----------------------------------
Daily insolation product from RMIB
----------------------------------

This spatially explicit product is generated at RMIB, to complement the official CM SAF fluxes at the Top of Atmosphere (RSF and OLR).

Inputs:
 * Total Solar Irradiance (TSI), in W/m²         : taken from C3S RMIB daily TSI composite ICDR
 * Solar Zenith Angle (SZA), in degrees          : calculated for each 5 minutes, by combining the exact time and the three-dimensional local normal vector in the center of the grid box.
 * Earth-Sun distance (d), in astronomical units : calculated for each day, using the Bretagnon method and coefficient look-up tables (Bretagnon and Simon, 1986).

Method (in a nutshell):
 * First calculate the 3D local normal vector in the center of each pixel in the grid; this is static information.
 * Iterate over years, months, days, and for each day:
    * calculate Earth-sun distance correction factor (d²)
    * iterate over all 0.25°x0.25° grid boxes (1440x720), and for every grid box:
        * iterate over all 288 temporal bins of 5 minutes each, and for every temporal 5-min bin:
            * calculate the SZA, and from that:
            * calculate the insolation as follows: insolation = MAX( 0.0, (TSI * cos(SZA) / d²) )
            * update the daily insolation value

Accuracy depends on spatial and temporal resolution; we found that a sub-daily resolution of 5 minutes, and a grid box size of 0.25°x0.25°, were sufficient for our purposes. However, this may still be improved at the cost of longer computation time. Note that the spatial resolution of the processing grid (i.e. 0.25°x0.25°) is not necessarily the same as the grid on which the results are saved to disk. At this moment, we offer two output resolutions:
* normal resolution of 0.25°x0.25° (1440x720)
* low resolution on CERES nested grid of 1°x1° (360x180)
It makes no sense to store results in a resolution finer than the computation grid, so hypothetically, if a grid of 0.05°x0.05° is desired, then the computation grid should be improved accordingly.

The TSI input can be updated or changed:
* CERES EBAF Ed2.8 Daily TSI (2000-recent): CERES_EBAF_Ed2.8_DailyTSI.txt
* C3S RMIB daily TSI composite ICDR v2.x (1979-recent): as used in the production of CLARA-A3 products from CM SAF (RSF and OLR): C3S_RMIB_daily_TSI_composite_ICDR_v2_latest.txt
* C3S RMIB daily TSI composite ICDR v3.x (1979-recent): as planned for the production of the CLAAS-4 products from CM SAF (RSF and OLR): C3S_RMIB_daily_TSI_composite_ICDR_v3.6.txt

The difference between v2 and v3 is an improved algorithm resulting in an offset of approximately 1-2 W/m²; this is significant, but on the other hand it doesn't mean that the CLARA-A3 products should be combined with the (improved) v3 TSI. It is up to the user to make a decision which TSI and hence which insolation product is best suited for his/her application or objective.


More information can be obtained with Tom Akkermans via email: tom.akkermans@meteo.be