Validation of environmental factors

Overview

The accuracy of Solargis satellite-derived albedo is evaluated by comparing it against ground albedo measurements at reference meteorological stations. Because publicly available albedo measurement networks are limited, regional validation is not feasible; instead, a set of high-quality stations representing variable geographical and land cover conditions is used to provide a general assessment of satellite-based albedo accuracy.

A key requirement for meaningful validation is that the landscape surrounding the albedometer is homogeneous at the scale of the satellite pixel (approximately 0.5 km for MODIS). This condition ensures that the spatial representativeness of both data sources, ground measurements and satellite estimates, is comparable. Where this condition is not met, additional sources of mismatch, such as intra-pixel land cover variability, may inflate the apparent error. This requirement significantly limits the number of stations suitable for validation.

The evaluation uses daily albedo values and reports the following statistical metrics: Bias, Mean Absolute Deviation (MAD), Root Mean Square Deviation (RMSD), Standard Deviation (STD), and Correlation Coefficient (CC).

Note: A negative bias indicates that the satellite-based estimate systematically underestimates albedo relative to the ground measurement.

Validation stations

Eight high-quality research meteorological stations in the United States are used for the evaluation. Stations belong to two publicly available networks: SURFRAD (Surface Radiation Budget Network) and AmeriFlux. The geographical distribution of the stations is shown in Figure 1, and their characteristics are listed in Table 1.

Figure 1: Geographical location of the ground albedo validation stations. Green markers indicate SURFRAD stations; orange markers indicate AmeriFlux stations.

Validation results

Table 2 shows the statistical metrics of the comparison between Solargis satellite-derived albedo and ground observations at each station.

* Sites not significantly affected by snow.

Interpretation of results

The validation results show that the performance of the satellite-based albedo estimate varies according to the characteristics of each station, particularly with respect to snow occurrence and land cover homogeneity.

• Snow-free sites (Desert Rock, Walnut Gulch Kendall Grasslands, Morgan Monroe State Forest, and Metolius Mature Ponderosa Pine) show the lowest bias values, ranging from -0.014 to approximately 0.001, indicating close agreement between satellite estimates and ground observations.

• Sites affected by ephemeral snow show higher bias. Fort Peck, where snow events are common, has a bias of -0.031. The higher deviation at these sites is driven by the difficulty of capturing short-duration snow events with satellite observations: snow is often accompanied by cloud cover that prevents the satellite from observing the surface, and rapid snow degradation further complicates the comparison.

• Bondville is a specific case where two factors combine to produce the highest bias (-0.058): the occurrence of snow events and significant intra-pixel variability of albedo, which means the satellite pixel does not adequately represent the local surface conditions observed by the albedometer.

In general, the bias across all stations is negative, indicating a tendency for the satellite-based model to slightly underestimate albedo relative to ground observations.

Note: For sites with homogeneous land cover and no significant snow influence, Solargis satellite-derived albedo shows very low bias, confirming its suitability for solar energy applications at such locations.

Further reading

• "Surface albedo and reflectance: Review of definitions, angular and spectral effects, and intercomparison of major data sources in support of advanced solar irradiance modeling over the Americas": C. Gueymard, V. Lara-Fanego, M. Sengupta, Y. Xie

• "How to use albedo for more accurate bifacial PV estimates"

• "Satellite-based albedo data"

• "Site adaptation of albedo data"

• "Solar irradiance ground measurements"