Using ship tracking methods to assist in quality controlling and bias adjusting meteorological observations in a marine environment

How global temperature has changed in the past is essential to both our understanding of the current climate and our ability to predict future change. The Marine Air Temperature (MAT) record that is derived from the International Comprehensive Ocean Atmosphere DataSet (ICOADS) provides a measure of global temperature which extends back to the mid 19^th Century. Although the Sea Surface Temperature record has a greater spatial coverage in the latter parts of the record the MAT record potentially has a lower bias and is an independent record of changes in global temperature. In records prior to the early and mid 19^th Century (data yet to be included in ICOADS) MAT becomes the dominant record type with measurements stretching back to the late 18^th Century. However, the MAT record faces two substantial issues; firstly the daytime measurements (which become dominant in the early part of the record) are contaminated by solar heating, secondly a large number of the observations in the ICOADS (90% in 1970) have no ship identity. In order to devise corrections for the daytime MAT, it is necessary to be able to differentiate between ship voyages (if not between the ships themselves). Since this has not been considered possible, current MAT datasets (i.e. MOHMAT43N) do not use daytime data except from observations with complete metadata. To create the bias adjustments, we must first identify the individual voyages using their kinematic properties. To do this we have employed a state-of-the-art tracking methodology to break up the mass of unidentified marine observations into ship tracks. This method also has the benefit of being able to differentiate between observing platforms based on the pattern of movement. Here we will present a test case of June 1998 demonstrating some of the potentially available improvements to the MAT record and hence to our understanding of changes in global temperature.