Abstract
Errors in numerical analysis using three observing stations are discussed for the large and small scale network. First, the differences between flux and advective form calculation are verified to be sufficiently small. Second, the errors caused by the drifting of balloons are shown to be considerably large in the smaller network.