Considerations for improving data quality of thermo-hygrometer sensors on board unmanned aerial systems for planetary boundary layer research

Abstract

Small unmanned aerial systems (UASs) are becoming a good candidate technology for solving the observational gap in the planetary boundary layer (PBL). Additionally, the rapid miniaturization of thermodynamic sensors over the past years has allowed for more seamless integration with small UASs and more simple system characterization procedures. However, given that the UAS alters its immediate surrounding air to stay aloft by nature, such integration can introduce several sources of bias and uncertainties to the measurements if not properly accounted for. If weather forecast models were to use UAS measurements, then these errors could significantly impact numerical predictions and hence influence the weather forecasters' situational awareness and their ability to issue warnings. Therefore, some considerations for sensor placement are presented in this study, as well as flight patterns and strategies to minimize the effects of UAS on the weather sensors. Moreover, advanced modeling techniques and signal processing algorithms are investigated to compensate for slow sensor dynamics. For this study, dynamic models were developed to characterize and assess the transient response of commonly used temperature and humidity sensors. Consequently, an inverse dynamic model processing (IDMP) algorithm that enhances signal restoration is presented and demonstrated on simulated data. This study also provides contributions on model stability analysis necessary for proper parameter tuning of the sensor measurement correction method. A few real case studies are discussed where the application and results of the IDMP through strong thermodynamic gradients of the PBL are shown. The conclusions of this study provide information regarding the effectiveness of the overall process of mitigating undesired distortions in the data sampled with a UAS to help increase the data quality and reliability.