1) This is from the node documentation for the fluid emitter node:
Fluid Jitter:
Determines whether emission is based on random (jitter) or fixed (no jitter) sampling.
The default, Jitter on, provides better anti-aliasting around the edges of the emitting volume. Jitter off works better with the Spring-Mesh solver (Ocean and Pond wakes), and reduces "sizzling" effects in the region of the emitter.
2) Fluids will behave very differently than particles with fields largely due to the incompressibility of the fluid solve. The force from the field adds into the velocity grid values and then the incompressibility condition is applied before propagating the density. One can lower the simulation quality, which will tend to allow more divergent and convergent (i.e. compression) velocities on the fluid, however the total density on the grid will then not be preserved. A divergent velocity can result in a spot of density becoming like an emitter. The fluid will also diffuse more. In general it is better to point fields, using more directional or volume axis fields. Volume bounds also help localize the effect of the field. In some cases you may only need to initialize the fluid start velocity with the field, after which you can disable the field. Or you could paint the start velocity by hand.
To best understand how it works enable the velocity grid display to see how the field interacts with the overall fluid. Also... some fields, like gravity, only effect velocity where there is density, while others may apply force to the "air" as well.
Duncan