Permafrost is a central feature of the Arctic terrestrial environment. It covers almost the whole of the Arctic, whether the ground contains bedrock, gravel, silt or soil – and it is perennially (but not permanently) frozen. It is also full of carbon, containing the roots, stems and leaves of plants, and animal remains that have been trapped for thousands of years.

With Arctic air temperatures rising at roughly double the global average, permafrost temperatures are also going up – it is already beginning to disappear at its southern fringes.

There are many possible consequences. The land can slump, causing houses to tilt and roads to crack. Ponds and lakes, once hemmed in by frozen ground, can drain and disappear. Microbes can resume their decay of the organic matter trapped in the soil and release carbon dioxide and methane – greenhouse gases – into the atmosphere, adding to the rate of global warming.

In advance of the International Conference on Permafrost, which takes place in Potsdam, Germany, next week, Arctic Deeply spoke with Antoni Lewkowicz, president of the International Permafrost Association.

Arctic Deeply: What is the relationship between permafrost temperature and thawing?

Antoni Lewkowicz: Permafrost is a very good thermometer for climate. Where the climate is cold, the permafrost is cold. We can see over the last 30 years or more that permafrost has been warming. But when its temperature gets close to zero, permafrost becomes a very poor thermometer. At that point, the excess energy from the atmosphere going into the ground is being used to melt any ice that’s in the permafrost.

You see the same thing when you defrost a turkey. If you stick a temperature logger into it – and I’ve done this – after you take it out of the freezer, the temperature gets close to 0C (32F) and then it sits there for tens of hours before suddenly passing through the freezing point. Permafrost does the same thing. The temperature will rise, sit there and then go through.

Arctic Deeply: Is it possible to figure out when that transition will happen?

Lewkowicz: Predicting when it will flip from the frozen state to the unfrozen state is not easy.

There is very little uniformity with permafrost. There is variability in the vegetation on the surface; spatial variability in the climate and hydrology; and variability in the ground – sand, silt, clay, gravel and bedrock. Then you can also have different levels of ground ice – or no ice at all. All that variability makes it harder for us to accurately predict the responses at the landscape scale. We would see the most rapid responses in the places that have the least ground ice. Bedrock will go through 0C (32F) and you would never know.

The global climate models (GCMs) don’t do a good job of predicting permafrost change. They’re not tuned to predict where the permafrost is. Based on field research, we know that some of the models are wildly inaccurate. We’re seeing better and better models created, but I think we still have a long way to go.

Arctic Deeply: There is a lot of carbon stored in permafrost. What happens as it thaws?

Lewkowicz: We are gaining knowledge very rapidly on that question. If we thaw out permafrost, we will have some kind of impact on the amount of carbon dioxide and methane in the atmosphere. But we didn’t really realize it was an issue until 2009. It turns out that there is twice as much carbon stored in permafrost in those top 2m (6.5ft) of soil as there is in the atmosphere. I’m sure there will be more surprises down the road.

Arctic Deeply: Where are the gaps in the science – especially if we are to keep warming well below 2C (35.6F)?

Lewkowicz: We need to develop better ways of detecting where permafrost change is taking place and we need better models to predict what will happen. The good news is that the most explosive ideas about carbon release have been more or less disproven. But we still don’t really have a good handle on where thawing permafrost will become wet and where it becomes dry. That is an important question. Wet areas will produce methane and where it becomes dry we’ll get carbon dioxide. And that is an important difference. [Methane is about 25-30 times more potent than carbon dioxide as a heat-trapping gas.] We can’t afford to wait.

We need to know where permafrost change is happening now, so that means better remote sensing tools. We also need to be able to make better predictions for the future, and that means better modeling tools.

Arctic Deeply: Are there policies that need to be considered?

Lewkowicz: It comes to the policy of science. We need to make sure that the funding we give goes to the people who can make the appropriate predictions about the rates of change over the next 50 years. If we are to keep global warming to 1.5C (34.7F), what is the percentage contribution that permafrost can make to that? The current numbers appear to be 10 percent. But that is equivalent to an awful a lot of cars. Ten percent, on the one hand, doesn’t sound like much, until you think about what people will have to do [to further reduce their greenhouse gas emissions]. There is no policy for the landscape, it doesn’t care. We’re the ones who need to take what the landscape is going to do into consideration.

This interview was edited and condensed for clarity and length.