Without going into a long discussion about what the wave function is, I will try to briefly answer your questions:
1) For atoms, the (approximate) solution to $\Psi$ is the product of two functions:
- the radial components (distance from the nucleus)
- the angular/spherical components (behavior going around the nucleus, like latitude and longitude)
2) Since the radial components of each component extend quite far, decaying to zero only at infinity, it would be very difficult to make a plot of the 3D object on a 2D surface that does not cover up the features that would be nearer the nucleus (further from eye), without using varying degrees of transparency.
The spherical component is usually solved analytically using complex numbers (i.e. a+b*i) but they can be transformed into real values by linear combinations. So, instead of two "channels" for real/imaginary, we can use only one "channel" (real) for each function, 1s, 3dxy, etc. Often, these are colored so that if the sign of the real function is positive, it is e.g. red, otherwise, e.g. blue.
A second feature of these types of graphics is called the "isosurface"--which means, plot only if the value of the function is equal to a chosen value. Sometimes the chosen value is taken to be "90% of the corresponding electron density contribution from this atomic orbital is enclosed" (which, b.t.w. is $\Psi^* \Psi$ or simply $\Psi^2$. Though $\Psi$ may be real/complex/negative/positive valued (hence need all sorts of colors and "channels" to represent it) the electron density is positive and real. This is very important, as complex quantities are not observable, but are rather mathematical constructs (vehicles) for arriving at things that are actually observable, such as the electron density.
But practically speaking, in order to see certain features like the doughnut in $dz^2$, people play a little fast and loose with precisely what the value of the isosurface is.
The shading on the surfaces that you have provided could be misinterpreted, because the colors do not correspond in any way to the value of the function being plotted, but rather seem to be an aid to the volumetric rendering of an isosurface that has already been set to either the absolute value or the square of the atomic orbital.
Finally, note that in your center drawing, which adds all of the individual atomic orbitals, that you are beginning to form a sphere. It had better, as atoms are, in fact, spherical!${}^{\dagger}$ The formal name of the functions are spherical harmonics, and the set of e.g. $p_x$, $p_y$, $p_z$ or the five $d$'s taken together form a spherical shape.
${}^{\dagger}$ assuming the nuclear quadrupole is zero