Abstract
The authors have discussed space-time descriptions of seawater qualities in coastal sea area in Osaka bay in the previous papers introducing the Auto-Regressive mathematical model with seasonal adjustment considerations. They showed that the seasonal adjustment introduced in the mathematical model makes it possible to estimate the trend and seasonal components of seawater qualities from data observed for many years. They also showed that the AR-components of short-period recurrent phenomena fluctuate randomly like noise components. To solve the problem, they introduced a probabilistic approach to-the AR-components, and they showed that it is possible to treat the AR-components with the noise components as random variables In the present paper, they discuss more precisely about the understanding and treatment of AR-components that are not described properly due to the effect of time interval of observations, and about the possibility of the space-time assimilation of measured data introducing a 2-dimensional interpolation technique. They show that the AR-components have the same probabilistic characteristics as that of noise components in the analysis model, and that 2-dimensional interpolation is applicable to the space-time assimilation of measured data.
