Evaluation of micro rain radar-based precipitation classification algorithms to discriminate between stratiform and convective precipitation

Abstract:

Evaporation of precipitation below cloud base is a crucial process in the water- and energy cycle. It generates for example cold pools that lead to convective organization. Still, the parameterization of precipitation evaporation is highly empirical in current general circulation models. In order to improve the parameterization of evaporation from convective rain a big data set of convective rain cases is needed to generate robust statistics. Since it is a large effort to manually discriminate between stratiform and convective cases, automated algorithms have previously been developed.

We present two approaches to discriminate between stratiform and convective precipitation based only on micro rain radar (MRR) observations to enable a wide spread and straightforward usage for ground-based remote-sensing sites. The first one is based on probability density functions to compute a confidence function f . On the basis of the return value of f which is between -1 and 1, a classification (between stratiform and convective) and a measure of the confidence of this classification can be derived. The sign of f determines the class assignment and the absolute magnitude of f assigns the confidence to the classification. The other method is an artificial neural network (ANN) classification. It is a multilayer perceptron with two hidden layers to classify a rain event to be stratiform, inconclusive or convective. Both methods use the maximum radar reflectivity per profile, the maximum of the observed mean Doppler velocity per profile, and the maximum of the temporal standard deviation (15min) of the observed mean Doppler velocity per profile from a micro rain radar. Training and testing of the algorithms were performed using a two year data set from the Jülich Observatory for Cloud Evolution (JOYCE) in Western Germany.

The algorithms performances are assessed by means of two cases (a winter and a summer rain event) and evaluated based on all rain events from an entire year (independent test data). Both methods agree well giving similar results. However, the results of the ANN are more reasonable since it is also able to distinguish an inconclusive class, in turn making the stratiform and convective classes more reliable.

In the future, it is also planned to apply the algorithms to different ground-based remote-sensing sites that have long-term MRR observations to create stratiform-vs-convective rain event climatologies.