Use of machine learning for the detection and classification of observation anomalies
Utilisation de l'apprentissage automatique pour la détection et la classification des anomalies d'observation
Dahoui, Mohamed
For the last few years, an automatic data checking system has been used at ECMWF to monitor the quality and availability of observations processed by ECMWF's data assimilation system (Dahoui et al., 2020). The tool is playing an important role in flagging up observation issues and enabling the timely triggering of mitigating actions. The system is performing well and has a good detection efficiency. However, its behaviour is less optimal when assigning a severity level to detected events. The statistical procedure used to assign the severity requires tuning, and the behaviour is different from one kind of observation quantity to another. As a result, occasionally less significant events can be communicated as considerable or severe. When the day-to-day variability is small, moderate changes can be interpreted as severe from a statistical point of view. Given that not every threshold violation is a problem, there is a need for an improved way of inferring severity.</p>
Accès à la notice sur le site du portail documentaire de Météo-France