Back to the page HelioClim databases icon Last updated - Dernière modification  21/11/08 Mines ParisTech - Armines

The HelioClim-3 Database - Versions and Roadmap

Versions

Version Number Date Comments
HC3. Beta Version February 2008 Known defects:
  • The irradiance is underestimated during the afternoon in sites located east of 30 deg in desertic areas with local relief. This is due to the combination of local shadowing and large albedo of the ground.
  • 2008-10-27: fixing a bug that occurred during October 2008. There was a confusion between irradiance and irradiation in the product HCHour for integration period shorter than 1 hour.
  • The irradiance may be at times slightly overestimated in Northern Europe for irradiance greater than 200 W/m2 and can be underestimated at times when the sun is low.

Roadmap

Work is underway to correct defects of the version beta. The methodology is to first correct for the major drawback, assess the performances of the new version and see whether the other drawbacks have been corrected or enhanced in this new version.

The cause of the major problem -underestimation in several desertic areas in Middle East- has been identified in summer 2008. The juxtaposition of very bright areas (illuminated sand) and very dark areas (shadowing due to local relief) within the same satellite pixel creates optical effects that are not well handled by the current version.

A correction algorithm has been devised in October 2008 and applied to data for Sede Boqer (Israel). Comparison with ground measurements shows large improvements and satisfactory results.

Next step is now to devise an alorithm for accurately identify affected pixels. Next step will be to apply the correction algorithm and further to check with ground measurements if this algorithm (i) is efficient and (ii) does not create drawbacks for other pixels.

Once this done, the new version will be implemented. It will be called version 2; there will be no version 1 for practical reasons.

Creation of consistent time-series of daily values from 1985 to today is also a concern to us. This may be done by adjusting HC-1 values on HC-3 values on the common period 2004-2005. Preliminary experiments show satisfactory results. More experiments are necessary and then, an operational method is to be devised. This may lead to re-design of databases.

Top of Page