Towards a new fundamental climate data record for microwave humidity sounders based on metrological best practice

Abstract

The goal is to produce a fundamental climate data record for microwave humidity sounders (SSMT2, AMSU-B, MHS) based on metrological best practice. This new data record overcomes three main problems that are present in the available operational data. First, the new data record is based on consolidated data, second, detailed uncertainty information is provided on pixel level and third, the inter-satellite biases are investigated and physical reasons for the inconsistencies are found (harmonisation).

The thesis describes the instruments and the respective satellite missions, introducing the main instrumental characteristics necessary for the recalibration process. The most important inter-satellite biases are presented that prevent the construction of consistent long time series so far. In a dedicated study, the noise performance in analysed. This noise analysis is used in the later production of the fundamental climate data record: the analysis method uses the Allan deviation instead of the standard deviation for in flight-noise estimate. In detail, the instrumental calibration is presented, along with each physical effect that influences the measurement. Also, the strict metrological application of a measurement equation is followed. Further following metrological best practice, the uncertainties associated with each physical effect are estimated and propagated to the final measurand of brightness temperature.

Executing the calibration and uncertainty propagation with newly coded processing tools, a new fundamental climate data record is obtained. This data record is based on consolidated data and it provides detailed uncertainty information. Moreover, it provides concise quality information for the users.

The next step of understanding and reducing inter-satellite biases is carried out with a sophisticated optimisation approach devised in the Horizon 2020 project FIDUCEO (Fidelity and uncertainty in climate data records from Earth Observations). First test results were produced in the framework of this thesis. Further refinement of the method is required, though. Independently, an analysis of the biases in the light of the lifetime evolution of the instruments is carried out. Here, the origin of certain inter-satellite biases can be assigned with high probability to the effect of radio frequency interference (RFI) affecting not only the AMSU-B instrument on NOAA-15, for which the effect is known, but also the AMSU-B on NOAA-16 and even the MHS on NOAA-19. Moreover, shortcomings in the operational calibration are revealed and corrected in the new fundamental climate data record.