Albedo, fAPAR, Land Surface Temperature, Light Use Efficiency (LUE), NDVI , Precipitation, Solar radiation, Statics and Weather data

Surface Albedo

Albedo from the land surface is the ratio of the radiant flux over the shortwave spectrum (approximately 200-3000nm) reflected from the earth’s surface to the incident flux. Similar to the different definitions of the “spectral reflectance” (BRDF, R-factor, hemispherical reflectance), the integrated albedo also comes in different versions, but for this project it suffices to find the hemispherical albedo. Albedo together with fAPAR play an important role in the radiative energy balance of ecosystems and in the estimation of the carbon balance.

WaPOR application

Albedo is produced at both levels with a dekadal temporal resolution. They are not published through WaPOR, but are used as input for the calculation of E and T. Surface albedo varies in space and time as a result of processes such as changes in solar position, snowfall and changes in vegetation cover. A typical range for albedo of land areas is 0.1 to 0.4

Table 6: Overview of the intermediate data component albedo

Methodology

The conversion from narrowband to broadband albedo is done using predefined weights for each sensor, similar to the approach described by Liang et al (1999). For the calculation of the albedo a specific weight wi is assigned to each available spectral band i. These weights compensate for the uneven distribution of the incoming solar radiation over the spectrum. The final albedo is thus computed as ∑w_j•r_i (summation over the i bands), with ri and wi the spectral reflectance and weight of the i-th band. .

The weights used are:
• Sentinel-2: b2 * 0.171 + b3 * 0.060 + b4 * 0.334 + b8 * 0.371 + 0.018
• Landsat 5: b1 * 0.116 + b2 * 0.010 + b3 * 0.364 + b4 * 0.360 + 0.032
• Landsat 7: b1 * 0.085 + b2 * 0.057 + b3 * 0.349 + b4 * 0.359 + 0.033
• Landsat 8 and 9: b2 * 0.141 + b3 * 0.114 + b4 * 0.322 + b5 * 0.364 + 0.018

Processing approach

Level 1 processing

Surface Albedo at Level 1 is based on VIIRS reflectances. The albedo time series is smoothened and interpolated to daily values by applying a pixel based temporal fill algorithm: the Whittaker smoother (Eilers, 2003, Eilers et al., 2017).

Level 2 processing

Surface Albedo is derived from Sentinel-2 L2A data simultaneously with the computation of NDVI. Sentinel-2 L2a reflectances are resampled to 100m prior to calculating albedo. The albedo time series is smoothened and interpolated by applying the Whittaker smoother (Eilers, 2003, Eilers et al., 2017) for every 100x100m pixel.

Level 3 processing

At Level 3 surface albedo is based on Landsat and Sentinel-2 reflectance data. Surface albedo is derived using the weight function described above. The missing data in the daily albedo time series (resulting from cloud masking) is then modelled using the gap-filling approach developed by Weiss, Daniel J., et al. (2014). This approach uses both spatial and temporal information within the albedo time-series to fill the missing pixels. The noise in the filled albedo time series is later smoothed out by applying the Whittaker smoother (Eilers, 2003, Eilers et al., 2017), a fast, flexible smoother technique.

Challenges

Similar to NDVI, albedo dekadal series are still perturbed by noise due to missing values and data errors, therefore the data are smoothed to remove noise and missing values.