Global carbon accounting relies heavily on what satellites can see. Remote sensing tools excel at measuring the greenness, height, and volume of live, aboveground vegetation. However, this reliance creates a systematic "aboveground bias." By equating visible canopy dynamics with total ecosystem carbon, current frameworks leave massive carbon pools completely unaccounted for.
To build accurate climate models, researchers are working to uncover these hidden pools. Understanding why this bias occurs—and how to fix it—requires looking at the ecosystem through three distinct blind spots: vertical, horizontal, and structural.
The Vertical Blind Spot: Roots and Soil
When we look at a forest, a grassland, or a wetland, more than half of the carbon story is hidden beneath our feet. Because belowground components are invisible to traditional remote sensing, they are routinely omitted from regional budgets.
Belowground Biomass
Root systems and rhizomes represent a massive fraction of total plant mass. In dryland ecosystems, for instance, root systems can account for nearly one-fifth of total tree biomass, as detailed in a large-scale sub-continental study [Sub-continental-scale carbon stocks of individual trees in African drylands]. In coastal and wetland systems, Otani et al. demonstrate that focusing exclusively on aboveground reed biomass completely misses the long-term carbon sequestered in extensive, buried rhizome networks. Even in agricultural monitoring, Ge et al. emphasize that root and shoot developments are often decoupled, meaning surface observations fail to reflect belowground realities.
The Soil Carbon Gap
Soil organic matter is often the largest carbon pool in an ecosystem, yet it is notoriously difficult to measure from space. As Zhu points out, even advanced multi-scale LiDAR frameworks remain blind to soil carbon stocks.
When models omit soil and litter, they miscalculate the true climate impact of land changes. For example, Liu et al. found that excluding soil carbon led to a significant underestimation of the long-term sequestration potential of commercial forests. Conversely, Fan et al. demonstrated that omitting soil and litter dynamics masks how severe forest disturbances (like fires and droughts) degrade the broader Siberian carbon sink, even when surface greening suggests recovery.
[ Satellite Sensors ] ── Sees only ──> Live Canopy & Aboveground Biomass
─────────────────────────────────────────────────────────────────────────
[ Invisible Earth ] ── Misses ────> Litter & Coarse Woody Debris
Rhizomes & Deep Root Systems
Soil Organic Matter Pool
The Horizontal Blind Spot: Fragmented Landscapes
The aboveground bias doesn't just cut off vertically; it also filters out entire landscapes geographically. Standard carbon monitoring protocols are optimized for dense, contiguous, closed-canopy forests. This creates a severe geographic bias against open or fragmented ecosystems.
Trees Outside Forests (TOF)
Brandt et al. demonstrated that traditional forest inventories systematically ignore trees scattered across agricultural mosaics, savannas, and urban areas. By mapping individual trees at a national scale using high-resolution imagery, the authors revealed a massive, hidden carbon stock that standard protocols miss. Similarly, Rasmussen et al. found that semi-arid farmlands support substantial woody vegetation that is completely overlooked by large-scale assessments focused exclusively on humid forests.
Arid and Semi-Arid Biomes
Open ecosystems like savannas, shrublands, and grasslands are frequently undervalued. Feng et al. argue that prioritizing closed-canopy forests leads to a systematic underestimation of these open biomes, which actually drive massive regional biomass fluctuations and dictate global carbon dynamics.
The Structural Blind Spot: Dead Wood and Non-Leaf Decay
Ecosystems are dynamic cycles of growth and decay, but aboveground-centric models often treat carbon fixation as a static endpoint. They frequently overlook the distinct pathways through which different plant components release carbon back into the atmosphere.
Dead Wood and Litter
Live trees are only part of the equation. Tong et al. and Cheng et al. highlight that failing to account for litter, understory shrubs, and dead wood creates an incomplete picture of forest carbon pools, skewing our understanding of net climate benefits.
Organ-Specific Decomposition
Carbon models frequently prioritize leaf decomposition because leaves are highly dynamic and visible. However, non-leaf components like stems and sheaths often make up the bulk of a plant's biomass. Tao et al. discovered that "standing litter" (dead plant matter still suspended in the air) contributes up to 70% of the emissions produced by ground litter, yet its unique decomposition rate is routinely ignored in carbon cycle projections.
Building Holistic Carbon Budgets
Why does closing these gaps matter? When we ignore belowground pools, fragmented landscapes, and non-leaf decomposition, our carbon accounting becomes an exercise in looking for keys only under the lamppost.
To fix this, researchers are pushing for integrated frameworks. For example, Xie emphasizes the need to couple vegetation productivity directly with litter decomposition and ecosystem respiration to evaluate true sequestration capacity. Without a full-pool accounting approach—one that tracks carbon from the deepest roots to standing dead wood—our strategies for ecological restoration and climate mitigation will remain fundamentally incomplete.
Key Foundational Work
Standing Biomass and Inorganic Nutrients Distribution for a Camellia japonica Stand at Mt. Cheon-gwan(Janghueng-gun, Jeonnam) (2005) — contributed 2 concept elements to this direction.
A framework for the cross-sectoral integration of multi-model impact projections: land use decisions under climate impacts uncertainties (2015) — contributed 1 concept elements to this direction.
Warmer and Wetter Soil Stimulates Assimilation More than Respiration in Rainfed Agricultural Ecosystem on the China Loess Plateau: The Role of Partial Plastic Film Mulching Tillage (2015) — contributed 1 concept elements to this direction.