There are reasons to Terraform and there are reasons to not Terraform.
For my answer, I take the term "Terraforming" as a generic term to create a passive closed-loop environmental system capable of sustaining human life. It's only external requirement is to supply the system with sufficient energy to drive the system. This includes massive Terraforming projects like making the surface of Mars habitable as well as creating passive closed-loop environments inside of asteroids or domed craters on various bodies (e.g. Mars and the Moon).
- Closed-loop - The system only requires an energy input (e.g. sunlight) to keep it running.
- Open-loop - The system requires constant replenishment of various components (oxygen, water, food, etc.).
- Active - The system requires machinery and/or human intervention to keep it running (e.g. the Closed Ecological Life Support System (CELRSS)).
- Passive - The system runs without machinery and/or human intervention (e.g. a biosphere).
Reasons to Terraform
- Maximize the space available for human settlement
- Minimize risks of environmental system failures
- For long-term habitability of the colony
The following sections discuss these 3 reasons but not in the same logical grouping.
By definition, humans evolved to operate in a "shirtsleeve environment". Operating under conditions outside of this bound (too hot, too cold, too much gravity, not enough pressure, etc.), reduces the productivity of humans. In some cases, it also has adverse affect on human biology.
Terra-forming provides a technique to create a shirtsleeve environment in as large an area as possible. For planetary bodies, this ultimately would mean terraforming the entire body. Before that though, the colonists will Terra-form smaller sections (e.g. domed craters).
Domed Lunar Craters:
As @o.m. pointed out. Humans require a pretty tightly controlled environment in which to live. You can either create these environmental conditions using an actively controlled environment or you can create a passively controlled environment that doesn't require human intervention.
Closed-loop active environment systems (e.g. something you'd put on a spacecraft) require constant supervision, maintenance, and other human intervention. If the responsible people get hit by a bus or the machines break and the people can't repair them - then everyone dies.
Even closed-loop active systems are large and massive. For trips of less than 1-2 years duration it makes more sense to just bring the supplies and forget all of that machinery and chances for problems.
The problem is you throw everything away after it is used. Which means everyone dies if you don't get regularly resupplied.
Closed-loop passive (aka Terraformed)
For very long-duration missions, especially ones in which it's possible the crew will forget how to maintain the environmental system, then you will want a closed-loop passive system.
Passive environmental systems function with or without human intervention. For long-term colonies or generation ships, some form of closed-loop system will be a requirement and making it passive will be the goal.
Reasons against Terraforming
- Project time
How much might it cost to redirect comets to strike Mars and increase its volatiles inventory? Hundreds of billions or trillions of dollars? Even the "richest" entity capable of running a project like this on Earth (the US Government) would find it difficult to finance such a goal. Give the time horizon mentioned below, it would be exceedingly difficult for other organizations to justify the risk.
The current cost combine with the risk would not be worth the distant reward. Even when you ask questions like "how much might a fully terraformed Mars be worth" are actually not important when the payout goes to the great x50 grandchildren of the person making the investment.
Terraforming even the most Earth-like planet (e.g. Mars) might take a thousand years or more. For even the most optimistic humans, this is well beyond our normal capacity to operate. We just don't think in terms of projects lasting that long.
Large corporations are arguably the best at long-term planning and execution and they aren't really capable of managing projects beyond a time-horizon of 10-20 years. By comparison, elected governments may plan for 10-20 year time frames by have problems with execution follow-through when elections replace the original planners.
After all the people reaping the rewards won't be born for more 50 generations. I suspect we won't see serious discussion of massive/costly Terraforming efforts until human life-expectancy gets to be close the length of time such an endeavor might take.
Passive environmental systems require much more "stuff" than the active ones. It won't make sense to use passive systems on anything intended to move around (e.g. most spacecraft) unless the mission requirements require it (e.g. a generation ship). Carrying the extra mass of the passive system would simply cost too much reaction mass to move.
Of course planets & moons will be likely targets for Terraforming. However, we might see Terraforming and development of passive systems for giant space habitats (e.g. O'Neill Cylinder habitats).
Interior of O'Neill Cylinder: