The Voluntary Observing Ships (VOS) scheme is an international programme comprising member countries of the WMO/IOC that recruit ships to take, record and transmit marine meteorological observations whilst at sea. The VOS Scheme is a core observing programme of the Ship Observations Team (SOT) in the Observations Programme Area of JCOMM. There are three types of ships in the VOS Scheme such as selected ships, supplementary ships and auxiliary ships. A selected ship is equipped with sufficient certified meteorological instruments for making observations, transmits regular weather reports and enters the observations in meteorological logbooks. Most of the VOS are selected ships. A supplementary ship is equipped with a limited number of certified meteorological instruments for making observations, transmits regular weather reports and enters the observations in meteorological logbooks. An auxiliary ship is without certified meteorological instruments and transmits reports in a reduced code or in plain language, either as a routine or on request, in certain areas or under certain conditions. Auxiliary ships usually report from data-sparse areas outside the regular shipping lanes.
Currently, VOS typically report every six or three hours interval, and make observations of surface wind speed and direction, air temperature, humidity, sea surface temperature (SST), atmospheric sea level pressure (SLP), cloud (including type, amount and height), wave and swell parameters and weather (including visibility) information. The data are sent to a meteorological service as soon as they are obtained, either by radio telephony to a coastal radio station, by telex over radio, or by INMARSAT-C. Around 5,000 ships are presently reporting marine meteorological parameters. Observations, such as sea ice and precipitation can also be reported. The temperature (air and SST), humidity and SLP are measured in-situ by meteorological instruments, whilst waves, clouds and weather types are estimated visually. Wind reports are a mixture of measurements and visual estimates. The observations are transmitted in real time and also recorded in paper or, with increasing frequency, electronic logbooks. The electronic logbook software is also used to format manual observations, calculate more uniformly (e.g. dewpoint, true wind) and perform simple quality control (Kent et al. 2009).
Automated weather stations (AWSs) are being installed on VOS in increasing numbers, resulting in more frequent observations. However, a systematic programme of intercomparison with the traditional observations to ensure data continuity in keeping with GCOS monitoring principles is presently lacking. Moreover, a full high-quality AWS is expensive and some national services install low cost systems making only a subset of the normal range of observations, typically SLP and one or two other variables. Some elements of the VOS report require manual input, typically the visual estimates. Convincing the observers that supplementing the reports with this vital information is worthwhile has proved challenging, and the introduction of AWSs has led to a marked decline in the proportion of reports containing these parameters. Adding the capability for manual input adds to both the cost and complexity of the systems and is not always judged to be cost-effective.
The surface ocean observing system has evolved rapidly over the past half century, from being primarily VOS-based through the 1960s, to comprising increasing numbers of moored and drifting buoy observations starting in the 1970s and particularly dominating the last decade. Kent et al. (2009) show an example of how the number of in-situ observations available in the International Comprehensive Ocean-Atmosphere Data Set (ICOADS) has changed over time for selected variables, with the impact of the drifting buoys clearly visible for SST. These in-situ observations have been complemented by satellite measurements that began in the late 1970s. To meet the needs of applications such as weather forecasting, VOS observations are transmitted in real time to the National Meteorological and Hydrological Services (NMHSs) who then share the observations with other services using the Global Telecommunications System (GTS). Some NMHSs keep an archive of the data extracted from the GTS, however, these can differ between services due to differences in data conversion and storage formats and the way in which the data are retrieved from the GTS. VOS data contain fairly large random uncertainties, but in many regions the mean uncertainty due to poor sampling is much larger (Kent and Berry 2008; Gulev et al. 2007). In well-sampled regions the random uncertainties in gridded datasets will be small as many observations can be averaged. Sampling by multiple platforms gives the potential for extensive quality assurance, including near neighbour "buddy checks" and analysis of outliers. Typically VOS grid box averages contain observations from multiple platforms, allowing random uncertainty and also ship-to-ship biases to be reduced by the averaging process.
Datasets and analyses based on ICOADS are highly cited in the literature and form an important resource for climate researchers, especially those interested in large-scale estimates of ocean-atmosphere exchange of heat, freshwater and momentum and multidecadal climate variability. Datasets using VOS observations— in many cases based on the ICOADS collection include SST, sea level pressure, air temperature and humidity, surface fluxes and surface waves. In addition, it should be noted that atmospheric model reanalyses, which are widely used for climate analysis, are heavily dependent on the assimilation of ship observations (Trenberth et al. 2009). National and international assessments of climate change, most prominently by the Intergovernmental Panel on Climate Change (IPCC), use VOS SST data in the assessment of global mean surface temperature changes. Confidence in the SST trend is increased by its consistency with the marine surface air temperature trend, which is an independent measurement. VOS are the major source of air temperature information over the ocean and also contribute to the monitoring of climate change, for example in the bias-adjustment of infrared satellite estimates of SST (e.g., Reynolds et al. 2005). VOS also provide a consistent record of cloud changes since 1949 and have been used to derive a century-long analysis of wave information. The continuing move to produce data products in a timely manner from VOS data should allow an enhanced climate monitoring role for the VOS, if sampling can be maintained or improved. However, VOS datasets are currently underutilized for calibration and validation. New higher-resolution datasets characterised by uncertainty estimates should have wide applications for calibration and validation.
Was this article helpful?