Читайте также:
|
Henson, Robert (2006) The rough guide to climate change. Penguin Books Ltd, 80 Strand, London WC2R ORL, pp. 217 – 232
Steady improvements in weather forecasting models
To focus on tracking the atmosphere
To undergo a rapid transformation
To be incorporated in ever-growing global climate models
The model’s complexity
Atmospheric and environmental variables
To depict long-term climate change
To attain equilibrium
To sift through the results, calculate yearly average and other statistics and produce graphics
To bolster confidence in the results
To project the climate of the future
MODELING THE FUTURE CLIMATE
Reproducing the atmosphere with computer models has been a key activity in computer science ever since the field was born. At first the emphasis was on weather prediction. By the 1960s, forecasters were using 3D models of atmospheric flow, or general circulation models, to gaze into the future of weather patterns several days out. Since then, breakthroughs in weather and computer science, together with Moor’s Law (the maxim that computer processor speed doubles about every eighteen months), have allowed for steady improvements in weather forecasting models and thus in forecast skill. Revelle and Seuss might have been happy to have a solid three-day forecast; now, ten-day outlooks are commonplace.
Models focusing on climate long lagged behind their counterparts in the weather field, and with good reason. The most obvious weather changes emerge from the interplay among a fairly limited set of ingredients in the atmosphere: pressure, temperature, moisture, wind and clouds. Weather is also shaped by other factors in the surrounding environment, such as ocean temperature, vegetation and sea ice. But these latter factors don’t change much over the course of a few days, so a weather model can focus on tracking the atmosphere while keeping the rest of the environment constant.
Climate models don’t have that luxury. As the weeks roll into months, vegetation thrives and decays. Sea ice comes and goes. Greenhouse gases accumulate. The ocean gradually absorbs heat from our warming atmosphere, sometimes releasing it is in giant pulses during El Niño events. In the very long term, even the topography of Earth changes. All of these variations feed into the atmosphere and influence weather – often so subtly that the effect isn’t obvious until it’s been playing out for years. For all of these reasons, it’s extremely difficult to go beyond a weather model’s focus on the atmosphere and depict Earth’s whole environment accurately in a global climate model.
Difficult, but not impossible. Climate modeling has undergone a rapid transformation in the last twenty years. As recently as the 1980s, the most sophisticated climate models portrayed the atmosphere and the land surface – and that was about it. Gradually, scientists have forged models that depict other parts of the Earth system. Many of these are incorporated in ever-growing global climate models. The best of the bunch now include land, ocean, sea ice, vegetation and the atmosphere, all interacting in a fairly realistic way. Global models are far from ideal; as skeptics often point out, there are still major uncertainties about climate processes that no model can resolve. However, the skill of recent global models in replicating twentieth-century climate, as we’ll see below, is one sign that the models are doing many things right.
Climate model operates in a parallel universe of sorts, one in which time runs far more quickly than in real life. Each model advances in time steps that move the simulated climate forward by anywhere from five minutes to a half hour, depending on the model’s complexity. At each time step, the model calls on formulas derived from the laws of physics to compute how all the atmospheric and environmental variables have changed since the last time step – the amount of sunlight hitting each spot on Earth, the convergence of winds, the formation and dissipation of clouds, and so forth. Each step only takes a few seconds to process, but it can still take a few weeks of dedicated time on a supercomputer to depict a century of climate. The output from all this number-crunching is saved in various formats along the way. After the processing is done, scientists can go back, sift through the results, calculate yearly average and other statistics and produce graphics.
In order to depict long-term climate change, scientists have to provide the model with input on how the environment is changing. One critical variable is fossil-fuel emissions. Modelers rely on two major types of simulations to show how an increase in greenhouse gases affects climate. (In both cases, carbon dioxide is traditionally used as a stand-in for all human-produced greenhouse gases in order to save on computing time and expense, though in recent years the most detailed models have also incorporated methane and other key gases.)
Equilibrium runs typically start off with an immediate, massive injection of carbon dioxide (for instance, enough to bring the airborne amount to twice the pre-industrial level, which is where we’re expected to be by later in this century). The model’s climate gradually responds to this extra CO2 and eventually attains equilibrium at some new (and presumably warmer) state. The idea isn’t to see how long the climate takes to respond – since we’d never add that much CO2 instantaneously in the real world – but where the climate ends up. To save on time and expense, these runs typically use simplified oceans.
Transient runs more closely resemble reality. In these simulations, CO2 is added in smaller increments that resemble the actual amounts added by human activity – typically 1 % per year, compounded year over year. In this case, scientists are interested as much in the process as the outcome. For instance, does the model climate warm smoothly, or does the temperature go up in fits and starts? The various global models have tended to show closer agreement for transient runs than for equilibrium runs – an encouraging sign, since the transient scenario is closer to how greenhouse gases increase in real life.
Of course, you can’t simply add CO2 to a global modal and leave everything else alone. There are also other processes that kick in as carbon dioxide warms Earth. Consider the sea ice that melts as the atmosphere warms up. The loss of that ice means less reflection of sunlight and an amplifying, positive-feedback loop that leads to even more warming. The positive feedback from water vapor also has to be portrayed accurately: the warming produced by the extra CO2 helps evaporate more moisture from the sea, and that water vapor raises global temperature further.
These environmental processes are quite different from the physics that drives day-to-day weather, and they can’t be added to a climate model in an easy or simple way. The original global climate models of the 1960s and 1970s couldn’t hope to account for these processes in a realistic fashion. But as computing power has improved, the models have grown steadily in sophistication and detail. They’ve incorporated a growing number of sub-models, dealing with each of the major climate-influencing elements of what scientists call the Earth system. When an atmospheric model is yoked to one or more of these sub-models, the new hybrid is called a coupled model (even when it has three or more components).
In a simple coupling, a sub-model runs side by side with a global model and passes it information – a one-way dialogue, as it were. For instance, a global model might be fed information on airborne nitrogen levels from a chemistry model, but the resulting changes in climate wouldn’t feed back into nitrogen production, as they would in real life. More realistic are integrated couplings, which allow two-way exchanges of information between model and sub-model. However, these can take years to develop.
Models also grow in the level of detail they provide. Model resolution – the fitness of the spacing between grid points – has gradually increased over the years. Today, some cutting-edge global models are beginning to resemble weather models. For instance, a team from Japan’s Meteorological Research Institute has begun to stimulate the frequency of tropical cyclones in a doubled-CO2 world using a model with a resolution on the order of 20 km, with 60 vertical layers.
Especially when the focus is on a weather or climate feature with restricted geographic extent, such as hurricanes, it makes sense to limit the model’s extra-sharp precision to the areas where it’s most needed. This can be done using a nested grid, whereby the globe as a whole is simulated at a lower resolution and the areas of interest are covered at higher resolutions.
Similarly, when studying how climate might evolve in a fixed area, such as parts of a continent, global model are often coupled to regional climate models that have extra detail on topography. This allows rainfall and other important features to be depicted with more precision.
Whatever its design, every global model has to be calibrated and tested against some known climate regime in order to tell if it’s functioning properly. The easiest way to do this is through a control run with greenhouse gases increasing at their twentieth-century pace. By simulating the last century of climate, scientists can see if a model accurately depicts the average global temperature as well as regional and seasonal patterns, ocean circulations and a host of other important climate features. Once it’s passed this test, a model can go on to project the climate of the future – or the distant past, the goal of paleoclimate modeling.
Another safety check is that the world’s leading models are regularly checked against each other through such activities as the Program for Climate Model Diagnosis and Intercomparison, conducted at the US Lawrence Livermore National Laboratory. When the models agree, that bolsters confidence in the results; if one model deviates markedly form the others, it’s a flag that more work is needed.
- Why is weather prediction considered to a key activity in computer science ever since the field was born?
- Why have scientists to provide the model with input on how the environment is changing?
- What have various global models tended to show?
The EU's contribution to shaping the future global climate change regime
The Kyoto Protocol
The ultimate goal
To stabilize the concentrations of greenhouse gases in the atmosphere
To avoid dire consequences from human interference with the climate system
To address the financing of actions by developing countries
A key priority
To build an effective global carbon market
The present commitments under the Kyoto Protocol to the United Nations Framework Convention on Climate Change (UNFCCC) are only the first step in addressing the climate change threat. The ultimate goal of the UNFCCC is to stabilise the concentrations of greenhouse gases in the atmosphere at a level that avoids dire consequences from human interference with the climate system.
The EU's objective is to ensure that global average temperature does not increase more than 2°C above pre-industrial levels. To avoid this, global emissions of greenhouse gases must peak before 2020 and then greatly decrease by 2050.
The necessary cuts in global emissions can be achieved only if all countries contribute their fair share according to their responsibility and capacity. And even if the temperature increase stays below 2°C there will still be a need for significant adaptation efforts by all countries.
International negotiations are under way to conclude a global agreement at the UN climate change conference in Copenhagen (December 2009) for the period after 2012. The successful conclusion of these negotiations is a key priority for the EU.
This Communication 'Towards a comprehensive climate change agreement in Copenhagen' sets out concrete proposals to achieve this goal. It addresses three key challenges:
- Targets by developed countries and appropriate actions by developing countries;
- The need to address the financing of actions by developing countries (both to mitigate greenhouse gas emissions and adapt to climate change);
- The need to build an effective global carbon market.
Дата добавления: 2015-11-16; просмотров: 61 | Нарушение авторских прав
| <== предыдущая страница | | | следующая страница ==> |
| Climate Change | | | The future of fossil fuels |