Simulation of global and hemispheric temperature variations and signal detection studies using neural networks

Abstract

The concept of neural network models is a statistical strategy which can be used if a superposition of any forcing mechanisms leads to any effects and if a sufficient related observational data base is available. In comparison to multiple regression analysis, the main advantages are that neural network models are an appropriate tool also in the case of non-linear cause-effect relations and that interactions of the forcing mechanisms are allowed. In comparison to more sophisticated methods like general circulation models, the main advantage is that details of the physical background, like feedbacks, can be unknown. Neural networks learn from observations which reflect feedbacks implicity. The disadvantage, of course, is that the physical background is neglected. In addition, the results prove to be sensitively dependent from the network architecture like the number of hidden neurons or the initialisation of learning parameters. We used a supervised learning backpropagationnetworkwith three neuron layers, as well as an unsupervised learning counterpropagation network. This latter network architecture consists of a so-called Kohonen layer, where the internal features of the data set in question are "learned", which is followed by a so-called Grossberg layer, where the "learned" internal representations are adjusted to the observed time series to be simulated. Both concepts (backpropagation, counterpropagation) are tested in respect to their ability to simulate the observed global as well as hemispheric mean surface air temperature annual variations 1874-1993 if parameter time series of the following forcing mechanisms are incorporated: equivalent CO2 concentrations, tropospheric sulfate aerosol concentrations (both anthropogenic), volcanism, solar activity, and ENSO (all natural). It arises that in this way up to 83 % of the observed temperature variance can be explained, significantly more than by multiple regression analysis. The implication of the North Atlantic Oscillation does not improve these results. On a global average, the greenhouse gas signal so far is assessed to be 0.9-1.3 K (warming), the sulfate signal 0.2-0.4 K (cooling), results which are in close similarity to the GCM findings published in the recent IPCC Report. The related signals of the natural forcing mechanisms considered cover amplitudes of 0.1-0.3 K.

Keywords

• neuron layers
• kohonen layer
• grossberg layer
• neutral networks
• regression analysis
• neuronale netzwerkmodelle
• regressionsanalyse