To avoid the most catastrophic impacts of climate change—and achieve a net-zero world—we need to not only reduce our emissions, but also remove residual carbon dioxide (CO2) from the air.

That part is clear. What’s less clear is how best to do it.

Enter carbon dioxide removal, or CDR, which refers to various methods of capturing carbon from the atmosphere and storing it for decades to millennia on land, in the ocean or in geological formations.

Scientists say CDR technologies are essential because the excess greenhouse gases, mostly CO2, we’ve already emitted into the atmosphere over the last 250 years need to eventually be removed to keep the global temperature increase between 1.5 and 2 degrees Celsius (the goal of the 2015 Paris Climate Agreement).

“There is no net zero without CDR: it puts the ‘net’ in net zero,” said Eve Tamme, managing director of Climate Principles, a climate policy advisory group based in the Netherlands. “CDR…needs to happen in addition to reducing emissions. It’s an ‘and and’ concept.”

CDR is also “unavoidable” in the short-term for emissions linked to hard-to-abate sectors like steel and cement, which are harder to decarbonize, United Nations scientists said last year in a major report. Vast amounts of CO2 will have to be captured in the medium and long-term as well to return the atmosphere to climate-safe levels.

Carbon removal methods vary in terms of how CO2 is extracted from the atmosphere, where the carbon is stored and for how long. They also vary on resource requirements, costs, hazards, development stages and price tags.

Some involve nudging nature to do what it has always done (nature-based solutions); others are novel technologies that have yet to be proven at scale (engineered solutions); and others are a mix of the two.

“Because it takes time to develop and scale-up, we need to be working on CDR technologies now so they can contribute by mid-century,” Howard Herzog, senior research engineer at the MIT Energy Initiative, said in June at a carbon removal summit in Switzerland organized by Climeworks, a major company in this field.

Speaking at the same event, Julio Friedmann, chief scientist at Carbon Direct, a carbon management firm, said he foresees these various CDR methods “competing for turf.”

“I am hoping we can all compete positively because we all know we have to grow everything at incredible speed and scale to get to where we need to go,” he said.

Today, the world emits about 36.6 billion metric tons of CO2 per year from fossil fuels and removes only about 2 billion metric tons annually, according to a report published early this year by scientists in Europe and the United States.

Out of that, 99% comes via conventional ways of removing carbon from the air, such as tree planting and restoration.

Only 0.1% comes from engineered technologies, precisely because they’re so early in their development. These solutions will need to scale up 30 times from current levels by 2030 and 1,300 times by 2050 to keep global warming in check, according to the report.

Despite their current minute contribution, this novel tech is causing the most controversy among various groups.

Environmentalists worry that engineered solutions meant to capture carbon from the air will throw a lifeline to fossil fuel companies and disincentivize decarbonization.

A key U.N. panel criticized carbon removal technologies in a draft guidance earlier this year, describing them as “technologically and economically unproven.” The language, which faced backlash from the carbon removal sector, raised questions about differing opinions within the U.N. agency and cast doubt on the tech’s potential.

To further complicate matters, some oil companies are using captured carbon to extract more oil, which blunts its climate benefits.

“I think it’s totally a fair concern to have that people are going to do this poorly,” Erin Burns, executive director of climate NGO Carbon180, said at the summit in Switzerland.

For his part, Friedmann said concerns that engineered carbon removal solutions could disincentivize cutting emissions have unfairly slowed down the development of the tech.

Another challenge in the CDR debate is how to account for its contribution.

Designing rules to properly monitor, report and verify that CO2 has actually been removed by a particular method (and is not counted twice, for example) is crucial considering the many differences between the various methods. The exercise has already led to tensions in the European Union.

Even as CDR tries to take off, Herzog reminded participants at the Switzerland summit of the world’s primary goal in the climate change fight: reduce emissions first.

“The best way to remove CO2 from the air is not release it into the air in the first place.”