I get your point of view: it feels like we are adding unnecessary complexity in the user interface. Plus, questioning why things are the way they are is generally a healthy and encouraged approach in UX.
However, regarding zoom buttons, they provide needed alternatives. Let's imagine various situations in which they might be needed.
Maybe it's an old mouse and the scroll is broken, maybe it's an uncommon device for which scroll cannot be used for the moment.
- The user cannot use the scroll
Often, one's is to experience frustration when browsing with a mobile on a page with a map and some content just underneath it. If only they thought about another way of zooming in/out in this particular case? Zoom buttons makes it easier to implement.
You may say, it's a technical problem, but let's imagine someone with a broken thumb and he just cannot pinch in or out, but would easily use zoom buttons.
- Scroll is not effective for a given task
Maybe the map is going to be used for certain tasks that require a special zoom level. Using zoom buttons can be a more reassuring option as they have fixed zoom levels.
My point is, it's hard to imagine all the situations in which an alternative might become useful. It's probably best to not test these assumptions and keep a convenient alternative interaction for zooming - which is already known by users.
Just a side note, it reminds me of the Curb cuts problem and how helping wheelchairs getting off and on sidewalks actually helps all people using wheels.