Climate scientists and science reporters are buzzing about a new graph showing data from the National Snow and Ice Data Center, which gives a unexpectedly bleak portrait of current sea ice area at the poles.

The conversation around the graph has been heated at times, and gives those with a casual interest in science a window into how controversial data gets analyzed and presented. While the science you learned in school might seem mostly fixed, the science of today is still a nerdy battleground.

The graph shows the amount and spread of ice on the surface of the ocean, called global “sea ice area” by scientists, dipping far lower than the established trends would have predicted. For most of 2016, you can see that global sea ice area was low, but near the established trends and closely mimicking their patterns. In the last few months however, it’s dropped off and away from these benchmarks.

This morning, one meteorologist rebutted the graph saying that the startling numbers were caused by a broken sensor in the Southern Hemisphere. That cleared things up for a moment, before others pointed out that the sensor had been repaired five months ago and he deleted the tweet.

Later, scientists and reporters expressed cautious confirmation that the graph was real and based on easily verifiable NSIDC data (data corroborated by Japan’s Aerospace Exploration Agency, which also tracks polar ice).

While the NSIDC has yet to comment on whether the graph accurately reflects their published data, they expressed to The Verge in an email that representing Antarctic and Arctic data in the same graph is not necessarily very useful:

“The combined number, while easy to derive from our online posted data, is not useful as an analysis tool or indicator of climate trends. Looking at each region’s ice extent trends and its processes separately provides more insight into how and why ice extent is changing. Sea ice in the Arctic is governed by somewhat different processes than the sea ice around Antarctica, and the very different geography of the two poles plays a large role. Sea ice in the Arctic exists in a small ocean surrounded by land masses, with greater input of dust, aerosols, and soot than in the Southern Hemisphere. Sea ice in the Southern Hemisphere fringes an ice-covered continent, Antarctica, surrounded by open oceans. While both regions are affected by air, wind, and ocean, the systems and their patterns are inherently very different. Moreover, at any point in time, the two poles are in opposite seasons, and so a combined number would conflate summer and winter trends, or spring and autumn trends, for the two regions.”

The problem scientists have to face here isn’t whether the data is real, but whether this is an appropriate way to represent it. A combined graph might not help us understand specifically what is happening in the two separate environments, but why is the data suddenly so shocking? The NSIDC said it was reaching out to an affiliate scientist at NASA to get a better answer for that.

In the meantime, the NSIDC’s graphs of Antarctic and Arctic regions separately also look pretty bleak.

The graph was originally tweeted by Zack Labe, a PhD student in UC Irvine’s department of earth system sciences, though it was created by a member of the Arctic sea ice forum. Zack also happens to have been my college roommate! I spoke with him briefly about why the tweet went viral and caused a storm of debate and how complicated these conversations can get.

This interview has been edited for clarity and length.

Where did this chart come from and where did the data behind it come from?

The data is available from the National Snow and Ice Data Center. Sea ice area is derived from sea ice extent (total region ice covered) and sea ice concentration (fraction of sea water covered by ice) using this satellite data from the NSIDC.

The figure itself is not mine, [but] I do tons of these types of plots and the data looked the same as mine.

So why is this causing such a hubbub on Twitter? What’s so alarming about this representation?

It’s global sea ice area — meaning both the Arctic and Antarctic. We typically associate recent low sea ice cover (from climate change and natural variability) in the Arctic, but not the Antarctic. In this case, both the Arctic and Antarctic are at record lows for the date.

Can you tell us a little bit about the conversations you’re having on Twitter right now?

I am mostly having a lot of questions about the data itself. Where is it from? How is it collected? Can we trust it? Fortunately, the NSIDC has an excellent website with information on their satellites and methods.

You tweet about data like this a lot, do you often have these conversations? Do you interact often with people who don’t even believe climate change is real? Does data seem to help with those conversations?

Science (climate change) communication is a challenge because it is clearly not working yet. Something isn’t clicking. So I’ve tried to find ways of sharing (at times) complex data in visualizing-pleasing graphics. My goal is to connect to people the science that is going on behind the scenes without any alarmism. Therefore, they can look at the data and also make an educated decision themselves.

So you have a consistent style and tone for your tweets?

I think so. I rarely refer to anything as caused directly by climate change. Earth’s climate / weather is very variable over short- and long-time scales, but there is now also this anthropogenic footprint acting to essentially nudge some of the climate system. My tweets / messages focus on data analysis in observations and models. I stay away from any alarmism or discussion on policy / activism.

Do people mostly respond courteously? Do you have productive conversations around these?

It depends. I routinely get fairly hateful comments that science is bogus — to which I do not respond. But some people have genuine questions asking about the data themselves. In that case, I always try to help.

What’s it like to have a tweet blow up like this one? How’s your morning been?

Stressful, haha. Given there was skepticism about the data, I wanted to make sure there were no issues with the science / information being presented.

These charts are based on official data, so why does the conversation get so complicated?

I think it is a good thing that people question data, particularly in cases of such abnormal statistics. Without questioning science, we can not learn from our mistakes / understanding. However, the issues arise when there is miscommunication about data and its true interpretation.

The graphs and their source data will surely be interrogated closely by scientists in days and weeks to come. If there’s anything to be gleaned from the current conversation, it’s that data representation of complex phenomenon can confuse and cause bickering even among people who really understand the principles. For the everyday observer, things are even harder, so the fact that these small bits of in-fighting and back-and-forth are happening in public (on Twitter) is really a service. You get to watch scientific debate unfold in real time, which not only gives you a better idea of what to believe but also lets you in on the questions that scientists and science reporters use to interrogate data all the time. A little bickering is a good thing.