Solar irradiance site-adaptation

  • Updated on Feb 10, 2025
  • Published on Jan 17, 2025
  • 4 minute(s) read
  • PC

In this document

When ground-measured solar irradiance data are available at the project site, data correlation is used for reducing mismatch and mitigating systematic issues in the satellite-derived data, especially when the magnitude of the deviation is invariant over time or has a seasonal periodicity.

Overview

Solar irradiance retrieved from satellite-based models have lower spatial and temporal resolution compared to on-site measurements. Accuracy-enhancement methods are capable of adapting satellite-derived DNI and GHI datasets (and derived parameters) to the local climate conditions that cannot be recorded in the original satellite and atmospheric inputs.

For the adaptation of satellite data to the conditions represented by ground measurements at the project site, two main approaches are typically employed. The first approach involves the adaptation of satellite-based GHI and DNI values, correcting the bias (systematic deviation), and aligning the cumulative distribution functions. The second approach focuses on the adaptation of the input parameters and data used in the solar radiation model, where more complex parameters, such as Aerosol Optical Depth and/or Cloud Index, are adjusted.

This adaptation is possible when measured data meets certain requirements regarding length and quality. The outcome of this process is the construction of a multi-year solar dataset with improved accuracy.

Scatterplots representing satellite-based vs measured DNI sample values, before and after site-adaptation

Adaptation of satellite-based GHI and DNI values

This approach is effective when deviations between ground-measured and satellite-derived data are small and stable. However, it becomes unreliable when significant inconsistencies exist. Two primary methods are commonly used:

  • Ratio Method: This method adjusts long-term monthly and annual averages of DNI and GHI by correcting systematic bias. Ratios between ground-measured and satellite-derived solar radiation are calculated for the overlap period and applied to recalibrate the long-term satellite dataset. While simple and easy to implement, this method only addresses mean bias and does not fully utilize the detailed information provided by ground measurements. Moreover, it does not account for distribution differences, limiting its ability to correct more complex discrepancies.

  • Fitting Cumulative Distribution Function: This method aligns the cumulative distribution functions (CDF) of satellite-derived and ground-measured data. By matching the frequency distribution of satellite data to that of ground measurements, it improves the representation of typical and extreme values. Additionally, the mean value of the satellite data is adjusted to match the ground-measured mean. This approach typically reduces root mean square deviation (RMSD) and enhances the accuracy of satellite-based data, making it more reliable for local applications.

Adaptation of input parameters of solar radiation model

This approach focuses on adapting the input parameters of solar radiation models to improve the accuracy of GHI, DIF, and DNI outputs. Two primary methods are employed:

  • Adaptation of clearness index: This method involves the adaptation of the clearness index, which represents the ratio of surface radiation to top-of-atmosphere radiation, as the key model input. The process begins by deriving input parameters from ground measurements using a solar radiation model. The clearness index is then calibrated using least-square regression (or a similar method) to align hourly or daily differences between ground-measured and satellite-based parameters. Finally, the site-adapted clearness index is used to recalculate GHI, DIF, and DNI values within the solar model. This method effectively reduces mean bias, RMSD, and KSI (differences in frequency distributions), making it a straightforward approach to enhance model performance. However, since it uses a single parameter for adaptation, its capacity to handle complex atmospheric effects is limited.

  • Adaptation of the model input data: This method substitutes the clearness index with more detailed parameters, such as Aerosol Optical Depth (AOD) or Cloud Index. This advanced approach follows the same steps but recalculates the entire model, ensuring consistency among GHI, DNI, and DIF components. By addressing seasonal and regional inaccuracies in AOD and cloud descriptions, it resolves specific challenges that are common in certain regions. This method reduces mean bias, aligning with the expected uncertainty of measurement instruments, while also minimizing RMSD and KSI values. Although more complex, it is particularly valuable in areas where aerosols or cloud cover significantly affect solar radiation, such as semi-arid and desert regions. When executed with expertise, this method offers a robust and comprehensive solution for site adaptation.

Requirements for measured data

The capability to perform site-adaptation of satellite data depends on several factors or requirements, particularly the quality and duration of the ground measurements:

High-quality meteorological instruments are essential for accurate measurements. Secondary standard pyranometers and first-class pyrheliometers are recommended for measuring GHI and DNI, respectively. While an RSR (Rotating Shadowband Radiometer) can be used as a substitute for a pyrheliometer, it introduces higher uncertainty in measured GHI and DNI values. The use of redundant instruments—ideally one for each component (GHI, DIF, and DNI)—enhances accuracy and reliability throughout the site adaptation process.

Regular maintenance, cleaning, and calibration of instruments are necessary to ensure reliable data. A robust quality control framework, combining automated checks and operator oversight, is crucial for identifying and addressing potential measurement errors. Only pre-qualified data should be used for site adaptation to maintain accuracy.

The availability of high-quality ground measurements for at least 12 months is optimal to capture seasonal variability and ensure robust site adaptation. For projects with tight timelines, a shorter period of 9+ months can suffice, although it may not fully account for all seasonal deviations. Data from shorter durations (e.g., 3–6 months) is less reliable and may lead to an inaccurate relationship between long-term satellite data and local measurements, resulting in suboptimal adaptation.

Further reading