In many instances, it is useful to view data as a collection of curves, particularly when the questions motivating the analysis relate to properties of curves. In this case it makes sense to view the fundamental datum as a curve and refer to the collection of curves as functional data. This work deals with spatial models of functional data, where the underlying smooth curves are assumed to belong to a reproducing kernel Hilbert space. The methodology developed here is illustrated using satellite data of phenological measurements over India.