In Geostatistics, vector data with two components, such as measures for wind field, electromagnetic field and ocean currents can be appropriately modeled by recalling the theory of complex-valued random fields. This is especially suitable for describing phenomena whose variables are expressed in the same unit of measurement and refer to homogeneous quantities.

In this paper, a first approach useful for taking into account the temporal profile of complex data is described. After introducing the fundamental aspects of the complex formalism of a random field indexed in time, a new class of models apt to include the temporal component is proposed. A case study on a vector data set regarding surface ocean currents is provided. In particular, these data, derived from high frequency radar systems, were collected during the 30th of April 2016, from 207 stations along the US East and Gulf Coast. The complex covariance function, indexed in time, is estimated and modeled, then it is used for prediction purposes. A numerical analysis is also proposed in order to evaluate the consistency of the time varying complex model.

在地统计学中，可以通过回顾复数值随机场的理论来适当地建模具有两个分量的矢量数据，例如风场，电磁场和洋流的度量。这尤其适用于描述其变量以相同的度量单位表示并且指的是均匀量的现象。

在本文中，描述了一种可用于考虑复杂数据的时间分布的第一种方法。在介绍了随时间变化的随机字段的复杂形式主义的基本方面之后，提出了一种新的易于包含时间分量的模型。提供了有关与表面洋流有关的矢量数据集的案例研究。尤其是，这些源自高频雷达系统的数据是在2016年4月30日从美国东部和墨西哥湾沿岸的207个站点收集的。对按时间索引的复协方差函数进行估计和建模，然后将其用于预测。还提出了数值分析，以评估时变复杂模型的一致性。