Structural Criteria for Quality Control of Oceanographic Data

Quality control (QC) of oceanographic profile data is a needed element in development and maintenance of the multipurpose World Ocean Database (WOD). Space-time climatic characteristics of the World Ocean are a vital tool for construction of criteria for QC. The procedure of calculating climatic characteristics consists of dividing an area of study into regular rectangles (grid) and calculating the statistical parameters of each rectangle. Statistical mean and deviation are usually used to estimate the quality of each measurement. The main problem in application of climatic statistics of a regular grid is due to the fact that if two or more water masses exist in a grid, then the average climatic characteristics of the area can't serve as a generalization of gridded data. To address this problem we propose the structural criteria (SC) for QC of oceanographic profile data that does not need to calculate climatic characteristics. Instead of dividing an area of study on a regular grid and calculating climatic parameters of rectangles, we calculate the shape of the area where parameters of sea water vary within very narrow limits. The Barents Sea was used as the test area for the developed SC for QC of water temperature data. The Barents Sea temperature varies from -2C to 18C, we plotted the area of the temperature distribution in two degree intervals. The shape of the area for each month is the "normal" of temperature distribution within a particular range. The same procedure was repeated for temperature ranges -2-0°C, 0-2 °C, 2-4°C, ..., 16-18 °C., This method was developed based on the shape of these areas, to allow us to estimate how incoming temperature data within a certain range relates to the shapes of "norms". Results of experiments on QC of 328,945 stations carried out in the Barents Sea for the years 1870-2015 are discussed.