SCANDIA - The NOAA AVHRR cloud classification algorithm at SMHI

At SMHI automatic cloud classification using NOAA AVHRR (see the NOAA KLM Users Guide for a description of the latest instrument) data has been run in routine since 1988. Currently two versions run in operational mode. One algorithm produce classified results in deferred (4 km) resolution covering the entire Scandinavian region and surrounding waters (example given to the left above). The other algorithm runs for two smaller regions and in full (1 km) resolution, covering southern and northern Sweden respectively (central and right image above). Since the algorithms are conceptually equal and based on the same training data set we will hereafter speak about the algorithms as one "model".

The function of the cloud classification model is to identify/determine the type of cloud or surface (or mixture thereof, when applicable) within each pixel. The design of the model is directed towards Nowcasting purposes, so the main focus is on the assignment of

• clear versus cloudy conditions
• low, middle, or high level cloudiness
• convective or stratiform cloud type
• precipitation

The classification model is described in detail in various publications (listed at the end). A summary description of the model and its history, and a brief discussion of some features of the perforemance is given here.

Contents

• History and basic description
• Analysis of low-level clouds and precipitation
• Mesoscale analysis and cloud climatlogies with SCANDIA
• Current status