Our need for plants can be broken down into three factors: amount of plant life, kind of plant life, and the efficiency with which we use it. All need consideration. Here I'm concerned with the amount of plant life in the biosphere.
In terms of percent of mass or volume, animals are an almost insignificant part of the biosphere, though we tend to attach considerable importance to ourselves. In turn, the percent of actual "living" carbon is small compared to the unoxidized, organic, but "mineral" carbon. This is in the form of coal, peat bogs and even the interior of trees, which is basically dead wood except for a thin layer just inside the bark. And again, the total organic carbon is small compared to inorganic, oxidized carbon in carbonate rocks such as limestone.
The amount of plant life is dependent on five main factors: Sunlight, water, minerals (mainly phosphorus), carbon dioxide and oxygen (to plants, a poisonous waste product).
On a worldwide scale, the amount of sunlight, CO2 and oxygen is fairly constant. But in many land areas, a shortage of water or minerals means little of the sunlight and CO2 can be utilized. We mine phosphorus and other minerals, and spread them on the land where they sooner or later get leached or flushed into the ocean. Geological deposits of phosphorus will be in constantly shorter supply.
There is a set amount of available oxygen at the Earth's surface, free in the atmosphere and combined in rocks, water and organic matter. It makes up about half of the Earth's crust. The forces of chemical oxidation and photosynthesis create an equilibrium. Oxidation, such as forest fires and rotting, gives us CO2. Excesses of CO2 become carbonate rocks. Photosynthesis fights back turning CO2 back into oxygen and unoxidized carbon.
We often assume that since oxygen is a necessity for animal life, more must be better. Perhaps in some cases, but only if it doesn't interfere with our greater need for plants. Any changes we make in the equilibrium will be small until we deal with the oxygen problem (though our changes could be big enough to trigger an ice age). If we slightly reduce the percent of atmospheric oxygen, and increase CO2, by cutting away, and burning the tropical rain forests, plants will grow a little more prolifically on the rest of the world (where there's sufficient water and minerals). If we remineralize and reforest the deserts and destroyed rain forests areas, and increase their oxygen output, forest fires will burn a little more easily and oxidation of all kinds will happen a little faster. If we change the equilibrium of the biosphere in one area (in the ways usually considered) it simply changes in the other direction over the rest of the world. Instead, we need to change the balance point by changing the proportion of elements in the Earth's "metabolism".
A few years ago, hearing that nuclear fusion would be such an environmentally safe power supply, I was looking for possible harmful effects that might result from it. It occurred to me that consuming hydrogen (from water) for fusion power plants would release free oxygen which would slowly change the above equilibrium, oxidizing the biosphere (over centuries). I was promptly informed that the solar wind removes much more hydrogen from the upper atmosphere than we could ever use without creating global heat dissipation problems. See The Solar System Still. But according to my latest research (which isn't the final word), we still have a few billion years worth of unoxidized carbon.
If the deserts of the world are really increasing, with no compensating increase in growth elsewhere, then perhaps we are losing hydrogen from the upper atmosphere a lot faster than we think. Could it be that in spite of all our hopes and fears about continuing life on Earth, our future here will be decided by forces virtually beyond our control? Is oxygen perhaps not only the final, but also more immediate, enemy of the biosphere? (More likely, minerals are simply leaching away too fast.)
Even if oxygen isn't diminishing the biosphere, could we do anything to enhance it? Virtually all of every oxydizable element on the Earth's surface is found naturally in oxide compounds. (Silicates, carbonates, nitrates, sulfates, phosphates etc. are first of all oxides.) The major industrial process of smelting ores is to separate metals from oxygen. Carbon, the basis of all life as we know it, is one of the major holdouts, because of photosynthesis. Today, organic, unoxidized carbon only makes up about one part in 10,000,000 of the total.
The only way I can imagine to deal intentionally with the excess oxygen is to reduce the total amount in the world environment. Our best efforts at present would cause only a very insignificant change in the equilibrium, probably not enough to cancel the upset caused by the solar wind. If we could manage to maintain the oxygen percent slightly below it's present level, plant life would begin to increase (assuming sufficient water and minerals), using up CO2 and producing more oxygen. This would upset the balance between CO2 and carbonate rocks, causing the rocks to break down, producing more CO2 to be consumed by the plants. Perhaps some of the other oxide minerals in the Earth's crust would begin to break down to less stable ones, making smelting easier.
Now to the nonsense part, how to get rid of oxygen (only). Starting with the most far out science fiction: Please don't suggest that we ship it away from the Earth in rockets. This is like using your gas-eating Cadillac to take cans to the recycling center. On the other hand, rocket engines for other purposes, that used oxygen for reaction mass, and spend much of their time away from the Earth, might help a little. We might invent matter transmitters and ship oxygen off to Mars, or better yet, a fusion power plant which would convert oxygen to iron while at the same time producing energy.
Onward to a just slightly more real possibility (with nothing standing in the way except the need for world-wide cooperation, and we know how that goes). We could store liquid oxygen in giant insulated tanks, underground perhaps, and miles in diameter. Every eight grams of oxygen stored would release three grams of carbon. This would produce (I'm guesstimating) about five grams of organic matter. Not much, but then how many grams of oxygen would, say, a cubic mile tank hold? I think about 4 trillion metric tons, for an increase of two and a half trillion metric tons of biosphere.
So far, I've pretty much overlooked problems of mineral shortages. It's felt by some scientists that this shortage may add to the greenhouse effect, and also possibly trigger the next ice age. Shortage of minerals mean less plant growth, and in turn, more CO2, higher temperatures more evaporation from the oceans and more precipitation, which at higher latitudes and higher elevations will end up as ice. With water and more minerals evenly distributed around the world, plant life could control the greenhouse effect by increasing the bioshere in response to increased CO2.
On the other hand some might say that we need another ice age. Ice ages make more minerals available, and during the times between them the minerals leach into the ocean. Some of the rock dust from glaciers, or "glacial flour", containing many of the minerals the biosphere needs, gets transported by winds to nearby lands. With it's greater surface area of rock dust over solid rock, the minerals it contains are more available to plant life. But I think we can do it better than by producing ice ages.
It's been suggested that we develop rock crushing and grinding machines on a massive scale to supply fresh minerals. This sounds very energy intensive. The glaciers took considerably more time and energy to do this than is available to us. The great majority of this rock dust must end up in the sea, in river deltas and river bottoms. It could be quarried and redistributed.
I also have an idea, as yet untested, for a cheap, efficient method of de-salting seawater, which would help relieve the above shortages. This would also be a good start on a process of getting minerals from seawater.
I like the idea I of "planetary symbiosis" (in which terraforming is one of the actual physical processes). But then lets go a step further and talk of Earth as a single organism in which the network of human brains is just the latest major step in the development of the more highly conscious super-brain, sometimes called Gaia.