The primary reason we need nuclear power is that it allows us to reduce our fossil fuel use.1 There's little chance it will ever eliminate fossil fuels, even for its own operation. Raw uranium will have to be mined, processed, transported. Nuclear waste will have to be vitrified and hauled to repositories. However, a 1-Gigawatt nuclear power plant requires 1 ton of enriched uranium to operate for a year. Getting the 1 ton means mining 200 tons of uranium ore. But a 1-Gigawatt coal-fired plant requires 2 million tons of coal for that year of output. There is much more transport involved in getting the coal to the plant by truck or railroad, and much more fossil fuel is used to power that transport. The coal plant itself, of course, burns a lot of fossil fuel and pours out oodles of carbon dioxide. It also pours out varying amounts of pollutants, fly ash, and radioactivity.

The Union of Concerned Scientists, citing EPA statistics for 2014, gives us this assessment of aggregate pollutants emitted by U.S. coal-fired plants that year: 2

Other options — wind, solar, hydro, geothermal — can also reduce our dependence on fossil fuels. But they are intermittent in space or time (or both) and so not as well suited as nuclear is to providing baseload power whenever and wherever needed. And they have other defects. Every source of power has defects. The key is to enlist smart engineering to minimize those defects and to combine various sources to best advantage in well-designed systems.

The second reason for keeping nuclear power a part of the energy picture is flexibility. This has two aspects. First, logistics: if you need to site a power plant in a place to which it's hard to transport fuel, nuclear power has a clear advantage. Today, the prime example is McMurdo Station in Antarctica. The other aspect is power density. A space mission away from the Sun quickly reaches a point where solar power is impractical. Current nuclear-powered spacecraft use radioisotopes, not fission. But very soon the power requirements will reach hundreds of kilowatts. Radioisotopes are feasible up to about one kilowatt; beyond that, they become prohibitively expensive. There's only one proven technology that can handle tomorrow's high-power space missions: the compact nuclear fission reactor.

The third reason to keep nuclear power alive is that it also keeps nuclear weapons technology alive. I know this is heresy in the current political climate, and I have no problem with a "build-down" in the numbers of deployed nuclear weapons. But if we hope to open up the solar system, we must preserve the ability to apply nuclear weapons. Actually, I believe that's vital even if we stay on Earth. The reason is the same: asteroid deflection. We will find one with our name on it — not soon, but someday. When we do, we might manage the deflection with conventional explosives, or with big rocket motors. But I believe nukes will prove necessary. Then, once we have moved out beyond the Moon, I foresee a need to boost thousand-ton payloads around the system ASAP, and for that the nuclear pulse rocket3 seems ideal. I also suspect asteroid mining will be aided by nuclear weapons, but that's a harder case to make. Nevertheless, we must not bind our successors.

Nuclear power is no panacea. It will never produce energy "too cheap to meter". It will always demand great vigilance in operation, and it will always produce radioactive wastes. The wastes, especially, are a formidable problem. But I believe we can and will solve it. With the costs of fossil fuels rising and their stocks dwindling, with man-made global warming looking more real every day, it is clear to me that nuclear power should be a part of our energy strategy.