Stable part of time series


i have a time series (let's say cash volume) designed like this:
Is there a statistical model I can use to estimate the stable part of my cash volume? In the figure the stable part is the sum of the dark blue and gray areas.
Thank you for your help,

Jim Castriotis
Does it need to be any more complex than the area of a rectangle plus the area of a triangle?
Yes. I have time on the x axis. I need to estimate stable volume in the y axis in the next time period. So i must consider the increasing trend too. I was thinking more of a time series decomoposition method but I dont know which to apply in practice