Forests in the United States that generate the carbon offsets bought by companies including BP and Microsoft are on fire as summer blazes rage in North America. Source: Financial Times
Corporate net-zero emission pledges rely on such projects to compensate for the carbon dioxide generated by companies that are unable to make sufficient cuts to their actual emissions.
In principle each offset represents a ton of carbon that has been permanently removed from the atmosphere or avoided. Offsets generated by projects that plant or protect trees, which absorb carbon, are among the most popular.
But forestry projects are vulnerable to wildfires, drought and disease—permanent threats that are being exacerbated by global warming.
“We’ve bought forest offsets that are now burning,” Elizabeth Willmott, Microsoft’s carbon program manager, told attendees at an event hosted by Carbon180, a non-profit organization that focuses on carbon removal.
In Washington and Oregon, at least two forestry projects used by companies including BP and Microsoft are ablaze.
Fires ignited by lightning in July swept through the Colville Indian Reservation in Washington, prompting evacuations and damaging a project that BP buys offsets from.
With some areas still burning, the extent of the damage is unknown.
The offset programs carry a “buffer” of credits that are not sold to any companies but are available to cover any shortfalls resulting from problems with a project.
The Colville project has generated more than 14 million credits since 2016, about 5 million of which have been used by buyers, according to data compiled by the Berkeley Carbon Trading Project.
BP alone bought 13 million offsets from the Colville project in a transaction valued at more than $100 million, according to the Washington Department of Natural Resources.
In December, BP also bought a majority stake in the carbon offset developer Finite Carbon, which was involved in the Colville project. This week the oil group said it was in “close communication with our counterparties as events continue to unfold to determine the extent of the impact.”
In Oregon, the Bootleg fire near Klamath Falls has caused widespread devastation, and burnt through part of a project, Klamath East, that provided Microsoft with offsets.
“Since the Bootleg fire started, we have had all hands on deck fighting the fire. It will be weeks before we can get in to assess damage,” said Patti Case, a spokesperson for the forestry group Green Diamond Resource Company, which developed the scheme covering Klamath.
Overall, Klamath East has generated nearly 800,000 offsets since 2017, about 140,000 of which have been used by buyers.
Microsoft bought a combined 240,000 offsets from Klamath East and Klamath West last year.
The tech giant was assessing how the disaster might affect how it buys offsets in future, Willmott said at the Carbon180 event. “We don’t want this to force us to pull out of investing in nature-based solutions,” she said. Instead, buyers must “get really smart about what the risks are.”
In California, the McFarland Fire is roughly 33 miles away from the Gabrych Ranch offsetting project, a distance that the Sydney-based investment manager New Forests, which developed the scheme, said was “not a threat.”
Given the risks from fire and drought, forestry offsetting schemes contributed about 10% to 20% of the credits they generate to the “buffer pool.”
Critics of the unregulated offsetting system have warned that buffer pools may be too small to compensate for the damage done by major fires.
“The concern is that the pool is not large enough to cover the increased risk of [the carbon benefits being reversed] with climate change over the full set of participating projects,” said Barbara Haya, research fellow at the University of California, Berkeley.
Danny Cullenward, policy director at CarbonPlan, a non-profit organization that has previously conducted analysis on soil carbon offsets with funding from Microsoft, said quantifying the carbon impacts of a specific fire, and therefore how many offsets to cancel, was complicated.
