In the race to scale durable carbon dioxide removal (CDR), most attention goes to engineered solutions—direct air capture, enhanced weathering, or BECCS. But a new 2026 study published in Proceedings of the National Academy of Sciences (PNAS) shows that one of the most extreme landscapes on Earth—the Taklamakan Desert—is being transformed into a measurable carbon sink.
The paper, “Human-induced biospheric carbon sink: Impact from the Taklamakan Afforestation Project”, provides direct observational evidence that large-scale ecological engineering can enhance carbon sequestration even in hyperarid environments.
For the carbon removal community, this is more than a greening story. It’s a data-driven case study in land-based CO₂ removal at planetary scale.
From “Biological Void” to measurable carbon sink
The Taklamakan Desert covers ~337,000 km² in western China and is one of the driest deserts in the world. Annual rainfall is typically below 50 mm, and more than 95% of the surface consists of shifting sand dunes.
Historically, it has been considered ecologically inert.
But over the past few decades, China’s Three-North Shelterbelt Program—sometimes called the “Green Great Wall”—has established forested belts around the desert’s margins to:
- Reduce sandstorms
- Stabilize dunes
- Protect infrastructure
- Combat desertification
The new study shows this intervention has done something else: it has increased net carbon uptake.
The measurement stack: satellites + biosphere modeling
What makes this study especially compelling for CDR professionals is its multi-layered measurement approach. Researchers combined:
- Vegetation cover (NDVI) from MODIS
- Photosynthetic activity (SIF) from TROPOMI
- Atmospheric CO₂ columns from Orbiting Carbon Observatory-2
- Carbon flux modeling using CarbonTracker and the MiCASA biosphere model
This combination allowed researchers to link vegetation growth directly to measurable atmospheric CO₂ drawdown—seasonally and over long-term trends.
This is critical: many land-based CDR claims struggle with MRV (measurement, reporting, verification). Here, we see atmospheric confirmation.
Seasonal carbon removal: A 3 ppm swing
The desert now shows strong seasonal dynamics.
During the wet season (July–September):
- Precipitation rises to ~16.3 mm/month (2.5× dry-season levels)
- Photosynthetic activity nearly doubles
- Vegetation index increases significantly
- Atmospheric CO₂ concentrations drop by ~3 ppm relative to dry season
In carbon terms:
- Net ecosystem exchange (NEE) becomes negative
- The biosphere absorbs more CO₂ than it emits
During the dry season (December–February), the system flips, and CO₂ levels rise.
This seasonal oscillation demonstrates active biospheric carbon cycling—not passive greening.
Long-term trend: structural shift toward a carbon sink
The most important finding is not seasonal variability—it’s trend.
Over ~25 years:
- Vegetation cover (NDVI) shows a statistically significant upward trend
- Photosynthetic activity (SIF) is increasing
- Net ecosystem exchange (NEE) shows a strengthening negative trend
In other words:
The region is gradually absorbing more CO₂ each year.
Notably, precipitation does not show a significant long-term increase. The greening signal aligns instead with afforestation efforts—indicating human-driven ecological transformation.
This matters because it separates climate variability from engineered land intervention.
Quantifying the removal potential
The study estimates that:
- The existing shelterbelt removes ~1.74 tCO₂ per hectare per year.
- If expanded across the entire Taklamakan (~337,000 km²), potential removal could reach ~58.7 million metric tons CO₂ per year.
- Hypothetically extending such afforestation across China’s land area (~9.6 million km²) could remove ~1.67 gigatons CO₂ per year—about 14% of China’s annual emissions.
These numbers are theoretical upper bounds, not policy prescriptions. Water constraints, biodiversity tradeoffs, and land use priorities all matter.
But the scale signal is unmistakable.
Why this is different from other “Great Green Wall” Projects
Afforestation initiatives in arid regions have often struggled—most notably the Sahel’s Great Green Wall.
The Taklamakan case differs in three ways:
- Multidecadal consistency (since 1978)
- Centralized governance and funding
- Integrated infrastructure + ecological strategy
The result appears to be the first large-scale example of a desert margin transitioning into a measurable atmospheric carbon sink.
Implications for the CO₂ removal sector
This study reframes how we think about drylands.
Deserts occupy more than 6% of Earth’s surface and have largely been excluded from CDR modeling. Yet semiarid ecosystems already drive much of the variability in the global land carbon sink.
Key takeaways for CDR strategists:
- They are highly responsive to precipitation pulses and human intervention.
- Satellite-based SIF + column CO₂ measurements can validate biospheric sequestration at scale.
- The carbon cycle is not fixed. With sustained intervention, regional sinks can be strengthened.
- Even modest increases in rainfall helped survival rates of planted vegetation in this case. Water constraints remain the limiting factor.
A broader perspective on nature-based CDR
The Taklamakan example doesn’t mean all deserts should be forested. Large-scale land transformation carries ecological, hydrological, and social risks.
But it does demonstrate something profound:
Human-led restoration can convert extreme landscapes into functioning carbon sinks—with measurable atmospheric impact.
For a world pursuing gigaton-scale carbon removal, this expands the solution space.
The desert is no longer just a symbol of ecological loss.
It may also be part of the carbon solution.