Climate Change and the Oceans


The Basic Problem

In recent years, scientists have become increasingly aware that the earth’s climate is changing. Below is a plot showing the earth’s global mean surface temperature for the past 140 years, roughly the period covered by good measurements of air temperature. Temperature has increased by about 0.8°C (1.5°F) over that time.

IPCC report

Plots like this raise many questions: What causes climate change? What explains the recent (last 140 years) rise in earth's surface temperature? Is all climate change natural, or do human activities account for some or most of the change? Does climate change happen gradually (so that humanity and ecosystems can evolve without too much stress), or does climate jerk quickly from one distinct "stable" state to another, too rapidly for social and biological systems to adapt? If human impact on climate is significant, is the overall effect of present activity negative or positive? If the effect is negative, are there any socially and politically plausible options for us to mitigate this impact?

Fortunately, in the last few decades we have developed our scientific understanding of how climate change occurs, and the links between different elements in climate. We have also developed many techniques to determine past climate conditions over even longer periods. These help us to understand whether the measured rate of recent climate change is unusual, or has happened before. However, the problem is very complex, and there are still many uncertainties as we try to answer the preceding questions.

We start by considering some different ways in which climate might change. These include natural cycles, sudden changes from one stable state to another (“regime shifts”), and human (“anthropogenic”) impacts.

Different ways in which Earth systems can change:


The earth's climate experiences many cycles. Some of the most prominent are:

  • Daily, or "diurnal"
  • Yearly, or "annual"
  • Interannual, like El Niño
  • Decadal, like the Pacific-North American, North Atlantic, and Arctic Oscillations
  • Millennial and longer period, including 'ice age' cycles

The most easily understood cycles are the shorter changes such as daily and seasonal cycles. These depend on well-known variations in the distribution of solar input that are a function of the earth’s rotation and orbit. At the other end of the spectrum, the well-known “ice age” cycles also appear to be modulated by very long period (20,000-100,000 year) cycles in the earth’s orbit (the so-called “Milankovitch cycles”).

The least well understood cycles have time scales of a few years (like El Niño) up to a couple of thousand years. These processes are difficult to model because they depend on highly complex feedback mechanisms between solar input, the ocean, the atmosphere, the cryosphere (sea ice, ice sheets, and glaciers), and possibly also the role of the earth's terrestrial and aquatic flora and fauna (including humans).

Sudden regime shifts

We gain some understanding of climate variability over the last half million years by looking at concentration of dust and gases, precipitation rates, and inferred temperatures in Greenland and Antarctic ice cores.


This record from the Vostok ice core in Antarctica (Petit et al., 1999; Nature) shows past CO2 concentrations and atmospheric temperatures going back over 400,000 years. Temperatures are derived from measured ratios of hydrogen and oxygen isotopes in the ice, which depend on temperature at the time the water evaporated from the ocean and at the time of precipitation. (This is called “isotope fractionation”.) The large excursions from glacial (cold) to interglacial (warm) are evident, covering 4 ice age cycles. Typically, the temperature rise from glacial to interglacial happens rapidly, while the decline back into an ice age happens more slowly, although with some smaller but still significant rapid temperature jumps occurring along the way.

Detailed analyses of these types of data make it clear that long-period orbital variability, which changes the annual mean and seasonal distributions of solar radiation (Milankovitch cycling), especially in polar regions, is the driver for the ice age cycles. However, the greenhouse effect plays a significant role in modifying the earth's climate response from what would be expected just from the known changes in incoming radiation.

We are now well in an interglacial period. In fact, the present interglacial (called the “Holocene”) has already lasted longer than any of the 3 previous interglacials. Based just on orbital forcing, and following previous ice age cycles, we are due for a long period of gradual cooling to the next ice age. However, note that past CO2 concentrations have never been as high as they are at present (currently above 389 ppm), which changes the basic conditions under which our climate operates and makes it difficult to base future predictions on past behavior.

You might think that long period variability (the 20,000-100,000 year ice age cycles) are the least important cycles of climate change from a societal viewpoint. After all, humanity should have a very long time to adapt to these changes. However, detailed analyses of ocean sediment and ice core records suggest that pre-industrial global climate might have jerked from glacial to interglacial conditions very quickly, perhaps in a few decades, that is, a single human lifetime. Could this happen again? As we noted, the expected climate behavior at the end of an interglacial warm period like the current Holocene is a generally slow cooling. But plots like the one above also suggest that, once some threshold in climate forcing is exceeded, climate can jump quite quickly to a new warmer state. We don’t know if this will happen. An optimistic view of the above plot is that the earth’s climate somehow keeps itself regulated in a fairly narrow temperature range. Perhaps it is simply not possible to force the climate to a significantly warmer state, and the end result of a cascade of processes might be to initiate the end of the Holocene through a long period of gentle cooling. However, because the CO2 concentrations are now significantly higher than at any time through the last 4 ice age cycles, we must consider the possibility that greenhouse warming will lead to a different response. All we really know is that climate response to quite subtle and slow changes in forcing associated with long-period orbital changes includes rapid jumps to new states, and that the rise in greenhouse gas concentrations is equivalent to a forcing that is neither subtle nor slow.

Changes brought about by perturbations to natural cycle

As will be shown later on, the present warming trend can only be explained by ocean and atmospheric retention of more heat (solar radiation) than is released to space by back-radiation. This measured warming is consistent with what we expect from changes due to the measured increase in greenhouse gases over the past century and, since no other credible mechanism has been suggested, it is attributed to the greenhouse effect.

Perturbations to natural cycles can occur in many ways. Major volcanic eruptions spew dust high into the atmosphere. The eruption of Mount Pinatubo (Philippines) in 1991 led to lower global temperatures (about a 0.5°C drop) for the next two years. Ash layers found in Antarctic ice cores and elsewhere point to several eruptions over the last several millennia that dwarf Pinatubo. There is also evidence of episodic massive outflows of lava lasting thousands of years that would also profoundly change the chemical balance of the atmosphere. Many scientists also believe that asteroids have provided many of the past climate shocks that led to mass extinctions, while other scientists have hypothesized that gamma-ray bursts from nearby supernovas could impact climate through ionization of the earth’s atmosphere.

In later pages we will present evidence that a more predictable perturbation is responsible for the recent climate change, the contribution of human activity to the buildup of greenhouse gases in the atmosphere.