Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1)

Georgy Ayzel,Tanja Winterrath,Maik Heistermann

doi:10.5194/gmd-12-1387-2019

Abstract

Abstract. Quantitative precipitation nowcasting (QPN) has become an essential technique in various application contexts, such as early warning or urban sewage control. A common heuristic prediction approach is to track the motion of precipitation features from a sequence of weather radar images and then to displace the precipitation field to the imminent future (minutes to hours) based on that motion, assuming that the intensity of the features remains constant (“Lagrangian persistence”). In that context, “optical flow” has become one of the most popular tracking techniques. Yet the present landscape of computational QPN models still struggles with producing open software implementations. Focusing on this gap, we have developed and extensively benchmarked a stack of models based on different optical flow algorithms for the tracking step and a set of parsimonious extrapolation procedures based on image warping and advection. We demonstrate that these models provide skillful predictions comparable with or even superior to state-of-the-art operational software. Our software library (“rainymotion”) for precipitation nowcasting is written in the Python programming language and openly available at GitHub (https://github.com/hydrogo/rainymotion, Ayzel et al., 2019). That way, the library may serve as a tool for providing fast, free, and transparent solutions that could serve as a benchmark for further model development and hypothesis testing – a benchmark that is far more advanced than the conventional benchmark of Eulerian persistence commonly used in QPN verification experiments.

Highlights

How much will it rain within the hour? The term “quantitative precipitation nowcasting” refers to forecasts at high spatiotemporal resolution (60–600 s, 100–1000 m) and short lead times of only a few hours
The absolute difference in performance between the Dense group models and the RV product appears to be independent of rainfall intensity threshold and lead time (Table 3), which implies that the relative advance of the Dense group models over the RV product increases both with lead time and www.geosci-model-dev.net/12/1387/2019/
We examined the performance of optical-flow-based models for radar-based precipitation nowcasting, as implemented in the open-source rainymotion library, for a wide range of rainfall events using radar data provided by the DWD

Summary

Introduction

The term “quantitative precipitation nowcasting” refers to forecasts at high spatiotemporal resolution (60–600 s, 100–1000 m) and short lead times of only a few hours. Nowcasts have become important for broad levels of the population for planning various kinds of activities. They are relevant in the context of early warning of heavy convective rainfall events and their corresponding impacts such as flash floods, landslides, or sewage overflow in urban areas. The heuristic extrapolation of rain field motion and development, as observed by weather radar, still appears to outperform NWP forecasts at very short lead times (Berenguer et al, 2012; Jensen et al, 2015; Lin et al, 2005). For an extensive review of existing operational systems, please refer to Reyniers (2008)

Objectives

Results

Discussion

Conclusion