Wireless sensor network can act as a distributed sampling system in precision farming applications which aim at estimating the spatiotemporal variability. Sensor nodes should be deployed referring to the communication range rc and the sampling range rs instead of the sensing range. In this paper, a kind of systematic random deployment based on three popular regular patterns-equilateral triangle, square and hexagon is explored to achieve both coverage and connectivity. The efficiency of different regular patterns of systematic random deployment is also compared, thus enabling to make a better deployment decision for different values of rc/rs.