EXPERT FLAGGING OF COMMERCIAL MICROWAVE LINK SIGNAL ANOMALIES: EFFECT ON RAINFALL ESTIMATION AND AMBIGUITY OF FLAGGING

Accurate detection of signal anomalies in the attenuation time-series from commercial microwave links (CMLs) is crucial for high quality rainfall estimates. Example causes of such anomalies include dew or ice on the antenna and multipath propagation. In a first effort to catalog examples of CML signal anomalies, four experts flagged suspicious segments in the time-series of 20 CMLs in Germany. The results show that the agreement between experts depends on the definition of the anomaly class. Removing the flagged anomalies increased the Pearson correlation coefficient between CML and radar rainfall estimates from 0.61 to 0.70 and reduced the BIAS by 40%. An implication of our study is that expert uncertainty is an important factor for the quality control of environmental sensor data.

Automated summary

Accurate detection of signal anomalies in commercial microwave links (CMLs) attenuation time-series is crucial for high quality rainfall estimates. Examples of anomalies include dew/ice on the antenna and multipath propagation. A study analyzing 20 CMLs in Germany found that removing flagged anomalies improved the correlation between CML and radar rainfall estimates, emphasizing the importance of considering expert uncertainty in quality control of environmental sensor data.