The Cosmological Frame

Volume VII — Planetary and Cosmic

THE PLATE

An observatory dome on a lodge at dusk, its windows lit, red cloud over the mountains. — Plate VII.10The Aperture, Facing Up.Nobody took the heat away. The tray simply had an unobstructed view of something very cold and very far off, and that was enough.

THE LETTER

This chapter is the largest frame in the edition and it is the one to hold most carefully, because the largest frame is where a book of this kind usually goes soft. The move is familiar: invoke the heat death of the universe, gesture at the Sun, say the word entropy in a tone of voice, and let the reader supply the argument. We are not going to do that. Every sentence here is going to be paid for in physics or in arithmetic, and where the frame cannot pay, this chapter will say so before you notice.

Let us start with the concession, since it decides how you should read the rest. Nothing in this chapter will change a decision you make this century. The thermodynamic limits on a planetary civilisation sit two to six centuries out at any plausible rate of growth, and at a three percent discount rate a certain payoff two hundred and thirty-one years from now has a present value of 0.000978 — one part in a thousand of itself. No committee has ever been moved by that and none should be. If you came for an operational instrument, the last movement has one, and it is about a roof.

So what is the frame for, if it does not bind?

It is for making two words mean something specific. This edition uses growth and limit on almost every page, and those words are rhetorical until somebody says what is growing and against what it presses. The cosmological frame answers both, exactly: what grows is the rate at which a system captures low-entropy energy and disposes of high-entropy waste, and what it presses against is not the supply of energy — which is effectively infinite — but the capacity of the sky to carry the waste away. That is a physical statement with numbers attached, and once you have it, a great many arguments you have heard about growth and limits resolve into arguments about which term somebody meant.

You will also get the reference frame for everything Volume IV says about regeneration. Regeneration is not a moral posture. It is a thing a planet is physically permitted to do because it is an open system — the one fact underneath every claim this edition makes — and this is the chapter where that permission is written out with the numbers in it.

— The Editors

DISCOVERY

What is already working

Begin where the frame has already paid: with the places where somebody treated the sky as an asset and got something real back.

The yakhchāl, and ice made above freezing. In the arid plateau of Iran, and later across northern India, ice was manufactured commercially for centuries without refrigeration and without the air ever dropping below zero. The method was shallow, thermally isolated trays of water, shaded by day and left open to a clear night sky, often with a straw-insulated bed beneath and a low wall to block the breeze. On a dry, cloudless night a surface with a clear view upward radiates into the atmospheric transparency window — the band between roughly eight and thirteen micrometres where the atmosphere is nearly invisible to infrared — and loses heat to a sky whose effective temperature is tens of degrees below the air touching it. Mehdi Bahadori's 1978 survey of Iranian passive cooling documents the architecture in detail. The operators had no thermodynamics and no word for entropy. They had noticed that a view of the sky is a resource, and they built a business on it.

The planet's energy accounts, now actually measured. For most of the period in which people argued about planetary limits, the planetary balance sheet was an estimate. It is no longer. NASA's CERES instruments measure the radiation leaving the top of the atmosphere; the Argo float array measures the heat accumulating in the ocean, which is where more than ninety percent of any imbalance ends up. These are two entirely independent instruments with no shared assumptions, and Norman Loeb and colleagues showed in 2021 that they agree — both showing Earth's energy imbalance roughly doubling across 2005 to 2019. Karina von Schuckmann's consortium assembled the same quantity from the heat inventory side and landed in the same place. A civilisation now keeps a double-entry account of its own planet's energy, verified by two methods that could have disagreed and did not. That is an achievement of the same kind as the invention of the balance sheet, and it happened in the working lifetime of people reading this.

The Montreal Protocol, which is the only completed planetary intervention we have. In 1987 a set of chemicals was identified as altering a global radiative property; the production of those chemicals was phased down by treaty; the atmospheric burden turned over and began falling; and the ozone column is now on a trajectory to recover. The WMO's 2022 assessment states the recovery timetable directly. The part less often reported is the energy side: the compounds involved were also potent greenhouse gases, and Rishav Goyal and colleagues estimate in 2019 that the Protocol avoided a substantial fraction of a degree of warming as a side effect nobody negotiated for. The precedent is not that it was easy. It is that the loop closed — intervention, measurement, verified reversal — at planetary scale, once, and the file is public.

Passive daytime radiative cooling, which is the yakhchāl with the physics written down. In 2014 Aaswath Raman, Shanhui Fan and colleagues published a photonic surface that reflects almost all incoming sunlight while emitting strongly in the eight-to-thirteen-micrometre window, and held it several degrees below ambient air temperature in direct midday sun, with no power input of any kind. The result is not a trick: it is a device that treats the cold of deep space as a thermodynamic reservoir it can reach through a hole in the atmosphere, and it works at noon. Commercial panels on this principle are now installed on supermarket and data-centre roofs, where they pre-cool condenser water and displace chiller electricity.

Four cases, four eras, one mechanism. Each found the same thing: the Sun is not the only half of the transaction, and the other half was free and unclaimed. Hold that, because it is where this chapter's arithmetic ends up.

THE ARITHMETIC

What works, what does not, and where the line sits

First, the budget. Exactly, with every term.

Total solar irradiance at Earth's mean orbital distance is 1,361 W/m², a figure tightened by Greg Kopp and Judith Lean in 2011 from the older 1,366. Earth presents a disc, not a sphere, to that beam:

  total solar irradiance                      1,361     W/m2
  Earth mean radius                       6,371,000     m
  disc area          pi r^2              1.2752e+14     m2
  intercepted        1,361 x disc        1.7355e+17     W   =  173,549 TW
  Bond albedo                                 0.294
  absorbed           x (1 - 0.294)       1.2253e+17     W   =  122,526 TW
  spread over the sphere  1,361 / 4          340.25     W/m2
  absorbed per square metre                   240.2     W/m2

At equilibrium the planet radiates exactly what it absorbs. Invert Stefan-Boltzmann and you get the temperature it must radiate at to do so:

  T_eff = (240.2 / 5.670374419e-8)^0.25  =  255.1 K  =  -18.0 C

Two numbers now stand next to each other and everything else in this chapter falls out of the gap between them. Energy arrives from a source at 5,772 K. The identical quantity of energy leaves at 255.1 K. Nothing is consumed. The books balance to the watt. What changed is the temperature at which the joules are carried, and that is the only thing that ever changes.

Second, the entropy export, which is what the planet is actually doing.

Radiation carries entropy as well as energy — for a blackbody flux, four-thirds of the energy flux divided by the temperature. So:

  entropy in    (4/3) x P / 5,772 K       2.8304e+13   W/K
  entropy out   (4/3) x P / 255.1 K       6.4035e+14   W/K
  ------------------------------------------------------------
  net export                              6.1205e+14   W/K   = 6.12 x 10^14
  on the bare Q/T convention                                   4.59 x 10^14
  ratio out / in  =  5,772 / 255.1              22.6   x
  per square metre of planet                    1.20   W/(m2 K)

The four-thirds cancels in the ratio, so the 22.6 holds whichever convention you were taught. Read it as photons and it becomes physical: mean photon energy scales with temperature, so one photon in at visible wavelengths leaves as about twenty-three photons in the infrared. The same energy, spread across twenty-three times as many degrees of freedom. That spreading is the entropy export, and it runs at six hundred and twelve trillion watts per kelvin, continuously, and has since the planet formed.

This is the physical fact underneath every claim this edition makes about regeneration, and it is worth stating without ornament. A closed system cannot build order; its entropy can only rise. Earth is not closed. It sits in a steady stream of low-entropy radiation and dumps high-entropy radiation into a sky that is effectively at three kelvin, and the difference between those two flows is a continuous entropy budget that terrestrial order is paid for out of. Forests, soils, coral, cities, balance sheets and the sentence you are reading are all local decreases in entropy funded from that account. Erwin Schrödinger named it in 1944 and called what an organism eats negative entropy; Ludwig Boltzmann had put it more plainly in 1886, saying that the struggle for existence is not a struggle for energy, which is everywhere, but for available energy. Axel Kleidon's work is the modern accounting of it, and it is where a reader who wants the full planetary entropy budget should go next. Chapter II.06 does the industrial-scale version of this argument — exergy, the Carnot factor, Gouy-Stodola — and it is not repeated here.

Third, what actually gets taken out of that flux, which is less than anyone guesses.

  global net primary production               104.9   Pg C/yr
  as dry biomass, at 0.45 kg C per kg      2.3311e+14  kg/yr
  at 17.5 MJ/kg                            4.0794e+21  J/yr
  the entire biosphere                          129   TW
  human primary energy, 2023, 620 EJ           19.6   TW
  ---------------------------------------------------------
  biosphere as a share of absorbed flux       0.106   %
  humanity as a share of absorbed flux       0.0160   %
  humanity as a share of the biosphere         15.2   %

Every living thing on Earth, added together, captures about a tenth of one percent of the sunlight the planet absorbs. Photosynthesis is a poor engine and has always been. Humanity runs at fifteen percent of the whole biosphere's throughput and at sixteen thousandths of one percent of the incoming flux. The ratio of flux to human consumption is Chapter I.01's 9,105 times and is cited here rather than re-derived.

Fourth: Landauer's limit, and the floor under computation.

Rolf Landauer showed in 1961 that erasing one bit of information in an environment at temperature T must dissipate at least k_B · T · ln 2. Not as an engineering difficulty — as a consequence of the second law, because erasure reduces the number of accessible states and that reduction has to be paid for somewhere. Charles Bennett's 1982 review established the complement: computation that is logically reversible has no lower bound, so it is the discarding of information that costs, not the computing. Antoine Bérut and colleagues measured the limit directly in 2012 and found the predicted floor.

  k_B T ln2  at 300 K                          2.87   zJ per bit erased
  k_B T ln2  at the planet's 255.1 K           2.44   zJ
  data centre electricity, 2024                 415   TWh
    as continuous power                        47.3   GW
    as a share of world primary energy         0.24   %
  bit erasures that budget would fund at the floor
    per second                             1.6490e+31  bits

Now the headroom. Take the most efficient large machine actually built — the Green500 leader of June 2024, at 72.7 GFLOP per watt — and assume, as an order of magnitude, a thousand elementary bit operations per floating-point operation:

  energy per FLOP                          1.3755e-11  J
  energy per elementary bit operation      1.3755e-14  J
  headroom above the Landauer floor         4,791,104  x   =  6.68 orders
  as halvings                                    22.2
  at Koomey's post-2000 rate of 2.6 yr per halving      58  yr
  at the pre-2000 rate of 1.57 yr per halving           35  yr

The thousand-operations assumption is doing real work and the chapter will not hide it: move it to a hundred or to ten thousand and the headroom moves by an order of magnitude and the sixty years becomes forty or seventy-five. What does not move is the shape. Jonathan Koomey's series shows energy per computation halving every 1.57 years to 2000 and every 2.6 years since. Somewhere around the middle of the next century, on any of those assumptions, the free lunch that has funded the entire information economy runs out, and after that every additional computation costs strictly more energy than the last one did. That is the first hard economic consequence in this chapter and it is inside the planning horizon of institutions that already exist.

Fifth: Kardashev, read as accounting rather than as a ranking.

Nikolai Kardashev proposed his scale in 1964 to classify the radio power a civilisation might use for interstellar signalling; Carl Sagan gave it a continuous form, K = (log₁₀ P − 6) / 10 with power in watts. It has since been flattened into a league table, which is a waste of a good instrument. Taken as accounting it says something precise: each rung names where a quantity of waste heat must be put.

  humanity, 19.6 TW                     K =     0.729
  Sagan Type I, 10^16 W                 K =     1.0
  the whole Sun, 3.828e26 W             K =     2.058
  the whole galaxy, ~7.7e36 W           K =     3.089
  factor from here to Type I                     509   x
  Type I as a share of absorbed flux            8.16   %

Humanity stands at 0.729. Note what the last line says. A Sagan Type I civilisation dissipates eight percent of the entire absorbed solar flux, on the surface of a planet whose whole radiative budget is that flux. Which is the cut.

Sixth — and this is the move the chapter exists for. The scarce thing is not the energy. It is the cold.

At equilibrium, temperature goes as the fourth root of power, so about the present point, dT/T = (1/4)(dP/P), and T_eff / 4 is 63.78 K. Every watt a civilisation dissipates — whatever its source, fossil, fission, fusion or photovoltaic — must ultimately leave as infrared, and the planet can only raise its output by raising its temperature.

  warming from today's 19.6 TW                 0.010   K
  dissipation that adds one whole kelvin       1,921   TW
  warming at Sagan Type I, 10,000 TW             5.2   K
  T_eff if the absorbed flux were doubled      303.4   K

  years from 19.6 TW at   1.0 %/yr   ->  1,921 TW in 461 yr · 122,500 TW in 878 yr
  years from 19.6 TW at   2.0 %/yr   ->  1,921 TW in 231 yr · 122,500 TW in 441 yr
  years from 19.6 TW at   2.3 %/yr   ->  1,921 TW in 202 yr · 122,500 TW in 384 yr
  years from 19.6 TW at   3.0 %/yr   ->  1,921 TW in 155 yr · 122,500 TW in 296 yr

Read the middle line. At two percent annual growth in energy use, a civilisation adds a kelvin of pure thermodynamic warming — with no greenhouse gas involved whatsoever, with a perfectly clean grid — in two hundred and thirty-one years. It reaches the full absorbed solar flux in four hundred and forty-one. Tom Murphy put this arithmetic into Nature Physics in 2022 and it has not been refuted, because there is nothing in it to refute; it is Stefan-Boltzmann and a compound interest table.

So the terms of the long-run budget constraint, each with the horizon over which it binds:

Term	Binds at	Horizon at 2 %/yr
Waste heat into a 255 K sky	1,921 TW for +1 K	231 yr
The solar flux itself	122,526 TW	441 yr
Efficiency headroom in computation	Landauer floor	58 yr
The Sun's whole output	3.828 × 10²⁶ W	1,545 yr
The Sun's main-sequence life	—	~5 × 10⁹ yr

The source is not the constraint. The sink is, and it binds first by two hundred years. Sunlight is not scarce; night is. The asset a planetary economy is actually short of is a clear view of a cold sky, and the whole of the ordering above is what makes the word limit mean a specific thing rather than a mood. This is also why Volume VII's chapter on space and materials matters: a ladder that runs past Type I cannot be climbed on a planetary surface at all, because the rungs are defined by heat rejection and the surface has only one sky.

Seventh: where this fails, stated before anyone else states it.

It does not bind a decision, and here is the factor. At a three percent discount rate, a certain payoff at the two-hundred-and-thirty-one-year threshold is worth 0.000978 of itself; at one percent, 0.0993; and at three percent the present value of anything falls below one part in a million at 461 years. Chapter VII.08 does the discount rate properly and this chapter borrows only the factor — but the factor is the whole negative. No instrument, covenant, budget or board paper written this century should contain a cosmological term, and a chapter that pretended otherwise would be asking you to misallocate capital on the strength of a nice sentence.

The frame is smaller than the error bar on the thing it would manage. Earth's energy imbalance is about 0.9 W/m², with a systematic uncertainty on the absolute value of roughly 0.4 W/m². Over the planet's 5.1006e+14 square metres:

  Earth's energy imbalance                      459   TW
  uncertainty on that number                    204   TW
  the entire human energy system               19.6   TW
  imbalance / human power                      23.4   x
  uncertainty / human power                    10.4   x

The whole human energy system is ten times smaller than the error bar on the planetary budget, and twenty-three times smaller than the imbalance itself. Which also settles a confusion worth settling: today's climate problem is not a waste-heat problem. Humanity's direct dissipation warms the planet by 0.010 K and will for centuries. The forcing is 23.4 times larger because it is not about energy released, it is about radiative gating — a change in what the sky is permitted to let out. The entropy frame is what makes that distinction crisp, and it is also what tells you the frame has nothing to add to the measurement.

It cannot price anything. An entropy account ranks no two goods. A gigajoule spent on antibiotics and a gigajoule spent on a lottery advertisement are thermodynamically identical and economically not comparable, and nothing in this chapter closes that gap. Thermodynamics sets the boundary of the feasible set and says nothing whatsoever about which point inside it to choose. Anyone who tells you the second law implies a policy has stopped doing physics.

Eighth, then, the term that is actually economic — and it is the only one policy moves.

  world GDP, 2023                       1.054e+14   USD
  world primary energy, 2023                  620   EJ
  value per gigajoule                         170   USD/GJ
  value per kilowatt-hour                    0.61   USD/kWh
  energy intensity improvement, recent        1.3   %/yr
  long-run real output growth, assumed        2.0   %/yr
  implied growth in energy use                0.7   %/yr
  years to the one-kelvin threshold at 0.7 %/yr      657 yr

That is the budget constraint written the way an economist can use it. Output growth minus energy-intensity improvement equals energy growth, and energy growth is the term the sky charges for. At the recent intensity trend the horizon stretches from two hundred and thirty-one years to six hundred and fifty-seven. Raise the intensity term to two percent a year and energy use stops growing altogether and the constraint never binds at all. That gap — two percent wanted against one point three delivered, a gap of seven-tenths of a point a year — is the single most consequential number in this chapter, and unlike everything else in it, it is a number that a firm, a grid operator, a building code or a procurement policy actually moves.

DREAM

What becomes ordinary

In the version of this that has already happened, an ordinary building has two thermal connections rather than one, and both are on the drawings. It takes heat in from the Sun, as buildings always have, and it has a designed, maintained, unobstructed aperture facing the sky through which it puts heat out. The aperture is a plant item with a service interval and a line in the asset register, and nobody thinks this is unusual any more than they think a flue is unusual.

Property law has caught up to it. A right to sky view is a recorded interest, conveyed and priced the way light, air and water rights already are in every jurisdiction that has ever had a tall building. When a neighbouring development would occlude a cooling aperture, that is a negotiation over a valued easement rather than a surprise. Developers price it into the land. It is not a new kind of right; it is the oldest kind, finally attached to the correct physical quantity.

Energy accounts carry a second column. Beside the kilowatt-hours consumed sits the grade at which they were consumed and the temperature at which they left, so that a facility manager can see at a glance that the building is dumping six-hundred-degree combustion heat to do a twenty-degree job — an observation that is currently available to nobody because the number is not kept. Chapter II.06 turns that column into a management instrument; here it is simply part of the furniture.

And the vocabulary has separated. When somebody says the economy is growing, everybody in the room knows to ask which of three quantities is meant — value, throughput, or dissipation — because they now routinely move in different directions. Value per gigajoule rising while gigajoules stay flat is the normal case rather than a contested claim, and a plan that proposes growth in value is not automatically suspected of proposing growth in heat.

None of this requires a discovery. The physics has been settled since 1884, the photonic materials have been in the literature since 2014, and the property instruments have existed since the first English window-light case. It requires that somebody put the aperture on the drawing and the right in the register.

DESIGN

The structure that gets there

Four things, in order, and the order matters because each one makes the next one cheap.

One: make the sink a named quantity. Every thermal system already has a rejection path and almost none of them have it written down as a specification. The design change is a single line in the energy model: at what temperature does this system reject heat, and to what. Once that line exists, the difference between rejecting to a thirty-five degree afternoon air mass and rejecting to a sky at minus fifteen becomes visible as what it is — a very large free variable sitting unmanaged in the middle of the plant.

Two: give the aperture a legal existence. Cooling by radiation depends on an unobstructed upward view, which is exactly the class of thing that gets built over by somebody who owed you nothing. The instrument is not new. Easements of light and air, solar access laws, restrictive covenants and airspace rights are all mature, litigated and conveyable. A sky-access easement is a solar easement pointed the other way, and it can be drafted from the same precedent by any competent property lawyer in an afternoon. Do this before installing the aperture, never after, for the same reason you agree a baseline before a pilot.

Three: buy the capacity, do not buy the panel. The reason radiative cooling sits in pilot projects rather than in portfolios is that it is sold as a capital item to buyers whose constraint is capital. Sold as displaced chiller kilowatt-hours under an availability contract, it competes against the electricity tariff instead of against the capital plan, and it meets an entirely different committee. The structure is in the last movement.

Four: govern the growth term where it is actually decided. The one-kelvin threshold at 1,921 TW is not a policy target and should never be written into one; the horizon is six to seven centuries at present trends and the discount factor empties it. What is decidable is the intensity term — the seven-tenths of a point a year between two percent output growth and one point three percent intensity improvement. That is decided in building codes, motor and transformer standards, process heat integration, data-centre siting and the ordinary industrial engineering of not throwing away work. The cosmological frame's only operational instruction is: work on the denominator, and it is the same instruction you would have got from the quarterly numbers. That agreement is a feature. A frame this large that contradicted the near-term arithmetic would be telling you something was wrong with the frame.

The sequencing exists because each step de-risks the next. Naming the sink costs an hour and produces the evidence for the easement; the easement costs a conveyance and makes the capacity contract financeable; the contract produces the metered data that turns the intensity term from an aspiration into a series.

DESTINY

How it holds when nobody is pushing

This holds for an unusual reason: it is the only part of this edition that cannot become obsolete. Stefan-Boltzmann will not be revised. The Landauer bound is a theorem about counting states, not an engineering estimate. A reader opening this chapter in two hundred years will find the constants unchanged and the arithmetic still running, which is a durability that no policy chapter in any volume can claim.

What sustains the practice is narrower and more fragile, and it is worth naming precisely. The sky-access easement holds because it is recorded against title, survives every change of ownership, and costs nothing to maintain. The aperture holds because it is on a service interval with a named owner. The intensity series holds because it is metered monthly and someone's number.

And here is where it fails.

It fails when somebody puts the cosmological term into the business case. A board paper that argues for a chiller retrofit on the grounds that civilisation reaches a thermal limit in the twenty-third century will be laughed out of the room, and deserves to be, and will take the good part of the argument with it. The horizon is the reason to know this, never the reason to fund anything.

It fails when the aperture is installed and the easement is not, and a development three years later takes the sky away with no recourse at all.

It fails when the frame is used as a licence to stop measuring. An entropy account that is not tied to a meter is a story, and a story with physics in it is more persuasive than a story without and no more true.

And it fails, most quietly, when the elegance of the frame becomes the point. The temptation in a chapter like this is to feel that having understood the planetary entropy budget one has done something. One has not. The budget was running before anyone understood it and will run afterwards; the only part that responds to attention is the seven-tenths of a point a year.

DELIGHT

What it feels like

There is a particular pleasure in the first clear night after you have understood this, and it is not awe — awe is cheap and the night sky has always given it away. It is closer to recognition, the way the interior of an engine looks after somebody has shown you where the work is done.

You look up and the darkness stops being empty. It is a hole in the atmosphere about five micrometres wide in wavelength, opening onto a reservoir at three kelvin, through which the entire planet is at that moment shedding six hundred and twelve trillion watts per kelvin of accumulated disorder — and every ordered thing beneath it, the forest, the coastline, the lit street, the particular arrangement of your own body, is being paid for out of that flow.

And then the small domestic pleasure that follows it: dew on a car roof in the morning when the air never reached the dew point, frost on a clear night that skipped the cloudy one, the cold that comes off a window at two in the morning. Those are not metaphors for the physics. They are the physics, at the scale of a windscreen, happening in your street every night, free, and nobody had ever told you what you were looking at.

OPERATIONALIZE THIS

At the level of finance

The frame does not finance anything. The sink does. Here is the instrument.

The structure: a sky-access easement plus a cooling-capacity availability contract.

A building owner grants a recorded easement over the upward view of a defined roof area. A provider installs and maintains radiative apertures on that area at its own cost. The apertures pre-cool the condenser water loop, displacing chiller electricity. The owner pays per metered kilowatt-hour-thermal of displaced load, at a fixed discount to the prevailing electricity tariff. No capital leaves the building owner and no argument about the cosmos appears anywhere in the document.

The mechanics.

Easement. Recorded against title, defined by roof polygon and a solid-angle obstruction limit, perpetual or matched to the contract term with an extension option. Drafted from solar-access precedent. This is the single most important page in the transaction; without it the asset can be shaded out by a neighbour.
Metering. Inlet and outlet temperature and flow on the loop, logged at fifteen-minute resolution, with an agreed method for imputing the displaced chiller electricity at the machine's own part-load curve. Agreed and signed before installation, by both the owner and whoever pays the electricity bill.
Payment. Per displaced kilowatt-hour-electric, at a stated discount to the tariff, with a floor and a cap so that neither side is exposed to a tariff excursion. Availability rather than capacity: no cooling, no payment.
Maintenance and soiling. The provider's obligation, with a measured performance floor below which payments abate. Soiling is the real operational risk on an upward-facing optical surface and it belongs to whoever can clean it.
Term. Matched to the aperture life, with a hard review at the point where cumulative payments equal installed cost.

The balance-sheet treatment. For the owner, an availability contract for a service, with no right to direct the use of an identified asset, is an executory service contract rather than a lease — expensed as incurred, no right-of-use asset, no lease liability. Draft it deliberately that way; a contract that hands the owner control of the panels drags a liability onto the balance sheet and loses the entire commercial point. The easement itself is an intangible with an indefinite life where perpetual. Get the lease-versus-service determination in writing from the auditors at term-sheet stage, not at year-end.

The counterparty. Large single-owner roofs with a continuous cooling load and a commercial tariff: data centres, cold stores, supermarkets, hospitals, pharmaceutical plants. Owner-occupied first, because a split incentive between landlord and tenant kills this faster than any physics.

The number that decides it. One figure, and it belongs on the front page.

  net radiative cooling power            40      W/m2          measured class
  useful duty cycle                      40      %             ASSUMED
  thermal yield  40 x 8,760 / 1,000 x 0.40   140.2  kWh-th/m2/yr
  displaced electricity at chiller COP 4    35.0  kWh-e/m2/yr
  installed aperture cost               100      USD/m2        ASSUMED
  hurdle rate                             7      %             ASSUMED

              hurdle x installed cost           0.07 x 100
   BREAK-EVEN TARIFF  = ----------------------- = ---------- =  0.20 USD/kWh
              displaced kWh-e per m2 per yr         35.0

At a commercial tariff of 0.25 USD/kWh the aperture returns 8.76 USD/m²/yr, a simple return of 8.8 percent and a payback of 11.4 years. Above roughly twenty cents a kilowatt-hour this clears an ordinary hurdle rate; below it, it does not, and no amount of thermodynamics changes that. Check your tariff first. If you are below the line, the honest answer is that this is not yet your instrument, and the chapter would rather tell you that than sell you a roof.

The first ninety days.

Day	Action	Artifact
1–15	Pull the tariff and the annual chiller kWh from the bill	One page: tariff, load, roof area
16–30	Compute the break-even tariff for your site and compare	The go / no-go number
31–45	Instrument the condenser loop; establish the baseline	The signed baseline
46–60	Draft and record the sky-access easement	The recorded easement
61–75	Term sheet; auditors confirm service, not lease	Accounting memo
76–90	Install a single-bay pilot; begin metering	Measurement log

APPRECIATIVE QUESTIONS

Twelve, for a room

Discovery — what is already working

Where in this organisation are we already using something that costs nothing and belongs to nobody — a slope, a prevailing wind, a cold night, a river — and who noticed it first?
Which of our systems rejects heat, and has anyone here ever seen the temperature it rejects at written down? What would it take to find out this week?
Think of a time we solved something by opening a path rather than adding power. What made that possible, and what did it teach us that we have not applied since?

Dream — what becomes possible

If every energy figure we report carried the grade it was used at as well as the quantity, what would we see on the first page that we cannot see now?
Imagine a version of our building or our plant that has a designed, maintained connection to the sky. What is it doing for us on a hot Tuesday afternoon?
If our board could distinguish growth in value from growth in throughput on a single slide, which of our current arguments would simply dissolve?

Design — what we build

What is the smallest place we could put a meter this month that would turn our energy intensity from an assertion into a series?
Which of our assets depends on a condition somebody else could remove without telling us — a view, an access, a right of way — and what would it cost to record it properly while relations are good?
Who would have to agree for a service contract here to be written as a service rather than a lease, and what do they need to see?

Destiny — how it holds

What would have to be true for the intensity number to still be metered and owned when everyone in this room has moved on?
Which of the things we understand well have we mistaken for things we have done? How would we tell the difference from the outside?
If someone opened our records in two hundred years, which of our numbers would still be checkable — and what would we have to write down now for that to be true?

WORKS CITED

Kopp, G. and Lean, J. L. (2011). "A New, Lower Value of Total Solar Irradiance: Evidence and Climate Significance." Geophysical Research Letters, 38, L01706.

Stephens, G. L., Li, J., Wild, M. et al. (2012). "An Update on Earth's Energy Balance in Light of the Latest Global Observations." Nature Geoscience, 5, 691–696.

Trenberth, K. E., Fasullo, J. T. and Kiehl, J. (2009). "Earth's Global Energy Budget." Bulletin of the American Meteorological Society, 90(3), 311–323.

Boltzmann, L. (1886). "Der zweite Hauptsatz der mechanischen Wärmetheorie." Lecture to the Imperial Academy of Sciences, Vienna; in Populäre Schriften (1905), Barth.

Schrödinger, E. (1944). What Is Life? The Physical Aspect of the Living Cell. Cambridge University Press.

Kleidon, A. (2012). "How Does the Earth System Generate and Maintain Thermodynamic Disequilibrium and What Does It Imply for the Future of the Planet?" Philosophical Transactions of the Royal Society A, 370, 1012–1040.

Kleidon, A. (2016). Thermodynamic Foundations of the Earth System. Cambridge University Press.

Field, C. B., Behrenfeld, M. J., Randerson, J. T. and Falkowski, P. (1998). "Primary Production of the Biosphere: Integrating Terrestrial and Oceanic Components." Science, 281(5374), 237–240.

Energy Institute (2024). Statistical Review of World Energy 2024. Energy Institute, London.

Loeb, N. G., Johnson, G. C., Thorsen, T. J. et al. (2021). "Satellite and Ocean Data Reveal Marked Increase in Earth's Heating Rate." Geophysical Research Letters, 48, e2021GL093047.

von Schuckmann, K., Minière, A., Gues, F. et al. (2023). "Heat Stored in the Earth System 1960–2020: Where Does the Energy Go?" Earth System Science Data, 15, 1675–1709.

World Meteorological Organization (2022). Scientific Assessment of Ozone Depletion: 2022. GAW Report No. 278, WMO, Geneva.

Goyal, R., England, M. H., Sen Gupta, A. and Jucker, M. (2019). "Reduction in Surface Climate Change Achieved by the 1987 Montreal Protocol." Environmental Research Letters, 14, 124041.

Bahadori, M. N. (1978). "Passive Cooling Systems in Iranian Architecture." Scientific American, 238(2), 144–154.

Raman, A. P., Anoma, M. A., Zhu, L., Rephaeli, E. and Fan, S. (2014). "Passive Radiative Cooling Below Ambient Air Temperature Under Direct Sunlight." Nature, 515, 540–544.

Landauer, R. (1961). "Irreversibility and Heat Generation in the Computing Process." IBM Journal of Research and Development, 5(3), 183–191.

Bennett, C. H. (1982). "The Thermodynamics of Computation — A Review." International Journal of Theoretical Physics, 21(12), 905–940.

Bérut, A., Arakelyan, A., Petrosyan, A., Ciliberto, S., Dillenschneider, R. and Lutz, E. (2012). "Experimental Verification of Landauer's Principle Linking Information and Thermodynamics." Nature, 483, 187–189.

Koomey, J. G., Berard, S., Sanchez, M. and Wong, H. (2011). "Implications of Historical Trends in the Electrical Efficiency of Computing." IEEE Annals of the History of Computing, 33(3), 46–54.

International Energy Agency (2025). Energy and AI. IEA, Paris.

Kardashev, N. S. (1964). "Transmission of Information by Extraterrestrial Civilizations." Soviet Astronomy, 8(2), 217–221.

Sagan, C. (1973). The Cosmic Connection: An Extraterrestrial Perspective. Anchor Press / Doubleday.

Dyson, F. J. (1960). "Search for Artificial Stellar Sources of Infrared Radiation." Science, 131(3414), 1667–1668.

Dyson, F. J. (1979). "Time Without End: Physics and Biology in an Open Universe." Reviews of Modern Physics, 51(3), 447–460.

Murphy, T. W. (2022). "Limits to Economic Growth." Nature Physics, 18, 844–847.

Krauss, L. M. and Starkman, G. D. (2000). "Life, the Universe, and Nothing: Life and Death in an Ever-Expanding Universe." The Astrophysical Journal, 531, 22–30.

Caldeira, K. and Kasting, J. F. (1992). "The Life Span of the Biosphere Revisited." Nature, 360, 721–723.

Lotka, A. J. (1922). "Contribution to the Energetics of Evolution." Proceedings of the National Academy of Sciences, 8(6), 147–151.

Ord, T. (2020). The Precipice: Existential Risk and the Future of Humanity. Bloomsbury.

Note on figures. Every figure in this chapter is computed in lib/verify/VII_10.py and printed with its inputs, units and sources by python3 lib/verify.py VII.10. The solar-flux-to-primary-energy ratio of 9,105× belongs to Chapter I.01 and is cited rather than re-derived; the exergy, Carnot and Gouy–Stodola apparatus belongs to Chapter II.06 and is not reproduced; the discount-rate argument belongs to Chapter VII.08 and only its present-value factors are computed here. The bit-operations-per-floating-point-operation factor, the radiative aperture's duty cycle, installed cost, chiller coefficient of performance and hurdle rate are stated assumptions, labelled as such in the module, and the text says on the page how far each result moves if they are wrong.