The carbon cycle, like the hydrologic cycle, is a variation on the concept of conservation of mass. It is the tracking of carbon in the Earth system as it is stored in various reservoirs and transferred from one reservoir to another. In the process of transferring or storing carbon, the total mass of carbon in the Earth system is approximately constant. In fact, cycles of many other chemical components of the Earth system (e.g., nitrogen, sulphur, etc.) are also studied to understand such topics as nutrient availability, eutrophication, and the life cycles of pollutants.

Reservoirs: The major reservoirs of carbon are the oceans (dissolved in seawater and precipitated as shells and carbonate deposits), the lithosphere (carbonate rocks and fossil fuels), terrestrial biomass (vegetation and animals), and the atmosphere. The global warming and energy resources debate focuses on how the atmosphere reservoir is changing through time as a function of a shrinking fossil fuel reservoir. However, other reservoirs are also changing in size, either exacerbating or mitigating increasing atmospheric carbon.

Atmosphere: Measurements of CO₂ concentrations in the atmosphere at Mauna Loa have clearly shown that the carbon dioxide is increasing through time. Seasonal variability (low CO₂ concentrations in the summer, high concentrations in the winter) is superimposed on an increasing annual trend. Bubbles preserved in glacial ice provide us with a record of CO₂ concentrations back to ~600 thousand years ago. These records show that modern CO₂ levels are anomalously high, particularly when compared with levels over the last 1000 years. They also show that CO₂ has long been associated with temperature changes and the growing or shrinking of glaciers. Bill Ruddiman argues that CO₂ concentrations in the atmosphere have been anomalously high since ~8000 years ago, around the onset of agriculture, while methane concentrations have been anomalously high since ~5000 years ago, when rice cultivation became prevalent. The jury is still out on whether his hypothesis is correct.

Fossil fuels: The size of the fossil fuels reservoir is shrinking through time. The major question that policy makers debate is how soon this reservoir will be depleted. Hubbert's Peak is the name given to the point in time when the rate that humans extract fossil fuels begins to decrease. The logic behind the peak is that we will eventually run out of fossil fuels to extract, so the rate of extraction must decrease at some point. Worldwide, this time is predicted to occur in the next decade or so. However, many economists argue that market forces will prevent the decline in mining petroleum and other natural resources, or will at least significantly delay the decline.

Land: Land-use change has affected the amount of carbon stored as land-based biomass. However, quantifying the net change in carbon is difficult. The area covered by a certain vegetation regime and the density of carbon in that given area dictate the amount of carbon stored by that vegetation regime:

Mass of carbon=(vegetation land area in m²)*(kg of carbon per m²)

The land area is not too hard to determine today, particularly with the advent of satellites that can image all of Earth's surface. However, past land use is difficult to reconstruct for the entire globe, even for historic times. The amount of carbon stored below ground in soils and as root systems is even trickier. In other words, there are huge error bars on reconstructions for how the land biomass reservoir has changed. It is likely getting smaller, but the amount of shrinking is hard to pinpoint.

Oceans: Like the land-based reservoir of carbon, the amount of carbon in the oceans and how that reservoir is changing through time are difficult to determine precisely. The ocean is taking up some of the CO₂ emitted from burning fossil fuels, with some estimates indicating that nearly half of the emissions have been absorbed by the oceans. If that CO₂ remains in surface waters, it can be transferred into the atmosphere fairly easily. However, carbon that sinks to the deep ocean can remain sequestered for millions of years, until plate tectonics (volcanism) returns that carbon to the atmosphere. One proposal for "solving" the rising concentrations of CO₂ in the atmosphere is to artificially sequester carbon in the deep ocean.