Global patterns of solar resource short-term variability based on Solargis Time Series data.

Read full document

Overview

With advancements in geostationary meteorological satellites, analyzing short-term solar resource variability on a global scale has become feasible. This study leverages 10-minute and 15-minute Global Horizontal Irradiance (GHI) time series data from the Solargis satellite-based model, combining inputs from five major satellite missions covering all continents. The research delivers the first global assessment of the magnitude and frequency of short-term solar resource variability (“solar ramp rates”) in the recent five-year period (2019–2023), mapped at a nominal 4 km grid resolution.

Understanding sub-hourly variability is crucial for grid operators, regulators, and PV project developers. It informs strategies for grid integration, storage sizing, forecasting uncertainty, and mitigation of intermittency impacts, especially in regions where ground-based measurements are sparse.

Key methodology

Data source

Solargis satellite-based GHI time series at 10-minute and 15-minute intervals, merged from five satellite missions to ensure global land-area coverage (60°N to 60°S).

Period analyzed

• 2019–2023 with 4 km grid spacing.

Analysis steps

• Calculate solar noon for every day and grid cell.

• Extract GHI data within ±3 hours of solar noon each day to focus on periods of high ramp potential.

• Compute differences between consecutive GHI time steps (solar ramps) and generate time series of ramp magnitudes.

• Derive statistics—such as standard deviation, 90th/99th percentiles, and threshold exceedances—describing ramp rate distributions.

• Combine regional data into seamless global maps and analyze spatiotemporal ramp patterns.

Reference to ground measurement limitations and spatial aggregation

The study considers the “spatial averaging effect,” showing that while point pyranometer data (1-min) detect sharper ramps, the variability seen in 15-min satellite data is a practical proxy for the output variability of utility-scale PV plants due to plant-level averaging.

Main findings

Strong regional and seasonal patterns

• Areas with the highest short-term GHI variability include equatorial Africa, Indonesia, Papua New Guinea, the northern Amazon, Central America, and parts of Madagascar—corresponding to regions of frequent broken/cloudy skies.

• High variability is observed year-round in tropical belts, while regions like southern Africa, Australia, and Latin America show seasonal drops in variability during local winters.

• Arid regions (Sahara, parts of China, and Russia) present consistently low ramp rates across all seasons due to low cloud occurrence.

• Zonal, region-specific seasonal patterns are visible in Europe, North America, and India.

Frequency of large ramps ("ramp rates")

• The northeast coasts of Latin America and Central America see the highest frequency, with GHI ramps >500 W/m² occurring more than 360 times per year (~daily), and >300 W/m² more than 1000 times per year (3–4 times daily).

• Significant ramp activity also appears in the coasts of Australia, Southeast Asia, the highlands of Central Asia, and select African coastal zones.

Validation and context

• Comparison with ground 1-min data shows that although high-frequency ramps are somewhat smoothed in 10/15-min satellite series, spatial aggregation approximates the effective plant-scale variability, supporting use for utility-scale PV grid planning.

• The dataset and maps produced are valuable for PV site assessment, grid management, and energy storage planning, and can supplement regions lacking dense measurement networks.

Applications and implications

• The findings help grid and system planners identify where short-term solar variability is most likely to challenge integration—guiding decisions on geographical plant distribution, storage deployment, and smoothing strategies.

• Regions with high ramp rates require careful analysis of forecast uncertainty and storage/balancing infrastructure.

• Satellite-based ramp statistics enable more robust technical and financial modelling—especially where ground data are sparse.

Outlook

• Ongoing improvements in satellite imaging and data models (higher temporal and spatial resolution) will further improve global ramp rate analysis.

• High-frequency ramp data will soon support even more precise technical design and robust integration of solar PV in complex grids, especially in high-variability regions.

Further reading

• “Global patterns of solar resource short-term variability based on Solargis Time Series data” by Juraj Betak, Martin Opatovsky, Konstantin Rosina, Marcel Suri.