Abstract:
Quality control is a particularly demanding problem for micrometeorological studies in complex environments. With the transition to electronic sensing and storage of climate data in high temporal resolution,
traditional approaches of homogenization are insufficient for addressing the small-scale variability and spatial
heterogeneity of the data. This problem can be successfully addressed by introducing a new class of control
procedures based on the physical and climatological relations between different climate variables. The new
approach utilizes knowledge about the interdependency of air temperature, precipitation, radiation, relative
air humidity, cloud cover, and visibility to develop empirical functions for determining the probability
margins for the co-occurrence of specific conditions in tropical mountains and deserts. It can also be applied to
other geographic settings by adjusting the parameters derived from the data itself. All procedures are integrated into a processing chain with feedback loops and combined with conventional logical and statistical
checks, which enables it to detect small errors that normally pass unnoticed. The algorithms are also adapted
to incorporate the short time steps of the original data to retain the potential for detailed process analyses.
