The Anatomy and Physics of ZDR Columns

Kumjian, M., Pennsylvania State University

Cloud Processes

Convective Processes

Kumjian MR, AP Khain, N Benmoshe, E Ilotoviz, AV Ryzhkov, and VT Phillips. 2014. "The anatomy and physics of ZDR columns: Investigating a polarimetric radar signature with a spectral bin microphysical model." Journal of Applied Meteorology and Climatology, 53(7), 10.1175/jamc-d-13-0354.1.


Example of a ZDR column in a deep convective storm over Oklahoma. Panel (a) shows a vertical cross section of radar reflectivity factor taken by the S-band polarimetric WSR-88D radar KOUN. Panel (b) shows the differential reflectivity, ZDR. The ZDR column is located at a range of about 76 km.


The simulated field of ZDR is shown in color shading, during the mature stage of the deep convective storm. Updraft contours (in increments of 10 meters per second) are overlaid. Pink markers indicate locations from which particle mass distributions (outset panels) are taken. The mass distributions are shown as log-log plots, with the liquid drop distribution in black, hail in blue, and partially frozen drops in dashed cyan.


Example of a ZDR column in a deep convective storm over Oklahoma. Panel (a) shows a vertical cross section of radar reflectivity factor taken by the S-band polarimetric WSR-88D radar KOUN. Panel (b) shows the differential reflectivity, ZDR. The ZDR column is located at a range of about 76 km.

The simulated field of ZDR is shown in color shading, during the mature stage of the deep convective storm. Updraft contours (in increments of 10 meters per second) are overlaid. Pink markers indicate locations from which particle mass distributions (outset panels) are taken. The mass distributions are shown as log-log plots, with the liquid drop distribution in black, hail in blue, and partially frozen drops in dashed cyan.

Observations of deep convective storms using dual-polarization radars frequently reveal columnar regions of enhanced differential reflectivity (ZDR). These so-called "ZDR columns" often extend from low levels to several kilometers above the environmental 0 °C level and represent the lofting of supercooled liquid water by the storm's updraft. Despite several decades of remote sensing observations of ZDR columns, very little data exist on the internal microphysical structure of these features. Further, the link between ZDR columns and cloud properties and processes such as vertical velocity, phase partitioning, and precipitation growth need to be elucidated. To explore these questions, a sophisticated bin microphysical model that is coupled with a polarimetric radar forward operator to simulate the complete life cycle of ZDR columns in a deep convective cloud was used.

Used is this study was the two-dimensional version of the Hebrew University Cloud Model (HUCM), a sophisticated spectral microphysical model that accounts for various hydrometeor types and microphysical processes. Notably, recent improvements to the treatment of time-dependent freezing of raindrops and wet growth of ice particles has been implemented. The HUCM is coupled with a polarimetric forward operator that computes fields of the simulated radar variables based on the microphysical output of the HUCM. The combined microphysical and polarimetric models reproduce realistic-looking signatures and their evolution, including ZDR columns. This provides confidence that the key physics are adequately captured by the model.

The complete life cycle of ZDR columns are simulated. Owing to the high-CCN ("continental") environment, initial raindrop formation occurs aloft in the updraft. Recirculation of these initial raindrops into the base of the moisture-rich updraft allows for rapid growth of the drops to large sizes. Those drops large enough to fall through the updraft descend to the surface, continuing their growth by collection of cloud droplets and smaller raindrops. Other drops are lofted to subfreezing temperatures, where they subsequently freeze and grow into hailstones via accretion of supercooled liquid water. Freezing of the larger drops occurs at the top of the observable ZDR column. The ZDR columns collapse when sufficiently large hailstones descend through the updraft. The contribution to the radar reflectivity factor of these hailstones overwhelms the contributions from liquid and partially frozen drops, causing the observed ZDR to decrease towards zero. However, large quantities of supercooled liquid water still exists in the updraft during the ZDR column collapse. Thus, the absence of a ZDR column does not indicate the absence of a deep mixed-phase region of the cloud.

The anatomy (internal hydrometeor structure) of ZDR columns is quantified. The unique hydrometeor distributions reflect the complex processes ongoing in deep convective updrafts. At low levels, particle mass distributions are dominated by large raindrops that can fall through weaker portions of the updraft, indicative of ongoing size sorting. Further aloft, the liquid particle mass distribution becomes bimodal, reflecting contributions from cloud droplets and raindrops within the updraft. At sufficiently cold temperatures, a mixture of liquid, partially frozen, and ice particles exist, revealing the gradual freezing of larger raindrops. Further aloft still, the mass distributions are dominated by hail.

The relationships between ZDR columns and various storm attributes are elucidated. It has been demonstrated a strong correlation between the updraft speed and the height of the ZDR column: taller ZDR columns indicate stronger updrafts. Further, the evolution of the ZDR column mirrors the evolution of the individual updrafts within the storm. As such, ZDR column evolution provides information about the storm's behavior. For example, sudden increases in ZDR column height are followed by increases in hail production and precipitation fallout at the surface at lags of about 15 minutes. Such a lagged correlation for ZDR columns is stronger than more conventional radar-based measures such as reflectivity factor echo top.

ZDR columns are ubiquitous in deep continental convective clouds. They provide important information regarding the vertical velocity, precipitation growth processes, and cloud life cycle. Future in situ observations may help verify the correlations and relationships found in this modeling study. Future work will also explore how varying aerosol concentrations affect the appearance of ZDR columns, possibly leading to a radar-observable fingerprint of aerosol effects on deep convection.