The problem with using just one is that geosynchronous orbital periods are one sidereal [1] day, not one solar day. So depending on the time of the year, the satellite would be in position in the wrong half of the day. On the other hand a satellite in Tundra orbit spends not half but more than half of its time in the designated area. Maybe a good compromise orbit could be found.