Dual-polarisation weather radars have become more common through the last 15 years, as operational networks in the USA, UK and across Europe have been upgraded with this capability. The distribution of dual-polarisation radars, and their wide spatial coverage, provide a wealth of information for understanding and nowcasting high impact weather, including quantitative precipitation estimation (QPE), hail detection and the
potential for use in data assimilation.
One variable from these radars is Specific Differential Phase (KDP). There have been many studies into the uses of KDP, including QPE, hydrometeor classification, and liquid and ice water content estimations, with KDP often preferred over reflectivity based measurements due to being less dominated by larger, but fewer, targets. However, KDP is not directly measured by the radar, rather it is derived from the total phase shift measured, which has a number of other contributors including backscatter differential phase and noise, and therefore estimating KDP is difficult.
To explore the uncertainty associated with this estimation, a number of different estimation methods are studied, and it is shown they all struggle to accurately estimate KDP, especially when there is noise present. Being a difference between two orthogonal observations, KDP is affected by the viewing geometry of non-spherical targets. This is confirmed through observations of ice hydrometeors through a range of elevation angles, and the effect of elevation angle on KDP is shown through a hydrometeor classification algorithm, where adjusting KDP at high elevation angles changes the output from the algorithm. Finally, output from a radar forward operator using high resolution model simulations is compared with radar observations to study how well simulated fields can recreate estimated KDP . The generalisation of ice in microphysical schemes and the forward operator mean that radar signatures present in the observations are not replicated in the output of the forward operator.
