That image shows a fan, which is easier than an arc. To get a fan, fire projectile 1 in the direction of the ship. Fire the other two projectiles at an angle (ship angle +/- T)

The arc is more complicated, since the side projectiles will have to gravitate toward the trajectory of the middle projectile. I worked out the vector algebra for this, but I was rushed so you'll want to check my work. Here's the first projectile:

The projectile is fired with x velocity vx and y velocity vy.

To get the side projectiles, imagine another force pushing them along a vector orthogonal to projectile 1's velocity vector.

f designates the fraction this secondary force acts at as a fraction of the "forward" velocity vector. This doesn't give you an arc yet--rather, a fan. To get the arc, we'll want to reduce the impact of this secondary force as time elapses:

tmax represents the time it takes for the two side projectiles to align themselves in the same direction as the first projectile, and t is the current time. So your velocity formulas are

x velocity of projectile 2 = x velocity of projectile 1 - (f * t)/tmax * y velocity of projectile 1

y velocity of projectile 2 = y velocity of projectile 1 + (f * t)/tmax * x velocity of projectile 1

x velocity of projectile 3 = x velocity of projectile 1 + (f * t)/tmax * y velocity of projectile 1

y velocity of projectile 3 = y velocity of projectile 1 - (f * t)/tmax * x velocity of projectile 1

Initialize t = tmax and then subtract from t until it gets to 0. Once t = 0, discard these formulas and set the velocities of projectiles 2 and 3 equal the velocity of projectile 1.

**Addendum.** Once you get this working, there's opportunities for some interesting projectile patterns. For example, allow t to go negative and the side projectiles will start to gravitate back toward the middle projectile. You can also replace (f * t)/tmax with f*sin(t) to get a wiggle pattern.