The whole setting group is called transition settings, and refers to the boundary layer transition point (from laminar to turbulent).
The forced transition trip locations are exactly that, expressed as a fraction of the chord, for the top and bottom surfaces. They are akin to placing a trip strip at those locations.
The Ncrit value is a measure of free flow turbulence and is used to simulate the transition location when no forced trip location is given. XFLR5 uses the same method as XFOIL, which is an adaptation of the $e^N$ transition theory. Note this linear theory loses its validity at a fluctuation level of 1 to 1.5 percent, which corresponds to an N-factor of 6.9 to 7.3, meaning it is not suitable for highly turbulent flows.
As to what this is important for, the turbulent or laminar nature of the boundary layer affects flow separation and viscous drag. A turbulent layer will generate more drag, but separate at higher airfoil angles of attack.
Where this gets really interesting is at low Reynolds numbers, where a laminar layer will detach, but become reattached if turbulence is introduced into the airflow ahead of the airfoil, such as by a turbulator. It will then remain attached if the turbulator is removed, displaying a hysteresis loop. The same effect will of course occur in the opposite direction, with a barely attached turbulent layer detaching and staying that way if the free flow turbulence decreases enough to cause it to turn laminar.
I must note that this effect is confined to low Reynolds numbers, on the order of $10^4$.