Most likely, what you want to do is this:
int h = 3; int w = 4;
double[,] ou = new double[h, w];
Do you see the difference?
What you tried to write could make sense if you do this:
double[][] ou = new double[h][];
ou[0] = new double[w];
uo[1] = new double [200];
uo[1] = new double [w * 2];
This way, each inner array of
double
can have different length, hence "jagged array".
Now, you can learn what Rod suggested but... why not simply reading C# manual (for example, from Microsoft help)?