10 Ups and Downs of Rain
How do you make rain? Two ingredients are needed: water vapor and rising motion. Water vapor is evaporated into air from land, ocean, or other bodies of water, or transpired from plants. By “rising motion,” we mean vertically-moving air currents, either warm air bubbles moving upward because they’re buoyant, or vast areas of air ascending from larger-scale forced uplift.
Strictly speaking, condensation occurs when humid air cools, enough to bring it to saturation. But in practice, the cooling that produces precipitation is almost always from rising air. Those familiar with mountain climates may already be aware of the requirement of upward motion for rainfall. Most mountains have one rainy side, known as the windward side, where the winds are forced upwards, and a dry, or leeward, side on the downward slope. In Washington State, for instance, winds consistently blow out of the southwest, from the Pacific Ocean. When the winds approach the Olympic and Cascade mountain ranges, air is forced upwards, leading to extremely heavy rainfall. In the Olympic Peninsula along the Pacific coast, rainfall is substantial enough to earn the region the classification of temperate rainforest, with rainfall totals up to 180″ (4.5 m) per year. Ferns, mosses, and giant conifers fill the lush ecosystem, and rain gauges measure the year’s accumulation in feet. On the other side of the slope just a few dozen miles away, the town of Sequim is a sunny oasis, receiving only 16 inches (0.43 m) of rain per year. Similarly, in eastern Washington, on the other side of the large mountain ranges, the climate is desert-like. The area right behind a mountain which experiences steady downward motion and little precipitation is known as the “rain shadow.”
So what causes precipitation patterns on a large scale? Observe the monthly precipitation climatology below to get familiar with the rhythms of rain, the pulsing of monsoons, and the shifts of storm tracks.
Click to view a higher resolution (0.25 degree) dataset of precipitation over land from the Global Precipitation Climatology Centre. Note the units of that plot are mm per month.
A first feature to draw your attention to is how narrow the precipitation bands are on Earth. Some of the rainiest spots over the ocean, like in the tropical East Pacific, are located only a few 100 kilometers from the driest ocean regions. On land, the middle of the Congo Rainforest is just a couple thousand km from the driest parts of the Sahara Desert, with dramatic changes in landscape across the Sahel region just south of the Sahara.
The Indian province of Meghalaya, located just north of Bangladesh among sharp mountain upslopes and high plateaus, has a name meaning “abode of the clouds.” Towns there like Sohra (Cherrapunji) and Mawsynram hold records for weather extremes, including the highest annual average rainfall (11.872 m or 467.4″). During the monsoonal months, humid air from the ocean is lifted up the mountain slopes, leading to massive rain rates.
Sohra receives only an inch and a half (3.8 cm) of rain per month in November through February, while in June and July it receives three and a half inches (8.9 cm) of rain per day on average. During a 48-hour stretch back in 1995, Sohra received 98 inches of rain (2.493 m), which is the most ever recorded in the world in a two-day period. Around 3 feet fell on the first day, and 5 feet came down on the second day. Imagine going away for the weekend and coming back to find that enough rain had fallen to completely fill up a swimming pool.
All across the tropics, the rainy season is called the monsoon, and in most locations it occurs in summer, because the hot continents give an extra boost to the warm, rising air currents. Sohra has a supercharged monsoon because it sits on a sharp upslope, and receives additional upward motion as well as a funneling of moisture from the mountains. While most tropical locations have substantially less rainfall than Sohra, months of intense monsoon rains followed by a long break with little rainfall are typical for much of the tropics.
The Hadley Circulation
There are no mountains of ocean water beyond a few meters of waves, so sharp rain bands across the ocean are clearly not due to topographic effects. We can understand a lot by examining the vertical motion of air however. Within the tropics, notice that many of the rainiest locations on a large scale are near the equator. Many deserts, on the other hand, are located at latitudes around 25 degrees north and south, including the Sahara and Arabian Deserts, the Namib Desert, the Great Australian Desert, and Southwestern North America. Each of these features, rainy or dry, are explained in part by the Hadley Circulation, a large scale circulation with rising air over the warmest equatorial ocean, and sinking around 25 degrees in each hemisphere.
The Hadley Circulation has rising motion located where ocean temperatures are warmest, where water vapor is plentiful as well. A convergence of surface winds feeds the rising, bringing in air that has humidified on its voyage from the subtropics with evaporation from the warm ocean surface. The subtropical desert areas have subsidence of air, diverging moisture and keeping the region dry.
If warm air rises and cold air sinks, why doesn’t the Hadley Circulation extend all the way to the poles, where air is coldest? The answer is the Earth’s rotation, which causes extremely fast winds to form as tropical air is carried towards the poles. This rapid acceleration of winds blowing out of the west (known as “westerlies”) can be understood by the conservation of angular momentum. The slow turning around its axis might seem small, but a 24 hour rotation period around the 40,000 km circumference at the equator amounts to a speed of 460 m/s. As winds accelerate, they become unstable to a type of fluid dynamical instability known as baroclinic instability, which is the process that creates the day-to-day weather in the midlatitudes. These storm systems take over, halting the Hadley circulation around 30 degrees on Earth.
How about on different, or imagined planets? On Saturn’s moon Titan, which has a comparatively slow rotation rate, the Hadley circulation is indeed global in scale. In a climate model, when an Earth-like planet is rotated faster and faster, the Hadley circulation gets narrower and narrower, and multiple bands of storm tracks fill the rest of the world. Animations of surface pressure in a set of simulations with varying rotation rate can be viewed in the middle of this page.
Midlatitude Storm Tracks
Midlatitude weather systems are critical for hydrology not only because they stop the Hadley circulation, but also because of the rains they bring. Baroclinic instability and accompanying processes like frontogenesis create strong rising motion within certain sectors of storms, namely the warm sector and along warm and cold fronts. There are also dry areas of baroclinic eddies, which are linked with downward motion. These midlatitude weather systems exist for approximately 5 day timescales, with 1000 km length scales, causing the day-to-day weather variability from 30 degrees latitude all the way to the polar regions.
It’s useful to consider the moisture budget to understand both climate regions of today, and future changes in the hydrologic cycle. In a steady state (which is a very good approximation for water vapor at timescales more than a few weeks), precipitation P is equal to evaporation E plus the moisture convergence C.
[latex]P = E + C[/latex]
The moisture convergence C can be understood as the winds (either steady or transient) bringing in water vapor. It only moves moisture around, and integrates to zero over the planet. Examined across latitudes, E > P in the subtropical desert regions, indicating a flux of moisture away from these bands (C < 0). Within the deep tropics, P > E, due to the convergence of winds by the Hadley Circulation (C > 0). P > E within the high latitudes as well, as midlatitude storms transport high moisture air from the subtropics into the higher latitudes.
The midlatitude storm tracks help to concentrate precipitation in locations like the North Atlantic, where it focuses rainfall onto Ireland, the United Kingdom, and Scandinavia. The North Pacific Storm Track brings rainfall across a vast region extending from Japan to Alaska and Cascadia. The Southern Hemisphere Storm Track exists largely over ocean, but brings moisture to locations like Aotearoa/New Zealand, southern Chile, South Africa and southwestern Australia. The storm tracks are named because the centers of the systems can be tracked across the ocean basins and land, and although any given day is chaotic, the tracks appear over preferred locations.
It’s important also to identify the regions that only occasionally encounter a midlatitude weather system. For the North Atlantic storm track, this includes the Mediterranean and North Africa, while for the North Pacific storm track, southern California and Baja California are only occasionally rained on. In the Southern Hemisphere, the Namib, the Atacama and the Great Australian Desert wouldn’t be so dry if the Southern Hemisphere storm track was located farther north. We’ll show soon that because the storm tracks are expected to shift poleward in a warmer world, many of these locations are forecast to dry.
Changes with Warming
Rain requires rising air and water vapor, and we know that water vapor is increasing rapidly with global warming. Because of higher humidity, the rainiest locations and weather systems across the world are becoming rainier. However, the full picture of hydrologic cycle changes with warming is quite complex, and there are many parts of the world where it’s not clear whether the future will be wetter or drier on average.
Examining plots from the IPCC reports shows that while temperature changes are broad in scale and have large amounts of scientific confidence over most of the world, precipitation has comparatively little scientific certainty. In part this is due to the fact that while temperature has relatively little spatial variability when stretching from the cold polar regions to the hot tropics, precipitation is organized into narrow bands. It’s hard for climate models to simulate these small-scale features, and it’s difficult to know whether the features might move with warming. Even the relatively weak confidence metrics in the IPCC reports, e.g., two-thirds of models having changes that exceed natural variability and 80% of models agreeing on sign of change, uncertainty dominates maps of long-term precipitation change.
It’s useful to start with some basic expectations. The precipitation change with warming can be usefully separated into two components:
- A thermodynamic change, associated with the increase in temperature and water vapor, and
- A dynamical change, associated with changes in upward motion (either the location or intensity of air currents).
We’ll show that the thermodynamic changes are more confidently expected, and is associated with wet areas getting wetter and dry regions persisting. The dynamical changes have less scientific certainty, but include poleward shifts of the midlatitude storm tracks, and shifts of tropical rainfall towards the warmer hemisphere.
Thermodynamic Changes
Phrased simply, at a constant relative humidity, moisture increases around 7% per degree, as dictated by the Clausius-Clapeyron equation, so for a given strength of upward motion, precipitation will increase at approximately the same rate. Indeed, the humidity content of the Earth has increased in nearly all regions except for a few land areas that have had less precipitation. Global average surface humidity has increased by approximately 0.4 g/kg in most datasets (IPCC AR6 WGI Fig. 2.13).
More moisture in the atmosphere means more convergence of moisture, which explains the tendency for the high latitudes to increase in precipitation. This increase can come from either more rain or more snow. Indeed, it is often too cold to snow, when the atmosphere doesn’t have enough water vapor. Snowfall totals are expected to increase in many regions even though less places will receive frozen precipitation as temperatures rise.
The tropics also have a tendency to increase precipitation with warming, but dynamical influences often confound the picture enough that there is not confidence in predictions. It’s roughly accurate to think of the individual features as getting wetter, but potentially moving around spatially.
Thermodynamic changes are also associated with increased drying. Imagine the atmosphere as a sponge, holding only the moisture that is allowed by the Clausius-Clapeyron equation, but often less than this, especially over desert areas. Increasing the temperature is like increasing the size of the sponge. Downpours increase, as the highest rainfall events have a larger amount of water vapor to dump from the sky. But when the sponge is less than full, it can suck up even more moisture from land surfaces. Thus many continents are expected to decrease their relatively humidity with warming, with exceptions only over the areas that increase their rainfall the most.
How much can the globally averaged precipitation can change with warming? It’s much less quickly than the 7%/degree change in water vapor, because precipitation and evaporation are constrained by energetics. As we showed in Chapter 8, latent heating/cooling is the primary way that energy is transferred from the surface to the atmosphere. Since the atmosphere is in energy balance beyond a timescale of a few seasons, latent heat release can only increase if other energetic terms can decrease an equal amount. In climate models, the radiative cooling of the atmosphere does increase with warmer temperatures, which allows latent heating to increase at a slow rate of 1-3%/degree. Similarly, evaporation can only increase if other surface budget terms increase. An increase in downward longwave radiation facilitates the increased surface latent heat flux.
Doesn’t a higher humidity content require larger rain rates? Nope. Water vapor content can increase without evaporation or precipitation increases if the lifetime of water vapor in the air increases. Imagine a bus going about its normal route. Even with the same number of passengers getting on and off the bus (representing evaporation and precipitation), the bus can be more crowded on some days — if the passengers stay on the bus for a longer time. A longer residence time for water vapor can be seen in climate models.
If global precipitation goes up slowly, while rainfall in the rainiest locations increases, there must be some regions that dry. These tend to be subtropical regions, where humidity is taken away more easily as temperatures rise, and over continents. “Wet gets wetter and dry gets drier” is actually a legitimate scientific expectation for the future.
Examining plots of precipitation minus evaporation compared with simple scalings shows that much of the hydrologic cycle change in climate models can be explained with this thermodynamic effect alone. However, there are many imprecisions, especially over land and on the equatorward side of the midlatitude storm tracks. We need to consider the potential for rainy regions to move in space.
Dynamical Changes
A dynamical shift with a reasonable amount of scientific confidence is that there will be a poleward shift of midlatitude storm tracks, and a widening of the tropics with global warming. The IPCC deems a Southern Hemisphere storm track shift with warming as “medium confidence,” while a Northern Hemispheric storm track shift and Hadley Circulation widening as “low confidence” (Fig. 8.21). In general, models predict a drying on the equatorward flank of storm tracks, although shifts are of different magnitude in different models. Tropical expansion and a poleward shift of the Southern Hemispheric storm track have each occurred in observations.
In the tropics there is significant model variability in predictions. This is in part due to the tendency for tropical rainfall to shift in response to heating patterns. If certain areas of the tropics warm up more, due to forcings (such as aerosols), feedbacks (e.g., from cloud-radiative effects), or ocean circulation, then rainfall patterns tend to be affected.
We’ll discuss many of the dynamical changes in the tropics later in the book, including El Niño/Southern Oscillation (in the Oceans section), and shifts in the Intertropical Convergence Zone and other deep tropical features (after first discussing atmospheric energy transports). We’ll also get into more detail about storm track shifts and variability in the atmospheric dynamics section, including how ozone depletion has caused much of the observed shift in the Southern Hemisphere.
