Dissipative Path Assembly: an introduction

1. The starting physical picture

The biosphere is a dissipative structure¹ maintained by the solar gradient². Shortwave radiation³ arrives from the Sun, the Earth system absorbs part of it, and the planet eventually re-emits the energy as longwave infrared radiation. At the planetary boundary this is an energy balance problem. Inside the boundary it is a path problem.

¹ A dissipative structure is a thermodynamic concept from Ilya Prigogine (Nobel 1977). It is any organised, far-from-equilibrium pattern that can only be maintained by a continuous throughput of energy — consuming it and expelling entropy to the surroundings. Classic examples: a candle flame, a hurricane, a living cell. The key point is that the structure exists because energy flows through it, not despite that flow. Prigogine showed that far from equilibrium, matter can self-organise into ordered states, making dissipative structures the physical foundation for understanding life.

² A gradient here means a difference in a physical quantity across space — here, the Sun’s surface (~5,500 °C) versus the Earth’s (~15 °C average surface, ~-18 °C effective radiating temperature). Because the Sun is so much hotter, the photons it emits carry far more free energy per joule than the infrared photons Earth radiates back. The “solar gradient” is that disparity: the Earth-Sun system is perpetually out of equilibrium, and that imbalance is the ultimate power source for everything the biosphere does.

³ The electromagnetic spectrum splits here by wavelength. Sunlight arrives mostly as visible light and near-infrared (roughly 0.3–3 µm wavelength) — called “shortwave” because the Sun’s ~5,500 °C surface emits peak radiation there (Wien’s law). Earth, much cooler, re-emits radiation in the thermal infrared (roughly 4–100 µm) — called “longwave.” The distinction matters because shortwave photons are energetic enough to drive photochemistry (photosynthesis), while longwave photons are not.

⁴ To be “thermalized” is to have energy converted into disordered thermal motion — heat — that cannot be recaptured for useful work. When sunlight warms a rock or heats the atmosphere, its ordered, directional photon energy becomes the random jostling of molecules. The rock later emits longwave radiation, but the quality (free energy content per joule) has dropped dramatically. Thermalization is entropy production.

⁵ Free energy (here, Gibbs free energy \(G\)) is the portion of a system’s total energy that can be converted into work under constant temperature and pressure conditions. The rest is locked up as disordered heat. When a bond is formed or broken in a molecule, the sign of \(\Delta G\) tells you whether the reaction releases usable energy (\(\Delta G < 0\), spontaneous) or requires an energy input (\(\Delta G > 0\)). Photosynthesis uses light to drive an energetically uphill reaction — it uses solar free energy to create chemical free energy stored in sugars.

⁶ In chemistry, “reduction” means gaining electrons (or equivalently, gaining hydrogen atoms, since H carries one electron). Carbon in CO₂ is in its most oxidised state (carbon has lost as many electrons as it can). Photosynthesis reduces that carbon — stuffs electrons back into it — to make sugars and other organic compounds. These reduced carbon compounds carry chemical free energy because oxidising them back (burning glucose, aerobic respiration) releases that stored energy. The biosphere is essentially a slow-burning fire, with the “fuel” replenished by sunlight.

⁷ Net primary production is the rate at which plants, algae, and other photosynthesisers accumulate new biomass, after subtracting the carbon they consume in their own respiration. Gross primary production (GPP) is the total amount fixed by photosynthesis; plants use roughly half of that themselves, leaving NPP as the net addition to the biological carbon pool available to everything else. Global terrestrial NPP is roughly 50–60 Pg (petagrams) of carbon per year; marine NPP adds a similar amount.

Most absorbed solar energy is thermalized⁴ quickly by the atmosphere, ocean, land surface, and clouds. A much smaller but crucial fraction is captured by photosynthesis and stored, temporarily, as chemical free energy⁵ in reduced carbon compounds⁶. This captured fraction is net primary production, or NPP⁷.

Once energy enters NPP, it does not immediately become heat. The incoming shortwave photons are absorbed by pigments, and their excitation energy is used to move electrons uphill: water is oxidized, carbon dioxide is reduced, and a small part of the incoming radiative free energy is converted into chemical potential in carbohydrates, lipids, proteins, cell walls, roots, wood, detritus, and microbial biomass. The photon itself is gone almost immediately. What persists is the disequilibrium it helped create: reduced carbon and associated biological structure held away from the oxidized, thermal equilibrium state.

From the planetary boundary, the accounting is still a photon-balance story. The Earth absorbs relatively few high-energy, low-entropy shortwave photons and emits many more lower-energy, higher-entropy longwave infrared photons. Biology does not change that final balance. It inserts a structured delay between absorption and re-emission. During that delay, solar-derived free energy is routed through biological structures:

sunlight -> photosynthesis -> plant/algal biomass -> consumers
         -> detritus -> microbes -> respiration -> heat + CO2 + H2O

The important point is not that biology prevents entropy production. It cannot. The point is that biology delays, routes, and structures entropy production. The entropic gradient in the original sunlight has been partly converted into a chemical gradient: reduced organic matter plus atmospheric oxygen, maintained in a world where the spontaneous direction is oxidation back to CO2 and H2O. A unit of captured solar free energy can spend seconds in a metabolite, days in phytoplankton biomass, years in wood, decades in soil organic matter, or centuries in long-lived forest carbon.

The release mechanism is respiration, combustion, fermentation, decay, and other oxidative or redox reactions. Enzymes and food webs step the stored chemical free energy down through ATP, heat, waste products, and finally CO2 and H2O. That heat is exchanged with the surrounding air, water, and land surface and is ultimately emitted to space as longwave radiation. In this sense the whole biological path is a temporary detour in the conversion of incoming shortwave radiation into outgoing longwave radiation: one high-exergy photon stream enters, many lower-exergy infrared photons leave, and the dwell time is the interval over which part of that exergy was embodied as living and dead biological structure.

This note calls that delay the entropic dwell time of the captured energy. It is the mean time that a unit of NPP-derived free energy remains inside organized biological matter before final dissipation.

Strictly, the same original photon is not tracked through the system. The useful quantity is the residence time of the chemical free energy created by photosynthesis. The photon is thermalized quickly; the biological free-energy equivalent can dwell.

The hierarchy of the argument is therefore:

1. The photon/free-energy balance is the fundamental thermodynamic reality: shortwave exergy enters, entropy is produced, and longwave radiation leaves.
2. Entropic dwell time \(D\) is the quantity this note wants to foreground. It asks how long captured free energy remains organized before final dissipation. This is the speculative step, but also the motivating one: dwell time may be a useful way to compare biospheres, technospheres, societies, and technologies as dissipative systems.
3. Energy-rate density⁸ is a practical power metric. It asks how much free energy flows through a system per unit mass. It is easier to estimate across stars, organisms, machines, cities, and economies than detailed pathway structure, so it is a good first-pass measure of dissipative intensity.
4. Assembly index is a practical structure metric. It is harder to measure than energy-rate density, and only feasible in specific chemical or material contexts, but it points more directly at the historically constructed pathways through which free energy flows.

⁸ Energy-rate density, associated especially with Eric Chaisson’s work, is power flow normalized by mass: free energy per unit time per unit mass. In SI units it is W kg⁻¹; Chaisson often reports it in erg s⁻¹ g⁻¹. The normalization matters because total power alone mostly measures size. A star, a forest, a mammal, a jet engine, and a city can all process large power flows, but energy-rate density asks how concentrated that flow is in organized matter. The main caveat is boundary choice: changing which mass is included can change the value substantially.

Energy-rate density and assembly index are therefore not replacements for dwell time. They are two handles on it. Energy-rate density says how intensely a system processes free energy; assembly index says how deeply constructed the matter in the pathway is. The conjecture is that the things we care about in the biosphere, and perhaps in technology and society, are created when the fundamental entropic balance is routed through high-flux, long-dwell, historically assembled pathways.

---

2. From energy balance to path length

In a simple thermodynamic box model, one might only ask how much power enters and leaves:

incoming solar power = outgoing infrared power

For the biosphere, that is not enough. Two ecosystems can capture similar amounts of solar energy through photosynthesis — similar NPP — while having very different internal path lengths.

An open-ocean phytoplankton system can turn over biomass in days to weeks. A forest can hold a large fraction of its captured carbon in trunks, roots, detritus, and soil for years to centuries. Both dissipate solar free energy, but the second system has a much longer biological residence path.

Let \(F\) be the NPP flux entering an ecosystem — the rate of photosynthetic energy capture, expressible as power per unit area (W m⁻²) by converting the standard carbon-mass units using the free energy of reduced carbon, roughly 40 kJ per gram of carbon fixed.

Once energy is captured by photosynthesis it moves through the ecosystem as a flux — a flow of chemical free energy from pool to pool, eventually respired back to heat. Let the ecosystem be divided into compartments⁹ or pathway classes \(i\): primary producers, herbivores, predators, detritus, microbial biomass, long-lived structural biomass, and so on. For each compartment, let \(f_i\) be the fraction of the total NPP flux that passes through it, and \(\tau_i\) the mean time that energy dwells there.

⁹ In ecosystem modelling, a compartment (or “pool”) is a conceptually bounded category of biological or chemical material — “living plant biomass,” “soil organic matter,” “dissolved organic nitrogen” — treated as if it were a single well-mixed reservoir for the purposes of tracking flows. Carbon and energy flows are tracked as transfers between compartments. The choice of compartment boundaries is a modelling decision, not a fact about nature, and matters enormously for what the model can and cannot see.

Then define the dissipative path length or mean entropic dwell time:

\[D = \sum_i f_i \tau_i\]

D has units of time. It is a flux-weighted mean first-passage time¹⁰ from photosynthetic capture to final respiratory dissipation.

¹⁰ Mean first-passage time is a concept from the mathematics of random walks and stochastic processes. If you have a particle bouncing around a network, the first-passage time is how long it takes before it first reaches a designated “exit” state. The mean first-passage time averages this over many realisations of the random walk. Here the “particle” is a unit of free energy, the “network” is the ecosystem’s compartment graph, and the “exit” is respiration back to CO₂ and heat. Using first-passage time rather than, say, average sojourn time is precise: we want the expected total duration from photosynthetic capture to final dissipation.

This is the first key idea:

The biosphere can be compared not only by how much free energy it dissipates per square metre, but by how long and through what kinds of structures that free energy dwells before becoming heat.

The phrase “path length” is intentional. It is analogous to transport through a network. Energy enters a graph of biological compartments and exits when it is respired. The longer and more reticulated¹¹ the graph, the longer the mean dwell time.

¹¹ From the Latin rete, net. Reticulated means arranged like a net or web, with many interconnected loops rather than a simple linear chain. A reticulated network has high connectivity — many alternative paths between any two nodes. In ecosystem terms, a food web with many cross-links, detrital loops, and microbial recycling pathways is more reticulated than a simple linear food chain.

---

3. Why assembly theory belongs in this picture

Assembly theory¹² assigns an assembly index to an object. Informally, the assembly index of an object is the minimum number of joining operations¹³ required to build it from elementary parts, allowing previously built substructures to be reused.

¹² Assembly theory is a framework developed by Lee Cronin, Sara Walker, and colleagues (formalized in Nature 2023) for quantifying how much “history” is encoded in an object’s structure. The core idea: the assembly index of an object is the minimum number of recursive joining steps needed to build it from elementary units, reusing intermediate subassemblies. A molecule with assembly index 15 cannot be assembled by accident in fewer than 15 steps; finding it at high abundance implies a process that repeatedly executes those steps — i.e., a process with memory, like biology or deliberate manufacture. The theory bridges chemistry, information theory, and evolution.

¹³ A joining operation in assembly theory is simply the bonding of two previously existing fragments to make a larger one. For molecules, this corresponds to forming a covalent bond between subunits. Crucially, the framework allows previously built substructures to be reused as building blocks: making two copies of a subunit and then joining them counts as fewer steps than building the whole from scratch twice. This recursion is what makes the assembly index sensitive to hierarchical complexity rather than just raw size or molecular weight.

For molecules, the assembly index captures a kind of construction depth. Simple molecules such as water or carbon dioxide have very low assembly index. More complex molecules such as amino acids, nucleotides, pigments, cofactors, peptides, and specialized metabolites have higher assembly index.

The central conceptual move in assembly theory is that high assembly index, when observed at high copy number¹⁴, is evidence for a process with memory. A single complicated object might be a rare accident. Many copies of a high assembly object imply a generating process that repeatedly finds and preserves that structure. In biology, that memory is evolution, inheritance, metabolism, and ecological selection.

¹⁴ Copy number is simply how many identical copies of a given molecule (or other object) are present in a sample. For assembly theory, copy number is conceptually essential: a single complicated molecule could plausibly arise by chance. Observing \(10^{10}\) identical copies of a molecule with high assembly index is strong evidence that some process repeatedly produces that exact structure — which, in biology, means evolved metabolic machinery. Low assembly molecules can accumulate by simple chemistry; high assembly molecules at high copy number require biology (or designed manufacture).

For the purposes of this framework, we need only a modest version of assembly theory:

1. Biological matter contains many molecules with nontrivial assembly index.
2. The distribution of assembly indices in a biological sample can, in principle, be measured by mass spectrometry¹⁵.
3. The average assembly index of matter in an ecosystem is not just a chemical descriptor. It is also a record of evolved pathway structure.

¹⁵ Mass spectrometry fragments molecules and measures the mass-to-charge ratio of the resulting pieces. For assembly-index estimation, the key measurement is the set of fragment masses: how many distinct subcomponents does the molecule break into, and how large are they? The Marshall et al. (2021) paper showed that molecules from biological sources systematically score higher on this metric than abiotic controls, validating the method as a life-detection strategy. The technique cannot give the exact assembly index of every molecule, but can provide the distribution of assembly indices across the detectable molecular fraction of a sample.

The biosphere is therefore not merely a carbon reservoir or an energy-flow system. It is a planetary-scale generator and maintainer of high-copy, high-assembly objects.

---

4. From object assembly to flow assembly

Assembly theory usually begins with individual objects: a molecule, a polymer, a manufactured component, perhaps a digital structure. Ecosystems are different because they are flow systems. The relevant question is not only “what is the assembly index of this object?” but:

What is the mean assembly index of the matter through which free energy is currently flowing?

That leads to a flux-weighted assembly index¹⁶. For each compartment \(i\), let \(a_i\) be the mean molecular assembly index of matter in that pool. Then define:

¹⁶ A flux-weighted average weights each compartment’s contribution by the fraction of total energy flow passing through it, not by how much matter it contains or how large it is. Compare: the biomass-weighted average would be dominated by whichever pool holds the most carbon (often soil organic matter or wood). The flux-weighted average is dominated by whichever pool processes the most NPP-derived energy per unit time. These can be very different: a small but fast-cycling microbial pool might carry most of the flux while holding little biomass.

\[\bar{A} = \sum_i f_i a_i\]

\(\bar{A}\) is the mean assembly index experienced by a unit of energy as it moves through the ecosystem. It is not the maximum assembly index present anywhere in the system, and it is not the biomass-weighted average over all matter. It is a flow-weighted average over the dissipative path.

This distinction matters. A rare high-assembly molecule present in tiny amounts should not dominate the metric. A huge pool of inactive material should not dominate unless significant NPP-derived flux actually passes through it.

The proposed metric is therefore:

\[\Phi_A = F\bar{A}\]

the assembly flux, with the form “power multiplied by construction depth.” It measures how much solar-derived free energy is being routed through how much molecular assembly.

---

5. The core hypothesis

The core hypothesis is that flux-weighted assembly depth and entropic dwell time are coupled:

\[\bar{A} \propto D\]

or, at the compartment level, \(\tau_i\) increases with \(a_i\).

This should not be read as a universal law for every molecule. Many high-assembly biological molecules are deliberately short-lived: ATP¹⁷, signalling peptides, mRNA¹⁸, reactive metabolites, and regulatory proteins can turn over rapidly. Conversely, some low-assembly inorganic or mineral structures can persist for geological times.

¹⁷ Adenosine triphosphate — the universal energy currency of living cells. ATP is generated by breaking down food or by photosynthesis, and spent to drive nearly every energy-requiring process in biology: protein synthesis, ion pumping, movement, DNA replication. Despite being chemically complex (a nucleotide with a triphosphate tail; assembly index ~12–15), ATP turns over extremely fast — a resting human cycles through their body weight in ATP roughly every day, and an exercising cell can regenerate ATP hundreds of times per second. It is high-assembly but deliberately short-lived.

¹⁸ Messenger RNA is the transient transcript of a gene — the molecular “working copy” that carries the genetic instruction from DNA to the ribosome for protein synthesis. mRNA molecules are typically degraded within minutes to hours after synthesis; this brief lifetime is a feature, not a bug, since it allows cells to rapidly adjust which proteins they make in response to changing conditions. mRNA is chemically complex (a polymer of nucleotides with specific sequence), so it has a nontrivial assembly index, but it is among the shortest-lived high-assembly molecules in the cell.

¹⁹ At the pool (or compartment) level rather than the molecular level means: rather than asking whether this ATP molecule lives longer than this glucose molecule, we ask whether the ATP pool as a whole (all ATP molecules in the system, with their average residence times) differs systematically in dwell time from the soil organic matter pool as a whole. The pool-level claim is robust to the fact that individual molecules of the same type vary in their fates; it only requires that the average over many molecules in a given ecosystem compartment follows a pattern.

The stronger and more defensible version is a pool-level hypothesis¹⁹:

In living ecosystems, the NPP-derived flux that reaches more historically constructed, more compositionally specific, higher-assembly biological pools tends to have a longer mean residence time before final dissipation.

In other words, the expected relationship is not:

every high-AI molecule lives longer than every low-AI molecule

but rather:

biological pools with higher flux-weighted molecular assembly tend to be
associated with longer ecological residence times

This formulation avoids a false molecular claim while preserving the central thermodynamic idea. Assembly index is being used as a measure of historically constructed path depth, not as a simple bond-stability proxy.

---

6. The biomass-NPP identity

The most useful empirical handle is the relationship between standing biomass and productivity.

Let \(B\) be standing biomass (carbon per unit area) and NPP the annual net primary production (carbon per unit area per year). Then:

\[\frac{B}{\text{NPP}} = \bar{\tau}\]

This is just dimensional analysis²⁰, but it is powerful. \(B/\text{NPP}\) is the mean turnover time implied by the standing stock and the flow through it. It tells us how long, on average, carbon remains in the ecosystem’s biological stock.

²⁰ Dimensional analysis is the practice of checking or deriving relationships between quantities by tracking their units. Here: biomass \(B\) has units of [carbon mass / area]; NPP has units of [carbon mass / area / time]. Their ratio \(B/\text{NPP}\) has units of [time] — specifically, the time it would take the NPP flux to accumulate the current standing stock, if nothing was lost. This is the mean residence time, or turnover time, of carbon in the system. No physics is assumed — it falls out of the units alone.

If the core hypothesis is correct, then:

\[\frac{B}{\text{NPP}} \approx \text{const} \times \bar{A}\]

This gives a first-order proxy for flux-weighted assembly index without requiring immediate mass spectrometry. Ecosystems with high biomass per unit productivity have long dwell times. If those dwell times correspond to flux through more deeply assembled biological structures, they should also have high \(\bar{A}\).

This is why forests, mangroves²¹, peatlands²², soils, and detrital systems are so important for the framework. They do not merely contain a lot of carbon. They extend the time over which solar-derived free energy remains in organized biological form.

²¹ Mangroves are salt-tolerant trees and shrubs that grow in tropical and subtropical coastal intertidal zones. They are among the most carbon-dense ecosystems on Earth: their root systems trap sediment and accumulate organic matter, and their soils (often called “blue carbon” stores) can hold centuries of accumulated carbon. They also support highly productive and diverse food webs in their root networks. Mangroves are therefore high on both axes that matter here: high NPP and very long carbon residence times.

²² Peatlands (bogs, fens, moors) are wetland ecosystems where dead organic matter accumulates faster than it decomposes. Cold, waterlogged, and often acidic conditions suppress microbial decomposition, allowing partially decomposed plant material (peat) to build up over thousands of years. Peatlands cover only about 3% of Earth’s land area but store roughly 25–30% of all soil carbon — a massive, long-dwell carbon reservoir. Draining or burning peatlands releases this ancient carbon rapidly.

---

7. Explicit examples: the dwell time spectrum

To ground the framework in concrete biology, it helps to walk through specific dwell times from the fastest to the slowest. The full spectrum spans nine orders of magnitude — from ATP recycled in two minutes to bog oak intact after eight thousand years.

7.1 The fast end: metabolic turnover

The fastest-cycling pools are the control and currency molecules of the cell.

ATP: recycled roughly every two minutes in an active cell. A resting human turns over their body weight in ATP every day. Despite being one of the more complex small molecules in biology (assembly index ~24 by Marshall et al. 2021), ATP is designed for rapid cycling — its value is not persistence but availability. This is the canonical exception to the pool-level hypothesis: high assembly index, extremely short dwell time.

mRNA: most transcripts have half-lives of four minutes to eight hours depending on cell type and measurement method.²³ The short lifetime is intentional — rapid mRNA degradation lets cells adjust gene expression within minutes. Like ATP, mRNA is high-assembly but short-lived by design.

²³ mRNA half-life is method-dependent in a way that matters. Older methods — adding transcription inhibitors and watching mRNA levels fall — gave median half-lives of 7–10 hours in mammalian cells. Newer metabolic labeling methods (adding 4-thiouridine, then pulling down the labeled fraction) correct for ongoing synthesis and give a median of 4–5 minutes in some studies. The true value for a typical mammalian cell is probably 4–8 hours, with a fast-cycling fraction under 2 minutes.

Phytoplankton cells: a typical cell divides every one to three days in bloom conditions. Biomass turns over roughly 100 times per year at the global scale. Individual cells are gone in days; their molecular components are recycled within the microbial loop.

Zooplankton: copepods complete a generation in one to three weeks at tropical temperatures, longer in cold water. Krill live one to two years.

7.2 The intermediate range: leaves, roots, soil microbes

Leaf litter decomposes on timescales set almost entirely by temperature, moisture, and lignin content.²⁴ In tropical wet forests, \(k \approx 2\)–4 yr⁻¹, giving a half-life of two to four months. In boreal forests, \(k \approx 0.1\)–0.3 yr⁻¹, giving a half-life of two to seven years. The same leaf chemistry — but the residence time varies by a factor of thirty depending on climate.

²⁴ Leaf litter decomposition globally follows first-order kinetics: \(M(t) = M_0 e^{-kt}\), where \(k\) is the decomposition rate constant and the half-life is \(\ln(2)/k\). The value of \(k\) is predicted globally by actual evapotranspiration (AET, a combined measure of temperature and moisture) and by litter chemistry — lignin content especially. AET alone explains ~46% of global variance in \(k\); adding lignin content explains ~78% in tropical systems (Zhang et al. 2008). The practical consequence: tropical leaf litter decomposes 10–30× faster than boreal litter with the same chemistry.

Fine roots turn over on timescales of months to twelve years, with finer roots cycling faster than coarser ones.

Microbial biomass in soil has a wide range: the active bacterial fraction turns over in days to weeks; dormant bacteria and fungi can be resident for months to years. Melanized²⁵ fungi are among the slowest-cycling microbial pools, their cell walls resisting decomposition via metal-chelating pigments and mineral bonding.

²⁵ Fungal melanin is a dark amorphous pigment polymer in fungal cell walls. It chelates metal ions, quenches free radicals, and binds to clay minerals and iron oxides in soil — all of which slow decomposition. Melanized fungal biomass has a larger recalcitrant fraction than non-melanized biomass, making ectomycorrhizal fungal necromass a disproportionate contributor to stable soil organic matter. The mechanism is analogous to lignin’s recalcitrance but via a different polymer class.

Soil organic matter is a continuum. The active fraction turns over in months to a few years. The slow fraction, associated with clay minerals and iron oxides, has mean residence times of decades to centuries.

7.3 The structural champion: wood

Coarse woody debris (CWD): after a tree dies, the fallen log decomposes at rates set by wood density and climate. High-density tropical hardwoods (\(k = 0.12\) yr⁻¹) have a half-life of six years even in warm, moist conditions. In boreal forests, CWD half-lives of 67 years have been measured. Waterlogged logs exit the active biological cycle almost entirely — bog oak from Irish raised bogs is routinely dated to five to eight thousand years old.

Living wood: a tropical hardwood tree lives a mean of roughly 186 years and can reach 500+. A boreal conifer commonly exceeds 500 years; Pinus longaeva holds individuals verified at 4,900 years. For the entire life of that tree, the carbon in the heartwood simply dwells. The mean residence time of wood carbon in tropical forest has been estimated at roughly 50 years; in boreal forest the figure is higher.

Why wood dominates the \(D\) calculation. Wood constitutes 70–90% of total tree biomass in mature forest. Foliage is less than 5%. So the question “how long does NPP-derived carbon dwell in a forest?” is almost entirely a question about wood. No other biological pool combines a large flux fraction with multi-decade to multi-century residence times.

The key molecule is lignin²⁶, the irregular aromatic polymer that stiffens wood cell walls. Cellulose can be hydrolysed enzymatically by many organisms. Lignin requires oxidative cleavage — a chemically demanding step achievable only by white-rot basidiomycete fungi²⁷ under aerobic conditions. In cold, anoxic, or acidic environments, lignin is nearly indestructible. The recalcitrant fraction of soil lignin has a mean residence time of 20–38 years even under temperate conditions.

²⁶ Lignin is an irregular, three-dimensional aromatic polymer cross-linked by both ether bonds and carbon–carbon bonds. Unlike cellulose (a regular repeating glucose chain), lignin has no repeating unit — it is formed by random radical polymerization of three monolignols. This structural irregularity means there is no enzymatic “zipper” that can disassemble it: degradation requires oxidative enzymes (laccases, peroxidases) that attack it unpredictably. Lignin constitutes 20–30% of wood dry mass and is the primary reason wood is so durable. Its assembly index would be very high — hundreds of joining steps — and it is produced at enormous scale, making wood the highest-copy-number, highest-assembly-index biological material in the terrestrial biosphere.

²⁷ White-rot fungi (mostly Basidiomycetes: Trametes, Phanerochaete, Pleurotus, and relatives) are the only organisms that efficiently degrade lignin. They do so using non-specific oxidative enzymes — lignin peroxidase, manganese peroxidase, laccases — that generate free radicals attacking the aromatic ring system. The name comes from the white color of wood after the brown lignin layer is removed. In anaerobic conditions (waterlogged soil, peat, lake sediments), white-rot fungi cannot operate — which is why waterlogged wood can persist for thousands of years.

7.4 Megafauna bone: the case for the runner-up

Bone does not beat wood in the flux-weighted calculation — the flux fraction \(f_i\) for bone pathways is tiny. But bone sits at the end of the deepest trophic pathways and represents the most deeply assembled matter in the active biosphere.

In the living animal. Bone is continuously remodeled: in humans roughly 10% of the skeleton turns over per year. In large cetaceans, allometric scaling of osteocyte density suggests substantially slower remodeling — the bone of a living blue whale may cycle on timescales of decades. Bowhead whales live to 200+, meaning their skeletal carbon can dwell for two centuries at the top of one of the longest food chains on Earth.

Whale fall.²⁸ When a large whale sinks to the seafloor, the succession runs through three phases:

²⁸ A whale fall is a whale carcass that has sunk to the seafloor. The sulfophilic phase, discovered in the 1980s by Craig Smith and colleagues, was surprising because it hosts a chemoautotrophic community comparable to hydrothermal vent communities — powered by sulfide from anaerobic lipid decomposition rather than geothermal sulfide. Over 100 species have been found only at whale falls; Osedax (“bone eater”) worms, discovered in 2002, drill into whale bone with root-like tissues to access the lipid-rich bone matrix directly.

• Mobile scavenger phase (months–1.5 years): sharks and hagfish strip soft tissue at 40–60 kg/day.
• Enrichment opportunist phase (months–4.5 years): invertebrates colonize exposed bone sustained by lipid seepage.
• Sulfophilic phase (decades; 50–100 years for a large whale): anaerobic bacteria decompose lipids embedded in the bone’s porous matrix, producing hydrogen sulfide that sustains a chemoautotrophic community comparable to hydrothermal vent faunas.

In the sulfophilic phase the whale bone is not a passive store — it is the energy source for an entire ecosystem. A large blue whale contains enough lipid in its skeleton to sustain that community for half a century.

The comparison. A rough sketch for a tropical forest plus adjacent ocean:

Pool	\(f_i\)	\(\tau_i\) (yr)	\(f_i \tau_i\)
Living wood	0.40	80	32.0
Leaf/fine root cycle	0.30	0.5	0.15
CWD	0.10	12	1.2
Slow soil OM	0.04	50	2.0
Megafauna	0.001	25	0.025
Whale bone + fall	0.00001	75	0.001
Total \(D\)			~35 yr

Wood contributes ~33 of the ~35 year total. Bone contributes ~0.001 yr. The conclusion is stark: wood dominates \(D\) because flux fraction × residence time is overwhelmingly concentrated there. What bone contributes is assembly depth — it raises \(\bar{A}\) even when its contribution to \(D\) is negligible, which is exactly why the two metrics are distinct and both necessary.

7.5 The geological fringe: sporopollenin

Sporopollenin²⁹ is the polymer of pollen exine walls. Its unique dual cross-linkage — ester bonds (acid-resistant) and acetal bonds (base-resistant) — makes it the most chemically inert biopolymer known. It survives the acetolysis that destroys cellulose, protein, and lignin, which is why fossil pollen analysis is possible at all. Fossil sporopollenin from Ordovician spores is ~475 million years old.

²⁹ Sporopollenin is the polymer of the outer wall (exine) of pollen grains and some plant spores. Its molecular structure resisted characterization for decades due to its extreme chemical inertness. Recent analysis confirmed a unique dual cross-linkage: phenolic monomers linked by both ester and acetal bonds. Ester bonds resist acidic conditions; acetal bonds resist basic conditions; together they defeat the acetolysis treatment (acetic anhydride in sulfuric acid) that destroys virtually all other biological polymers, including cellulose, protein, and lignin. Palynology — the entire discipline of fossil pollen analysis — depends on this inertness.

For the framework’s purposes, sporopollenin is not the champion: pollen is produced in tiny amounts relative to wood, and pollen grains exit the active biological pathway almost immediately. But it proves that biological polymers can achieve geological-timescale persistence through chemistry alone — no mineral protection, no freezing required.

---

8. Four idealized ecosystem types

The framework separates power, dwell time, and assembly depth. This is easiest to see by comparing idealized ecosystem regimes.

7.1 High power, short dwell time

Open-ocean phytoplankton systems can have significant areal³⁰ or global productivity, but low standing biomass because cells turn over quickly and grazing/respiration pathways are fast.

³⁰ “Areal” simply means per unit area — per square metre or per hectare. Open ocean has lower NPP per square metre than most land ecosystems, but because it covers 70% of Earth’s surface, its total global NPP is comparable to terrestrial NPP. Distinguishing areal (local) productivity from total (global) productivity is important: a system can be globally significant while being locally sparse.

• high \(F\)
• low \(B/\text{NPP}\), short \(D\)
• low-to-moderate \(\bar{A}\)

Such systems dissipate free energy rapidly. They are powerful but shallow in biological dwell time.

7.2 Low power, long apparent dwell time

Deserts can have low NPP but surprisingly long B / NPP because sparse woody or belowground biomass may persist relative to the small annual flux.

• low \(F\)
• high \(B/\text{NPP}\), moderate \(D\)
• limited assembly flux \(\Phi_A = F\bar{A}\)

This is why dwell time alone cannot be the full metric. A system can retain matter for a long time because the flux is small, not because it is a powerful planetary dissipative pathway.

7.3 High power, long dwell time

Tropical forests, mangroves, and other high-biomass productive ecosystems combine high NPP with large standing stocks and layered food, detrital, root, fungal, and microbial networks.

• high \(F\)
• high \(B/\text{NPP}\), long \(D\)
• high assembly flux \(\Phi_A = F\bar{A}\)

These are the strongest biospheric examples of dissipative path assembly. They process substantial solar-derived power while keeping it resident in complex biological structures for long periods.

7.4 Low power, short dwell time

Strongly degraded systems can have both low NPP and low biomass retention.

• low \(F\)
• low \(B/\text{NPP}\), short \(D\)
• low assembly flux \(\Phi_A = F\bar{A}\)

Such systems are thermodynamically simple in this framework. They neither capture much free energy nor maintain long biological pathways.

---

9. Relation to entropy production

Maximum entropy production, or MEP³¹, is often stated as the idea that nonequilibrium systems³² tend to organize so as to degrade imposed gradients as effectively as constraints allow. In Earth-system language, life helps dissipate the solar gradient by coupling radiation, water, carbon, nutrients, and surface chemistry.

³¹ Maximum entropy production is a proposed non-equilibrium thermodynamic principle: that open, driven systems tend to evolve towards states that dissipate their driving gradients as fast as their internal constraints allow. In Earth-system science, Kleidon and others have argued that climate, hydrology, and life can all be understood partly through MEP. The principle is controversial — it is not a law with a rigorous derivation from first principles (unlike the second law of thermodynamics), but it has predictive power in several systems. The framework here deliberately does not depend on MEP being a strict law; it uses MEP as context while making its own, more conservative claims.

³² A nonequilibrium system is one that is not in thermodynamic equilibrium — it has ongoing flows of energy or matter through it, and gradients (of temperature, concentration, pressure, chemical potential) that have not been flattened by dissipation. The Earth-biosphere system is quintessentially nonequilibrium: sunlight constantly drives it away from the equilibrium state it would reach in the dark (a cold, oxidised, lifeless rock). Equilibrium thermodynamics (the familiar entropy-maximisation picture) applies to closed systems tending toward rest; nonequilibrium thermodynamics is needed to describe systems like ecosystems, where structure is maintained by throughput.

The framework here should not be phrased as a naive identity:

MEP = maximum biomass

or:

MEP = maximum assembly index

Those statements are too simple. Entropy production is a rate. Biomass is a stock. Assembly index is a construction-depth metric. Dwell time is a residence time. These quantities are related, but they are not interchangeable.

The stronger version is:

For the NPP-derived part of the Earth system, biological entropy production can be characterized by both a power term and a path-structure term.

The power term is \(F\) (NPP flux). The path-structure term is \(\bar{A}\) or \(D\). The combined quantity is:

\[\Phi_A = F\bar{A}\]

or, if the assembly-dwell relationship holds:

\[\Phi_D = F D\]

This says that a biosphere is not only dissipating watts. It is dissipating watts through historically constructed pathways of finite depth and duration.

From this point of view, a tropical forest is not thermodynamically interesting only because it has high NPP. It is interesting because high flux is coupled to long residence paths, layered biological memory, and high-copy molecular assembly. The forest is a structured delay line³³ for solar exergy³⁴.

³³ A delay line is an electrical engineering term for a circuit element that transmits a signal with a time lag — it holds the signal in transit rather than passing it through instantly. The metaphor here is precise: the forest does not destroy solar exergy immediately upon capture, it routes it through a series of biological structures that hold it for years to centuries before the energy finally exits as heat. The forest is, in this sense, a very slow, very complex delay line for the solar energy signal.

³⁴ Exergy is the maximum useful work that can be extracted from a system as it comes to equilibrium with its surroundings. It is sometimes called “available work” or “work potential.” Unlike energy (which is conserved), exergy is destroyed by irreversible processes — it is the quantity that entropy production consumes. A photon from the sun carries a lot of exergy because it is far from equilibrium with the cool Earth environment; a longwave infrared photon emitted by the Earth carries much less exergy. Biological processes spend exergy to build and maintain structure.

---

10. The food web is not just a trophic ladder

It is tempting to picture the system as a simple chain:

plants -> herbivores -> carnivores -> megafauna

That picture is useful pedagogically, but it is incomplete. In many ecosystems, most energy does not travel upward through large animals. It passes through:

roots
mycorrhizae
wood
leaf litter
dissolved organic matter
soil organic matter
microbial biomass
fungal networks
sediments
detrital food webs

These detrital³⁵ and microbial pathways may dominate dwell time. A whale, shark, or large predator is a spectacular high-level structure, but a forest soil or mangrove sediment may be more important for the mean entropic dwell time of the system.

³⁵ Detritus (from Latin deterere, to wear away) is dead organic material: fallen leaves, dead wood, animal carcasses, faeces, shed skin, root exudates, dead microbial cells. The detrital food web (or decomposer food web) is the network of organisms that feed on this dead matter — bacteria, fungi, mites, springtails, earthworms, beetles, and many others. In most ecosystems, the detrital pathway carries more energy than the “grazing” pathway through herbivores and carnivores: the majority of NPP dies without being eaten alive.

³⁶ A graph in mathematics is a set of nodes (vertices) connected by edges. The food-web-as-graph perspective treats each biological pool as a node and each energy or carbon flow as a directed edge. Graph theory provides tools for analysing path lengths, loop structures, connectivity, and flow distributions that are not available when thinking of ecosystems as simple ladders or chains. The mean first-passage time D is a graph-theoretic quantity: it averages over all paths through the network from photosynthetic entry to respiratory exit.

The framework should therefore be graph-based³⁶, not ladder-based. Each node is a biological or biogeochemical pool; each edge is a flux; each node has a residence time and an assembly-index distribution.

In graph language:

Element	Role
nodes	biological pools or molecular compartments
edges	carbon/energy fluxes between pools
exit	respiration to CO₂, H₂O, and heat
\(D\)	mean first-passage time to exit
\(\bar{A}\)	mean assembly index along paths to exit

This graph formulation is especially useful for the biosphere because it can handle food webs, soils, sediments, dissolved organic matter, and microbial loops in one language.

---

11. What mass spectrometry would measure

The mass-spectrometry program should not try to assign one assembly index to an entire ecosystem. It should measure distributions.

For each sampled compartment i, mass spectrometry can estimate a distribution:

P_i(a) = abundance distribution of molecular assembly indices

From that distribution one can compute:

a_i = mean assembly index of compartment i

or, more cautiously:

a_i = mean assembly index of the detectable molecular fraction of compartment i

The ecosystem-level quantity then requires flux weighting:

\[\bar{A} = \sum_i f_i a_i\]

This requires ecology as well as chemistry. The mass spectrometer gives assembly distributions. Field measurements and ecosystem models give biomass, NPP, turnover times, respiration rates, grazing rates, litter fluxes, soil carbon residence times, and trophic transfer efficiencies³⁷.

³⁷ Trophic transfer efficiency is the fraction of energy (or carbon) at one trophic level that is successfully incorporated into the next level up. Classic ecological estimates put it around 10% per trophic step — a rough rule that means large predators are rare, because so little of the original primary production makes it that far up the chain. Most NPP is respired by the organisms that consume it directly, excreted, or passed into the detrital pool. The 10% figure varies substantially across ecosystems and trophic links.

The clean experimental design is:

1. Choose a well-characterized ecosystem with known carbon and energy budgets.
2. Define compartments in the ecological flux model.
3. Measure molecular assembly distributions in those compartments.
4. Estimate residence times and flux fractions for the same compartments.
5. Test whether \(a_i\) predicts \(\tau_i\) at the pool level.
6. Compute \(\bar{A}\), \(D\), and \(\Phi_A = F\bar{A}\).

The decisive test is not whether every high-assembly molecule is long-lived. The decisive test is whether flux through high-assembly biological pools systematically increases the mean dwell time of NPP-derived free energy.

---

12. What would count as evidence

The framework makes several concrete predictions.

Prediction 1: pool-level assembly tracks residence time

Across ecological compartments, the mean assembly index of biological matter should correlate with mean residence time after controlling for obvious confounders³⁸ such as temperature, moisture, oxygen, mineral protection, and nutrient limitation.

³⁸ A confounder (in statistics and experimental design) is a variable that correlates with both the predictor and the outcome you’re trying to study, creating a spurious apparent relationship. Here: if cold temperature both slows decomposition (increasing residence time) and happens to correlate with ecosystems that have high-assembly biological molecules, then temperature would confound the assembly–residence relationship. The prediction is only interesting if assembly index predicts residence time over and above what temperature, moisture, oxygen, and other environmental controls already explain.

³⁹ A log-linear relationship between two variables \(x\) and \(y\) means that \(\log(y)\) is a linear function of \(x\) — equivalently, that \(y = e^{a + bx}\), so equal increments in \(x\) produce equal multiplicative changes in \(y\). Many natural processes are log-linear (or power-law) rather than linear: residence times in ecosystems span orders of magnitude (seconds for ATP, years for wood, millennia for peat), so a linear relationship with assembly index would be implausible across that full range. A log-linear model makes a more modest claim: each unit increase in assembly index multiplies the residence time by a constant factor.

The expected relationship may be log-linear³⁹ rather than linear:

\[\log \tau_i = \alpha + \beta\, a_i + \text{controls}\]

or a simple power law:

\[\tau_i = k\, a_i^\gamma\]

The exact scaling is empirical. The important claim is monotonic association at the pool level.

Prediction 2: high-biomass productive ecosystems have high assembly flux

Ecosystems such as tropical forests, mangroves, and peat-forming wetlands should have high \(\Phi_A = F\bar{A}\) because they combine high NPP with long residence paths and chemically rich biological pools.

Prediction 3: fast-turnover systems can have high NPP but low dwell depth

Open-ocean microbial systems may process large fluxes while maintaining low standing biomass and short residence times. They should rank lower in D and often lower in \(\bar{A}\) than their NPP alone would suggest.

Prediction 4: restoration can increase path length without merely raising NPP

Restoring trophic structure, detrital retention, root biomass, soil carbon, or marine nutrient recycling may increase \(D\) and \(\bar{A}\) even when NPP changes only modestly. This is important because it distinguishes dissipative path assembly from simple productivity.

---

13. The strongest version of the idea

The strongest version of the framework is not that life “maximizes complexity” in a vague sense. It is more specific:

The biosphere dissipates solar free energy through a network of molecular structures whose construction depth is measurable. The mean dwell time of NPP-derived free energy in that network is coupled to the flux-weighted assembly index of the matter through which the flux passes. Productive high-biomass ecosystems are therefore high-power, long-dwell, high-assembly dissipative pathways.

This gives a physical interpretation to ecosystem complexity:

complexity is not just diversity
complexity is not just biomass
complexity is not just NPP
complexity is not just molecular size

complexity is solar-derived flux routed through copied, historically
constructed molecular structures for finite residence times

The key quantities are:

Symbol	Meaning
\(F\)	NPP flux (power per unit area)
\(D\)	mean entropic dwell time
\(\bar{A}\)	flux-weighted mean assembly index
\(\Phi_A = F\bar{A}\)	assembly flux
\(\Phi_D = F D\)	dwell-weighted dissipative power

If \(\bar{A}\) and \(D\) are empirically coupled, then assembly theory supplies a chemical and historical basis for an ecological thermodynamic quantity. It turns “the biosphere as a dissipative structure” from a metaphor into a measurable program.

---

14. Why this is useful for talking about the biosphere

This framework is useful because it separates three things that are often blurred together.

First, it separates power from storage. NPP is a flow. Biomass is a stock. Their ratio is a residence time.

Second, it separates residence time from chemical construction depth. Long-lived carbon is not automatically high-assembly, and high-assembly matter is not automatically long-lived. The relationship has to be tested.

Third, it separates entropy production rate from dissipative pathway structure. Entropy production tells us that gradients are degraded. Dwell time and assembly index tell us how the degradation is internally organized.

This matters for comparing ecosystems. A hectare of ocean, grassland, forest, mangrove, peatland, cropland, and desert can be compared in at least five dimensions:

Dimension	Meaning
NPP flux \(F\)	how much free energy is captured per time
energy-rate density	how much free energy flow is concentrated per unit biomass or system mass
standing biomass \(B\)	how much biological stock is maintained
entropic dwell time \(D\)	how long captured energy remains organized
flux-weighted assembly \(\bar{A}\)	how deeply constructed the flux pathways are

Energy-rate density is especially useful when the comparison extends beyond ecology. It lets a bacterium, a tree, a mammal, a computer, a factory, and a city be placed on a common dissipative scale: not total power, but power concentrated through organized matter. Its weakness is that it does not say what the pathway is. A high value could mean rapid shallow throughput or deeply maintained organization. That is why it needs to be paired with dwell time and, where measurable, assembly index.

The framework also gives restoration ecology a thermodynamic language. Restoration is not merely the recovery of biomass or species counts. It can be described as the reconstruction of long-dwell, high-assembly dissipative pathways.

---

15. Minimal mathematical summary

Let an ecosystem receive NPP flux \(F\). Divide the biological flow network into compartments \(i\).

Flux fractions (the inequality allows for rapid respiration, unmeasured losses, or export outside the compartment set):

\[\sum_i f_i \leq 1\]

Mean entropic dwell time:

\[D = \sum_i f_i \tau_i\]

Flux-weighted mean assembly index:

\[\bar{A} = \sum_i f_i a_i\]

Assembly flux:

\[\Phi_A = F\bar{A}\]

Dwell-weighted dissipative power:

\[\Phi_D = F D\]

Empirical bridge hypothesis — \(\tau_i\) is some increasing function of \(a_i\), environment, and ecology:

\[\tau_i = g(a_i,\, \text{environment},\, \text{ecology}), \quad \frac{\partial \tau_i}{\partial a_i} > 0\]

at the level of biological pools, after accounting for environmental and ecological controls.

If this bridge holds approximately, then \(\bar{A} \propto D\) and \(\Phi_A \propto \Phi_D\), and assembly theory becomes a way of measuring the internal path structure of biospheric entropy production.

---

16. Main caveats

Several caveats are not peripheral; they are part of the framework.

First, individual molecules are the wrong level for the residence-time claim. The claim should be tested on flux-weighted pools. ATP and mRNA are high-assembly but short-lived because biology uses them as fast control and currency molecules.

Second, trophic level is not a reliable proxy for assembly depth. Plants, fungi, soils, sediments, and detrital pools can contain chemically rich and long-lived matter. The graph matters more than the ladder.

Third, high residence time can arise from environmental protection rather than intrinsic molecular assembly. Cold, dry, anoxic, acidic, or mineral-protected conditions can slow decomposition. These controls must be included rather than ignored.

Fourth, copy number matters. Assembly theory interprets high assembly index differently when many copies exist. Ecosystem-scale measurements should therefore combine assembly index with abundance, biomass, and flux.

Fifth, the NPP-derived subsystem is only one part of planetary entropy production. The atmosphere, ocean circulation, hydrological cycle, and surface radiation field dissipate vastly more power than biological NPP. The claim here is about the structured biological channel inside the larger Earth-system energy budget.

---

17. Suggested discussion framing

For a physics audience, the cleanest framing is:

1. Start with the Earth as a nonequilibrium radiative system.
2. Identify NPP as the biological free-energy input.
3. Treat the ecosystem as a dissipative network with residence times.
4. Define entropic dwell time D as a flux-weighted mean first-passage time to respiration.
5. Introduce assembly index as a measurable construction-depth statistic for molecules.
6. Define \(\bar{A}\) as the assembly index experienced by NPP-derived flux.
7. State the empirical bridge: high-assembly biological pools tend to extend dwell time.
8. Test the bridge with paired mass-spectrometry and ecosystem-flux measurements.

The shortest version is:

NPP is solar free energy entering biological matter. Ecosystems differ in how long that free energy dwells before respiration and in how deeply assembled the molecular structures are along the way. Dissipative path assembly proposes that these two quantities are coupled and jointly measurable.

---

18. Key references

The starting reference set is the same as in framework_note.md:

Topic	Reference
Assembly theory	Sharma, Walker, Cronin et al. (2023), Nature 622:322
Empirical molecular assembly by MS	Marshall et al. (2021), Nature Communications 12:3033
Biosphere biomass census	Bar-On, Phillips & Milo (2018), PNAS 115:6506
Anthropogenic mass	Elhacham et al. (2020), Nature 588:442
Life as gradient dissipation	Schneider & Kay (1994)
Earth-system free energy and MEP	Kleidon (2012)
Energy rate density and complexity	Chaisson (2010)
Global NPP	Field et al. (1998)

---