Interesting thread. The ...

Interesting thread. The responses to date leave me wondering about the data side of the question.  As noted, there are both historic planning methods issues, and also need to seriously re-evaluate relative to climate change.  

There is a tension between these in that the choice of long term baselines implies that you are smoothing out recent notable changes in precipitation regimes.  For the US, this is pretty well summarized in the National Climate Assessment here:

https://science2017.globalchange.gov/chapter/7/

But there are some interesting datasets and tools out there which could be used to construct an unbiassed answer at trans-national scale.

Global Reservoir and Dam (GRanD) database

http://www.gwsp.org/fileadmin/downloads/GRanD_Technical_Documentation_v1_1.pdf

WorldClim - Global Climate Data
Free climate data for ecological modeling and GIS at 1km downscale in GIS formats (also available on Google Earth Engine)

http://worldclim.org

Global "Hand" flow accumulation over terrain dataset

https://code.earthengine.google.com/ac07d0488c6de32f35372c3a2491f35a

I don't know if anyone has taken the time to start piecing this together.  In theory, you could use the datasets above to look systematically at historic and projected future changes to the watersheds feeding at least a few thousand important dams.