Reconstruct Across Datasets Window

This window lets you fill gaps in one or more datasets using the other datasets as reference, and even lengthen datasets.

With the controls in this window you can specify:

the data columns you wish to reconstruct
the data columns you wish to use for reference in the MCP-based reconstruction mechanism
the data columns from which you wish to calculate patterns in the pattern-based reconstruction mechanism
the MCP algorithms you wish to employ, and some settings such as number of direction sector divisions
a specific date range in which you wish to reconstruct the datasets
whether you wish to reconstruct data flagged to exclude, or only to reconstruct gaps
whether you wish to apply a flag to the synthesized data

Once the reconstruction process finishes, you can click Show Results to see detailed results in the Reconstruction Results window.

Click Test Uncertainty to see the Reconstruction Uncertainty Test window where you can investigate the accuracy with which the reconstruction process synthesizes data in your workbook.

Extending Datasets

If you choose to reconstruct a specific date range, you can even extend one or more datasets beyond their actual periods of record. The screenshot below shows an example where the user is about to extend a 3-year met tower dataset ten years back into the past using reanalysis datasets as reference:

The window will let you enter any date range within the aggregate period of record of the datasets you have chosen to provide MCP reference data. In the example above, the reanalysis datasets extend back to 2001 but no further, so the user could decide to reconstruct the 'Tower57' dataset back as far as 2001, but no further.

The article on the process of reconstruction across datasets explains the reconstruction process in greater detail.