You can tune climate models to match history (hindcasting) you can also tune them to predict any future you wish, however predicting the real climate is not that easy, for a start climate like weather is a chaos system, we can predict weather for a few days accurately after that it becomes more uncertain, the same with climate we can predict for a year or two, after that anything can happen.
Not one single climate model predicted the 15yr pause we are having, so I for one would not place any faith in them.
http://wattsupwiththat.files.wordpress.com/2013/10/cmip5-90-models-global-tsfc-vs-obs1.jpg
Models are the basis of all modern science. Physics, for example, is almost entirely built on mathematical models of reality that are tested against observation and experiment. Newton's law of gravitation is a mathematical model of a phenomenon we call gravity that does a good job of matching our observations for a wide range of, although not every, situation.
The question is, if you develop a model of a physical system, how do you validate it? There are two ways, both of which are intertwined. You can develop a model, then perform experiments to see how well you model predicted the outcome. In the case of climate models, that means taking conditions now and using them to generate probable outcomes in the future. The second way is to look at data you've already taken, and see how well the model matches what actually happened.
If the model does a good job in replicating existing data, then your confidence in the model increases. So climate models should do both. They should be able to replicate proxy data and what we think the climate was like in the past, and should therefore also be able to predict the climate of the future. All of this is, of course, not absolute but to within reasonable margins of error.
Additional: Well, we can look at the fossil record and geology. The solubility of iron in water is a function of its oxygen concentration. Therefore we can start to see iron deposits in geological layers whose concentration tells us what the O2 levels were. We can tell the concentration of CO2, for example, by examining the deposition of limestone and calcite which depends on the concentration of CO2 absorbed in water. We see the formation of the first blue-green algae in the fossil record about 1.6 billion years ago. They lowered the level of CO2, because their photosynthesis released oxygen, which resulted in a cooling of the planet and increased glaciation, which we also see in the geological record. So there are lots of ways we can get a fairly good idea of what the climate was like 1.6 billion years ago!
Not really, no. Past climate models are generally based on observed data - ice cores, tree rings, ocean sediments and the like. You can tell what the overall climate is like from these, just as you can get an idea of climate by looking at local plant life and the like. Future climate models are not based on observing the effects of future climate (we obviously cannot do that), but rather on predicting based on what we know of current and past climates.
So short answer, no. Models of past climates are based on observing the effects of climate, not predicting climate.
We learn about the dynamics of climate change through studying the history of climate change. Predicting is just a guess, but history is written in stone (or, in this case, ice). Scientists studying ice cores from the Antarctic can see climate change year-to-year for at least the last 10,000 years!
Global warming deniers point out that climate has always changed on earth, it changes naturally. And that's true, but only by studying past climate change can we see that it's never changed nearly as quickly as it has over the last century, and in the last decade the change has accelerated even more.
Using hindcasting with climate models is an empirical way to determine senesitivity of climate to certain parameters, like the Sun, aerosols and carbon dioxide.
it's a way of verifying models. same is done with weather forecasts. look up hindcasting.
Is this the same as predicting the future in some respects?
Models are the basis of all modern science. Physics, for example, is almost entirely built on mathematical models of reality that are tested against observation and experiment. Newton's law of gravitation is a mathematical model of a phenomenon we call gravity that does a good job of matching our observations for a wide range of, although not every, situation.
The question is, if you develop a model of a physical system, how do you validate it? There are two ways, both of which are intertwined. You can develop a model, then perform experiments to see how well you model predicted the outcome. In the case of climate models, that means taking conditions now and using them to generate probable outcomes in the future. The second way is to look at data you've already taken, and see how well the model matches what actually happened.
If the model does a good job in replicating existing data, then your confidence in the model increases. So climate models should do both. They should be able to replicate proxy data and what we think the climate was like in the past, and should therefore also be able to predict the climate of the future. All of this is, of course, not absolute but to within reasonable margins of error.
Additional: Well, we can look at the fossil record and geology. The solubility of iron in water is a function of its oxygen concentration. Therefore we can start to see iron deposits in geological layers whose concentration tells us what the O2 levels were. We can tell the concentration of CO2, for example, by examining the deposition of limestone and calcite which depends on the concentration of CO2 absorbed in water. We see the formation of the first blue-green algae in the fossil record about 1.6 billion years ago. They lowered the level of CO2, because their photosynthesis released oxygen, which resulted in a cooling of the planet and increased glaciation, which we also see in the geological record. So there are lots of ways we can get a fairly good idea of what the climate was like 1.6 billion years ago!
Not really, no. Past climate models are generally based on observed data - ice cores, tree rings, ocean sediments and the like. You can tell what the overall climate is like from these, just as you can get an idea of climate by looking at local plant life and the like. Future climate models are not based on observing the effects of future climate (we obviously cannot do that), but rather on predicting based on what we know of current and past climates.
So short answer, no. Models of past climates are based on observing the effects of climate, not predicting climate.
We learn about the dynamics of climate change through studying the history of climate change. Predicting is just a guess, but history is written in stone (or, in this case, ice). Scientists studying ice cores from the Antarctic can see climate change year-to-year for at least the last 10,000 years!
Global warming deniers point out that climate has always changed on earth, it changes naturally. And that's true, but only by studying past climate change can we see that it's never changed nearly as quickly as it has over the last century, and in the last decade the change has accelerated even more.
Using hindcasting with climate models is an empirical way to determine senesitivity of climate to certain parameters, like the Sun, aerosols and carbon dioxide.
it's a way of verifying models. same is done with weather forecasts. look up hindcasting.
Is this the same as predicting the future in some respects?
