AGW is a theory. Models are created to test the theory. The models and raw data must be peer reviewed.
The observations are that there has been negligible warming for the last 17 years... it does not support the hypothesis.
This is where critical thinkers adjust the theory. The limited scale, variables and natural adjustments make it highly difficult.
http://www.nature.com/ngeo/journal/v7/n3/full/ngeo2098.html
Despite continued growth in atmospheric levels of greenhouse gases, global mean surface and tropospheric temperatures have shown slower warming since 1998 than previously 1, 2, 3, 4, 5 . Possible explanations for the slow-down include internal climate variability 3, 4, 6, 7 , external cooling influences 1, 2, 4, 8, 9, 10, 11 and observational errors 12, 13 . Several recent modelling studies have examined the contribution of early twenty-first-century volcanic eruptions 1, 2, 4, 8 to the muted surface warming. Here we present a detailed analysis of the impact of recent volcanic forcing on tropospheric temperature, based on observations as well as climate model simulations. We identify statistically significant correlations between observations of stratospheric aerosol optical depth and satellite-based estimates of both tropospheric temperature and short-wave fluxes at the top of the atmosphere. We show that climate model simulations without the effects of early twenty-first-century volcanic eruptions overestimate the tropospheric warming observed since 1998. In two simulations with more realistic volcanic influences following the 1991 Pinatubo eruption, differences between simulated and observed tropospheric temperature trends over the period 1998 to 2012 are up to 15% smaller, with large uncertainties in the magnitude of the effect. To reduce these uncertainties, better observations of eruption-specific properties of volcanic aerosols are needed, as well as improved representation of these eruption-specific properties in climate model simulations.
http://web.mit.edu/newsoffice/2014/how-the-ocean-reins-in-global-warming.html
In recent years, a hot topic in climate science has arisen over the fact that climate models vary widely in their representation of ocean heat uptake. The oceans in some models absorb more or less heat in high-latitude regions such as the North Atlantic and Southern Ocean; some store heat at different depths. According to two new papers published in Geophysical Research Letters, those details matter a great deal to the predictions of global warming over the coming centuries.
There are other even greater sources for the differences in climate predictions across models, such as cloud responses to greenhouse gases, Kostov notes, “but all aspects of the climate system are important, and we have to take into account the role of the ocean in order to improve our predictions for future warming.”
These results led the MIT group to conclude that models must better represent the AMOC and its future changes, based on real-world measurements that extend over time and geographical location. Unfortunately, there is not a long record of observations in the AMOC, thanks to the enormous technical difficulty of probing the ocean’s deep layers. However, Kostov notes his excitement that a few large-scale oceanographic projects, including U.K. RAPID and U.S. CLIVAR, have started to continuously monitor how the circulation varies with depth in an effort to fill this scientific void.
Ultimately, the study critiques how the field uses observations in estimating the climate’s sensitivity to greenhouse gases. "A common way to calculate climate sensitivity simply combines recent observations of global surface temperature changes, heat uptake, and greenhouse gas forcing," says Armour, "which misses the details of how heat is getting into the ocean. One implication is that we can’t actually estimate long-term warming from present-day observations unless we take into account how the pattern of ocean heat uptake might change with time."