A method to correct sampling ghosts in historic near-infrared Fourier Transform Spectrometer (FTS) measurements

The Total Carbon Column Observing Network (TCCON) has been established to provide ground-based remote sensing measurements of the column-average dry air mole fractions of key greenhouse gases. To ensure the network wide consistency, biases between Fourier Transform spectrometers at different sites have to be well controlled. In this study we investigate a fundamental correction scheme for errors in the sampling of the interferogram. This is a two-step procedure in which the laser sampling error (LSE) is quantified using a subset of suitable interferograms and then used to resample all the interferograms in the timeseries. Timeseries of measurements acquired at the TCCON sites Iza a and Lauder are used to demonstrate the method. At both sites the sampling error histories show changes in LSE due to instrument interventions. Estimated LSE are in good agreement with sampling errors inferred from lamp measurements of the ghost to parent ratio (Lauder). The LSE introduce retrieval biases which are minimised when the interferograms are resampled. The original timeseries of Xair and XCO2 at both sites show discrepancies of 0.2–0.5% due to changes in the LSE associated with instrument interventions or changes in the measurement sample rate. After resampling discrepancies are reduced to 0.1% at Lauder and 0.2% at Iza a. In the latter case, coincident changes in interferometer alignment may also contribute to the residual difference.