We don't know what we pretend to think we know

Among critics of the anthropogenic global warming hypothesis, one of the central criticisms is that all the concern is based on the results of computer models which are trying to emulate the interaction and net effect of a large number of extremely complex, dynamic systems. Many of those component systems are poorly understood and, indeed, little studied. AGW could be a real issue but we simply do not know. The model forecasts are nice window-dressing but we also know they are exquisitely sensitive to both unconscious bias in their design as well as flawed assumptions about the operands of the constituent complex systems.

We know the climate is always changing. We know there are multiple non-human contributing causes. We know there are multiple human contributing causes. What we don't know with any degree of accuracy is how all the complex systems interact with one another and what their net effect is. It makes good sense to continue reducing emissions (of all sorts), reducing consumption, etc. But in terms of bankable forecasts in which we can have great confidence and therefore also balance costs and benefits, we simply aren't there yet no matter how loudly the baseless arguments are shouted.

As is illustrated in Climate warming experiment finds unexpected results by Emily Pontecorvo. Pontecorvo is reporting on an experiment which was being conducted to test one of the model input assumptions, that a warmer environment increases the release of CO2 into the environment. The thought was that a warmer environment will accelerate the decomposition of biomass, thus accelerating the release of CO2. That is a reasonably critical component of the forecasts. An error in our understanding of that particular process bodes ill for forecast accuracy and reliability.

We can't take the results of this experiment at face value as it was halted due to the damage from hurricane Maria, but it is sufficient to raise further doubts about our understanding of the core systems of AGW.

Roe said there are few empirical studies of how tropical forests will respond to climate change. She set out to address this gap in June of 2017, when she and her research team travelled to El Yunque National Forest in Puerto Rico. They landed at a site called TRACE—the Tropical Responses to Altered Climate Experiment.

TRACE is the first-ever long-term warming experiment conducted in a tropical forest. It was established by the US Forest Service in 2016 for research like Roe's. The site consists of three hexagonal plots of land enclosed by a ring of infrared heaters raised four meters above the ground, and three more plots enclosed by fake heaters that are used as the "control" forest.

Roe collected leaves from the plots, dried them out in the lab, and then returned them to the plots randomly. In addition to the native plants, she also included black and green tea, and popsicle sticks to represent woody biomass, to see how different materials would respond to the warming.

The heaters were programmed to continuously heat the plots to four degrees higher than the ambient temperature of the forest. The experiment was supposed to run for a full year, but at the beginning of October, Hurricane Maria swept across the island, destroying the TRACE sites.

Roe was back in Virginia when the storm struck. She had collected samples from the first few months of the experiment, and they were already showing signs of significant decomposition, so she decided to go ahead with the analysis based on what she had. And the results were not what she thought they would be.

"We would expect that microbes tend to work faster, like their metabolisms increase, with warmer temperatures," Roe said. "So we would expect to see an increase of activity of microbes and other decomposers to decompose the litter."

But instead of seeing faster rates of decomposition, Roe observed the warming produced a drying effect in the plots, which slowed decomposition. "What we found is actually it went the other way because moisture was impacted so much," Roe said. Moisture in the litter from the treatment sites was reduced by an average of 38 percent.

Roe pointed out that the increase in frequency and severity of storms in the region could amplify this effect. Hurricane Maria reduced significant portions of the tree canopy in El Yunque, allowing a lot more sunlight to reach the forest floor that can dry up the litter.