Part II. The quest to figure out the origin of rain: weather in digital worlds

This is a story about how our climate models were built - how the water cycle was added to a model of air movement, how heat behavior was integrated in, and how pertinent properties of soil and vegetation were put in. (Watching how something is built is hopefully educational.) Its a story of the emergence and detection of the small water cycle and the phenomena of soil and vegetation creating rain, with surprisingly many of the key characters in the greenhouse global warming movement involved. Its a story that leads to a Dutch rain map that might just revolutionize how states and countries come together to create water security and tame climate chaos (see last section).

[Part I of this post is about how meteorologists originally came to the conclusion that rain only came from the ocean, before they changed to their current thinking that rain comes from both ocean and land. It is about Budyko’s development of an equation for the small water cycle and its water and heat flow. It is also about the various stories of rain disappearing after deforestration.]

Air as a fluid

We lived submerged under an ocean of air.

This was the key insight that emerged in conversations Vilhelm Bjerknes and fellow scientists, that led Bjerknes to search for an algorithm for weather. Vilhelm Bjerknes diffident yet enthusiastic, possessed of a sharp mathematical sense as well as a high tightly curled head of hair, had grown up in Norway studying how fluids move, how they circulate, how they flow around objects, how they shear, and how vortices interact and birth other vortices. As he discussed fluid flow in conversations with fellow scientists, he realized this hydrodynamic behavior applied to the atmosphere also.

Air obeyed the laws of fluid dynamics. Add in some thermodynamics, and the behavior of this atmospheric ocean could be encapsulated into a set of equations, equations that Bjerknes came up with in 1904, equations that were destined to be known as the primitive equations, equations that would lay the foundation for atmospheric science.

If you put this fluid-like air onto a giant sphere, rotate the sphere, and then add heat at the equator, curious earth-like atmospheric behaviors will emerge. A cell of circulating air will arise - the air will rise near the equator, then flow poleward, descend around the 30 degree latitude, and then flow back towards the equator nearer the surface. This is the Hadley cell, a large scale atmospheric circulation on earth. This comes out of the primitive equations.

These equations would be too complicated to solve until the advent of computers decades later.

Forecasting the weather

John von Neumann had a vision, a vision that involved computers, a concept most people did not quite grasp in the 1940s. He thought computers would revolutionize weather prediction. Von Neumann, a polymath who helped lay the mathematical foundations of quantum mechanics, game theory and cybernetics, wanted to lay the foundations for this new field also. In the 1940s he recruited to his project Jule Charney, a man who would later be behind the famous ‘Charney report’ that brings the carbon greenhouse effect to the world’s attention.

Jule Charney had written his PhD thesis on mathematizing key atmospheric circulation phenomena. He was a man fierce in his search for intellectual clarity, at times vigorously argumentative in conversation, and also, able to enjoy a good laugh. Handed a visionary numerical weather forecasting project, Charney set out to figure out logistical issues like how important heat and friction are in the earth-atmosphere system, and how fine the grid size of the simulation must be before you could make a forecast.

He oversaw, in 1950, the first successful numerical weather prediction. It was done in a simulated world with grid cells 700km apart, with numbers run through a computer that appeared as a room full of connected vacuum tubes sending electrical signals to each other. It predicted the weather 24 hours ahead in time.

Adding water to air

Joseph Smagorinsky wanted to take simulation past just weather, which was for a few days, to simulating climate. He was the director of the US Weather’s Geophysical Fluid Dynamic Laboratory, and he had a talent for attracting creative scientists to the organization. In 1958 he brought a 27 year old physicist, named Syukuro Manabe from Japan to work on one of the first large scale global circulation models. Manabe would many decades later go on to win the Nobel prize for his carbon greenhouse global warming model.

Manabe had studied physics in Japan, and while slower than his fellow students at picking up the physics, he was very good at applying it to meteorology. He spent a lot of time contently drawing weather contours on graph paper. Although humble and polite, he found it refreshing coming to the USA, where people could openly disagree with each other, unlike in Japan. This kind of behavior he found useful for research. Manabe was endlessly curious and a daydreamer, and with this new computer tool he found a digital freedom to explore his many questions about climate and weather.

In order to model the climate properly, Smagorinsky, Manabe, and Schriffer had to figure out how to add the water cycle to their simulations. [Manabe 1965] .

Water was tricky to model. Dry air had beautifully neat equations, the so called primitive equations, that modelled how it moved. But water changed phases, it was sometimes liquid, sometimes vapor, sometimes ice, sometimes snow. The phase transitions were not easily so easily integrated in with dynamical equations. How the water and air mixed was not clear. Convection was difficult to model. It happened on a scale much smaller than global climate model grid size. How convection brought water vapor upwards was hard to simulate.

So they had to jerry rig the situation.

They resorted to an approximation that allows them to model the upward motion of air hydrostatically rather than hydrodynamically. The trick is called a convective adjustment. After they develop the method, convective adjustments are repeated in climate model after climate model for decades.

But approximations can be dangerous if not used correctly. Anastasia Makarieva will later argue that these convective adjustments, if used wrongly can lead to problems with global warming calculations, causing them to leave out the effect of land use change [Makarieva 2023] [video 1 , video 2 ]

After figuring out how to put convection into their model, it was necessary to figure out how to model the rain interacting with the land. Manabe poured over books about soil and water. He eventually settled on Budkyo’s bucket model. The rain would fill the soil to a certain level, and then, with the soil saturated, the rest of the water would spill into runoff, and flow back to the ocean. It was a decent approximation.

His simulations produced a rain-belt around the tropics, and a belt of less rain in the subtropics of the western US. Results which matched reality.

Manabe’s 1969 climate model was arguably the first that had the potential for precipitation recycling to occur in it - it allowed water to go from land to sky, and from sky to land. Whether precipitation recycling actually was happening in his simulations is another matter. Manabe did not peer under the hood of his climate model to look for the small water cycle. At least not yet.

Desertification and the disappearing rain

The rain was disappearing in certain parts of earth.

The earth lives submerged below a vast mass of fluid-like air. Air that can be humid in parts, dry in other parts. Air that could pick up water from the ground, and air that could dump water from its skies. The question arose - why was air getting dryer above certain continental expanses? Deserts, defined by having low annual rainfall of less than 100mm (4 inches), were increasing in China, India, Australia, North America, and South America, expanding by 10% since the 1920s.

The Sahara because of its severity, was the canonical example of desertication. The Sahel, a jagged chain of countries - Mauritania, Mali, Niger, Nigeria, Chad, Sudan, and Entrea that run through the top half Africa, lands of steppes and savannahs, lands covered in elephant grass and lemon grass and dotted with acacias, baobab, and palms was discovering that their vegetation was disappearing, and wind swept desert sands were encroaching.

Was the loss of vegetation leading to the loss of rain?

“In the 1970s, what we knew was based primarily on the network of meteorological observations that were used in weather forecasting. These gave us a very limited view of the general circulation of the atmosphere and very little understanding of the role of the water cycle.” says meteorologist Julia Walker.

One person interested in figuring out the cause of the alarming spread of desertification in the Sahara was Jule Charney. With the pioneering work of the first weather forecast model behind him at Princeton, Jule Charney is now heading the atmospheric and oceans department at MIT. He is interested to see if loss of vegetation is causing the decrease of rain.

Charney with his colleague Pater Stone, make a simple approximation to simulate vegetation. Land surfaces with vegetation have a lower albedo, land surfaces without vegetation have a higher albedo. Running their models, they find that vegetation creates rain, through a usage of their lower albedo. This vegetation was being eaten up by animals, leading to Charney to think that the overgrazing of animals, might be creating the drought in the Sahel. [Charney 1975] [Charney 1977]

More sophisticated observational studies later study the statistics of the rainfall in the Sahel. Yan Yu and her colleague’s research finds that Charney got the albedo mechanism wrong, but the basic causal chain of vegetation causing rain is correct. Precipitation recycling does occur, their observational statistics show, about one fifth of the rain comes from evapotranspiration in the local area. Interestingly, the vegetation makes more frequent rains, but not bigger rains. [Yan Yu 2017]. [D’Odorico 2103] Axel Kleidon and Klaus Fraedrich would later run models where the whole world was vegetated, or the whole world was world. With the whole world vegetated, there was 50% more rain, and many places were several degrees less from the surface evapotranspiration cooling [Fraedrich 1999]. Others would go onto find similar results [Medeiros et al., 2008] [Neale & Hoskins, 2000] {Williamson & Olson, 2003]

I find it very intriguing that Jule Charney proposed in his 1977 paper that land use change is causing climate change. And that two years later, he heads the writing of the Charney report, a report that references Manabe’s and Wetherald’s global warming model, and that will bring the carbon greenhouse effect to the world’s attention.

Jule Charney is a proponent of what Millan Millan calls the two legs of climate change. He believes that both land use change (through the water cycle), and carbon emissions create climate change. That the carbon part of his work blew up more, is probably in part due to the more quantifiable contribution that carbon emissions makes to climate change, that models at the time could find.

Julia Walker was also approaching the problem of desertification in the Sahara. She is a young British meteorologist, who loved the applied side of physics - "Meteorology just seemed to me to be the epitome of physics in action. You could look out of the window and see a rainbow, or you could see clouds form, and you knew exactly what was going on to create what you were seeing." Later in life she will become (with the name Julia Slingo following marriage) the Chief Scientist at the MET, Britain’s national weather service, where she will find herself smack dab in the middle of the climate wars.

Her approach to figuring out what is happening in the desert, is to do an experiment and see what happens if extra water is put into the soil. What she finds, along with her colleague, P.R. Rowntree, is that the extra water will cool the environment when it evaporates. This lessens temperature gradients with greener lands to the south, which then lessens winds.

In the model water vapor ascends, and then come back down as rain. Multiple times. The initial moisture anomaly will have an imprint for several weeks before dissippating to neighboring lands and oceans.

This is the small water cycle! I looked through a lot of the climate model papers at that time, and there are some conference proceedings I have been unable to access, and maybe there are some papers I did not find, but as far as I can tell Walker and Rowntree are the first modellers to explicitly find evidence for the small water cycle.

A statistical analysis is done of observational rainfall data in the Sahel, and they find a correlation between different rainfall events, suggesting that water is recycled between different downpours.

They suggest that deeper rooting vegetation can have access to more water and thus increase evapotranspiration and contribute to the small water cycle. Deeper rooting vegetation increases rain more than shallow rooting vegetation. Walker and Rowntree write in their paper “soil moisture, providing a positive feedback to any tendency to aridity, may be of importance to climatic stability, both with respect to local rainfall and - because of the albedo influence - to the global heat balance” [Walker Rowntree 1977]

Timescale of precipitation recycling

Emergence is when a system behaves in ways that were not programmed in from the beginning. A snowflake has intricate figures that were not programmed in. It arises out how the water molecules interact with each other and the changing temperature of the air. A school of fish can move in complex patterns, with no fish leading the group or with a master plan, rather each fish follows some simple rules of keeping a distance from other fish, not too far and not too close.

That a rain event, or soil moisture anomaly can have a ripple effect on local rain for a time scale of several weeks and months, even sometimes over a year, is an emergent behavior. Its not programmed into our meteorological model a priori.

It’s a phenomena that is intriguing for a number of reasons. First, if you see this phenomena, it means that rain can originate from the land. Second this phenomena can be used to aid in forecasts. For instance, if satellite images show a large area of soil moisture, then forecasts can be adjusted for three months down the line accordingly. Third, it means drought and heat waves can hang around for this timescale too, hopping from one place to another.

Meteorologist Jerome Namias [1959] speculated that the early spring rains in Texas were leading to late spring rains through this precipitation recycling phenomena. Now this phenomena had appeared in Walker and Rowntree’s digital world.

Others would follow suit and discover similar things in their models.

In 1984 Manabe, Yeh and Wetherald would find that an anomalous amount of moisture could last for two to three months. In the mid and high latitudes it could sometimes last even more than a year. Their paper directly addresses the existence of a small water cycle and its role in this - “the large increase of evapotranspiration from the ground resulted in an increase of water vapor content in the atmosphere, which in turn caused an enhancement of precipitation”. If the soil is able to hold more water, it can create more precipitation. They too have discovered the small water cycle now! As the water evapotranspires it cools the surface of the earth, by 5 or 6 degrees in many parts of the globe, and it warms the upper atmosphere. [Yeh, Manabe, Wetherald 1984].

Like in the case with Charney, I find it quite intriguing, that central figures in the carbon greenhouse movement, Manabe and Wetherald, was also quite involved with researching what I had thought was a more alternative worldview of water. They are proposing rain comes from the land also, not just from the ocean. They are saying that the soil’s ability to hold water can affect rain! They are placing the small water cycle as a significant part of the worlds global water cycle. And they are saying that parts of the earth’s surface cool because of water evapotranspiration, even if they haven’t drawn the conclusion that water evapotranspiration can cool the earth globally. Perhaps the alternative worldview of water, while still alternative in how far it has propogated into mainstream culture, is not quite so alternative in terms of who is proposing it.

Other researchers continue exploring the topic of how long these precipitation recycling persist for. Matei Georgescu, Roni Avissar and colleagues discovered three month long persistances of precipitation anomalies in the Mississippi. [Georgescu 2003] . This result is in contrast with what the two Johns Hopkins professors, Benton and Estoque had predicted back in 1950s about how there would be very little precipitation recycling in the Mississippi region (as described in the part I of “The quest to figure out the origin of rain”).

Paul Dirmeyer and Randall Koster [Koster 2006] pinpointed soil moisture anomalies through a detailed global soil moisture map, and tracked them as they persisted for a season, and used this information to make noticeable improvements to daily weather forecasts. Julio Herrera-Estrada, Francina Dominguez and colleagues tracked a drought as it moved over a several month period from California to the Midwest, in a simulation of what happened in the US in 2012. [Herrara 2019]

The rain in Spain

The Spanish were concerned. Spain was losing its rain. A member of the European commission asked the meteorologist Millan Millan, who already had been collecting a lot of meteorological data about the area, to investigate.

He found that deforestration, degradation and the paving over of the land were the culprits. Grasslands have less evapotranspiration that forests. Paved land has less evapotranspiration than grasslands. There was no longer enough evapotranpiration to add to the moisture blowing in from the ocean, to push the humidity over the saturation point. In order for rain to form, the air needed water contributions from both ocean and land, but the land contribution had been going down for decades.

A US forest service person, listening to Millan talk when he came to San Diego, said Millan’s results suggested that California, which had a similar Mediterranean climate to Spain, would experience similar troubles. The person guessed that in 20 odd years, California would dry out and start experiencing fires. A prediction, that years later, would unfortunately come true.

Millan Millan also found that the lack of rain in Spain, could lead to floods in the rest of Europe. Weather data collected showed that when moisture was unable to rain out, it would recirculate, along with pollutants, to form atmospheric layers, accumulating large amounts of water, hanging out above the coastal edge. All that moisture would then get blown over Europe and seed huge torrential rains [Millan 2005]. Land degradation can thus lead to the loss of smaller rains, while also creating larger rains. A result which points towards land restoration as a way to build a buffer against climate extremes.

“Water begets water, soil is the womb, vegetation is the midwife” Millan eloquently professes. The water stored in the soil, released by vegetation, is needed to create the waters of rain. He named the two causes of climate change the ‘two-legs of climate change’ - greenhouse emissions and land-use change. Millan continues to work to bring governmental and public attention to the leg of land-use change.

A map of the small water cycle

Hubert Savenije, a Dutch hydrologist, had also been researching precipitation recycling in the Sahel, building on the work of Charney, Walker and Rowntree. He wrote “Desertification, land degradation and the occurrence of droughts are problems of a global dimension which are primarily of anthropogenic nature… Research demonstrates that impacts resulting from land-use change are more important than the greenhouse effect to explain regional climate change, occurrence of droughts and desertification. The most important mechanism, in this respect, is the feedback of moisture to the atmosphere through evaporation from vegetation which is required to sustain continental rainfall…Particularly in semi-arid continental regions this problem is predominant… In this light, it is remarkable that in the international research environment so little attention is paid to the analysis and quantification of moisture feedback processes, whereas enormous research efforts are dedicated to global warming and global climate model predictions, which will not yield any tangible actions to be taken locally.” [Savenije 1995]

Savenije was approached by his Masters student, Ruud van der Ent, to make a map of the small water cycle. Climate models have become much more sophisticated in 2010, they have different evapotranspiration rates for different vegetation, different moisture holding abilities for different soil types, different frictional abilities for different forests to slow winds. And they have more detailed weather input from satellites and other meteorological apparatuses. With this technology available to them, Savenije and Van der Ent create a map of evaporation recycling, and its converse, a map of precipitation recycling. [van der Ent 2010]

The map below is of evaporation recycling, which shows how much evapotranspiration in an area comes back down on the same continent. About 50%-60% of the evaporation on the US west coast comes back down in the US and in Canada. Western Europe is supplying China with rain; 50-70% of the evaporation in western Europe comes back down over Europe and China. The Amazon supplies the rest of South America with rain, with about 80% of the evaporation of the Amazon comes back down on the continent. Parts of eastern Africa supplies the Congo with rain, and the Congo rainforest then supplies the rest of Africa with rain. 80-90% of the evaporation from the Congo rainforest comes back down in Africa. (You can see Savenije giving a nice talk about this topic here.)

The map below is of precipitation recycling, which shows how much precipitation in an area comes from evaporatranspiration on the same continent. With the exclusion of parts of the west coast, about 40% of the rain in North America comes from land. The Colorado River gets about 40% of its rain from land evapotranspiration, which suggest that eco-restoration of the US west coast could help increase the rain over the Colorado River and grow its dwindling waters, helping with water scarcity issues. The northern half of China is getting a significant percentage of it rain from western Europe. That rain helps the Chinese grow enough food in more dry areas. This suggests the Chinese may want to encourage Europe to eco-restore more of its lands so that northern China has more rain. The middle of South America gets a large portion of its rain from the Amazon. This suggests the countries there put pressure to stop Amazonian deforestration. The Sahel is getting a decent part of its rain from the Congo basin. If this was realized more, then Sahel anti-desertication efforts might also focus on the Congo - stopping the logging there, halting the conversion of forests to farms, and helping locals there find alternatives to chopping down trees for subsistence.

Benjamin Holzman speculated in 1937 that water that evapotranspired in the US, would get sucked into dry air masses from the Artic, Canada, and Southwest, and blown out to sea before it could rain. (see part I of “The quest to figure out the origin of rain). Now with sophisticated climate models that could simulate these air masses, it is shown that his analysis, while advanced for his time, was incorrect. At least half of the evapotranspiration in the US comes back down in the North America. Holzman’s declarations had helped wrongly shift public opinion to believe that the origin of rain was only from the ocean. Perhaps rain maps like these, maps that reflect modern climate science understandings, if given more attention, if broadcasted in blogs, youtube videos, and podcasts, if spread by permaculture classes, environmental organizations and climate groups, could shift public opinion to understand that the true origin of rain is both ocean and land, not just from the ocean.

As governments come to understand that their country’s rain and water security is dependent on countries upwind of them, then governments will start coming to the negotiating table to discuss why they need their neighboring countries to restore their land. Jessica Keune and Diego Miralles, of Ghent University in Belgium have been proposing that the European countries come together to form a watershed precipitation recycling network [Keune 2019]. Governments already come to the table to hammer out agreements about river usage upstream and downstream of themselves. The next step is to come together to discuss accords that reflect the new understandings of the origin of rain, that reflect the impact of land restoration on rain downwind, accords that could help birth a new era of water security. ……………………………………………..

……………………………………………..

The evolution of research is not linear, it is multidimensional, going down many streams. It is definitely true of climate science. There were so many threads and themes to explore and coalesce around as I went through the climate modelling papers. I had to pick and choose, sometimes writing one up and then discarding it, to put them into a coherent narrative.

This is a reader supported publication, if you would like to contribute financially, much gratitude.

Thank you to all of you who have supported me in various ways.

…..

References

Doherty, Ruth, J. Kutzbach, J. Foley, and D. Pollard. "Fully coupled climate/dynamical vegetation model simulations over Northern Africa during the mid-Holocene." Climate Dynamics 16 (2000)

D’Odorico, Paolo, Abinash Bhattachan, Kyle F. Davis, Sujith Ravi, and Christiane W. Runyan. "Global desertification: Drivers and feedbacks." Advances in water resources 51 (2013)

Fraedrich K, Kleidon A, Lunkeit F. A green planet versus a desert world: estimating the effect of vegetation extremes on the atmosphere. J Clim (1999)

Keune, Jessica, and D. G. Miralles. "A precipitation recycling network to assess freshwater vulnerability: Challenging the watershed convention." Water Resources Research 55, no. 11 (2019)

Makarieva Anastassia M., Nefiodov Andrei V., Rammig Anja, Nobre Antonio Donato Front. For. Glob. Change, 20 July 2023 Sec. Forest Management
Volume 6 - 2023 | https://doi.org/10.3389/ffgc.2023.1150191

Manabe, Syukuro, Joseph Smagorinsky, and Robert F. Strickler. "Simulated climatology of a general circulation model with a hydrologic cycle." Monthly Weather Review 93, no. 12 (1965)

Manabe, S., and R. T. Wetherald, 1967: Thermal Equilibrium of the Atmosphere with a Given Distribution of Relative Humidity. J. Atmos. Sci., 24, 241–259, https://doi.org/10.1175/1520-0469(1967)024<0241:TEOTAW>2.0.CO;2

Millán, M. M., and Coauthors, 2005: Climatic Feedbacks and Desertification: The Mediterranean Model. J. Climate, 18, 684–701, https://doi.org/10.1175/JCLI-3283.1

Milly, P. C. D., and K. A. Dunne, 1994: Sensitivity of the Global Water Cycle to the Water-Holding Capacity of Land. J. Climate, 7, 506–526

Van der Ent, Rudi J., Hubert HG Savenije, Bettina Schaefli, and Susan C. Steele‐Dunne. "Origin and fate of atmospheric moisture over continents." Water Resources Research 46, no. 9 (2010)

Wang G, Eltahir EAB. Ecosystem dynamics and the Sahel drought. Geophys Res Lett 2000

Yu, Y., Notaro, M., Wang, F. et al. Observed positive vegetation-rainfall feedbacks in the Sahel dominated by a moisture recycling mechanism. Nat Commun 8, 1873 (2017). https://doi.org/10.1038/s41467-017-02021-1