Firstly, let me state that I accept the scientific concensus on AGW.
The scientific method, in simple terms, involves developing a theory that fits observations. That theory is then either reinforced, modified, or rejected on the basis of futher observations. The theory of AGW is accepted by the scientific community because it is the theory that best fits the observational data. It does not mean that AGW cannot be wrong or that it cannot be disproved, merely that it is currently the theory with the best explanatory power. Within any field of science there will be disagreements, data that seems contradictory, and bits that are not yet well defined, well-understood, or need to be examined in more detail. AGW, like any good theory, should allow us to make predictions that can either be verified or falsified by experiment.
AGW explains a number of observations. It explains why we are observing a CO2 increase, why this has occured since the 1950s, and why the rate of change is as fast as it is - the burning of fossil fuels by humans. One of the predictions of AGW is that, if human activities are responsible, we should observe certain ratios of carbon isotopes in our atmosphere. This has been confirmed experimentally. When an analysis of the expected warming effect of this CO2 is made, it fits temperature data reasonably well. The application of the theory predicts other effects, such as changes in the pH of oceans, changes in hurricane frequency and strength, etc which also have been observed. Whether or not one agrees with AGW, one cannot disagree with the ability of that theory to provide a explanation and reasonably good fit to the available observational data.
In science, fitting the observations is all that matters. To date, no alternative theory has been proposed that has been shown to match CO2 data and temperature data. These 'alternatives' have not made predictions we can test. Therefore, science must go with AGW until a viable alternative theory is formulated, presented, discussed, shown to match the observations, and make predictions. What many people forget is that alternative theories for the same observational data do not simply have to match one another in explanatory power. They must ALSO show why the other alternatives are not valid. In the case of an alternative to AGW, that theory must also explain such things as:
1. Why is the CO2 level rising? If it is a natural response to warming, then what is the mechanism? If it is something to do with plants, then what is changing in the plant-world to cause it? If not plants, then what else is causing it? What data can you present to validate that conclusion?
2. Why are human emissions of CO2 not having the warming effect predicted by AGW? If the amount we are pumping is simply too small to have an effect, where is the mathematical model showing this? How does this model fit the observational isotope data?
and so on. An alternative to AGW must show why AGW is wrong.