Data Quality Control Based on Self-Consistency

Abstract

Abstract Conducting meteorological measurements, one is always confronted with a wide variety of different types of errors and with the decision of how to correct data for further use, if necessary. The selection of an adequate quality control (QC) procedure out of a wide range of methodologies depends on the properties of the observed parameter such as spatial or temporal consistency. But the intended data application (e.g., model-independent data analysis) or the availability of prior knowledge also has to be taken into account. The herein-presented self-consistent and model-independent QC process makes use of the spatial and temporal consistency of meteorological parameters. It is applicable to measurements featuring a high degree of autocorrelation with regard to the resolution of the observational network in space and time. The presented QC procedure can mathematically be expressed as an optimization problem minimizing the curvature of the analyzed field. This results in a matrix equation that can be solved without needing to converge iterations. Based on the resulting deviations and, if applied, on their impacts on the cost function, station values are accepted, corrected, or identified as outliers and hence dismissed. Furthermore, it is pointed out that this method is able to handle complicated station distributions, such as clustered stations or inhomogeneous station densities. This QC method is not only an appropriate tool for case studies but also for model validation and has been proving itself as a preprocessing tool for operational meso- and micrometeorological analyses.