Eight of the many reasons such studies are fatally flawed:

First point. Increased CO2 leads to higher crop yields. Even the pessimists quote increases from 8-25% from CO2 fertilization by itself. I've seen studies with values ranging from 15-45%. Higher CO2 = higher plant growth.

Second point. These results are based on models such as CERES-Maize, using predicted temperature/rainfall changes. But if you look at the results, you'll see something odd. Across the entire world, every single nation shows declining production from climate change. Every single one. Climate change predicts some areas will warm, some will cool. Some will get slightly more rain, some less. But such models assume that every spot in the planet is already at optimum grain growing conditions. If more rain is predicted, they use a negative function for flood, rot, etc damage. If less rain is predicted, they reduce the yield also -- but don't add in increases due to less rain losses. The assumption is laughable on its face. Even if these changes DID mean less production globally, there would still be many, many areas for which a little more(less) rain or a little higher(lower) temperature would be BETTER for agriculture. In short, the assumption that every square inch of the planet is already "optimum" for growing food is tripe.

Third point. Most studies don't attempt to dispute that, in at least some nations (the US, Canada, Argentina, Australia, etc) that total grain production will, when CO2 fertilization is considered, be **higher**. Their negative effects are only when losses in Africa and other third-world nations are added in. And, as I've demonstrated, those values are highly unrealistic.

Fourth point. The entire premise of this modelling is based on another model, a GCM. Is that reliable? We already have substantial evidence to demonstrate that climat sensitivity is being grossly ovestated. Climate models have continually downscaled their sensitivity over time. Hansen/GISS in the 1980s was claiming sensitivities of 6C+. Today, the IPCC says 3C..and other research says less than half that.

Fifth point. The models predict future conditions, but assume year-2000 mitigation technologies (even lower, for third-world nations). In 2080, will farmers be better equipped to deal with, say, a 10% loss in rainfall than they are today? Even the simple expedient of a few large-scale desalinator plants can essentially eliminate aridity effects in most nations.

Sixth point. Model fails entirely to consider increases in crop yields due to biotechnology, or even simple plant husbandry (breeding for better yields). We've seen astronomical gains from this in just the last 100 years alone. There's no reason to believe they'll stop overnight.

Seventh point. The simple transfer of First-world farming methods to the Third World would increase global food production by several times what these predicted losses are. Currently, the First world produces 2-3 tons food/acre/year. Most of the Third World about half that...simply because they don't use our high-yield strains and farming methods.

Eigth point. The amount of cultivated land in most First-World nations runs between 5 and 30%. In the US, its only 20%...and has been declining for many years. We need less and less land each year to feed more and more people...despite 150 years of climate change. But even ignoring all the other objections, the simple expedient of farming a slightly larger percentage would more than compensate for all the hypothesized losses.

How many more points do you need?

In the 1800s, predictions were rife that the world would be starving by 1900. In 1960, a flurry of scientists and government bodies claimed with total assurance that food production would soon decline and, by the year 2000, billions would be starving to death.

Today, we see the same fearmongering. Deja vu all over again. Memories are short.