | **Interval** | **Duration in hours** | **Irradiation** | | --- | --- | --- | | 1 hour | 1 | 400 Wh/m² | | 15 minutes | 0.25 | 100 Wh/m² | ### Solar radiation components Our models provide three key components of solar radiation: - **Direct Normal Irradiance (DNI). ** Solar irradiance received directly from the sun by a surface perpendicular to the sun’s rays. DNI is particularly relevant to concentrating solar technologies and systems that track the sun. - **Diffuse Horizontal Irradiance (DHI). ** Solar irradiance received by a horizontal surface from sunlight scattered by the atmosphere and clouds, excluding direct sunlight. - **Global Horizontal Irradiance (GHI). ** Total solar irradiance received by a horizontal surface. GHI combines DHI with the horizontal component of direct solar irradiance and is widely used in photovoltaic performance modelling. These components can also be expressed as irradiation when accumulated over a specified period. We provide GHI, DNI and DHI for both **Clear Sky** and **Cloudy Sky** models. Our models build on established methods, including Simplified Solis and DIRINT, with adaptations developed for the OpenWeather service and evaluated against [reference datasets](file:///Users/stanislav/Downloads/solar-irradiance-technology.html#reference). ## OpenWeather Clear Sky model The OpenWeather Clear Sky model estimates the solar radiation reaching the Earth’s surface under cloud-free conditions. It accounts for the position of the sun, site elevation and relevant atmospheric conditions to calculate clear-sky GHI, DNI and DHI. The model builds on [Simplified Solis](https://pvlib-python.readthedocs.io/en/stable/reference/generated/pvlib.clearsky.simplified_solis.html), with adaptations developed for the OpenWeather service. These adaptations address the timing of calculations within each reporting interval and the treatment of solar angles around sunrise and sunset. Clear-sky estimates provide a baseline for understanding the solar resource at a location. They also serve as the starting point for the OpenWeather Cloudy Sky model, which estimates the effect of clouds on solar radiation. ### Model inputs The OpenWeather Clear Sky model uses the following data to calculate clear-sky **Global Horizontal Irradiance (GHI)**, **Direct Normal Irradiance (DNI)** and **Diffuse Horizontal Irradiance (DHI)**: - **Date and time** (required input) — determine the sun’s position for the requested period. - **Geographical coordinates** (required input) — identify the location and support solar geometry calculations. - **Sunrise and sunset times** (retrieved automatically from OpenWeather data) — define the daylight period and support adjustments near the horizon. - **Height above sea level** (retrieved automatically from OpenWeather data) — specifies the elevation of the requested location. - **Surface pressure** (retrieved automatically from OpenWeather data) — describes atmospheric pressure at the location. - **Near-surface air temperature** (retrieved automatically from OpenWeather data) — represents local near-surface atmospheric conditions. - **Solar zenith angle** (retrieved automatically from OpenWeather data) — describes the angle between the sun’s position and the local vertical, affecting how sunlight reaches a horizontal surface. **OpenWeather adaptations** **Interval timing.** Solar radiation varies throughout each reporting interval as the sun moves across the sky. A calculation made only at the beginning or end of an interval may therefore be less representative of conditions during that period. For hourly calculations, OpenWeather applies a 30-minute backward shift to use the solar position at the interval midpoint. This helps align the solar geometry used in the calculation with the period represented by the returned data. **Sunrise and sunset.** Irradiance calculations are particularly sensitive when the sun is close to the horizon. At these low solar elevations, small differences in solar geometry can produce unrealistic calculated values.OpenWeather uses sunrise and sunset times to adjust the treatment of solar angles during these periods. These corrections help limit anomalous values as the sun rises above or falls below the horizon. ### Model outputs The model returns **clear-sky GHI, DNI and DHI**, representing the global horizontal, direct normal and diffuse horizontal components of solar radiation under cloud-free conditions. ## OpenWeather Cloudy Sky model The OpenWeather Cloudy Sky model estimates the solar radiation reaching the Earth’s surface with the effect of clouds included. It combines clear-sky radiation estimates with cloud information and atmospheric conditions to calculate cloudy-sky GHI, DNI and DHI. The model uses a cloud-transmittance correction developed by OpenWeather, together with a modified [DIRINT](https://pvlib-python.readthedocs.io/en/stable/reference/generated/pvlib.irradiance.dirint.html) approach for estimating direct normal irradiance from global horizontal irradiance. Clear-sky estimates provide the baseline, while the cloud correction accounts for the reduction in solar radiation caused by clouds. ### Model inputs The OpenWeather Cloudy Sky model uses the following data to calculate cloudy-sky **Global Horizontal Irradiance (GHI)**, **Direct Normal Irradiance (DNI)** and **Diffuse Horizontal Irradiance (DHI)**: - **Clear-sky GHI** (calculated automatically by the Clear Sky model) — provides the baseline before cloud attenuation is applied. - **Dew-point temperature** (retrieved automatically from OpenWeather data) — provides information about atmospheric moisture. - **Cloud cover** (retrieved automatically from OpenWeather data) — supports the calculation of cloud transmittance. - **Date and time** (required input) — determine the sun’s position for the requested period. - **Geographical coordinates** (required input) — identify the location and support solar geometry calculations. - **Height above sea level** (retrieved automatically from OpenWeather data) — specifies the elevation of the requested location. - **Surface pressure** (retrieved automatically from OpenWeather data) — describes atmospheric pressure at the location. - **Near-surface air temperature** (retrieved automatically from OpenWeather data) — represents local near-surface atmospheric conditions. - **Solar zenith angle** (retrieved automatically from OpenWeather data) — describes the angle between the sun’s position and the local vertical, affecting how sunlight reaches a horizontal surface. As in clear sky model, most of the input data is taken from the [OpenWeather datasets and products](https://openweathermap.org/api). **OpenWeather adaptations** **Cloud-transmittance correction.** Cloud cover, expressed as a percentage, is converted into a coefficient representing the estimated transmission of solar radiation through clouds. This coefficient is applied to clear-sky GHI to calculate GHI with cloud attenuation included. Higher cloud cover generally produces greater attenuation. In cloud-free conditions, the cloud correction does not reduce clear-sky GHI. **Direct and diffuse components.** After cloud attenuation is applied, the model uses a modified DIRINT approach to estimate DNI from GHI. Solar geometry, surface pressure and dew-point information support this calculation. The direct and diffuse components describe how the radiation reaching the surface is distributed between direct sunlight and scattered light. ### Model outputs The model returns **cloudy-sky GHI, DNI and DHI**, representing the global horizontal, direct normal and diffuse horizontal components of solar radiation with the estimated effect of clouds included.  ## Solar irradiance forecasts OpenWeather provides solar irradiance forecasts **up to 15 days ahead**, combining numerical weather prediction data, local correction methods and satellite-based short-term forecasting. ### **General forecast** The general forecast uses numerical weather prediction data together with the OpenWeather solar irradiance models to estimate GHI, DNI and DHI. Improved correction methods, informed by **local site data**, help reduce systematic forecast errors and better represent local irradiance conditions. ### **Short-term satellite-based forecast** For supported locations across **Europe, the Middle East and Africa**, satellite observations further improve forecasts **up to three hours ahead**. Successive satellite images are used to estimate cloud movement and forecast its effect on solar irradiance. These short-term forecasts are **refreshed every 15 minutes** as new satellite observations are processed. Longer-range forecasts continue to use numerical weather prediction data with the corrections described above. | **Timing** | **Meaning** | | --- | --- | | **Overall forecast horizon** | Up to 15 days ahead. | | **Satellite-based forecast horizon** | Up to three hours ahead within supported regional coverage. | | **Short-term refresh frequency** | Every 15 minutes for the satellite-based short-term forecast. | | **Output interval** | The interval parameter controls aggregation: 15m, 1h or 1d. | The 15-minute refresh frequency applies to the satellite-based short-term forecast. Refresh frequency and satellite observation age are separate: imagery may have been acquired earlier because data delivery and processing take time. ## Release notes **10 November 2026 — Satellite-based short-term forecasts & history precision improvement** Satellite-based improvements to [Solar Irradiance](https://openweathermap.org/api/solar-irradiance) & [Solar Panel Energy Prediction](https://openweathermap.org/api/solar-panels-and-energy-prediction-2) for supported locations across Europe, the Middle East and Africa: - Improved historical accuracy: more accurate historical solar irradiance and energy generation estimates. - Improved real-time accuracy: better representation of current cloud conditions and their effect on solar irradiance and energy generation. - Improved short-term forecasts: more accurate forecasts for the next three hours, refreshed every 15 minutes using updated satellite observations. --- **28 July 2026 — Improved forecast accuracy** Improved [Solar Irradiance](https://openweathermap.org/api/solar-irradiance) & [Solar Panel Energy Prediction](https://openweathermap.org/api/solar-panels-and-energy-prediction-2) forecast accuracy across Europe through corrections based on observations from local ground-based measurement stations. --- **27 April 2026 — Satellite-enhanced solar irradiance algorithm** Released a new satellite-enhanced solar irradiance algorithm supporting [Solar Irradiance](https://openweathermap.org/api/solar-irradiance) & [Solar Panel Energy Prediction](https://openweathermap.org/api/solar-panels-and-energy-prediction-2) , improving the radiation estimates used by both services. - Improved accuracy: hourly GHI RMSD reduced by up to 50%, depending on location. - Lower data latency: real-time data delay reduced from up to six hours to a maximum of one hour. - Regional coverage: satellite enhancements introduced across Europe, Central Asia and Africa. - Faster responses: calls to the Solar Irradiance endpoint became twice as fast. - Full backward compatibility: no changes to API endpoints or request logic.