The signals are received by being bounced off of a layer of moisture in the sky, that arcs from horizon to horizon (dipping slightly in the center). At and near the equator, the curve of this moisture means they must point straight up to avoid being scattered, while elsewhere there is less of a curve.
If that was the case it would be really unreliable as you are dependant on special weather conditions to provide the exact amount of thickness to provide the angle. If it was the normal answer of satellites in space then all you have to worry about is cloudy weather.
I do not need special weather conditions, the moisture is always there. The composition of air does not change so much depending on the day. I am not talking about specific weather.
You are acting as if reflection and refraction are unknown. They are well studied and understood phenomenon. The amount of signal reflected from passing from air to water is minimal, there are even calculators you can use online which will tell you the the exact amounts based on variables.
If the signals we are receiving are actually ground based and reflected by moisture, how do satilite receivers work under water? Would the surface of the ocean not reflect the signal back?