East Asia’s Forests Are Losing Their Appetite for Carbon Dioxide

For decades, scientists have counted on the world’s vegetation to quietly absorb a large share of the carbon dioxide that humanity pumps into the atmosphere. Every year, plants pull somewhere between 112 and 169 petagrams of carbon out of the air through photosynthesis, a vast and largely invisible service that slows the buildup of greenhouse gases. But a new study of East Asian ecosystems suggests that this service is not guaranteed to continue at the same pace. An international team of researchers, publishing in the journal Earth System Dynamics, has found that the sensitivity of ecosystem productivity to rising atmospheric carbon dioxide has been steadily weakening across East Asia since the late 1990s, and that the reasons for the decline differ sharply from one sub-region to the next.

The research, led by Yun-Soo Na and Sang-Wook Yeh of Ewha Womans University in Seoul, together with Jong-Seong Kug of Seoul National University and Young-Min Yang of Jeonbuk National University, set out to answer a deceptively simple question: how strongly does vegetation productivity in East Asia respond to the year-to-year growth rate of atmospheric carbon dioxide? The answer matters because East Asia is a region of extraordinary climatic and ecological diversity, spanning the monsoon-drenched forests and croplands of China, Korea, and Japan as well as the drier grasslands and barren expanses of the interior. Understanding how these contrasting landscapes respond to a changing atmosphere is essential for projecting the future of the global carbon cycle.

To quantify that response, the team drew on twelve Dynamic Global Vegetation Models from the TRENDY version 13 ensemble, the same suite of models that underpins the annual Global Carbon Budget, covering the period from 1959 to 2023. Their methodology was meticulous. They first removed the long-term trends from both gross primary productivity, the measure of carbon absorbed by plants through photosynthesis, and the atmospheric carbon dioxide growth rate, using a nonlinear smoothing technique called LOWESS. They then performed linear regressions within 20-year moving windows, defining the slope of each regression as the sensitivity of productivity to carbon dioxide growth. The approach captures the combined ecosystem response to carbon dioxide, climate variability, and land-use change together, rather than isolating the fertilizing effect of carbon dioxide alone.

The first striking result concerned net ecosystem production, the net balance between carbon taken up by photosynthesis and carbon released by respiration. Eleven of the twelve vegetation models simulated a negative sensitivity, meaning that when atmospheric carbon dioxide grew faster, East Asian ecosystems tended to take up less carbon, and vice versa. After applying statistical screening based on the most recent two decades of data, the researchers retained nine models for detailed analysis. The spatial pattern of productivity sensitivity that emerged was anything but uniform: central-eastern China, Korea, and Japan behaved very differently from the northern and western parts of the region, prompting the team to divide East Asia into a monsoon region and a non-monsoon region for all subsequent analysis.

When the researchers tracked sensitivity through time, a clear and troubling trajectory appeared. In the non-monsoon region, net ecosystem production sensitivity remained negative throughout the record, briefly approaching zero in the late 1990s before turning more negative again. In the monsoon region, sensitivity started out weakly positive but gradually declined, crossing into negative territory around the 1990s. Gross primary productivity told the same story with even greater amplitude: positive sensitivity in the monsoon region weakened and flipped negative in the late 2000s, while the non-monsoon region drifted back toward strongly negative values. In both regions, the inverse relationship between carbon dioxide growth and carbon uptake has been intensifying, suggesting that faster atmospheric carbon dioxide accumulation now goes hand in hand with weaker photosynthetic carbon absorption across East Asia.

Why is this happening? The answer, it turns out, depends on where you look. In the non-monsoon region, the decline in productivity sensitivity tracked soil moisture sensitivity almost perfectly, with a correlation coefficient of 0.85 and soil moisture explaining 72 percent of the variance in productivity sensitivity. This is a region dominated by grasslands and barren terrain, where water is chronically scarce. When the carbon dioxide-associated response of soil moisture weakened, vegetation activity weakened with it. The analysis of transpiration added a further layer of detail: reduced stomatal conductance accompanied by reduced transpiration did not translate into meaningful gains in soil water in this dry environment, so the plants gained no compensating benefit and photosynthesis remained constrained.

The monsoon region presented a strikingly different picture. There, where rainfall is comparatively abundant, soil moisture sensitivity actually increased over time, moving from negative to positive values, while productivity sensitivity moved in the opposite direction, yielding a significant negative correlation of minus 0.65. The real culprit was light. Photosynthetically active radiation, the portion of sunlight that plants use for photosynthesis, showed a distinct declining trend in the monsoon region, and productivity sensitivity fell in step with it, with a positive correlation of 0.72 and radiation explaining 52 percent of the variance. In other words, even when water was plentiful, dimming skies limited the plants’ capacity to convert rising carbon dioxide into additional growth. The researchers also found that this radiation control strengthened over time: when they split the record into early and late periods, the explanatory power of radiation in the monsoon region rose from 21 percent to 59 percent.

Vegetation structure appears to underpin much of this regional contrast. Using satellite-based land cover data from the MODIS instrument, the team showed that the non-monsoon region is dominated by grasslands and barren surfaces, while the monsoon region is dominated by croplands and woody savannas, ecosystems with fundamentally higher baseline productivity and different water-use strategies. These structural differences, combined with contrasting hydroclimatic regimes, help explain why the same global driver, the rising atmospheric carbon dioxide growth rate, produces such divergent local responses. The findings were robust to methodological choices: the researchers repeated their analysis with moving windows of 10, 15, 25, and 30 years and with nonparametric Spearman correlations, and the main patterns held.

The implications reach far beyond East Asia. Global studies have already suggested that the carbon dioxide fertilization effect on vegetation has been weakening in recent decades, and that the positive contribution of carbon dioxide to productivity trends has roughly halved since the 2000s. This study adds a crucial regional dimension, showing that the weakening is governed by different limiting factors in different places, water in the dry interior, light under the monsoon skies. If the sensitivity of ecosystems to carbon dioxide growth continues to decline, the terrestrial carbon sink may absorb proportionally less of future emissions, leaving more carbon dioxide in the atmosphere and accelerating warming. The authors argue that future carbon cycle projections and carbon budget assessments must explicitly account for regional climatic constraints, particularly soil moisture and radiation, rather than treating the land surface as a homogeneous sponge.

There is also a message here for policy. Carbon neutrality strategies, the researchers suggest, should be designed with regional specificity, because the capacity of natural ecosystems to buffer human emissions depends on local climate systems and vegetation characteristics that vary enormously even within a single continent. The study’s authors are careful to note that their sensitivity measure reflects statistical associations within the coupled Earth system rather than isolated causal effects, and that correlation analysis alone cannot fully establish causality. Even so, the convergence of evidence from twelve independent vegetation models, satellite land cover observations, and six decades of atmospheric measurements paints a coherent picture: East Asia’s ecosystems are becoming less responsive to rising carbon dioxide, and the planet’s carbon arithmetic may be shifting as a result. As atmospheric carbon dioxide continues its climb, the question of whether forests, grasslands, and croplands can keep pace is no longer academic, and this study suggests the answer will be written region by region, in soil moisture and sunlight.

Subject of Research: Regional sensitivity of East Asian ecosystem productivity to atmospheric CO2 growth

Article Title: Regional contrasts of ecosystem productivity sensitivity to atmospheric CO2 growth across East Asia

Article References: Regional contrasts of ecosystem productivity sensitivity to atmospheric CO2 growth across East Asia. (n.d.). https://doi.org/10.5194/esd-17-1381-2026

Image Credits: AI Generated

DOI: 10.5194/esd-17-1381-2026

Keywords: carbon cycle, East Asia, gross primary productivity, atmospheric CO2 growth rate, soil moisture, photosynthetically active radiation, monsoon, TRENDY vegetation models, carbon sink, land cover, climate-vegetation feedback, Earth System Dynamics