In August 2025, I published Post #190: “Arctic Vapor Plumes and Earth’s Energy Imbalance”. I argued that hydroelectric dams across the subarctic were injecting enough moisture into Arctic winter air to produce 17–29× the radiative forcing of CO₂ doubling — a missing piece of the Arctic amplification puzzle. I built a six-step energy balance across three scenarios and estimated 1.58 to 8.78 zettajoules per winter season.

Then I did what you’re supposed to do with a hypothesis. I tested it.

The paper is now published: Shahid & Yates (2026), “Observational Detection of Persistent Boundary Layer Modification by Subarctic Hydroelectric Dam Discharge,” ESSOar.

Then I did what you’re supposed to do with a hypothesis. I tested it.

The paper is now published: Shahid & Yates (2026), “Observational Detection of Persistent Boundary Layer Modification by Subarctic Hydroelectric Dam Discharge,” ESSOar.

Here’s what happened.

Subscribe now

The Hypothesis

Post #190 asked whether dam-generated vapor plumes could help explain the persistent gap between modeled and observed Arctic ice loss. The logic was straightforward:

1. Dams release warm water into freezing dry air.

2. The resulting vapor plumes trap outgoing longwave radiation.

3. The forcing is large — Sokolowsky et al. (2020) put it at 60–100 W/m².

4. Dozens of dams across the Arctic could produce a hemispheric-scale effect.

I used a coefficient from Doyle et al. (2011) — 12 W/m² per kg/m² of precipitable water to convert moisture anomalies into radiative forcing. The numbers were enormous. I was ecstatic.

Testing a Hypothesis Is Different From Writing a Blog Post

A blog post asks you to tell a compelling story. A test asks you to try to destroy the story.

So I flipped the approach. Instead of working backward from the conclusion, I worked forward from physics. The plan: derive predictions from first principles before looking at any observations. Then test each prediction with an independent dataset. If I’m right, the data should speak with one voice. If I’m wrong — and this is the important part — the data has dozens of ways to tell me so, and I need every one of them to fail before I claim a positive result.

Three dams on two continents. Twenty weather stations. Seven independent observational methods. Ten months.

Here is how that pipeline was built.

1. Energy balance; the physics predictor

So before we look at any measurement, can we predict what should happen?

If you leave a warm bath in a cold room, two things happen. The air above the water gets warmer. You can feel it on your hand. And the room gets damper. Steam rises invisibly and condenses on cold surfaces. That’s two channels for heat to escape the water: one you feel as warmth, one that hides inside water vapor. A warm river sitting in freezing air is the same thing, just 150 square kilometers of it.

A framework exists for computing exactly this. It’s called a bulk aerodynamic energy balance, and engineers have used it for decades to design cooling towers and weather models. For sensible heat: Q_H = ρ c_p C_H U (T_w − T_a). For latent heat: Q_L = ρ L_v C_E U (q_s − q_a). With stability-corrected transfer coefficients from Large & Pond (1982), mean DJF wind of 4.5 m/s, water temperature of 3.5°C, and air temperature of −16°C, the numbers fall out: Q_H = 192 W/m², Q_L = 107 W/m², total 299 W/m². Over 150 km² and 90 days of winter: 44.8 GW of thermal power — seven times the dam’s electrical output. The Bowen ratio β = Q_H/Q_L = 1.79 is the constraint that ends up mattering the most. Sensible heat dominates latent heat nearly 2 to 1.

Hold onto that number. It comes back.

2. ISD-Lite; the ground truth

Physics predicts +4°C downstream and +0.061 kg/m² of extra moisture. So how do you actually check? You need weather data going back decades. Where do you find it?

Every airport, weather office, and research station around the world has thermometers, wind sensors, humidity sensors, and cloud observers running around the clock. They write down what they see every hour. Someone has been collecting all of those records for decades and putting them into one big archive.

Such an archive exists. It’s called ISD-Lite, the Integrated Surface Database, maintained by NOAA’s National Centers for Environmental Information (Lott 2004). It provides quality-controlled hourly records as 8-element tuples (air temperature, dewpoint, sea-level pressure, wind direction and speed, total cloud cover, 1-hour and 6-hour precipitation) from over 14,000 stations worldwide, going back to 1901 at some sites. For the Krasnoyarsk analysis we pulled records from the 20 stations within 200 km of the dam that had at least 20 years of pre-dam and 20 years of post-dam coverage: 73 years of hourly data per station, roughly half a billion measurements.

3. Difference-in-Differences (isolating the dam’s effect)

But the Arctic has been warming anyway. If I just compare temperatures before and after the dam, I can’t tell what was the dam and what was climate change. So how do you separate one from the other?

Imagine you want to know if a new rule at school, “everyone must drink milk at lunch,” actually makes kids taller. You can’t just measure kids before and after the rule, because kids grow anyway. So you compare two classrooms: one that got the rule, one that didn’t. You measure both before and after. Then you subtract twice: the milk classroom’s growth, minus the other classroom’s growth. Whatever’s left over is what the milk actually did, separated from normal growth.

A technique exists for exactly this. It’s called Difference-in-Differences, or DID, and it comes from econometrics (Card & Krueger 1994). The formal estimator is DID = (ȳ_treated,post − ȳ_treated,pre) − (ȳ_control,post − ȳ_control,pre). For each of the 20 Krasnoyarsk stations we compute DJF climatologies for the pre-dam (1951–1971) and post-dam (1973–2024) epochs, then difference against an upstream control station at Abakan. We wind-condition (keeping only hours with wind from the dam toward the station) to isolate advected plume effects from ambient flow. Standard errors are block-bootstrapped to preserve autocorrelation. The key identifying assumption is parallel trends in the absence of treatment, which we probe by varying the pre-dam baseline window (1966–71, 1967–71, 1968–71) and showing the DID moves by less than 0.1°C.

4. IGRA radiosondes: the vertical structure

ISD-Lite stations measure the air at the ground. But plumes rise. If moisture is concentrated in the lowest 1 km of the atmosphere, the boundary layer, how do you see it?

Twice a day at thousands of sites worldwide, meteorologists release a helium balloon with a small sensor box attached. The balloon rises to 30 km, sending back temperature, humidity, and pressure readings all the way up. You end up with a vertical slice through the atmosphere at that moment.

Such a dataset exists. It’s called IGRA v2, the Integrated Global Radiosonde Archive (Durre et al. 2018). We used three stations around Krasnoyarsk: Enisejsk (278 km downstream, WMO 29263), Abakan (251 km upstream, control, WMO 29862), and Krasnoyarsk proper (37 km from the dam, WMO 29563). For each sounding we integrate specific humidity from 1000 to 925 hPa to get boundary-layer precipitable water. The result at Enisejsk: a DID of +0.514 kg/m² in the lowest kilometer after the dam was built (p < 0.001). The vertical structure is what matters. A 51% increase below 925 hPa, but only 23% in the 700–500 hPa layer. That’s the fingerprint of a surface source. If the moisture were being blown in from distant weather systems, the increase would be spread evenly through the column. It isn’t. It’s stacked at the bottom, right where the warm river is.

5. AIRS: moisture from orbit

Balloons only go up at specific stations. If we want to see moisture across the whole 300 km downstream corridor, not just points, how do we do that?

A NASA satellite called Aqua orbits 700 km above Earth. On board is an instrument that can “see” water vapor invisibly, by measuring the infrared light the atmosphere emits at specific wavelengths. Different amounts of water vapor give different infrared signatures. It’s thermal vision for steam.

That instrument is AIRS, the Atmospheric Infrared Sounder, a 2,378-channel grating spectrometer (Chahine et al. 2006). We used AIRS Level-2 support product V7 retrievals of column-integrated precipitable water at native 45-km footprint resolution, averaged to a 1° × 1° grid. For each DJF from 2002–2023, we computed the downstream-minus-upstream PW anomaly over the Krasnoyarsk corridor. The 22-year mean: ΔPW = +0.062 kg/m², positive in 18 of 22 individual years (p < 0.01 by binomial test). The physics-based energy balance predicted +0.061 kg/m². The match is within 2%.

6. ERA5: an independent verification

AIRS is a satellite retrieval with its own assumptions. What if the satellite is wrong in some systematic way we haven’t noticed?

Weather scientists around the world feed every measurement on Earth such as ships, airplanes, weather stations, balloons, satellites, into a giant computer simulation of the atmosphere. The simulation fills in the gaps between measurements and gives you, in effect, “the weather everywhere, every hour, since 1940.” That’s called a reanalysis. If our specific finding also appears in a completely different, independently-built product, it’s not a quirk of any one dataset.

Such a reanalysis exists. ERA5, produced by the European Centre for Medium-Range Weather Forecasts (Hersbach et al. 2020), assimilates roughly 24 million observations per day via 4D-Var into the IFS model, at 0.25° × 0.25° horizontal resolution with 137 vertical levels. Crucially, ERA5 does not assimilate Landsat thermal imagery or AIRS precipitable water directly, giving methodological independence from the satellite evidence chain. We extracted hourly TCWV, 2-m temperature and dewpoint, 10-m winds, and surface downwelling longwave for the 1940–2024 DJF climatology. Downstream ΔTCWV = +0.019 kg/m², directionally consistent with AIRS but smaller, exactly as you’d expect given that ERA5’s grid cells are wider than the 30-km plume itself.

7. CERES: can we see heat escaping to space?

If the corridor is genuinely warmer, it should be radiating more heat to space. Can we measure that directly from orbit?

NASA has another set of satellites carrying instruments specifically built to measure the energy balance of the planet: sunlight coming in, Earth’s heat going out, at every location, every month, for the past 20 years.

Those are the CERES instruments (Clouds and the Earth’s Radiant Energy System; Loeb et al. 2018), flying on Terra and Aqua. We used the EBAF Edition 4.2 Level-3 product for monthly DJF climatologies of top-of-atmosphere outgoing longwave radiation, surface downwelling longwave, and low-cloud fraction. Downstream low-cloud cover is +1.4% vs upstream, consistent with fog and stratus being generated by the plume. But that same cloud cover attenuates the direct surface-emission signal we’d otherwise see. The plume creates its own cloud blanket, which muffles the thermal signal from space. In a way, the dam is hiding its own evidence, which is itself evidence.

8. Landsat: can we see it with our eyes?

Everything so far has been inference, averages, statistics, model agreements. But can we actually see the plume? Not infer it. See it.

The Landsat satellites take thermal photographs of Earth at 100-meter resolution. That’s sharp enough to resolve a single parking lot from 700 km up. If you pull winter images of the Krasnoyarsk corridor, the warm river glows against frozen terrain.

These are Landsat 8 and 9 Collection 2 Level 2 Science Products, acquired via the Microsoft Planetary Computer. Cloud-free DJF scenes from 2015–2023 show a thermal anomaly of +10 to +25 K relative to the surrounding snow-covered landscape, extending 350–700 m wide along the channel and persisting at least 90 km downstream. And there’s a twist: close to the dam itself, the steam fog is so thick it actually hides the warm water from the satellite. Only 0.45 km² is detectable above 0°C, out of 150 km² known to be open water. The fog is opaque to thermal imaging. That’s not a measurement problem. It’s evidence. The fog exists because the evaporation is intense. The evidence is the obscuration.

9. HYSPLIT: where did the air actually come from?

OK. We have warming at stations downstream. We have moisture in the boundary layer. We have plumes visible from space. But how do we PROVE the air we’re measuring actually traveled over the dam? It could be that the air just happened to warm up because of some large-scale weather pattern.

Imagine you want to find where a smell in your house is coming from. You follow the air backward from your nose. Which window is open, which room has airflow. A computer can do this for the atmosphere at a continental scale. Given a location and a time, it can trace a parcel of air backward for 72 hours, showing you exactly where it was.

That program is HYSPLIT, the Hybrid Single-Particle Lagrangian Integrated Trajectory model (Stein et al. 2015), operated by NOAA’s Air Resources Laboratory, driven by GDAS1 meteorology. We computed 72-hour back-trajectories at three altitudes (500, 1000, 1500 m AGL) from the three IGRA stations, for five high-signal events and five control events, 30 trajectories total. The result was unexpected: 100% of trajectories in both groups passed within 50 km of the dam. Mean back-bearing ~219° for both groups, SSW, matching the reservoir geometry. The air always comes from the dam. What differs between a strong-signal day and a quiet day isn’t trajectory. It’s how active the river surface is at the moment the air passes over. The source is permanent. What varies is its intensity.

What I Found

The plumes are real. Unambiguously, measurably, visibly-from-space real.

Krasnoyarsk Dam maintains a 250-kilometer ice-free corridor through central Siberia in winter. The energy balance calculates 44.8 gigawatts of continuous thermal power, seven times the dam’s electrical capacity. The predicted moisture enhancement at 50 km downstream: 0.061 kg/m². AIRS satellite observed: 0.062 kg/m². The predicted temperature anomaly at 94 km: +3.9°C. Station Difference-in-Differences observed: +4.0°C.

Three predictions, three observations, three matches within rounding error.

Robert-Bourassa in Quebec replicated the signal on a different continent: +12.8 W/m² at 24 km (p = 0.0008). Churchill Falls, with only 6 km² of open water and the nearest station 196 km away, returned a clean null, exactly as the physics predicts. The method doesn’t produce false positives.

So the plumes exist, they’re persistent, they’re large, and they’re detectable by every instrument we pointed at them.

But Post #190’s central claim — that the moisture pathway drives a massive radiative forcing — was wrong.

10. Brutsaert: why my first answer was wrong (and a novel finding I didn’t set out to make)

Given all this, with the plume real, moisture elevated, and warming measurable, why was Post #190’s radiative forcing estimate still wrong by an order of magnitude?

I had used a “rule of thumb” for how much more ground-level warmth you get when there’s more water vapor overhead. The rule said: every 1 kg/m² of extra water vapor gives you an extra 12 W/m² of downwelling longwave radiation. But that rule came from measurements in the polar desert, the driest air on Earth, with precipitable water around 1 kg/m². Krasnoyarsk is not a polar desert. It has about four times as much water vapor as the place the rule was measured. When air already has a lot of water, adding more doesn’t do much, because the greenhouse effect of water vapor saturates. It’s like pouring more into a nearly full glass.

A correct framework for computing this exists. It’s the Brutsaert (1975) clear-sky emissivity formula: ε = 1.24 (e_a/T_a)^(1/7). It gives you a downwelling longwave sensitivity ∂DLW/∂PW that is a strongly nonlinear function of background water vapor. Doyle et al. (2011) derived ≈ 12 W/m²/(kg/m²) under polar-night conditions. I used that number. When I recomputed the sensitivity over Krasnoyarsk’s actual DJF phase space (temperatures between −25°C and −5°C, PW between 2 and 6 kg/m²), the coefficient drops to 5.4 W/m²/(kg/m²). At PW > 2 it falls below 8. At PW > 4, the Krasnoyarsk average, it falls to 5–6.

And then I went looking through the literature to double-check my work, and noticed something I hadn’t expected. The Doyle coefficient has been cited and applied in Arctic radiative studies for over a decade. But I couldn’t find a paper that had actually mapped its sensitivity across the subarctic PW/T phase space, the conditions where most of the Arctic research community is now doing its radiation work. The polar-desert value had quietly become a default, applied to environments it was never calibrated for.

That single miscalibration cut my estimated moisture forcing in half before the decomposition analysis even began. But fixing it meant producing a sensitivity surface across the Arctic DJF phase space, mapping ΔDLW per kg/m² as a function of background T and PW, that, as best I can tell, didn’t exist in the literature. The correction itself turned into a contribution. A wrong answer that forced me to do the work no one had quite gotten around to.

The Decomposition: The Part I Didn’t See Coming

The Brutsaert correction explains about half of my original overestimate. But when I separated the station signal into its temperature and moisture components, expecting to find maybe a 60/40 thermal-to-moisture split, given the Bowen ratio of 1.79, I instead found this:

At every significant downstream station. At all three dams. On two continents.

The dam warms the air faster than it humidifies it. The moisture injection is real, 6,400 kg of water vapor per second of evaporation, and it shows up everywhere you look. Fog frequency triples. Boundary-layer humidity jumps 44%. AIRS sees the column enhancement in 18 of 22 winters. But moisture expresses through fog and precipitation, not through enhanced downwelling longwave radiation.

The actual moisture-driven DLW increase? About 0.5 W/m². Not 60–100. Not 17–29× CO₂. Half a watt.

What the Paper Actually Says

The effect is real but spatially confined. The thermal plume decays to undetectable levels within ~300 km along the discharge corridor and negligible distances perpendicular to it. The affected area per dam, roughly 300 km by 30 km, is a corridor, not a hemisphere.

We inventoried 127 subarctic hydroelectric dams above 50°N using the GRanD database. Combined estimated winter thermal injection: 300–900 GW. That’s a large number, comparable in magnitude to global electricity consumption. But it’s distributed across thousands of kilometers of river corridors on three continents, with each plume confined to its own ~300 km strip. The total altered boundary-layer area is less than 1% of the Arctic domain.

Dam discharge plumes are an anthropogenic analog to a geothermal hot spring or urban heat island: concentrated thermal disturbances with significant local effects (fog, permafrost destabilization, ecosystem disruption, transport hazards) but negligible contribution to hemispheric Arctic amplification.

What I Learned

Post #190 worked backward from the conclusion I wanted. I started with the gap between modeled and observed ice loss, found a mechanism that could theoretically fill it, and applied a coefficient that made the numbers work. That’s not science. That’s storytelling with equations.

The paper worked forward from what I could measure. Energy balance first. Predictions before observations. Seven independent datasets, each capable of falsifying the hypothesis. Intercontinental replication. A null control site. Leave-one-out cross-validation on the plume model. Sensitivity analysis on every parameter.

The plumes exist. The forcing is thermal, not radiative. The scale is regional, not hemispheric. Post #190’s headline was wrong in the way that mattered most: the mechanism doesn’t scale the way I assumed.

But the question was worth asking, the investigation was worth doing, and the answer, that subarctic dams create measurable, persistent, previously undocumented boundary-layer modifications along hundreds of kilometers of river corridor, is a contribution to the literature that stands on its own.

The paper is here: Shahid & Yates (2026), ESSOar

The dam question was the right question. The first answer was the wrong answer. The second answer is the real one.

Code: https://github.com/R3GENESI5/arctic-dam-vapor-plumes
Zenodo DOI: 10.5281/zenodo.18893188

Share Regenesis

Leave a comment

Read more

Most climate models focus heavily on carbon dioxide and rising air temperatures. They track emissions, simulate atmosphere and ocean dynamics, and try to forecast where things are headed. But even with all that, there’s a consistent mismatch. The ice, especially around Greenland, is vanishing quicker than expected.

That gap made me pause.

I didn’t think the models were broken. But maybe they were missing something small but important. Something close to the surface.

So I asked a simple question: what have humans actually built in the Arctic in the last 50 or 60 years?

The answer: a lot of hydroelectric dams. Dozens of them, across cold northern rivers in Russia, Canada, and Scandinavia. These dams store water and then release it, often during winter, into freezing, dry Arctic air.

Now, Arctic air in winter is extremely dry. Scientists call it an “Arctic blue desert” because there’s barely any water vapor in it. It’s that dryness that helps keep the region cold and water vapor is one of the most powerful greenhouse agents, and even small amounts can trap heat.

When a dam releases warm water into that dry air, it creates vapor plumes, big ones, visible from space. That added moisture doesn’t just float away. It traps outgoing heat, slows cooling, and adds extra energy into the local atmosphere.

So I ran the numbers.

1 | The Data That Didn't Add Up

Let’s start with the elephant in the room.

Observed ice loss exceeds what global climate models predict by significant margins. The Arctic is warming 2-4 times faster than the global average, yet our best theoretical frameworks consistently underestimate the pace of change.

This discrepancy isn't small. As documented by Stroeve and Notz (2018), Arctic sea ice extent has been declining at approximately 13% per decade since satellite monitoring began, while climate models predicted much slower rates. The gap between observation and theory suggests missing forcing mechanisms in our understanding of Arctic climate dynamics.

The breakthrough came from an unexpected source: systematic analysis of atmospheric moisture intrusions documented in satellite data.

2 | The Sokolowsky Discovery: 60-100 W/m² of Missing Heat

The key observational foundation comes from Sokolowsky et al. (2020), who used satellite data to document persistent moisture intrusions into the Arctic during winter months. These events showed downwelling longwave radiation anomalies of 60-100 W/m² lasting for days to weeks - enormous amounts of heat being trapped in the atmosphere and redirected back to the surface.

To understand the significance of these numbers, consider that the global average radiative forcing from all human CO₂ emissions is approximately 3.5 W/m². These Arctic moisture events were delivering 17-29 times that intensity to specific regions during winter, when the Arctic should be losing heat to space most rapidly.

The crucial insight from atmospheric analysis revealed these moisture intrusions weren't random weather patterns. They originated from specific point sources: major hydroelectric installations in Arctic and subarctic regions.

3 | My Research Methodology: Building on the Sokolowsky Foundation

This is where I stepped in to connect the dots. The Sokolowsky observations provided the missing piece, direct satellite evidence of the energy magnitudes involved but, no one had systematically linked this to infrastructure development or quantified the cumulative climate effects. I developed a comprehensive six-step energy balance methodology to bridge this gap, treating the Sokolowsky range as the observational anchor for what became the first quantitative assessment of dam-induced climate forcing.

My approach started with a fundamental systems question:

if individual moisture intrusion events can deliver 60-100 W/m² of additional heating to specific Arctic regions, what happens when you have seventeen major installations operating continuously for decades, each creating persistent vapor sources where natural humidity should be near zero?

To answer this, I needed to move beyond isolated observations to systematic energy budget analysis.

I constructed three literature-validated scenarios that scale systematically from the Sokolowsky baseline while accounting for the cumulative and enhancement effects that emerge when multiple installations operate as a distributed system. The mathematical framework treats each dam installation as a persistent vapor source operating within the broader Arctic atmospheric system, accounting for transport mechanisms, persistence duration, and spatial concentration effects that amplify individual contributions.

Read more

Read more

To understand the implications of altered freshwater timing, we first need to understand the system it threatens. So before we proceed, let us ask: What is the AMOC?

The Atlantic Meridional Overturning Circulation (AMOC) is the Atlantic branch of the global ocean conveyor belt, a planetary-scale system that moves seawater to redistribute heat, salt, and nutrients. Surface currents such as the Gulf Stream carry warm, salty water from the tropics northward. As this water cools at higher latitudes, it becomes denser due to temperature loss and increased salinity (from sea ice formation). The dense water then sinks 2–3 km into the deep ocean, flows southward along the ocean floor, and eventually resurfaces in other parts of the world.

This overturning spans the Atlantic from the Southern Ocean to the Nordic Seas. It is massive in scale, moving 17 million cubic meters of water per second (about 6,800 Olympic swimming pools each second) and transporting roughly 1.2 petawatts of heat. In the North Atlantic alone, this is equivalent to fifty times the global human energy use. The heat released near Greenland and Iceland helps explain why the Northern Hemisphere is approximately 1.4°C warmer than the Southern. Without the AMOC, Western Europe would be significantly colder.

So, how does it work?

The AMOC is thermohaline, meaning it is driven by both temperature and salinity differences. Warm, salty water moves north, cools, and becomes dense enough to sink. The cold deep water then flows southward and eventually upwells, closing the loop. A single parcel of water may take around a thousand years to complete this journey.

Salinity is the linchpin of this circulation. North Atlantic surface waters remain salty because the overturning process continuously brings in salt from the subtropics. The circulation sustains itself through this feedback. However, if surface salinity drops due to rainfall, ice melt, or increased river discharge, the density decreases, sinking slows, and the entire system weakens.

Now, why does it matter?

The AMOC is often described as Earth’s climate engine. It moves heat from the tropics northward and transports carbon-rich deep water southward, thus regulating climate zones and enabling carbon sequestration. A weaker AMOC leads to more CO₂ in the atmosphere, more volatile rainfall and storm patterns, and disruptions to agriculture and water security.

Monitoring systems like the RAPID array (26.5°N) and OSNAP (in the Labrador and Irminger Seas) have shown year-to-year variability and a notable decline around 2009–2010. Data suggests the AMOC was stronger in the 1990s and has weakened by approximately 15% since 1950. Models forecast continued weakening as warming and freshwater inflow increase.

This brings us to the critical point: too much freshwater can destabilize the AMOC. Once salinity drops below a threshold, the feedback loop breaks. The dense water stops sinking, the conveyor slows dramatically, and the system may collapse into a sluggish state that resists reversal.

Now, let us ask: Where is the freshwater coming from?

When Hans Neu warned in 1982 that regulating freshwater flow into the ocean could destabilize global circulation, it was seen as speculative at best, fringe at worst. After all, the logic was quiet, not outrightly catastrophic. It didn’t predict a flood, or a drought, or a war. It predicted a change in density, a shift in flow, a redrawing of the planet’s invisible gradients.

But what Neu saw, and what we can now measure with unnerving precision, is this: the ocean’s memory is written in salinity and timing. Disrupt either long enough, and you don’t just change the coasts. You alter the deep water. You unravel the planet’s circulatory system.

Here is how it unfolds…

Read more

Read more

Read more

We …

Read more

Where We've Been, and Where We're Headed

So far, we've explored two critical threads. The first traced how the logic of enclosure, the drive to contain, control, and regulate, became a dominant force in shaping landscapes and climate. The second focused on its physical consequences, especially through hydroengineering: mega-dams, artificial reservoirs, and rigid control of river flows that altered the natural water cycle.

We saw how these structures do more than store water. They reshape climate patterns, disrupt nutrient flows, and lock ecosystems into engineered schedules. Once disturbed, these patterns ripple outward, affecting entire regions and beyond.

From this foundation, we now move into one of the most consequential zones of intervention: the Russian Arctic.

This next thread focuses on Siberia, the subarctic, and the long-standing project to make them warm and wet. We will explore how Soviet dam-building, designed to modify climate, triggered a series of effects still unfolding today. These include thawing permafrost, weakening of Atlantic currents, nutrient depletion in Arctic waters, and the buildup of heat in the atmosphere.

These impacts are not remote or abstract. They are directly connected to food insecurity, polar warming, and the slowing of global ocean circulation.

This is not simply a historical case. It is a lesson. When large-scale water control systems are built, the consequences do not stay local. They stretch across entire continents and seasons, with long-term effects.

We now turn to the Russian hydrological network, not as an isolated experiment, but as a key part of the machinery that is reshaping Earth's climate systems.

Whenever I use the word DOMES, know that it was coined by Stephen Kasprzak in his book Arctic Blue Deserts, this is already mentioned in the first episode. It stands for Domes of Moisture Emissions, which is admittedly a bit of a mouthful, but captures the idea well.

I. The Soviet Plan to Warm the Arctic

Read more

Read more

Thanks for tuning in.

Let’s begin not with satellite data, but with your skin.
You sit in a sauna. The air is dry.
Warm, yes, but not yet immersive.
Then someone pours a ladle of water on the stones.

Ssssshhh.

A hiss. Then a swirl.
The heat doesn’t increase. It changes.
It stops being something you feel on your skin,
and becomes something you feel within.

What just happened?

The water, evaporating, carried with it a hidden payload i.e., 2.45 million joules per kilogram of energy, as latent heat inside each gram of vapor. It didn’t disappear. It traveled (invisible, unmeasured, underestimated) until it condensed again, releasing all its energy instantly, locally, and often unpredictably.

This isn't an analogy for the climate system.

It is the exact mechanism by which water, when transformed into vapor, becomes one of the most powerful atmospheric transport agents on Earth.

Phase Change: The Overlooked Engine of Global Energy Transport

Climate models correctly emphasize radiative imbalance as a central driver of global warming. But in their precision, they often miss the mechanisms by which energy actually moves not just as radiation, but as vapor. Water, through its phase transitions, may be the most under-credited force in the planetary heat engine.

Evaporation is energy storage.
Condensation is energy release.

Each gram of vapor carries 2.45 million joules per kilogram until it finds a place to condense. That moment of condensation does more than form clouds. It alters the structure of the sky.

Here’s why this matters:

• Hurricanes explode over warm oceans because evaporating seawater loads the air with latent heat, fueling rapid intensification upon condensation.

• Tropical forests self-cool by transpiring vapor, which draws in moist air and regulates local climates.

• Steam burns deeper than fire by dumping latent heat directly into tissues, illustrating the raw power of phase change.

But the bigger picture emerged with Anastassia Makarieva and Victor Gorshkov’s biotic pump theory (introduced in their 2007 paper in Hydrology and Earth System Sciences). Forests aren’t passive players in the climate they actively regulate it. Through massive transpiration (evapotranspiration), trees release water vapor that condenses over the canopy, creating low-pressure zones. This “pump” draws moist ocean air inland, sustaining rainfall far from the coasts. It’s physics meets ecology: the forest’s vast leaf surface evaporates more water than an equivalent ocean area, amplifying condensation and pressure drops.

This is not just theory. The model predicts pressure differences, wind vectors, and rainfall patterns with uncanny accuracy in forested regions. The implication is profound: remove the forest, and you not only reduce evapotranspiration, you collapse the mechanism that pulls moisture from sea to land.

Peter Bunyard brought this theory to life with hands-on experiments, detailed in his 2017 and 2019 papers in DYNA and a 2025 Substack analysis. In a sealed glass chamber, he simulated atmospheric conditions: humid air at the bottom, a cooled upper surface to trigger condensation.

The results were striking. Cooling alone (without condensation) produced negligible airflow, less than 0.01 m/s. But when condensation kicked in:

• Airflow surged to over 1.0 m/s, sustained and directional.

• This wasn’t buoyancy from rising warm air. It was an implosive vacuum effect: Water vapor condensing into liquid shrinks by a factor of about 1,700 in volume, creating a local low-pressure “pull” that draws in surrounding air.

Bunyard quantified the forces:

• Latent heat release: ~2.26 kJ/g (disperses broadly, warming the air).

• Implosive cooling: Just 0.17 kJ/g; but it’s this concentrated volume collapse that drives circulation, much like the vacuum in a Newcomen steam engine.

In the chamber, this generated micro-winds. Scaled up, it mirrors real-world observations: Amazonian forests pull in 10 m/s winds from the Atlantic, transporting moisture thousands of kilometers inland.

And yet, in engineered systems (particularly, those where winter inversions form and air cannot rise) we release vapor without the forest’s pump. Without the canopy. Without the living condensation nuclei that guide the rhythm of ascent and return.

In nature, condensation completes the cycle aloft: moist air climbs, cools, and releases its latent heat high above the ground, drawing new air inward and renewing circulation.

But in inversion-bound systems, the story collapses near the surface. Condensation occurs within a few tens of meters, inside a lid of cold air. The released heat cannot escape upward; it spreads sideways, trapped beneath the inversion, warming but not breathing. It is a heartbeat muffled by its own fog.

So when the sky grows saturated yet cannot resolve its excess,
we witness its paralysis as fog, as inversion, as a stagnant pressure field that silences rain and weakens circulation.

Anastassia Makarieva calls this the biotic heart of the planet—
Peter Bunyard has shown it can beat, and stop.

And the question remains: in the name of clean energy, are we quieting that heart in places where the air itself cannot lift?

But Before the Vapor — The Carbon

Not all emissions come from smokestacks. Some rise from drowned shorelines—where trees rot underwater, peat ferments in the shadows, and long-stored carbon begins to stir.

Hydropower is often sold as clean. Low-carbon. A benign partner in our planetary repair. But the reality beneath the surface is more complicated especially, in cold, boreal latitudes where decomposition slows, and methane waits like a delayed fuse.

In these regions, reservoirs behave less like batteries and more like biogeochemical reactors. Organic matter (trapped under still, anaerobic water) doesn’t simply dissolve. It transforms. Microbes feast. Methanogenesis begins. And in the absence of oxygen, CH₄ (not CO₂) becomes the dominant emission.

And methane, as we know, is no minor character.
It’s more than 80 times more potent than CO₂ over a 20-year horizon.

Above the surface, a second emission pathway unfolds: warm water discharged from the dam interacts with -30°C Arctic air, releasing massive amounts of latent heat as vapor. This seasonal pulse of energy (uncounted in carbon inventories) travels downwind, contributing to Arctic amplification. Together, these two flows gas below, vapor above form an infrastructural climate feedback that standard models rarely capture, but nature already feels.

Here’s where the story tightens:
Some of the very same dams implicated in vapor-driven warming also carry some of the highest carbon footprints in the hydropower sector

For instance …

• Brisay–Caniapiscau: Estimated lifecycle emissions of 2,265 gCO₂e/kWh, according to Scherer & Pfister (2016), which assessed the full biogenic carbon footprint of hydropower reservoirs in boreal zones.

• Robert-Bourassa: Produces around 400 gCO₂e/kWh, as noted by Barros et al. (2011) and confirmed in Hydro-Québec’s internal assessments reported by McCully (1996).

• Eastmain-1: In its first year, emitted at coal-equivalent levels (approaching 1,000 gCO₂e/kWh), before declining to natural gas levels (~400 gCO₂e/kWh) by year three, as documented in Teodoru et al. (2012).

These aren’t obscure or outdated installations. They are the crown jewels of Quebec’s hydroelectric fleet, feeding power into New York, Ontario, and beyond, under the banner of “clean” energy.

And their geography matters. Because these are the very systems that sit beneath the Arctic vapor corridor, the same plume that traces its way toward Greenland.

What this suggests is unsettling but increasingly undeniable:

• Some of the most significant contributors to regional Arctic amplification may be dual-mode emitters.

• They release long-lived greenhouse gases through their reservoirs — and short-cycle vapor through their turbines.

• And both forms of emission—gas and phase-state—interact with the climate system in fundamentally different, yet mutually reinforcing ways.

One warms the planet slowly, persistently. The other modulates the thermal engine of the sky.

Read more

Read more