I can express z=sqrt(1-...) but in a general case...
In the general case I usually use isosurfaces to plot implicit functions. Since I am completely new to scilab I do not know if it is just a lack of my ability to find a corresponding method in scilab of if there is none to plot isosurfaces.
If there is none and there is interest in having one I could adapt the Octave function I wrote a few years ago and which is since then part of their software to work in scilab and also the implicit function for plotting the zero level surfaces of 3d functions.
I tested it now: Take the function contour3d.sci from the link above. It
seems to be an old function which uses a few deprecated function calls.
You need to change line 36 and replace xbas() with gcf() in addition you
need to replace in line 82 sort with gsort.
Now you can run the following simple test
function val = isotest(x, y,z)
val = x.^2+y.^2+x+y+x.*y+z.^2-1;
x = linspace(-2, 2, 31);
y = x;
z = x;
[X Y Z] = meshgrid(x,y,z);
val = isotest(X, Y, Z);
contour3d(X, Y, Z, val, 0)
f = gcf();
f.color_map = [1 0 0]; // make the black surface red