On the Statistical Structure of Hydrometeorological Fields in the North Atlantic

We study the possibility of detection of statistically significant trends as a characteristic of low-frequency variability of hydrometeorological fields in the North Atlantic on the basis of relatively short time series. We use the monthly data of the Russian Hydrometeorological Center for 1957–1990 and analyze their statistical structure. It is shown that significant linear and quadratic trends can be detected in the fields of most hydrometeorological characteristics of the North Atlantic. The residual signal is approximated by using the first-order autoregressive model. Typical values of the coefficient of autoregression vary within the range 0.3–0.6 (for different hydrometeorological characteristics of the North Atlantic). The maximum correlation is observed for the fields of sea-surface temperature and humidity, whereas the minimum correlation is typical of the fields of wind velocity and the difference between the temperatures of water and air.