Perhaps the biggest scientific, political, and practical challenge facing the world today is how to rapidly decarbonize our energy infrastructure. The goal is to do it fast enough to avoid total global warming in excess of 1.5 degrees Celsius, which we will probably fail to do. At the very least we want to avoid going over 2.0 degrees Celsius, which is more realistic but will still be a challenge. We are not on a path to achieving either goal right now. We are currently at about 1 degree increase over pre-industrial average surface temperature.

There is a lot of healthy debate about what is the best path to net-zero carbon energy, and it’s fascinating to watch. I personally think we should do everything. We don’t want to put our chips down on one answer and find out in 30 years it was the wrong choice. We really have just one shot left to avoid 2.0 C, so we should spread out bets out as much as possible. Also, we should use each source where optimal, rather than try to have one or a few energy sources fit all circumstances. This means investing in renewables, many forms of grid storage, carbon capture, nuclear power, nuclear isotope batteries, artificial leaf technology to make hydrogen, maybe even some biofuel, and improved energy efficiency. One thing is certain – we need to get off fossil fuels as quickly as possible. This also means the transportation sector needs to wean off fossil fuels.

In this debate there are many solar advocates. I, too, am a fan of solar, I just don’t think we should count on solar alone (which would require grid storage we don’t currently have). Solar is currently a cost-effective option, and is only getting better as the technology continuously incrementally improves. The problem is, as we try to push intermittent renewable energy further and further, that cost effectiveness goes away. This is because you need overproduction and/or grid storage, plus significant grid upgrades, to handle the intermittent and distributed nature of sources like solar.

If we try to push solar to be a significant portion of our energy supply, the question then becomes – where are we going to put all these solar panels? First we should pick the low-hanging fruit – places that are environmentally safe and also produce the power where it is used. The obvious answer here is rooftops, but how much of our energy needs would rooftop solar cost? One analysis found:

In total, they estimate that there are a little over 8 billion square meters of suitable roofs in the US. Cover that in solar panels, and you would produce about 1,400 terawatt hours of electricity each year—about two-thirds of which would come from small residential buildings. The total production is equal to nearly 40 percent of the total electricity currently sold by utilities in the US.

That’s not bad. That may be the only solar we need, if we combine it with other energy sources. Even 40% solar penetration would need serious grid storage, but that can be paired with the solar panels in a distributed fashion also. Add in wind farms where they are optimal, hydroelectric, geothermal, and new small modular nuclear reactors and we might be there. But what if we want to push solar further? What about solar farms where the sun shines the brightest? This runs into issues of environmental impact and land use. Although this is likely to be far better than strip-mining and burning coal, anything we do big is likely to have a big impact.

One proposed partial solution is floating solar farms on large lakes. The question is, will this have a net positive or negative impact on the ecology of the lake, taking into account the degree of warming that has already occurred? The short version is that there is some good news, but concerns remain and we need more study.

The good news is that the floating solar farms will block some of the sun from heating the surface water, which will counteract some of the negative ecological effects of global warming. They will also prevent evaporation, again compensating for the increased evaporation from increased average temperatures.

However, there are also potential negative effects having to do with the temperature gradient at different depths of the lake. It is not yet clear what the net effect would be, and of course would depend on the percentage of the lake surface which is covered by solar panels.

I think this study further supports that the best approach to energy production is to spread it out as much as possible – to different techniques and different locations.  There will not be any one solution, partly because of unintended consequences when you try to significantly increase the scale of any one technology.

The obvious location for solar panels is in the desert, but even there it can be disruptive to local fauna and flora. The Sahara, which is mostly sand, would be a good choice, but it would be challenging to distribute that energy to he world. So again we are left with spreading solar around as much as possible.

What about space solar? NASA explored this idea, but eventually abandoned it. However, recently the Chinese government has been looking at space-based solar. Nothing is final yet, but they are contemplating developing a commercial-scale space based solar farm by 2050. In any case – this would involve orbital solar panels that then beam energy down to stations on the Earth’s surface.

The basics are that solar satellites would be assembled in geosynchronous orbit. At this distance they would rarely be in shadow, would not have to contend with atmospheric blockage, and so would get lots of sunlight. This could be converted to microwave energy that is then beamed to a fixed receiver on Earth. Such satellites could generate a lot of power, and while the startup costs would be in the billions, would ultimately be fairly cost effective. One downside is that the ground receiver would have to be several km in diameter, so this does not save us from land usage. Repair would also be prohibitively difficult.

An alternate system would use lasers rather than microwave beams. These would be highly efficient and require smaller receivers, but raise concerns of weaponizing space-based lasers.

It is probably unlikely that we will get significant energy from space-based solar, but it is not outside the realm of possibility.

For now we are not even close to maximizing rooftop solar, so we could choose to do that before we need to consider where to put massive solar farms. But in practice both are happening simultaneously.