Characterizing Mesoscale Pressure Features with Bias Corrected Smartphone Pressures

Abstract

With over a billion smartphones capable of measuring atmospheric pressure, a global mesoscale surface pressure network based on smartphone pressure sensors may be possible if key technical issues are solved, including privacy and bias correction. To overcome these challenges, a novel framework was developed for the anonymization and bias correction of smartphone pressure observations (SPOs) and was applied to billions of SPOs from The Weather Company (IBM). Bias correction using machine learning reduced the errors of anonymous (ANON) SPOs and uniquely identifiable (UID) SPOs by 43% and 57%, respectively. Applying multi-resolution kriging, gridded analyses of bias-corrected smartphone pressure observations were produced for an entire year (2018), using both ANON and UID observations. Relative to MADIS analyses, ANON and UID smartphone analyses reduced domain-average pressure errors by 21% and 31%. To demonstrate the utility of SPOs, a comparison between MADIS and smartphone pressure analysis was performed by tracking and characterizing pressure features from 2018. Smartphone pressure features lasted, on average, 25-min longer, traveled 25-km further, and exhibited larger amplitudes than features observed by MADIS. With over 87% of observed pressure features associated with convection, the climatology of surface pressure features largely reflected the geographic, seasonal, and diurnal variation of organized mesoscale convection. Phase relationships between pressure features and other surface variables were consistent with those expected for meso-highs and wake-lows. This result suggests that SPOs could enhance convective analyses and forecasts by better resolving mesoscale structures and features, such as wake-lows and meso-highs, under-observed by existing surface networks like MADIS.