The Random forest is a non-continuous model which is represented by the horizontal lines in the graph. We get more stairs as compared to a single decision tree. We have a lot more of splits on the whole range of levels and therefore a lot more intervals of the different levels.

So each straight horizontal line here separated by the vertical lines is one split.

Consider the position level = 6.5, here the random forest took 10 different trees on which the salary of the 6.5 level would be and then it took the average of all the different predictions of the salary of the 6.5 level made by different trees in the forest.

A whole range of levels is spitted into more intervals and that is because the random forest is calculating many different averages of its Decision Tree predictions. Here no. of trees in Random Forest have no effect on the no. of steps (splits). More no. of trees just mean that the average of these votes will converge to the same ultimate average.