Second most obviously, because they want more power than the river can generate.
Third most obviously, because they want to leave some energy in the river.
[1] https://www.oecd-nea.org/ndd/reports/2010/nea6861-comparing-...
They can also pose a substantial danger in the case of failure, see e.g. https://en.wikipedia.org/wiki/Vajont_Dam (~2000 casualties, vs Fukushima Daiichi's estimated death toll of ~500).
(I'm a big fan of hydro power, just saying it also has its draw backs)
Also I would assume, the higher up the mountain the dam is build the smaller the impact due to less flora and fauna.
My point: Once you go into small and hard-to-quantify large-population effects, your error margins become ridiculous and you can pull out any conclusion out of your hat...
https://www.research-collection.ethz.ch/bitstream/handle/20.... (page 83, German)
And similar to air travel, people think it's unsafe because of constant fear mongering and focus on incidents that are statistically insignificant. There were more people who died at Fukushima because of the evacuation procedures than the expected long term death toll from radiation exposure (and, notably, there were 0 deaths directly caused by the incident).