Free-space optical (FSO) communication systems have currently a restricted range, because of atmospheric effects
which reduce their application range. The goal of the SCALPEL project is to study the feasibility of long range FSO
systems (goal: 20 km), i.e. to estimate how dedicated devices could enhance the range of FSO communication systems,
for example by changing the link's wavelength for a better atmospheric transmission and weaker turbulence effects,
and/or by using an innovative adaptive optics device to compensate, at least partially, turbulence perturbations.
In this paper, we study how the atmosphere constrains the system design in terms of transmission and turbulence.
We show that the system cannot work unless it has a full-wave adaptive optics correction, and that a wavelength around
4 μm presents several advantages toward the usual wavelength, i.e. 1.55 μm. A first design of the system is then
presented, including the source and the sensor.