Solar and wind power are intermittent. When the sun goes down, or when the wind stops, solar and wind power facilities cease to generate power. Sometimes we can predict when that happens, as for example when the sun sets. But sometimes it happens without much warning: a nest of large dark clouds will roll in and cover an entire state, or the usual prevailing winds vanish for days on end.
We've built an industrial society based on the assumption of 24/7 electricity based mostly on burning coal for our electrical power: coal-fired power plants can run day and night. In order to replace coal-fired electricity with renewables, we need to find a way to make renewable power as readily available. But except for large hydro (which can't be explanded much more) and geothermal (which isn't particularly widespread), renewables just aren't as dependable.
And you can't just put power into the grid when the sun's shining for use later. Electricity is either on or off.
Which means that to replace those 24/7 coal (and oil and natural gas) power plants with solar and wind, we need to find a way to store that power. We don't have it yet.
In order to store the power in electricity, it has to be turned into some other kind of energy. This is relatively straightforward, though not always cheap. There are a few basic methods that have been proposed for renewable energy storage, and we'll probably end up using most of them in combination.
Pumped Water Storage: currently the most feasible method of storing renewable electricity, this method involves pumping water uphill to a reservoir when the sun is up or wind is blowing, then letting it run back downhill through a set of turbines to regain some of the power used to pump it. A main advantage is that can be done using present-day, widely available technology. It's also reasonably efficient, the turbines recovering as much as 75% of the energy used to pump the water. Siting is a problem, though, as a large-scale pumped water storage facility would require two adjacent reservoirs with one significantly higher than the other. If the local topography doesn't lend itself to the design then implementing this strategy could be prohibitively expensive. On the other hand, it does offer some potential for smaller scale energy storage in small communities, using water tanks as reservoirs.
Similar means of storing electrical power as mechanical energy include compressed air stored in mines, and flywheels -- which are generally useful only in strictly limited applications.
Hydrogen: We heard a lot about this possibility about ten years ago, when former California governor Schwarzennegger was touting his proposed "Hydrogen Highway." Hydrogen is produced by splitting water using electrical energy; if it's captured and stored, it can be burned to power turbines or -- in the slightly more science fictiony scenario -- run through fuel cells to produce electricity.
Despite its reputation, hydrogen doesn't burn clean. Air is four-fifths nitrogen, and air pumped into a hot engine that's burning hydrogen fuel mixes nitrogen, oxygen, and hydrogen at high temperatures, allowing formation of smog precursors like oxides of nitrogen. Most hydrogen-powered vehicles use fuel cells rather than internal combustion engines, limiting that drawback somewhat. And though public fears of Hindenberg-style disasters are probably overblown, as gasoline is a far more dangerous fuel than hydrogen, hydrogen does explode very readily when it comes in contact with open air and an ignition source, so hydrogen leaks in tunnels and parking garages would be problematic.
Mostly, though, hydrogen just isn't an efficient way to store electricity. It takes far more power to split water to make hydrogen than you get combining hydrogen with oxygen to make water. Water is a very tightly bound molecule. (That's why it's good for putting out fires.) When it comes to energy storage -- and powering vehicles, for that matter -- hydrogen just doesn't stack up compared to other available methods.
Batteries: We've each of us used batteries to store power every day, from the hazardous lead-acid battery in your vehicle to the slightly less messy (and much lighter) cells in your phone, laptop and almost every other appliance in your house these days. People who've maintained "off-the-grid" houses over the last 30 years or so generally stow a bank of batteries in an out-of-the-way spot, charge them during the day and run off the stored power at night.
Some utilities do the same thing on a larger scale, though the technology gets more expensive and more industrial the bigger you get. Sodium-sulfur batteries are used in some places for grid storage; running at temperatures like 300°F and made up of at least one metal -- sodium -- that explodes on contact with air, these are not things you want in your basement.
One idea touted with increasing frequency is the notion of using plug-in electric cars as grid storage. Once electric cars gain popularity, a certain percentage of them will be plugged in for charging during business hours, when solar panels and windmills pump energy into the grid. At night, a large percentage of those cars will be plugged in in garages at home, and could conceivably feed that stored power back into the grid. Batteries that can be used in vehicles have limited lifespans, though, and increasing the frequency of charge cycles will lower that lifespan.
Thermal Storage: some solar thermal facilities on the drawing board plan to use a heat-retaining fluid medium, such as molten saltpeter or other salts, to keep the plant's turbines operating long after the sun goes down. The idea is that concentrated solar heat during the day will melt the salt, which will then be stored in an insulated tank. Initial results indicate that heat sufficient to turn turbines could be stored for as long as a week.
While technically feasible, solar thermal storage may not yet be economically feasible. Adding thermal storage to a solar power tower greatly increases the cost of construction and operation, thus increasing the cost of energy from that plant. The extra hours of power production each day would be fed into the grid at night, when electricity demand is lowest and prices per kilowatt-hour are lowest as a result. Thus expensive solar thermal storage power would have to compete with cheap off-peak power, leaving the commercial viability of the practice open to question.