Excessive Convective Precipitation Formation Caused by Warm-Phase Bulk Microphysical Scheme Biases: A Multi-Platform Observational Diagnosis

Abstract

Recent work has demonstrated that convection-permitting models often produce a greater number of precipitating convective cells compared to radar observations. Congestus-depth cells contribute most to the bias, which are colloquially termed “popcorn convection.” This study explores warm-phase microphysical contributions to the bias using simulations spanning a field campaign hosted in the mountains of central Argentina. Observations from a C-band radar, satellite, in situ aircraft, and surface disdrometers are used to evaluate a 3-km grid spacing, 6.5-month simulation using bulk aerosol-aware microphysics and case studies with higher model resolutions. Compared to observations, clouds precipitate too frequently (by up to a factor of 2.5), particularly for congestus-depth clouds. These cells precipitate under high aerosol loadings (>4,000 cm⁻³) and most are purely liquid upon initiation of precipitation. Comparison with in situ aircraft observations shows that simulations shift too much bulk liquid mass to precipitation-sized drops, especially for low liquid water contents (LWCs), indicating potential issues with parameterized autoconversion. This leads to high-biased in-cloud radar reflectivity of 8–14 dBZ for LWCs < 10⁻¹ g m⁻³, producing radar reflectivity signals that contribute to convective cell identification and cell-number biases. Case studies using 500 and 100 m horizontal grid spacings do not ameliorate these biases, implying that under-resolved updrafts and entrainment are not first-order culprits. Attributing warm-phase microphysical processes to convective cell-number and precipitation frequency biases motivates a more detailed understanding of precipitation formation pathways in continental convective clouds, their sensitivity to aerosols, and novel methods capable of relieving modern structural limitations in bulk microphysics schemes.