midtskogen wrote:You gave a figure, 2C, without specifying its meaning, so I had to make assumptions.
You can always ask for clarification. But either way that is fine I will give my best feedback. I will agree with you when I think your assumptions and intuitions are accurate, or help you to correct them if they are not.
I also want to say that I am not trying to put you down -- you are very intelligent, have more world experience than I do, and I deeply respect your insights and contributions to the discussion. I agree with your motivation for looking for tests and potential falsifiability of these projections. I also encourage you to try your own hand at doing a literature review to see for yourself what the spread of expert opinion on this subject is, how that has emerged, what the uncertainties are, and how these models are constructed and tested.
It's a projection, which looks similar to what we've seen over the past 50 years or so. One which I think is very bold for the Arctic. The 1990-2015 temperature increase looks very similar to the 1920-1935 increase, which was followed by a cooling from 1945-1970. I think it's bold to declare that history wont repeat this time.
You can suggest that history would repeat itself and the Arctic would cool, but think about sources of confidence for and against it. The idea is contradicted by an overwhelming consensus of models, the magnitude of thermal energy input to the system compared to internal variations, and the physics of the ice-albedo feedback. Notice that virtually every climate model agrees with these spatial trends (in particular, the stippling pattern represents more than 2-sigma deviation from internal variability and >90% agreement in the sign of the change across the models).
Remember, in the early 20th century the magnitude of warming due to the greenhouse forcing was about the same magnitude as internal variability. But through the 21st century under moderate to high emissions scenarios, the greenhouse forcing is about an order of magnitude greater. I showed a figure demonstrating this on page 1, and I recommend looking at it again and thinking about what it implies. In my opinion this is one of the most important figures in all of climate science.
The point is, in order to give a 2C figure, we must know what the distribution will be. You indicate that we do. However, would you consider it falsified if the Arctic fails to produce significant warming by 2035?
Depending on the concentration pathway we follow (the figure I showed is for RCP 4.5), absolutely this can be falsifiable. The higher the emissions scenario, the more significant the predicted change, and therefore the greater the confidence in rejecting the model should there be an observation of insignificant warming.
You can also test these projections through other metrics. On page 2 I discussed an example of how observations are consistent with climate model predictions, based on a meteorological definition for change in mild weather. You may notice the spatial distribution closely resembles that of positive or negative impacts of climate change on agriculture.
In other words, this is observational evidence that supports the spatial distribution of change predicted by climate models.
Let's look at real data
Sure! You are correct to conclude that the economic factors, land use change, and agricultural development over this timeframe are more important than the climatological impacts over this degree of warming, and so you should not expect to see obvious decline in production in these data. Simply put, we have been farming more, and grown more efficient in how we farm. So far this more than offsets climate impacts.
If you perform a literature review, you will find that these economic and agricultural growth factors are included in the projections, and when you examine the trends, a business as usual approach does not achieve food security. Food security can only be achieved if climate sensitivity happens to be on the extreme low end of the probability distribution (possible but an unwise basis for policy-making) or by employing strategies for adaptation and/or mitigation, which do exist.
A better approach to testing observable climate effect on crops would be to consider the change in crop yield compared to expectation with and without the climate changes, using agricultural models tuned by experiment. What you will find is that the climate changes are indeed already impacting global yields.
The impacts also depend a lot on the region and on the type of crop. For example:
Would a yield in 15-20 years that is higher than today be sufficient to falsify?
In general, almost but not quite. It depends on the emissions pathway we follow and how much emphasis humanity places on adaptive strategies. Under RCP8.5, 2040 is when we can expect roughly 50% probability of passing 2C. So I would expect to observe these impacts to just begin to dominate at the end of this period if we follow high emissions with no adaptation. I would also expect a more clear negative trend in the tropics -- particularly Africa which is known to be the most vulnerable -- and positive trend in the high latitude. A failure to observe those latitudinal trends would be a more powerful falsification, because globally they are nearly balanced at this warming threshold.
How is this all sitting with you so far? I think you are raising good questions and I hope you continue to think about it and peruse available literature and ask questions if you have them. I hope you will also take time to look at what expert opinion is with regards to how to approach the problem of achieving food security. There are quite a few papers that I have read over the years and I'd be happy to provide you with suggested reading material on both sides of the spectrum.