The 2021 Southwest monsoon season, which still has two weeks to go, has dropped record-setting rainfall and wreaked havoc, triggering flash floods, mudslides and dust storms in multiple states. The monsoon’s erratic nature has been on display, drenching one neighborhood, or region, while surrounding areas go dry.

Both hyperactive and hyperlocal, this year’s monsoon has revealed the difficulty of forecasting the features of this seasonal wind shift but also shown the substantial progress made in recent years.

It once was almost futile to forecast the monsoon in any dimension: where, when and how much. Accuracy is improving for short-term and long-range outlooks, thanks to advances in science and technology. Still, in many ways, the monsoon remains a mystery.

Chaotic and frustrating

The chaos of the monsoon still frequently blows past the limits of current predictive powers, frustrating forecasters tasked with warning of flash floods, dust storms and the like. A day that looks primed to explode instead passes quietly after cloud cover drifts into the area, preventing solar heating of the land. Or evening falls on a rainless Phoenix, only to have a dead-of-night deluge open up because colliding overhead are storms that first brewed up many miles away.

Often adding to the challenge is the region’s large number of recent transplants from other states.

“Newcomers don’t quite understand why we can’t tell them exactly when it will rain and where,” said Amber Sullins, chief meteorologist at Phoenix’s ABC15 News. “They’re coming from places with cold fronts, where you can easily say, ‘Hey, between 3 to 6 p.m., we’ll have a line of thunderstorms move right through the city.’

“The weather here doesn’t work like that at all.”

Long-range forecasts are even more challenging.

Just look at the National Weather Service’s forecast, made in mid-June, for precipitation during the span of July to September: equal chances of above-average rainfall, a normal amount or another “non-soon” with little storm activity.

As Michael Crimmins, professor of climate science at the University of Arizona, admitted in a mid-spring podcast, “It’s tough to make a good monsoon forecast.”

Crimmins later told Capital Weather Gang: “That was an understatement, too.” During July, Tucson’s official gauge took in 8.06 inches of rain — the wettest calendar month in the city’s recorded meteorological history, which dates back more than 125 years.

A calm yet stormy future

As with all weather systems now, it’s essential to understand how the monsoon acts in a world of extreme climate change. A colleague of Crimmins’s says if you want a preview, check out present-day hurricanes.

“We know for sure that at that top end of the scale, those most intense hurricanes are getting more intense. … Monsoon storms here in Arizona are similar to that paradigm,” said University of Arizona professor Christopher Castro. “Maybe the frequency is decreasing because we’re getting less of those upper-air triggers to help the storms get organized.

“But when they do happen, they have a moister atmosphere to work with — more moisture and more energy.”

Kernel of truth

To explain what makes the monsoon so difficult to predict, an often-used analogy is a pan of unpopped popcorn on a hot stove: The kernels represent humid air while the burner is the sun heating the ground. Now, which kernel will pop first?

While that comparison is educational for demonstrating the monsoon’s randomness, over the years scientists have learned that some areas of the burner, so to speak, run hotter than others. One consistent trigger is mountainous terrain, explaining why Tucson gets more monsoon rainfall than Phoenix, as the Old Pueblo is nestled in the foothills of the Santa Catalina and Rincon mountains.

“Topography always wins with the monsoon; that’s your most reliable guarantee,” Crimmins said. “Then everybody else has to fight for the scraps.”

Hi-Rez, higher accuracy

Every advancement in computers and modeling is an advancement in predicting the monsoon. Paul Iñiguez, science and operations officer at the Weather Service in Phoenix, recalls old modeling could only get 20-kilometer resolution, far too spread out to correctly mimic monsoon storms, so it was necessary to “fudge the physics” for a correct forecast.

Now, though, the Weather Service uses models with resolutions as high as three kilometers. Even more, Iñiguez added, the models are run more often, compared with years past when forecasters could get such information only twice a day.

Back at the University of Arizona, a researcher has ratcheted down models’ resolution even tighter — 1.8 kilometers — for a monsoon forecasting system of startlingly high quality.

“Frankly, it’s still shocking to me sometimes that it works as well as it does,” said Mike Leuthold, who runs Arizona regional WRF (Weather Research and Forecasting) model forecasts.

Leuthold said there was no “Eureka!” moment for his modeling, only the recognition some 15 years ago that it might be possible to forecast the monsoon’s organized convective systems or complexes of thunderstorms. From there, improvements have followed advances in computing horsepower.

Sullins was effusive in her praise of Arizona WRF: “They’ve been instrumental in helping us get better monsoon forecasts across Arizona and, really, across the Southwest over the last decade.”

Still, Leuthold noted, Arizona WRF’s forecasts are only as good as the data it has to work on: garbage in, garbage out.

“The trick is to determine if the model run is going to be right or not,” he said. “So I spend a lot of time in the morning looking at the way the models initialized and making sure that clouds are in the right place or not, making sure that the amount of water vapor in the air is correct.

“A lot of people would just use a model without looking at the initial conditions, and that’s a recipe for disaster.”

Monsoon research’s roots

Relatively speaking, the monsoon is a new field of study, with the first major undertaking starting in 1990. Before then, there were arguments about whether the monsoon was really a monsoon, as the word, derived from Arabic, refers to a shift of wind direction.

Meteorologists in Arizona knew enough about the annual surge of humid air from the south to develop a way of tracking its arrival: the primitive but effective guideline of three straight days with a dew point reaching at least 55 degrees. But was the source of moisture the Gulf of Mexico or the Gulf of California?

Setting out to answer that question and more was the Southwest Area Monsoon Project (SWAMP), which went as far as to enlist a hurricane hunter plane from the National Oceanic and Atmospheric Administration for an inside look at the monsoon’s thunderstorm systems. Heading the project was Bob Maddox, who had left as director of the National Severe Storms Laboratory in Oklahoma for southern Arizona.

Maddox said they soon noted surges of low-level moisture moving rapidly up the Gulf of California and into the lower Colorado River Basin and southern Arizona. These surges often led to several days of thunderstorms in southern Arizona, and occasionally in far southeastern California and southern Nevada.

With time, Maddox added, it became clear these surges could be triggered by three different features: a westward-moving zone of low pressure embedded within tropical trade winds; a complex of the thunderstorms, known as a mesoscale convective system, moving from the mountains of Mexico over the Gulf of California; or a tropical system passing southwest of the southern end of Baja.

Still, moisture from the Gulf of Mexico does play a part. Research from Marty Ralph, director of the Center for Western Weather and Water Extremes, first noted the Continental Divide has a 155-mile low stretch between the Rocky Mountains and the Sierra Madre, east of the Arizona-New Mexico border and extending south into Mexico. Here, the elevation never tops 8,200 feet — about 3,000 feet lower than average along the Divide. That subtle gap is important, because most of southeastern Arizona’s wettest monsoon days occur in conditions with easterly water vapor transport from the Gulf through that gap on the previous day.

(Because that geographic feature was unidentified until Ralph’s research, he got to name it. It’s now known as the Chiricahua Gap, taking its identifier from the Chiricahua Apaches native to the region.)

Monsoon mysteries

But all these decades later, Maddox admits many questions remain unanswered. When a Gulf surge is forecast, how deep will it be? After a mesoscale convective system forms, how long will it last and where will it move?

Unfortunately, the educated grasping for answers is more than an academic pursuit, because every bit of knowledge gained can only improve forecasts. And if forecasts are wrong, the consequences can be deadly.

Said Sullins: “If you say there’s a chance for storms this day but nothing happens, and the next day it’s the same, by day three, people are like, ‘Yeah, right. It hasn’t happened the last two days. I’ll be fine.’ But then it does pan out on day three, and they go out and get themselves into a situation where now they’re stuck in a flooded wash. …

“It becomes an issue of trust.”