Solcast vs Meteonorm
Meteonorm is a long-running commercial source of lower quality, lower cost irradiance data with a focus on the use of measurement data, supplemented with lower-spec satellite data. Compared with Meteonorm, Solcast is bankable, has lower uncertainty, has more data features, and is more open and easier to integrate with.
About Meteonorm
Meteonorm is a Windows software application and associated database made available by Swiss weather company Meteotest. Users install the application, purchase annual licenses, and download data from the application. Meteonorm has primarily focused on aggregation and extrapolation of surface measurements of varying quality, and more recently has also incorporated lower resolution satellite inputs.
Commercial and Technology
Meteonorm and Solcast both provide global irradiance and weather data, but are very different offerings.
Solcast, unlike Meteonorm, is bankable, and has an API for both Timeseries and TMY with open documentation. Meteonorm tends to be cheaper overall.
Data Features and Capabilities
Solcast |
Meteonorm |
|
|---|---|---|
| Bankable? | ||
| Free trial with instant access and data download? | ||
| Download wait time | 1-3 seconds | 1 minute |
| Comprehensive, global, independent validation | ||
| Finest time resolution of satellite-based irradiance | 5 minutes | 60 minutes |
| Real time data available? | ||
| Excludes older, less reliable satellites |
Source: Meteonorm Handbook part I: Software & Meteonorm Handbook part II: Theory
Inputs and Algorithms
Meteonorm uses a fundamentally different way to model the solar resource, based on interpolation of station data as a leading driver, and synthetic TMY generation. Constructing a global database using weather stations is challenging, due to data quality issues with the measurements (siting, calibration, cleaning, etc), and the uneven geographic distribution of stations. Clouds and irradiance have poor spatial autocorrelation, which is why things can change markedly in as little as 5 to 10km away from the closest surface measurement site, sometimes even less. Meteonorm has supplemented its database with satellite data, however this has been done using only hourly and 3-hourly inputs (compared to 5 to 15 minutes for Solcast and Solargis). The most granular solar resource data used by Meteonorm from ~1300 weather stations is monthly averages, using these to statistically generate TMY hourly values. This synthetic generation of a typical year dataset results in loss of the coherence between solar irradiance and air temperature. As the performance of solar power systems varies with solar irradiance and air temperature, use of synthetic hourly dataset increases the uncertainty of solar energy simulations.
Validation and Accuracy
Meteonorm and Solcast are both independently validated, globally, Meteonorm at 80 sites and Solcast at 207 sites. Solcast significantly outperforms Meteonorm on all measures, by a large margin, across GHI and DNI, and across bias spread and RMSE.
Meta analysis of large global validation studies: GHI results
| Solcast | Meteonorm | |
|---|---|---|
| Performed by | DNV | IEA PVPS |
| Year published | 2023 | 2023 |
| No. of sites | 207 | 85 |
| Mean Bias | +0.33% | -0.67% |
| Bias Std. Dev. | ±2.47% | ±6.99% |
| 80% CI Bias (10% to 90%) | -2.84% to 3.50% | -9.63% to +8.29% |
| 90% CI Bias (5% to 95%) | -3.74% to 4.40% | -12.17% to +10.83% |
| Mean nMAD (nMAE) | 10.33% | 12.55% |
| Std. Dev. nMAD (nMAE) | ±3.72% | ±3.88% |
| Mean nRMSD (nRMSE) | 15.99% | 17.80% |
| Std. Dev. nRMSD (nRMSE) | ±5.74% | ±5.56% |
Meta analysis of large global validation studies: DNI results
| Solcast | Meteonorm | |
|---|---|---|
| Performed by | DNV | IEA PVPS |
| Year published | 2023 | 2023 |
| No. of sites | 117 | 85 |
| Mean Bias | +1.50% | +0.27% |
| Bias Std. Dev. | ±5.75% | ±15.61% |
| 80% CI Bias (10% to 90%) | -5.87% to 8.86% | -19.74% to +20.28% |
| 90% CI Bias (5% to 95%) | -7.96% to 10.95% | -25.40% to +25.94% |
| Mean nMAD (nMAE) | 19.97% | 27.76% |
| Std. Dev. nMAD (nMAE) | ±5.94% | ±7.42% |
| Mean nRMSD (nRMSE) | 31.51% | 38.11% |
| Std. Dev. nRMSD (nRMSE) | ±9.99% | ±10.28% |
References
Forstinger, A., et al. (2023). Worldwide benchmark of modelled solar irradiance data (2023 PVPS Task 16): Solar resource for high penetration and large-scale applications. ResearchGate. Cuevas-Agulló, E., et al. (2023). A new global high-resolution solar resource dataset. Zenodo.