Forecast
Water and global warming
|
| Forecasts come from models |
A laboratory big enough to represent the atmosphere of the Earth in a realistic way, does not exist. In order to forecast how the climate on Earth will behave in the future, scientists have developed computer models of the atmosphere. In these computer models, the processes that occur in our atmosphere are expressed in mathematical equations. These processes are, for instance, air and ocean dynamics (wind, or ocean currents) or atmospheric chemistry. Unfortunately, it is not possible to describe our entire Earth by equations, so models never give a fully realistic representation of what is happening on our planet. First of all, the capacity of our computers is limited, and secondly, we do not yet understand every process that is relevant to our climate well enough to express it in an equation.
Water vapour plays a big role in the atmosphere and we have to understand this role fully, if we want to describe it in a model. Water vapour is the most important greenhouse gas, but it is not evenly distributed in the atmosphere. Its influence on the climate varies on such small scales that our models, cannot represent the impact correctly as they are limited by computer power.
Our surroundings as box in a global model |
A trip into model calculations....
Modellers divide the atmosphere in boxes. In each box we assume a constant value for each chemical and meteorological variable (for instance relative humidity, wind velocity, radiation) are assumed. This means that within one box these values are evenly distributed; no difference can be distinguished inside this box.
|
Origin: MPI Mainz How far can we go in climate modelling?
Three examples:
A 1200 km x 1200 km
about 14.000 boxes
B 500 km x 500 km
about 57.000 boxes
C 200 km x 200 km
about 500.000 boxes
|
The increasing capacity of computers during the past 20 years has enabled us to use three-dimensional models. In these models, one can look at the whole atmosphere, or at just a part of it. For example, we could look at only Europe, or only the upper or lower part of the atmosphere.
We distinguish two types of global models:
Global Circulation Models (GCM)
In these models, all meteorological parameters like wind velocity, relative humidity, temperature and snow coverage, are calculated by the model itself. These models are not very different from the models that are used to forecast tomorrow's weather, except that they use bigger boxes, so the resolution is much coarser. Weather forecast models usually only consider a part of the atmosphere, for instance only over Europe, while GCMs cover the entire atmosphere. GCMs can also be linked to atmospheric chemistry models.
Chemical Transport Models (CTM)
These models have meteorological values as input, and use these to calculate the transport, chemical reactions and other removal of chemical substances in the atmosphere. The model is cheaper to run because the meteorological values are given to the model and do not have to be calculated. This means that computer power can be used to describe the chemistry in more detail. The figure on the left shows different horizontal resolutions for such models and the corresponding number of boxes.
We can see, that a model with a resolution of 200 x 200 km (this means that the same data are taken for London and Dover, and for Fort Williams and the top of the Ben Nevis) already leaves us with half a million boxes. This model uses enormous amounts of computer capacity. If we consider, that the weather in Fort William may be sunny, whereas on the Ben Nevis there may be snow and thunderstorm, we can get a feeling of just how much trouble it will take to calculate water vapour and cloud formation correctly in global climate forecasts.
|
Then how do we model clouds?
We have seen that a model can not predict every single cloud in the atmosphere because of the coarse resolution. In order to understand clouds in a model, a more precise model performs calculations for small areas. In this process, the modellers try to understand what is the relationship between cloud-related parameters, like cloud cover and transport of chemical compounds in clouds, and the other model values. This process is called parameterisation. In summary, the atmosphere is simplified in a model, which assumes average values for a big area and checks with a more precise model to ensure that the result is still more or less right for a smaller area.
Through these single model studies, the modellers try to make sure that both the complex description (precise model) as well as the simple description (global climate model) are correct. The model results of, for example, air humidity also have to agree with the measured values. Often, however, the model has to be simplified so much, that the result has a big uncertainty.
|
When more greenhouse gases, like carbon dioxide and methane, are emitted into the atmosphere, the global warming makes the water vapour content of the atmosphere rise. This is simply due to the fact, that warmer air can contain more gaseous water, so more water evaporates from the oceans, and not because more water vapour is emitted.
Because of the increasing water vapour concentration, the greenhouse effect is enhanced again. This leads to a self-amplifying effect, which is called a "positive feedback" on the climate system. According to the latest results, a warming of 1.4 to 5.8°C is expected. This corresponds to a radiation density of 3.5 to 4.4 W/m2. For comparison: the radiation density of sunlight on the Earth is about 1370 W/m2. According to this model, human emission of carbon dioxide has made the radiation density on Earth rise by 1.5 W/m2. If the other greenhouse gases methane, nitrous oxide and the CFC´s are taken into account, this number amounts to 2.4 W/m2.
The first climate models have calculated global warming including the positive feedback of water vapour and came up with changes that are now considered too large. In the meantime, it has become clear that higher water vapour concentrations in the atmosphere can also exert a "negative feedback", so a cooling, on climate change. If more water vapour leads to more clouds, then more solar radiation is reflected back into space, without being able to be trapped in the atmosphere by the greenhouse gases. This impact depends strongly on the geographical distribution of clouds, on the altitude at which they form, and even on the size of the droplets. Because of this complexity, the effect still cannot be modelled correctly. If the cloud effect is taken into account, different models (which deal with small-scale processes in different ways) give different results for water feedback, varying between -3 and 3 W/m2. Even measurements of the reflection of clouds by the means of radiation measurement equipment on board of satellites do not agree.
In addition to this, there are processes, that may well be only small-scale, but may have big impacts on the water vapour distribution and therefore on the climate. All together, current models show a mean warming of the atmosphere due to the human emission of greenhouse gases. The compensating processes do not seem strong enough to cancel this warming. The limited knowledge of, among other things, the water feedback effects makes it hard to predict future developments in our climate in a reliable way.
|
Summary
Both the short-term weather forecasts and the long-term climate projections are given by computer models. In these models, the atmosphere is divided into thousands of "boxes", wherein the meteorological and chemical properties are calculated. Because of limited computer capacity, the number of boxes has to be limited, so the global models have a very coarse resolution. This resolution, together with a limited insight in the complex feedback effects that water has on the climate, the water impact on human-induced climate change is hard to simulate in a model. Even though most models predict a temperature increase on Earth of 1.4 to 5.8°C, the uncertainties are still very high.
|
|
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
|
