Looks like you have multiple such curves in your data. And they don't have an exact profile as the one that you show.
![peng_liu_1-1640097365774.png peng_liu_1-1640097365774.png](https://community.jmp.com/t5/image/serverpage/image-id/38576i4C6210F46F9B9F3B/image-size/medium?v=v2&px=400)
![peng_liu_2-1640097448445.png peng_liu_2-1640097448445.png](https://community.jmp.com/t5/image/serverpage/image-id/38577i21F4B99E285794A7/image-size/medium?v=v2&px=400)
![peng_liu_3-1640097474070.png peng_liu_3-1640097474070.png](https://community.jmp.com/t5/image/serverpage/image-id/38578i8B678CCE82EFCC99/image-size/medium?v=v2&px=400)
So questions:
- What is the definition of "Top" given the real data?
- Do you want the all Top's of individual curves or, a "Top" of all curves? Definition.
Here are a couple of possible approaches.
First one. If your definition can define one and only one point on individual curves, or all curves, express your definition as a formula, which returns 1 for match the definition, 0 otherwise. I have tried a couple of my imaginations, and seems such a rigorous definition is not trivial.
Second one. If the desire is just a rough quantity of some sort of "bottom" at the "top", seems even a histogram can be an interesting approach.
![peng_liu_8-1640098954529.png peng_liu_8-1640098954529.png](https://community.jmp.com/t5/image/serverpage/image-id/38584i07C26D69E8B56B2B/image-size/medium?v=v2&px=400)