Quality control of oceanographic in situ data from Argo floats using climatological convex hulls

Abstract

A new method of identifying anomalous oceanic temperature and salinity (T/S) data from Argo profiling floats is proposed. The proposed method uses World Ocean Database 2013 climatology to classify good against anomalous data by using convex hulls. An n-sided polygon (convex hull) with least area encompassing all the climatological points is constructed using Jarvis March algorithm. Subsequently Points In Polygon (PIP) principle implemented using ray casting algorithm is used to classify the T/S data as within or without acceptable bounds. It is observed that various types of anomalies associated with the oceanographic data viz., spikes, bias, sensor drifts etc can be identified using this method. Though demonstrated for Argo data it can be applied to any oceanographic data. • The patterns of variation of the parameter (temperature or salinity) corresponding to a particular depth, along the longitude or latitude can be used to build convex hulls. • This method can be effectively used for quality control by building Convex hulls for various observed depths corresponding to biogeochemical data which are sparsely observed. • This method has the advantage of treating the bulk of oceanographic in situ data in a single iteration which filters out anomalous data.

Keywords

• Climatological convex hulls for outlier detection