27.10.03 · earth-science / soil-science-pedology

Soil carbon cycling and the rhizosphere microbiome: mycorrhizae, biochar, and the largest terrestrial carbon sink

shipped3 tiersLean: none

Anchor (Master): Dokuchaev 1883; Jenny 1941; Waksman 1927; Schlesinger 1977 Annu. Rev. Ecol. Syst. 8:51; Parton et al. 1987 Soil Sci. Soc. Am. J. 51:1173; Post et al. 1982 Nature 298:156; Lehmann 2006 Mitig. Adapt. Strateg. Glob. Change 11:403; Tarnocai 2009 Global Biogeochem. Cycles 23:GB2027; Smith 2008 Phil. Trans. R. Soc. B 363:789; Melillo 2002 Science 298:2173 / 2011 Phil. Trans. R. Soc. B 366; Todd-Brown 2013 Biogeosciences 10:1717

Intuition Beginner

Soil is not just dirt. It is the largest store of carbon on land. Globally, soils hold roughly 1,500 to 2,400 gigatons of carbon, about three times the carbon now in the atmosphere as CO2. Plants pull carbon dioxide out of the air through photosynthesis; when leaves, roots, and whole plants die, that carbon enters the soil as litter. Microbes break the litter down, releasing some carbon back to the air and converting the rest into soil organic matter. A fraction of that matter persists for centuries to millennia. Soil is the planet's slow-moving carbon vault.

The biggest unknown is the feedback to climate. As the climate warms, microbial decomposition speeds up, which could release more CO2 and accelerate warming. At the same time, warmer temperatures and higher CO2 can speed up plant growth, taking up more carbon. Which effect wins? In temperate forests the answer seems to be that decomposition accelerates at first, then settles. In permafrost regions the picture is far more alarming. The frozen soils of Siberia, Alaska, and Canada hold 1,300 to 1,600 gigatons of carbon, nearly as much as the atmosphere. As they thaw, ancient carbon becomes food for microbes.

One proposed tool for putting carbon back into soil is biochar, a form of charcoal made by heating plant material without oxygen. Biochar resists decomposition for centuries, locking carbon in the ground while often improving soil fertility. Mycorrhizal fungi, which live on plant roots and trade soil nutrients for plant sugars, are another key player: they stabilise carbon in soil and shape how ecosystems respond to warming. Soil carbon cycling matters because it is among the largest uncertainties in climate models and one of the few climate levers we can pull through land management.

Visual Beginner

The soil-carbon cycle runs through three pools with very different lifetimes. The active pool turns over in months, the slow pool in decades, and the passive pool in millennia. Only the passive pool matters for long-term climate.

Pool	Lifetime	Main components	Typical share of soil organic carbon
Active	Months to a few years	Fresh litter, dissolved organics, living microbial biomass	~2 to 5%
Slow	Decades to a few centuries	Partly decomposed matter, clay-bound and aggregate-protected fragments	~30 to 50%
Passive	Centuries to millennia	Chemically recalcitrant humus, charred carbon, deep mineral-bound carbon	~45 to 60%

The same litter input sustains all three pools, but the decomposition rate constants differ by roughly four orders of magnitude between active and passive. This spread in lifetimes is what makes the passive pool a climate vault.

Worked example Beginner

Since 1991, researchers at Harvard Forest in Massachusetts have run one of the longest experiments in ecology. Buried cables heat sections of temperate forest soil by 5 degrees C above the surrounding ground, day and night, year after year. The question is simple: does warmed soil release carbon faster, and does that effect persist?

Step 1. In the first decade, soil respiration, the CO2 released by microbes as they decompose organic matter, rose by about 30 percent in the heated plots compared with unheated controls. The warmer soil was breathing out carbon much faster, consistent with the prediction that warming accelerates decomposition.

Step 2. Around year ten, the pattern reversed. The heated plots began respiring less than the controls, not more. By 2007 the warming effect on respiration had nearly vanished. The explanation, reported by Melillo and colleagues in 2011, was that the easily decomposable carbon in the heated plots had been used up. Only chemically tougher, slower-to-decompose carbon remained.

Step 3. The take-home is a phenomenon called acclimation. Short-term warming spikes carbon loss; long-term warming may not, at least in temperate forests, because the labile carbon pool exhausts itself. Whether permafrost regions behave the same way is unknown and far more worrying, because their carbon stocks are vast and still frozen.

What this tells us: soil-carbon feedbacks are not linear. Different soils respond to warming on different timescales, and a decade of data can flip the sign of the answer.

Check your understanding Beginner

Exercise 1 (easy, multiple choice).

Which reservoir holds the most carbon today?

A. The atmosphere, as CO2

B. Soil organic carbon in the top metre

C. The surface ocean, as dissolved CO2

D. Living terrestrial vegetation (plants and animals)

Hint

Recall the budget figures from the Intuition section: soil organic carbon is roughly 1,500 to 2,400 Gt C, while the atmosphere holds about 875 Gt C and living biomass about 550 Gt C.

Answer

Option B. Soil organic carbon in the top metre (~1,500 to 2,400 Gt C) is the largest active terrestrial carbon pool, about three times the atmospheric stock of ~875 Gt C and roughly three to four times the ~550 Gt C stored in living biomass. Option A underestimates the atmosphere by ignoring that soil carbon dwarfs it. Option C undercounts because the surface ocean alone holds less than the deep ocean and exchanges with the atmosphere on yearly timescales. Option D is the smallest of the four.

Exercise 2 (easy, true-false).

Converting native grassland or forest to conventional tillage agriculture typically increases soil organic carbon over the first few decades.

Hint

Think about what tillage does to soil: it breaks up aggregates, exposes protected organic matter to oxygen, and removes the perennial root inputs that built the soil carbon in the first place.

Answer

False. Cultivation typically loses 20 to 30 percent of native SOC within decades and up to 50 to 70 percent in warm, moist regions. Tillage physically disrupts soil aggregates, exposing previously protected organic matter to microbial decomposition; aeration accelerates oxidative breakdown; and removal of perennial root inputs cuts the litter flux that sustained the slow and passive pools. Regenerative practices (cover crops, reduced tillage, manure, agroforestry) can rebuild SOC on the same decadal timescale, but conventional tillage agriculture is a net source of soil carbon, not a sink.

Formal definition Intermediate+

Definition (soil organic carbon and the soil-carbon cycle). Soil organic carbon (SOC) is the carbon held in soil organic matter (SOM): plant and animal residues in various stages of decomposition, microbial biomass, and humic substances. The soil-carbon cycle comprises (i) primary production fixing atmospheric CO2 into plant biomass, (ii) litterfall, root turnover, and rhizodeposition transferring that carbon to the soil, (iii) microbial decomposition releasing CO2 via heterotrophic respiration, and (iv) stabilisation of a residual fraction as SOM through physical protection inside aggregates, chemical binding to mineral surfaces (clay-organic complexes), and chemical recalcitrance of complex aromatic structures. Globally the cycle stores $C_{soil} \approx 1500$ – $2400 Gt C$ in the top metre, against $C_{atm} \approx 875 Gt C$ in the atmosphere ^{[Schlesinger 1977]}.

Definition (CENTURY pools). Following Parton et al. (1987) ^{[Parton 1987]}, soil organic carbon is partitioned into three pools with distinct turnover times: the active pool ( $τ_{a} \sim 1$ – $5 yr$ ; microbial biomass, labile substrates), the slow pool ( $τ_{s} \sim 20$ – $50 yr$ ; physically and biochemically protected SOM), and the passive pool ( $τ_{p} \sim 1000$ – $5000 yr$ ; chemically recalcitrant humus and charred carbon).

Definition (rhizosphere, mycorrhizae, priming). The rhizosphere is the 1 to 2 mm of soil immediately surrounding a living root, hosting up to $1 0^{10}$ microbial cells per gram of soil. Mycorrhizae are symbiotic associations between fungi and plant roots: arbuscular mycorrhizae (AM) penetrate the root cortex and colonise roughly 80 percent of land plant species, supplying phosphorus in exchange for plant-fixed carbon; ectomycorrhizae (ECM) sheath the root surface and dominate temperate and boreal forest trees. The priming effect is the stimulation (or suppression) of old-SOM decomposition caused by fresh carbon inputs such as root exudates.

Counterexamples to common slips Intermediate+

"Soil carbon is permanent." No. Decomposition continually returns SOC to the atmosphere. Conversion of grassland or forest to arable agriculture typically loses 50 to 70 percent of native SOC within decades, by exposing protected organic matter to oxygen and breaking up aggregates.
"Warming always increases soil-carbon loss." Short-term yes, long-term it depends. The Harvard Forest experiment shows acclimation in temperate forests once the labile pool is depleted. Permafrost regions, with vastly larger vulnerable stocks, may not acclimate on policy-relevant timescales.
"Biochar solves climate change." No. Biochar resists decomposition and improves some soils, but feedstock supply and land area limit global potential to roughly 1 Gt C per year, against current emissions of about 10 Gt C per year.
"Mycorrhizae are a minor component." No. AM fungi colonise roughly 80 percent of land plants and dominate phosphorus uptake; ECM fungi control most of the slow-pool carbon in boreal and temperate forest soils.
"Agriculture builds soil carbon." Mostly the reverse. Conventional tillage reduces SOC; only regenerative practices (cover crops, reduced tillage, manure, agroforestry) rebuild it, and slowly.
"The '4 per 1000' target is easily achievable." Contested. Increasing global SOC by 0.4 percent per year would offset a meaningful fraction of fossil-fuel emissions, but requires coordinated land-management change across every agricultural region on Earth.
"Soil microbes are interchangeable." No. Community composition controls decomposition rates, priming direction, and the partitioning of litter carbon between respiration and stabilisation.

Key result: the CENTURY soil-carbon framework Intermediate+

Theorem (linear-reservoir steady-state inventory; Parton et al. 1987). Consider $n$ soil-carbon pools with first-order decomposition kinetics,

\frac{d C _{i}}{d t} = u_{i} - k_{i} C_{i}, i = 1, \dots, n,

where $u_{i} \geq 0$ is the rate of litter input allocated to pool $i$ and $k_{i} > 0$ is the decomposition rate constant. Then (i) each pool equilibrates exponentially to $C_{i}^{*} = u_{i} / k_{i}$ with relaxation time $τ_{i} = 1/ k_{i}$ , (ii) the total steady-state soil carbon is $C^{*} = \sum_{i} u_{i} τ_{i}$ , and (iii) the input-weighted mean residence time of the whole soil profile is $\overset{τ}{ˉ} = C^{*} / \sum_{i} u_{i} = \sum_{i} (u_{i} / u_{tot}) τ_{i}$ .

Proof. Each equation $\dot{C}_{i} = u_{i} - k_{i} C_{i}$ is a linear first-order ODE with constant coefficients and forcing $u_{i}$ . The homogeneous solution is $C_{i}^{(h)} (t) = A_{i} e^{- k_{i} t}$ ; a particular solution is the constant $C_{i}^{(p)} = u_{i} / k_{i}$ . The general solution with initial condition $C_{i} (0)$ is

C_{i} (t) = \frac{u _{i}}{k _{i}} + (C_{i} (0) - \frac{u _{i}}{k _{i}}) e^{- k_{i} t} .

Because $k_{i} > 0$ , the exponential term decays and $C_{i} (t) \to u_{i} / k_{i} = C_{i}^{*}$ as $t \to \infty$ , with relaxation time $τ_{i} = 1/ k_{i}$ . This establishes (i). The linear-stability Jacobian $d \dot{C}_{i} / d C_{i} = - k_{i} < 0$ confirms global asymptotic stability. Summing over pools yields the total inventory $C^{*} = \sum_{i} C_{i}^{*} = \sum_{i} u_{i} τ_{i}$ , which is (ii). Dividing by $u_{tot} = \sum_{i} u_{i}$ gives $\overset{τ}{ˉ} = C^{*} / u_{tot} = \sum_{i} (u_{i} / u_{tot}) τ_{i}$ , the input-weighted average of the pool turnover times, which is (iii). $□$

For the CENTURY-default parameterisation, taking $τ_{a} = 3 yr$ , $τ_{s} = 30 yr$ , $τ_{p} = 2000 yr$ and allocating litter as $u_{a} = 0.55 u_{tot}$ , $u_{s} = 0.40 u_{tot}$ , $u_{p} = 0.05 u_{tot}$ , the input-weighted mean residence time is $\overset{τ}{ˉ} = 0.55 \times 3 + 0.40 \times 30 + 0.05 \times 2000 = 113.65 yr$ . The passive pool, despite receiving only 5 percent of litter input, holds $C_{p}^{*} / C^{*} = (0.05 \times 2000) /113.65 \approx 0.88$ of the steady-state inventory. Slow and passive pools together hold over 95 percent of all soil carbon. The passive pool dominates the climate response: small fractional changes in $k_{p}$ (through warming) move carbon on century-to-millennium timescales, while changes in $k_{a}$ equilibrate within years and matter only for the interannual CO2 flux.

Corollary (warming feedback on steady-state SOC). Suppose decomposition rate constants scale with temperature as $k_{i} (T) = k_{i} (T_{0}) Q_{10}^{(T - T_{0}) /10}$ with $Q_{10} \approx 2$ (rates double per 10 degrees C of warming), while inputs $u_{i}$ are held fixed. Then the steady-state inventory under warming $Δ T$ is

C^{*} (T_{0} + Δ T) = C^{*} (T_{0}) Q_{10}^{- Δ T /10} .

Proof. Substituting $k_{i} (T_{0} + Δ T) = k_{i} (T_{0}) Q_{10}^{Δ T /10}$ into (i) gives $C_{i}^{*} (T_{0} + Δ T) = u_{i} / [k_{i} (T_{0}) Q_{10}^{Δ T /10}] = C_{i}^{*} (T_{0}) Q_{10}^{- Δ T /10}$ . Summing over pools carries the pool-independent factor through. $□$

For $Δ T = 5 °C$ (the Harvard Forest treatment) and $Q_{10} = 2$ , the predicted steady-state loss is $1 - 2^{- 0.5} \approx 0.29$ , or roughly 30 percent of the initial SOC. The observed early-phase respiration increase of about 30 percent is consistent with this transient prediction; the observed late-phase acclimation reflects depletion of the labile pool and is not captured by a single-pool linear model.

Bridge. The linear-reservoir theorem builds toward the permafrost-feedback analysis of 27.07.01, where the same $Q_{10}$ scaling applied to a 1330 to 1580 Gt C stock with no acclimation buffer produces the dominant positive carbon-cycle feedback in CMIP6. The foundational reason the linear reservoir is the right starting point is that decomposition is, to first order, a first-order reaction in substrate concentration, and this is exactly the structure Parton et al. (1987) exploited to make CENTURY the operational backbone of every major land-carbon model for thirty years. The framework generalises to inter-pool transfers (active feeding slow, slow feeding passive) and to microbial-explicit extensions such as MIMICS and CORPSE that resolve the priming effect the original model misses, and appears again in 27.05.04 as the terrestrial counterpart of the Martin-curve compartmental model for the ocean biological pump. The bridge is that both land and ocean carbon storage are governed by the same topological structure: a small number of pools with widely separated turnover times, with the climate response quantified through the rate constants $k_{i}$ and the input partition ${u_{i}}$ .

Exercises Intermediate+

Exercise 1 (easy, multiple choice).

In the three-pool CENTURY model with turnover times $τ_{a} = 3 yr$ , $τ_{s} = 30 yr$ , $τ_{p} = 2000 yr$ and litter fractions $u_{a} = 0.55$ , $u_{s} = 0.40$ , $u_{p} = 0.05$ , which pool holds the largest share of steady-state soil organic carbon?

A. The active pool, because it receives the largest fraction of fresh litter

B. The slow pool, because it combines a moderate input with a moderate turnover time

C. The passive pool, because its long turnover time amplifies even a small input

D. All three pools contribute equally to the steady-state inventory

Hint

The steady-state pool size is $C_{i}^{*} = u_{i} τ_{i}$ . Compute each product and compare.

Answer

Option C. The steady-state sizes are $C_{a}^{*} = 0.55 \times 3 = 1.65$ , $C_{s}^{*} = 0.40 \times 30 = 12$ , $C_{p}^{*} = 0.05 \times 2000 = 100$ (in arbitrary units of input flux). The total is $113.65$ , so the passive-pool share is $100/113.65 \approx 0.88$ , i.e. about 88 percent of the total SOC. Despite receiving only 5 percent of the litter input, the passive pool dominates because its turnover time is two to three orders of magnitude longer than the others. This is the central structural insight of the CENTURY framework: inventory is governed by the product of input and turnover time, so slow pools accumulate disproportionately.

Exercise 3 (medium, multiple choice).

The priming effect refers to which of the following?

A. The initial burst of microbial respiration when fresh litter first contacts soil

B. The stimulation (or suppression) of old soil-organic-matter decomposition caused by inputs of fresh carbon such as root exudates

C. The temperature dependence of microbial respiration, quantified by $Q_{10}$

D. The physical protection of organic matter inside soil aggregates

Hint

The key word is "old": priming concerns how new carbon inputs change the fate of carbon that was already stored in the soil, not the decomposition of the new inputs themselves.

Answer

Option B. The priming effect is the change in the decomposition rate of existing (often old, stabilised) SOM caused by the addition of fresh, energy-rich carbon substrates such as root exudates or dissolved organic carbon. Root exudates provide microbes with an energy source that can stimulate the production of extracellular enzymes, which then also attack older SOM, accelerating its decomposition. The effect can be positive (old SOM decomposes faster) or negative (old SOM is protected by new carbon inputs), depending on soil type, microbial community composition, and nutrient availability. Option A describes the decomposition of the fresh litter itself, not priming. Option C describes the temperature sensitivity. Option D describes a stabilisation mechanism, not priming.

Exercise 4 (medium, numeric).

Using the warming-feedback corollary with $Q_{10} = 2.0$ , what is the ratio $C^{*} (T_{0} + 5 °C) / C^{*} (T_{0})$ of steady-state SOC after 5 degrees C of warming to the initial steady-state? Give your answer as a decimal to three significant figures.

Hint

Apply the formula $C^{*} (T_{0} + Δ T) / C^{*} (T_{0}) = Q_{10}^{- Δ T /10}$ with $Δ T = 5$ .

Answer

$0.707$ . $Q_{10}^{- Δ T /10} = 2^{- 5/10} = 2^{- 0.5} = 1/ 2 \approx 0.7071$ . So warming of 5 degrees C reduces the predicted steady-state SOC by about $1 - 0.707 = 0.293$ , or 29.3 percent, provided inputs stay constant. This matches the observed early-phase respiration increase in the Harvard Forest warming experiment and gives the transient upper bound for the warming feedback. Long-term acclimation of the labile pool reduces the realised loss in temperate forests, but the corollary is the baseline against which the acclimation correction is measured.

Exercise 5 (medium, short answer).

Contrast arbuscular mycorrhizae (AM) and ectomycorrhizae (ECM) on three axes: (i) how they interface with the plant root, (ii) which plant groups they dominate, and (iii) the main nutrient they supply to the host.

Hint

AM fungi physically penetrate the root cortex; ECM fungi form a sheath around the root surface without penetrating. AM dominate most herbaceous plants and tropical trees; ECM dominate temperate and boreal forest trees.

Answer

(i) Root interface. AM fungi form intracellular structures called arbuscules inside the root cortex cells, providing a large surface area for nutrient-for-carbon exchange directly between fungus and plant cell. ECM fungi form a dense fungal sheath (the mantle) around the outside of the root tips and grow between (but not inside) the cortical cells in the Hartig net, mediating exchange extracellularly.

(ii) Plant groups. AM fungi colonise roughly 80 percent of land plant species, including nearly all herbaceous plants, most crops (maize, wheat, rice, soy), and the majority of tropical trees. ECM fungi associate predominantly with temperate and boreal forest trees: pine, spruce, fir, oak, beech, and birch are classical ECM hosts. The global biogeographic pattern is that tropical forests are AM-dominated, boreal forests are ECM-dominated, and temperate forests are mixed.

(iii) Main nutrient supplied. AM fungi are especially effective at supplying phosphorus, which they access via extensive extraradical hyphae that extend centimetres beyond the root depletion zone. ECM fungi supply both nitrogen and phosphorus, and are particularly important in mobilising organic nitrogen from SOM via extracellular enzymes, which is one reason ECM-dominated forests accumulate thicker organic layers. In exchange, both symbioses receive plant-fixed carbon (typically 10 to 20 percent of net photosynthate), which is a major flux into the soil-carbon cycle.

Exercise 6 (medium, multiple choice).

Which statement about biochar is correct?

A. Biochar decomposes on the same timescale as fresh plant litter, within years

B. Biochar is produced by aerobic decomposition of crop residues, as in standard composting

C. Biochar is produced by pyrolysis, heating biomass in the near-absence of oxygen, and resists microbial decomposition for centuries to millennia

D. Biochar already accounts for over 50 percent of global terrestrial carbon storage

Hint

Biochar is a charcoal-like substance. The defining process is pyrolysis: thermal decomposition of biomass under oxygen-limited conditions, which rearranges the carbon into aromatic structures that microbes cannot easily break down.

Answer

Option C. Biochar is produced by pyrolysis of biomass at temperatures typically between 350 and 700 degrees C in the near-absence of oxygen. Under these conditions the carbon rearranges into fused aromatic ring structures that are highly resistant to microbial decomposition, with mean residence times in soil of centuries to millennia. Lehmann et al. (2006) proposed biochar as a carbon-negative technology: the pyrolysis also produces syngas and bio-oil that can be used as fuel, while the biochar itself sequesters carbon. The global technical potential is bounded by feedstock supply to roughly 0.5 to 2 Gt C per year. Option A is wrong because biochar is far more recalcitrant than litter. Option B describes composting (aerobic, mineralises carbon to CO2 within months), the opposite of pyrolysis. Option D overstates the current contribution: biochar accounts for a tiny fraction of terrestrial carbon storage today, though it is being scaled up.

Exercise 7 (hard, numeric).

A permafrost region contains $1500 Gt C$ . Suppose that under a high-emissions scenario, 2 percent of this stock thaws and decomposes over the 21st century (100 years), with no compensating new uptake. What is the average annual flux of carbon to the atmosphere from this permafrost source, in $Gt C yr^{- 1}$ ? How does it compare with current fossil-fuel emissions of about $10 Gt C yr^{- 1}$ ?

Hint

Two percent of 1500 Gt C is the total cumulative release. Divide by 100 years to get the average annual flux. Then take the ratio to 10 Gt C per year.

Answer

Average annual flux: $0.3 Gt C yr^{- 1}$ , or about 3 percent of current fossil-fuel emissions.

Cumulative release over the century: $0.02 \times 1500 = 30 Gt C$ . Spread uniformly over 100 years: $30/100 = 0.3 Gt C yr^{- 1}$ . As a fraction of current fossil-fuel emissions: $0.3/10 = 0.03$ , i.e. about 3 percent. The cumulative release (30 Gt C) is equivalent to about 3 years of current fossil-fuel emissions.

The number looks small, but three caveats make this the dominant carbon-cycle feedback in CMIP6. First, this is a conservative 2 percent scenario; high-warming pathways release 5 to 15 percent this century and continue for millennia. Second, a fraction of the release is methane (CH4) rather than CO2 under anaerobic conditions, and CH4 is roughly 30 times more potent as a greenhouse gas over a century. Third, the flux is purely additive: it sits on top of, not in place of, continued fossil-fuel emissions. The permafrost feedback cannot be turned off by policy, only limited by reducing the warming that drives it.

Exercise 8 (hard, short answer).

The "4 per 1000" initiative, launched at COP21 in Paris in 2015, proposes increasing global soil organic carbon stocks by 0.4 percent per year. Given global SOC of approximately $1500 Gt C$ , compute the implied annual sequestration rate in $Gt C yr^{- 1}$ and compare to current fossil-fuel emissions of about $10 Gt C yr^{- 1}$ . Then discuss two practical obstacles to achieving the target globally.

Hint

For the rate: multiply 1500 Gt C by 0.004 per year. For the obstacles, think about measurement (how do you verify a 0.4 percent change in soil carbon?), permanence (can sequestered carbon be re-released?), and land competition (does soil-carbon sequestration compete with food production?).

Answer

Implied sequestration rate: $1500 \times 0.004 = 6 Gt C yr^{- 1}$ . This is about 60 percent of current fossil-fuel emissions ( $10 Gt C yr^{- 1}$ ) and would, if achieved and sustained, offset a substantial fraction of the anthropogenic perturbation. The figure is the policy motivation for the initiative.

Obstacle 1: measurement, reporting, and verification (MRV). A 0.4 percent annual change in SOC corresponds to roughly 4 to 8 tonnes of carbon per hectare on cropland, depending on initial stocks. Detecting this against natural spatial variability (which can exceed 50 percent within a single field) and interannual variability (driven by weather) requires dense sampling or proximal sensing that is expensive and slow. Standard soil coring detects changes only on timescales of 3 to 5 years; remote-sensing and model-based MRV are improving but remain imperfect. Without credible MRV, soil-carbon credits cannot be traded at scale, and farmers have no incentive to invest in the practices.

Obstacle 2: permanence and reversibility. Carbon built up over a decade of cover cropping can be re-released in a single year of conventional tillage, a drought that kills the cover crop, or a wildfire that combusts the litter layer. The mean residence time of the sequestered carbon is typically in the active and slow pools (years to decades), not the passive pool (centuries to millennia), so the storage is reversible on the same timescale it was built. Climate-driven increases in decomposition under warming can further erode the stocks, so achieving 0.4 percent gains in a stable climate does not guarantee 0.4 percent gains under ongoing climate change.

Obstacle 3 (also major): land competition and saturation. The practices that build soil carbon, cover crops, reduced tillage, manure, biochar, agroforestry, rotational grazing, compete with food production for land, labour, and water. Soils also approach a carbon-storage ceiling set by clay content, climate, and mineralogy, so sequestration rates decline as the soil approaches saturation. The 6 Gt C per year figure assumes globally coordinated adoption on every agricultural hectare, which is not consistent with the institutional and economic realities of farming in most countries.

Advanced results Master

Dokuchaev's founding of pedology and Jenny's CLORPT equation

Dokuchaev (1883) ^{[Dokuchaev 1883]} established that soil is an independent natural body, not merely a geological substrate, formed by the joint action of climate, organisms, relief, parent material, and time. Jenny (1941) ^{[Jenny 1941]} formalised this as the state function $S = f (c l, o, r, p, t)$ , the CLORPT equation, treating soil properties as a deterministic function of five forming factors. The CLORPT framework remains the conceptual backbone of pedology and is the reason soil-carbon stocks can be predicted globally from climate, vegetation, and topography data. Both contributions shifted pedology from descriptive geology to a quantitative predictive science.

Schlesinger's global soil-carbon budget

Schlesinger (1977) ^{[Schlesinger 1977]} compiled the first coherent global budget for terrestrial detrital carbon, fixing the modern estimate of global SOC at roughly 1500 to 2400 Gt C, against 875 Gt C in the atmosphere and about 550 Gt C in living terrestrial biomass. The budget established that soil is the largest active terrestrial carbon pool and that the soil-atmosphere CO2 flux through heterotrophic respiration is roughly 60 Gt C per year, an order of magnitude larger than fossil-fuel emissions. Small fractional changes in the soil reservoir therefore produce carbon-cycle signals comparable to the entire anthropogenic perturbation.

The CENTURY model (Parton et al. 1987)

Parton, Schimel, Cole, and Ojima (1987) ^{[Parton 1987]} built the compartmental soil-carbon model that became the operational core of nearly every land-surface model for thirty years. The model partitions SOC into active, slow, and passive pools with turnover times of years, decades, and millennia, and couples them to litter inputs partitioned by quality (metabolic versus structural) and to environmental scalings of decomposition by temperature, moisture, and soil texture. Calibration against long-term field trials in the Great Plains demonstrated that the three-pool structure captures observed SOC dynamics across cultivation, grazing, and climate gradients. The CENTURY framework is the load-bearing structure behind the CMIP5 and CMIP6 land-carbon projections.

Agricultural soil-carbon loss (Post et al. 1982)

Post and colleagues (1982, Nature) ^{[Post 1982]} documented that conversion of native grassland and forest to arable cultivation loses on average 20 to 30 percent of native SOC and up to 50 to 70 percent in warm, moist regions, within decades, by physical disruption of aggregates, aeration of previously anoxic microsites, and removal of perennial litter inputs. The result quantified the reversibility of soil carbon storage and underpins the modern premise that regenerative practices, cover crops, reduced tillage, manure application, and agroforestry, can rebuild SOC on the same decadal timescale, though never to the full native-state inventory under annual cropping.

Biochar as a carbon-negative technology (Lehmann 2006)

Lehmann and colleagues (2006) ^{[Lehmann 2006]} proposed pyrolysis of biomass to produce biochar, a charcoal-like substance whose aromatic carbon structure resists microbial decomposition for centuries to millennia. Co-applied to soil, biochar both sequesters carbon and often improves water retention, cation exchange capacity, and crop yields on highly weathered tropical soils. The global technical potential is bounded by feedstock supply to roughly 0.5 to 2 Gt C per year, a meaningful but not decisive contribution against current emissions of about 10 Gt C per year. Biochar is one of the few negative-emissions technologies with field-trial validation and commercial deployment at small scale.

The permafrost-carbon feedback (Tarnocai 2009)

Tarnocai and colleagues (2009) ^{[Tarnocai 2009]} produced the authoritative inventory of the northern circumpolar permafrost-carbon stock, revising it upward to 1330 to 1580 Gt C in the top three metres, of which roughly 800 Gt C is in the upper one metre and is the most vulnerable to thaw this century. Thawing exposes organic matter frozen for tens of thousands of years to microbial decomposition, releasing CO2 under aerobic conditions and the more potent CH4 under anaerobic. The resulting permafrost feedback is the largest single positive carbon-cycle feedback identified in CMIP6, and the "compost bomb" instability, in which warming drives respiration faster than vegetation can re-sequester the released carbon, is one of the few candidate tipping points in the climate system.

The Harvard Forest soil-warming experiment (Melillo 2002, 2011)

Melillo and colleagues ^{[Melillo 2002]} established the longest-running soil-warming experiment in the world at Harvard Forest, Massachusetts, starting in 1991. Buried cables maintain heated plots at 5 degrees C above ambient. Soil respiration rose by about 30 percent in the first decade, then declined to control levels by year ten as the labile carbon pool was exhausted, an acclimation reported in Melillo et al. (2011). The finding constrains the form of the warming feedback in temperate forests: short-term carbon loss, long-term equilibration on the labile pool's exhaustion timescale. Whether permafrost soils, with vastly larger vulnerable stocks, exhibit the same acclimation is an open empirical question that the active permafrost-warming experiments are now addressing.

CMIP5 soil-carbon projection uncertainty (Todd-Brown et al. 2013)

Todd-Brown and colleagues (2013) ^{[Todd-Brown 2013]} compared the CMIP5 earth-system models and showed that global SOC stocks by 2100 vary by more than 600 Gt C across models, a spread comparable to the entire cumulative fossil-fuel emissions over the same period. The dominant source of uncertainty is the decomposition-temperature sensitivity, particularly the $Q_{10}$ value assigned to the passive pool. This is the operational reason the soil-carbon feedback is the largest single uncertainty in twenty-first-century carbon-cycle projections, and it motivates the prioritisation of permafrost warming, deep-soil carbon, and long-term decomposition experiments as the empirical frontier.

Synthesis. Soil, viewed as a carbon reservoir, is the foundational reason that terrestrial ecosystems hold roughly three times more carbon below ground than above, and the framework that quantifies this reservoir, the linear-reservoir decomposition of Parton et al. (1987) ^{[Parton 1987]}, is exactly the structure that makes the inventory predictable from inputs and turnover times. The central insight is that the passive pool, with turnover in millennia, dominates the steady-state inventory and therefore the climate response: small fractional changes in $k_{p}$ under warming move carbon on the timescale that matters for atmospheric CO2. Putting these together, the global budget of Schlesinger, the agricultural-loss measurements of Post, the permafrost inventory of Tarnocai, and the warming-experiment data of Melillo all close on the same compartmental framework, and the framework generalises to microbial-explicit extensions (MIMICS, CORPSE) that resolve the priming effect the original CENTURY model misses.

The pattern appears again in 27.05.04 as the terrestrial mirror of the ocean biological pump: both systems are governed by a small number of pools with widely separated turnover times, and the bridge is the recognition that any climate-change response of either reservoir must be quantified through the rate constants $k_{i}$ and the input partition ${u_{i}}$ . This pattern recurs across every coupled carbon-cycle model, from CMIP5 to CMIP6, and identifies the soil-microbial temperature sensitivity as the single largest empirical uncertainty in the twenty-first-century carbon budget.

Full proof set Master

Proposition 1 (linear-reservoir steady-state inventory). Restated. For $n$ first-order pools $\dot{C}_{i} = u_{i} - k_{i} C_{i}$ with $u_{i} \geq 0$ and $k_{i} > 0$ , the steady-state inventory is $C^{*} = \sum_{i} u_{i} / k_{i}$ and the input-weighted mean residence time is $\overset{τ}{ˉ} = C^{*} / u_{tot}$ .

Proof. Each equation $\dot{C}_{i} = u_{i} - k_{i} C_{i}$ is a linear first-order ODE. The equilibrium is found by setting $\dot{C}_{i} = 0$ , which gives $C_{i}^{*} = u_{i} / k_{i}$ . Linear stability: the Jacobian $d \dot{C}_{i} / d C_{i} = - k_{i} < 0$ , so the equilibrium is globally asymptotically stable and the relaxation time is $τ_{i} = 1/ k_{i}$ . The explicit solution with initial condition $C_{i} (0)$ is $C_{i} (t) = C_{i}^{*} + (C_{i} (0) - C_{i}^{*}) e^{- t / τ_{i}}$ , confirming exponential relaxation. Summing over pools gives $C^{*} = \sum_{i} u_{i} / k_{i} = \sum_{i} u_{i} τ_{i}$ . Dividing by $u_{tot} = \sum_{i} u_{i}$ yields $\overset{τ}{ˉ} = \sum_{i} (u_{i} / u_{tot}) τ_{i}$ , the input-weighted average turnover time. $□$

Proposition 2 (passive-pool dominance). In a three-pool CENTURY system with turnover times $τ_{a} ≪ τ_{s} ≪ τ_{p}$ and litter fractions $u_{a}, u_{s}, u_{p}$ summing to 1, the passive-pool share of steady-state SOC is

\frac{C _{p}^{*}}{C ^{*}} = \frac{u _{p} τ _{p}}{u _{a} τ _{a} + u _{s} τ _{s} + u _{p} τ _{p}} .

For CENTURY-default values $(u_{a}, u_{s}, u_{p}) = (0.55, 0.40, 0.05)$ and $(τ_{a}, τ_{s}, τ_{p}) = (3, 30, 2000) yr$ , this share is approximately $0.88$ .

Proof. From Proposition 1, $C_{p}^{*} = u_{p} τ_{p}$ and $C^{*} = u_{a} τ_{a} + u_{s} τ_{s} + u_{p} τ_{p}$ . The share is the ratio

\frac{C _{p}^{*}}{C ^{*}} = \frac{u _{p} τ _{p}}{u _{a} τ _{a} + u _{s} τ _{s} + u _{p} τ _{p}} .

Substituting the defaults: numerator $= 0.05 \times 2000 = 100$ ; denominator $= 0.55 \times 3 + 0.40 \times 30 + 0.05 \times 2000 = 1.65 + 12 + 100 = 113.65$ . Ratio $= 100/113.65 \approx 0.880$ . Because $τ_{p}$ exceeds $τ_{a}$ by a factor of roughly $1 0^{3}$ , even a small $u_{p}$ makes the passive pool dominant in inventory; the precise share exceeds 50 percent whenever $u_{p} τ_{p} > (1 - u_{p}) max (τ_{a}, τ_{s})$ . $□$

Proposition 3 (warming-feedback scaling). If decomposition rate constants scale as $k_{i} (T) = k_{i} (T_{0}) Q_{10}^{(T - T_{0}) /10}$ and inputs $u_{i}$ are temperature-independent, then the steady-state inventory scales as

C^{*} (T_{0} + Δ T) = C^{*} (T_{0}) Q_{10}^{- Δ T /10},

independently of pool structure.

Proof. From Proposition 1, $C_{i}^{*} (T) = u_{i} / k_{i} (T)$ . Substituting $k_{i} (T_{0} + Δ T) = k_{i} (T_{0}) Q_{10}^{Δ T /10}$ gives

C_{i}^{*} (T_{0} + Δ T) = \frac{u _{i}}{k _{i} ( T _{0} ) Q _{10}^{Δ T /10}} = C_{i}^{*} (T_{0}) \cdot Q_{10}^{- Δ T /10} .

Summing over pools,

C^{*} (T_{0} + Δ T) = i \sum C_{i}^{*} (T_{0}) \cdot Q_{10}^{- Δ T /10} = (i \sum C_{i}^{*} (T_{0})) Q_{10}^{- Δ T /10} = C^{*} (T_{0}) Q_{10}^{- Δ T /10} .

The factor $Q_{10}^{- Δ T /10}$ is pool-independent, so the scaling is universal across the linear-reservoir family. For the Harvard Forest parameters $Δ T = 5 °C$ , $Q_{10} = 2$ , the predicted long-term inventory loss is $1 - 2^{- 0.5} \approx 0.293$ , or 29 percent, a transient upper bound modulated by acclimation of the labile pool and by possible temperature dependence of the input partition ${u_{i}}$ . $□$

Connections Master

Soil science and pedology 27.10.01 pending. This unit is the depth expansion of the chapter's survey anchor. Where 27.10.01 pending introduces soil formation, profiles, classification, and the basics of soil chemistry, this unit zooms in on the single most policy-relevant soil property, its organic carbon stock, and the biogeochemical machinery that builds and loses it. The CLORPT framework of Jenny (1941) introduced in 27.10.01 pending is the load-bearing structure for predicting soil-carbon stocks globally from climate, vegetation, and parent material, and the CENTURY compartmental model developed here is the quantitative refinement that turns CLORPT's qualitative state function into a predictive, time-evolving simulation.
The biological pump: ocean carbon cycle and climate regulation 27.05.04. This unit is the terrestrial mirror of the ocean biological pump. Both systems move carbon out of contact with the atmosphere by partitioning organic matter into pools with widely separated turnover times: the marine snow fast-sinking flux and the Martin-curve remineralization profile on the ocean side, and the active, slow, and passive SOM pools on the land side. The Martin curve and the CENTURY decomposition share the same mathematical structure, a power-law or exponential attenuation of carbon with depth or time, and both are governed by a single dominant parameter (the Martin exponent $b$ in the ocean, the passive-pool turnover time $τ_{p}$ on land). Both systems are also the principal uncertainties in the coupled carbon-cycle response to climate change.
Climate change: evidence, impacts, mitigation 27.07.01. Soil carbon is the single largest terrestrial carbon sink and one of the largest feedback uncertainties in CMIP6 projections. The warming-feedback corollary developed here, $C^{*} (T_{0} + Δ T) = C^{*} (T_{0}) Q_{10}^{- Δ T /10}$ , is the quantitative input that couples soil biology to atmospheric CO2 on policy-relevant timescales, and the permafrost-carbon feedback (Tarnocai 2009) is the dominant positive feedback identified in 27.07.01. The "4 per 1000" initiative, cover cropping, reduced tillage, and biochar deployment are the land-management levers evaluated in 27.07.01 using the CENTURY-derived sequestration potentials estimated here.
Permian-Triassic mass extinction 27.08.04. Mass-extinction events provide the deep-time test of soil-carbon feedbacks. The end-Permian crisis, driven by Siberian-Traps CO2 release, includes a soil-carbon signal in the form of massive organic-matter burial and erosion pulses visible in the isotopic and sedimentological record. Both this unit and 27.08.04 treat soil as a coupled biogeochemical system whose response to abrupt perturbation is governed by the same first-order compartmental kinetics, and the Permian-Triassic soil disruption is the geological-scale analogue of the permafrost "compost bomb" analysed here. The pattern links modern climate-projection uncertainty to the rock record.

Historical & philosophical context Master

Vasily Dokuchaev's 1883 monograph on the Russian chernozem ^{[Dokuchaev 1883]} founded modern pedology by establishing that soil is an independent natural body, the product of climate, organisms, relief, parent material, and time acting in concert, rather than a geological substrate modified by weathering. The chernozem, the deep black steppe soil of the Ukrainian and Russian plains, was Dokuchaev's type specimen: its fertility and depth could not be explained by the underlying bedrock alone but required the full five-factor synthesis. The Dokuchaev framework, formalised as the CLORPT equation, remains the conceptual foundation of pedology.

Hans Jenny's 1941 monograph Factors of Soil Formation ^{[Jenny 1941]} recast Dokuchaev's qualitative factors as a quantitative state function $S = f (c l, o, r, p, t)$ , opening pedology to mathematical prediction. The CLORPT framework made it possible to estimate soil-carbon stocks from climate and vegetation data alone, and Jenny's later work (Jenny 1980) extended the synthesis to soil-ecology systems. Selman Waksman's 1927 Principles of Soil Microbiology ^{[Waksman 1927]}, which earned him the 1952 Nobel Prize for the discovery of streptomycin, established the microbial-biochemical view of soil organic matter turnover that underlies all modern decomposition models.

William Schlesinger's 1977 review in the Annual Review of Ecology and Systematics ^{[Schlesinger 1977]} compiled the first coherent global terrestrial-carbon budget and fixed the modern scale of the soil reservoir at roughly three times the atmospheric. Parton, Schimel, Cole, and Ojima's 1987 paper in the Soil Science Society of America Journal ^{[Parton 1987]} introduced the CENTURY model, the compartmental framework that operationalised Schlesinger's budget for simulation. Johannes Lehmann's 2006 review ^{[Lehmann 2006]} opened the modern era of biochar research, and Charles Tarnocai's 2009 inventory of permafrost carbon ^{[Tarnocai 2009]} revealed that the northern circumpolar stock rivals the atmosphere in size and is the largest single positive carbon-cycle feedback. Jerry Melillo's Harvard Forest soil-warming experiment ^{[Melillo 2002]}, running since 1991, remains the principal empirical constraint on the form of the warming response, and Todd-Brown et al. (2013) quantified the resulting CMIP5 projection spread. The "4 per 1000" initiative, launched at COP21 in Paris in 2015, is the policy expression of this lineage.

Bibliography Master

@book{Dokuchaev1883,
  author    = {Dokuchaev, V. V.},
  title     = {Russian Chernozem},
  year      = {1883},
  note      = {Selected chapters translated by N. Kaner, Israel Program for Scientific Translations, Jerusalem, 1967},
}

@book{Jenny1941,
  author    = {Jenny, Hans},
  title     = {Factors of Soil Formation: A System of Quantitative Pedology},
  publisher = {McGraw-Hill},
  address   = {New York},
  year      = {1941},
}

@book{Waksman1927,
  author    = {Waksman, Selman A.},
  title     = {Principles of Soil Microbiology},
  publisher = {Williams \& Wilkins},
  address   = {Baltimore},
  year      = {1927},
}

@article{Schlesinger1977,
  author  = {Schlesinger, William H.},
  title   = {Carbon balance in terrestrial detritus},
  journal = {Annual Review of Ecology and Systematics},
  volume  = {8},
  pages   = {51--81},
  year    = {1977},
}

@article{Parton1987,
  author  = {Parton, William J. and Schimel, David S. and Cole, C. Vernon and Ojima, Dennis S.},
  title   = {Analysis of factors controlling soil organic matter levels in {Great Plains} grasslands},
  journal = {Soil Science Society of America Journal},
  volume  = {51},
  number  = {5},
  pages   = {1173--1179},
  year    = {1987},
}

@article{Post1982,
  author  = {Post, Wilfred M. and Emanuel, William R. and Zinke, Paul J. and Stangenberger, Alan G.},
  title   = {Soil carbon pools and world life zones},
  journal = {Nature},
  volume  = {298},
  pages   = {156--159},
  year    = {1982},
}

@article{Lehmann2006,
  author  = {Lehmann, Johannes and Gaunt, John and Rondon, Marco},
  title   = {Bio-char sequestration in terrestrial ecosystems---a review},
  journal = {Mitigation and Adaptation Strategies for Global Change},
  volume  = {11},
  pages   = {403--427},
  year    = {2006},
}

@article{Tarnocai2009,
  author  = {Tarnocai, Charles and Canadell, Josep G. and Schuur, Edward A. G. and Kuhry, Peter and Mazhitova, Galina and Zimov, Sergey},
  title   = {Soil organic carbon pools in the northern circumpolar permafrost region},
  journal = {Global Biogeochemical Cycles},
  volume  = {23},
  pages   = {GB2027},
  year    = {2009},
}

@article{Smith2008,
  author  = {Smith, Pete and Martino, Daniel and Cai, Zucong and Gwary, Daniel and Janzen, Henry and Kumar, Pushpam and McCarl, Bruce and Ogle, Stephen and O'Mara, Frank and Rice, Charles and Scholes, Bob and Sirotenko, Oleg and Howden, Mark and McAllister, Tim and Pan, Genxing and Romanenkov, Vladimir and Schneider, Uwe and Towprayoon, Sirintornthep and Wattenbach, Martin and Smith, Jo},
  title   = {Greenhouse gas mitigation in agriculture},
  journal = {Philosophical Transactions of the Royal Society B},
  volume  = {363},
  pages   = {789--813},
  year    = {2008},
}

@article{Melillo2002,
  author  = {Melillo, Jerry M. and Steudler, Paul A. and Aber, John D. and Newkirk, Kathleen and Lux, Heidi and Bowles, Frederick P. and Catricala, Charles and Magill, Alan and Ahrens, Timothy and Morrisseau, Stephanie},
  title   = {Soil warming and carbon-cycle feedbacks to the climate system},
  journal = {Science},
  volume  = {298},
  pages   = {2173--2176},
  year    = {2002},
}

@article{Melillo2011,
  author  = {Melillo, Jerry M. and Butler, Sarah and Johnson, Jennifer and Mohan, Jeffrey and Steudler, Paul and Lux, Heidi and Burrows, Erin and Bowles, Frederick and Smith, Riley and Scott, Lindsey and Vario, Chelsea and Hill, Timothy and Burton, Andrew and Zhou, Yi-Ming and Tang, Julian},
  title   = {Soil warming, carbon-nitrogen interactions, and forest carbon budgets},
  journal = {Proceedings of the National Academy of Sciences},
  volume  = {108},
  number  = {23},
  pages   = {9508--9512},
  year    = {2011},
}

@article{ToddBrown2013,
  author  = {Todd-Brown, Katherine E. O. and Randerson, James T. and Post, Wilfred M. and Hoffman, Forrest M. and Tarnocai, Charles and Schuur, Edward A. G. and Allison, Steven D.},
  title   = {Causes of variation in soil carbon simulations from {CMIP5} {Earth} system models and comparison with observations},
  journal = {Biogeosciences},
  volume  = {10},
  pages   = {1717--1736},
  year    = {2013},
}

@book{Schlesinger1997,
  author    = {Schlesinger, William H. and Bernhardt, Emily S.},
  title     = {Biogeochemistry: An Analysis of Global Change},
  edition   = {3rd},
  publisher = {Academic Press},
  address   = {San Diego},
  year      = {2013},
}

Prerequisites

27.10.01

Tier anchors

beginner: Kristin Ohlson, The Soil Will Save Us (2016); USDA-NRCS 'Soil Health for Educators' materials; IPBES Land Degradation and Restoration Assessment (2018)
intermediate: Schlesinger, Biogeochemistry: An Analysis of Global Change (3e, 2013); Stockmann et al. 2013 Agric. Ecosyst. Environ. 164:21; FAO 'Recarbonizing Global Soils' technical report (2020)
master: Dokuchaev 1883; Jenny 1941; Waksman 1927; Schlesinger 1977 Annu. Rev. Ecol. Syst. 8:51; Parton et al. 1987 Soil Sci. Soc. Am. J. 51:1173; Post et al. 1982 Nature 298:156; Lehmann 2006 Mitig. Adapt. Strateg. Glob. Change 11:403; Tarnocai 2009 Global Biogeochem. Cycles 23:GB2027; Smith 2008 Phil. Trans. R. Soc. B 363:789; Melillo 2002 Science 298:2173 / 2011 Phil. Trans. R. Soc. B 366; Todd-Brown 2013 Biogeosciences 10:1717

References

Dokuchaev, V. V. — Russian Chernozem (1883), translated by N. Kaner, Israel Program for Scientific Translations (1967) · Part I; soils as independent natural bodies
Jenny, H. — Factors of Soil Formation: A System of Quantitative Pedology, McGraw-Hill, New York (1941) · Ch. 1-2; the CLORPT state function
Parton, W. J., Schimel, D. S., Cole, C. V., and Ojima, D. S. — Analysis of factors controlling soil organic matter levels in Great Plains grasslands, Soil Science Society of America Journal 51, 1173-1179 (1987) · Full text; the CENTURY model pool structure and decomposition rate constants
Schlesinger, W. H. — Carbon balance in terrestrial detritus, Annual Review of Ecology and Systematics 8, 51-81 (1977) · Full text; the global soil-carbon budget
Lehmann, J., Gaunt, J., and Rondon, M. — Bio-char sequestration in terrestrial ecosystems: a review, Mitigation and Adaptation Strategies for Global Change 11, 403-427 (2006) · Full review; biochar as a carbon-negative technology
Tarnocai, C., Canadell, J. G., Schuur, E. A. G., Kuhry, P., Mazhitova, G., and Zimov, S. — Soil organic carbon pools in the northern circumpolar permafrost region, Global Biogeochemical Cycles 23, GB2027 (2009) · Full text; the permafrost-carbon stock inventory
Smith, P., Martino, D., Cai, Z., Gwary, D., Janzen, H., Kumar, P., McCarl, B., Ogle, S., O'Mara, F., Rice, C., Scholes, B., Sirotenko, O., Howden, M., McAllister, T., Pan, G., Romanenkov, V., Schneider, U., Towprayoon, S., Wattenbach, M., and Smith, J. — Greenhouse gas mitigation in agriculture, Philosophical Transactions of the Royal Society B 363, 789-813 (2008) · Review; climate-smart soils and the 4 per 1000 framework
Melillo, J. M., Steudler, P. A., Aber, J. D., Newkirk, K., Lux, H., Bowles, F. P., Catricala, C., Magill, A., Ahrens, T., and Morrisseau, S. — Soil warming and carbon-cycle feedbacks to the climate system, Science 298, 2173-2176 (2002) · Full text; the Harvard Forest soil-warming experiment

Estimated time

beginner: 25m
intermediate: 50m
master: 85m