Hydro plants don't USE water in the consumptive sense - they extract power as the water flows by. Losses if any are due to evaporation in storage. However, to be practical, they require storage and head, which affect how much power is produced at any given time, and of course storage so they can keep generating when it's not raining or melting snow. A major part of SF's Muni power comes from hydro generated by the Hetch Hetchy aqueduct, which runs by gravity from Hetch Hetchy reservoir to Crystal Springs - the hydro potential being one reason for choosing that water source in the first place. That said, you do need water to generate hydro power (small e.g. eligible renewable, or large), and in a drought you have less of it.

For those whose exposure to steam power is mostly through viewing UP 844 ... steam power plants cover the base (constant) electrical load. Regardless of how powered (nuke plants are steam plants, as are coal, oil, and some older natural gas plants) they use some water consumptively (whatever is lost during condensing, which is more for inland plants that don't have a convenient body of water to dump the heat to). Coastal plants use pass-through condensing of the exhaust steam, which may not consume water per se but does warm it to where the bugs and fish have problems. The latter has been an issue, and has shut down several plants where the power companies didn't want to construct different condensers or didn't have the space/money to do it. We will have base load plants for a long time - that's what keeps the lights on at night. Unlike most steam locomotives, where the water goes through the boiler only once before exhausting to the atmosphere, most power plants condense the used steam and recirculate the water to the boilers - it's more efficient that way.

There are many other types of power plants, including modern gas-powered peaking plants using turbines or engines that don't use any water (or very little - geothermal plants mostly are still steam plants and use some water during condensing). We need those, too, to cover cases where the sun doesn't shine on the solar plants, or the wind suddenly dies.

As far as hydro percentages are concerned - you're right, and Margaret needs to do a bit more research. They vary by area, as I noted earlier with respect to SMUD v. statewide averages. It's been worse since the drought hit. Of course, some imported power (practically all of what comes from the Pac NW) is hydro too, but it's hard to figure how much because the data are often buried in the "import" category where some coal also hides.