(This essay first appeared in the anthology Ground/Water: The Art, Design and Science of a Dry River, edited by Ellen McMahon; Ander Monson; Beth Weinstein, University of Arizona Press, 2012.

Rillito
Meaning little river. Fair enough. The Rillito gathers out of its tributaries to run twelve miles along the southern slope of the Santa Catalina Mountains before flushing into the Santa Cruz River. Twelve miles—there it is. Except that it doesn’t run. It’s dry most of the year. Except that this little, twelve-mile river’s watershed drains almost a thousand square miles—from the Santa Catalina to the Rincon to the Santa Rita mountains, from Sonoran desert scrub to Rocky Mountain aspen, from three to nine thousand feet; a thousand square miles of Southern Arizona’s runoff into the little, twelve-mile river. Except that this little river’s main tributary, which is actually three different courses—Pantano Wash out of Las Cienegas Creek out of Gardner Canyon Creek—stretches six times the Rillito’s length to its headwaters. And the story of the little river is not the twelve miles or the thousand miles but the confluences—the Rillito formed of the confluence of Pantano Wash and Tanque Verde Wash and fed—to list just those off the southern slope of the Catalina range—by Nanini Wash, Casa Adobes Wash, Citrus Wash, Coaster Wash, Pima Wash, Village Wash, Friendly Wash, Race Track Wash, Via Wash, Campbell Wash, Camino Real Wash, Finger Wash, Valley View Wash, Flecha Caida Wash. But a confluence isn’t solely the merging of two flowing bodies of water, or in the Rillito’s case, two nonflowing watercourses, but a concourse, a coming together of people or things, of people and things. The Rillito is a confluence—made of the actual physical junctures of tributaries and the more nebulous confluence of river and city.

Chaos and Causation
Every instant of our individual and collective existence we affect our surroundings. We live in a world of systems and a single event, a single change, or myriad, even unrelated events can coalesce and ripple throughout an entire system, perhaps dying out, perhaps intensifying with each successive ripple. The slightest perturbation in the initial stages of a system’s trajectory may lead to significantly different outcomes. And so the butterfly effect—the displaced air from a butterfly wing rippling outward, shifting weather patterns, creating a hurricane. And so chaos theory, the study of systems so sensitive to change that the slightest shift skews the entire system’s dynamics. And so the Rillito River: if not a chaotic system, contains chaotic elements; if the smallest perturbation—the cow path, the death of a single cottonwood—didn’t shift the system, the Rillito shifted nonetheless. Such finer distinctions are, perhaps, beside the point when considering that the earth beneath Tucson is subsiding into the void of an overtapped aquifer or that Tucson consists of a virtually contiguous slab of concrete laid over 100 square miles of alluvial plain. The story of Tucson and the Rillito is the story of cause and effect, the story of humanity’s need for stability overlaid on an inherently unstable system.

Archeoriverdynamics
The mountain ranges ringing Tucson have been crumbling for over five million years, filling the basin between them with over five thousand vertical feet of alluvium. The accrued alluvium has been leveled into a gently sloping plain by the windshieldwiper-like action of the Rillito’s ancestors. In more humid parts of the world, watercourses run perennially, and have predictable, continual adjustments between process and form—they flood at rather predictable times and the floods shape the stream rather predictably into valleys. They are stable systems, there in their little valleys. Not so the water courses along Tucson’s alluvial plains. They flowed fitfully atop the basin floor, fickle to the weather—in drought they’d wither and die, in flood they’d spasm like a loose fire hose. Occasionally they’d incise, in full or in part, and within the incision they’d meander, forming cutbanks, downcutting further, until a big flood would debouch from the incision and gouge a new channel, or flow in sheets across the plain.

Knowing a River: Senses
Tucson is a city of bridges. Fifteen bridges cross the Rillito River. We Tucsonans know the Rillito more through the brief glance at the open stretch of beige and brown and white as we bridge than any deeper associations or reflections between aridity and human presence. But we can scramble off the bridges; we can know this river through immersion. Every day, hundreds of Tucsonans walk, run, and bike the paths that run parallel to the Rillito. People play with their dogs and fly kites in the riverbed. This is knowledge gained through the blazing sun on the neck and forearms, the Cooper’s hawk sweeping low overhead, the rich dust in the nostrils. The satisfying crunch of gravel underfoot; the fresh anthills spilling up smoothed granite, diorite, schist, and gneiss; the clunks of concrete and gnarled rootballs woven with nests of flood debris. This is knowledge gained through sounds: Home Depot intercom, sirens, air-force bombers rumbling into Davis-Monthan Air Force Base, bat squeaks, coyote yips, the aggressive chirp of a hummingbird on a dead branch. This is knowledge through the shed detritus of the city, the flotsam and jetsam of diaper bags, pool toys, cigarette butts, bones, batteries, rebar, tires, mangled shopping carts, broken glass bottles, spectacles, unbroken plates, dead pigeons, dolls, and the thousands of plastic water containers in various states of photovoltaic decay.

Anatomy of a Flood: Rain
It rains. Not much and not very often, about twelve inches a year in the Tucson Basin. There are winter storms, steady in rain, but most rain arrives with the monsoons, the violent summer pulsing of rain, lightning, and thunder. The monsoons roll off the Gulf so laden with rain they barely float. Rain collapses out of their black, flat bellies as though thermals were thorns. The first rains of the season splatter and slide off the hardpan soils. By the end of the season, or often by the end of a single storm, the rain runs off the now-saturated soil. Not all runs off, but enough—one inch of rain over one square mile amounts to 17.38 million gallons of water. The Rillito watershed is almost one thousand square miles. It’s not uncommon for two to three inches to fall in a single storm; a single storm once shed ten inches on the Santa Catalina. It rains, it floods.

A River City: Experience
Tucson as we know it is a strikingly recent phenomenon. In 1900,7,531 people lived in the city of Tucson. A century later, the city housed 486,699, with the greater metropolitan Tucson area at least twice that number. The desert ecologist Gary Paul Nabhan claims this general “Sun-Belt” population spurt is the single largest human migration in human history. Many came from the East, where rivers run with water and inundate when they flood. Between the 1930s and 1960s, when the population ballooned with 300% annual growth rates and the city sprawled across the alluvial plain, there were no major floods. The rivers barely flowed. Thus when the Rillito and Santa Cruz rivers flooded in 1983, Tucsonans told the Arizona Daily Star that the floods were “very strange and frightening”; that “we didn’t realize this could happen”; or “I didn’t think it could do this.” Even the Daily Star, having reported Tucson floods since 1879, headlined, “This is a desert?” Afterwards, the Visitors Bureau reassured the citizenry: “The 100-year flood has come and gone, so, by all rights, Tucsonans should enjoy another century of great Southwest weather.” We believed it because it matched our memory: we have been here a short time and we have short lives and shorter memories. Nobody remembers the mesquite bosques and beaver dams along the Rillito. There is no time-lapse reel detailing desertification, the slow incision of the riverbed, the water slipping through the sands, the land subsiding beneath our feet.

Anatomy of a Flood: Entrenchment
The Rillito, from confluence to confluence, start to finish, is an entrenched system. No more the sliding across the plain, the great leveling and shifting—the river’s fixed banks are now wreathed in concrete or armored with rip-rap and spaced wide-enough apart for the big floods to pass. Between the banks is the river. There is one main channel—the channel of the low-volume flows, called the thalweg—and splintering off the main, and rejoining, and splintering off again, are side channels, sought by bigger flows when breaching the overstuffed main. The big, bank-to-bank floods will wipe the palimpsest riverbed clean; lower, more constant flows will rescribe a new thalweg, and midsize flows will reestablish the sundry braids. So confined, the river works its channel. Smaller floods tend to deepen the channel; they’ll rub against many of the bends and abrade the banks, but they generally scour the riverbed itself. Large floods tend to widen the river’s banks: slamming against the bends, rasping along in helical abrasion. They bore into and burrow under the banks, sweeping away the slumped earth, scooping out the concave meanders in cumulative causation—the greater the curve of the meander, the greater the river rubs against it, accentuating the meander. A big flood, eating at a developed meander, can erode hundreds of feet of earth.

A City River: Headcuts
In 1889, when the Santa Cruz River still flowed perennially, Sam Hughes, a local merchant and farmer, dug a large irrigation canal that intercepted the river’s subsurface flow near downtown Tucson. This was common practice at the time, though most would armor their canals to ensure they wouldn’t divert the entire course of the river. Either Hughes failed to do this or the huge flood that ripped through the following year uprooted the armor. It was a flood such as the old-timers reckoned they hadn’t seen in twenty-five years. The waters cut into Hughes’s canal—which was, again, deeper than the riverbed—and the flood poured off the drop into the canal. Such a waterfall is called a knickpoint—a sudden, often dramatic change in a river or stream’s gradient. The flood quickly and consistently eroded the lip of this waterfall, moving the knickpoint upstream—a process called headcutting. The headcutting rippled throughout the river system. Floods no longer fingered through the braids in thin sheets but swerved deep into the headcut, eroding its lip farther upstream. The river below the headcut, focused and powerful, began to incise into an arroyo. With the river running deeper within the earth, the groundwater table dropped to match. The canals that had intercepted the previous surface flow were left, literally, high and dry. The lowering of the Santa Cruz’s riverbed formed a knickpoint at the confluence of the Rillito, which slowly migrated upstream, lowering of the bed of the Rillito.

A City River: Thirst
The Rillito is not one story, except perhaps that of the climate. The river ran dry, wet, then dry; the river incised, aggraded, then incised again; the river shifted one way then back again, dictated by the ebbs and flows of drought and downpour. In 1858, a local historian described the Rillito as flowing intermittently, lined by cottonwoods, willow, mesquite bosques and cattail cienegas, broken and stilled by numerous beaver dams. The Rillito floated atop the groundwater table—the shallow aquifer suspended between the Tucson Basin’s alluvial fill. The Rillito was, is, the groundwater table—the surface flow only the visible manifest, the tip of the iceberg. In wet spells the aquifer buoyed the surface flows; in dry spells the river shrank into the earth after the diminished aquifer. The Rillito is a story of thirst—in 1891, we replaced irrigation canals with mechanical pumps that tapped the shallow aquifer. Within fifty years we were pumping out more than was being naturally replenished; every year thereafter saw the water table decline substantially. The aquifer dropped thirty-seven feet between 1942 and 1952. The cattail cienegas dried and died. The beavers were long gone. The groundwater receded past reach of cottonwood and willow roots, past the mesquite bosques, whose taproots plunge eighty feet deep. By 1990, we were sucking 70 billion gallons out of the aquifer a year; in some places, by 2008, the groundwater had dropped 250 feet. Not only do we now have to drill deeper and deeper wells, and not only will we eventually sap the aquifer, but the ground beneath our feet settles into the spaces once saturated by water, the subsidence itself irreversibly affecting the aquifer.

Knowing a River: Thought
When is a river a river? Or not a river? Is a dry riverbed a river? Is it a river only in the few times a year it flows? Or is that a flood? What if it never flows, if we diverted even the monsoon floods into catchment basins and settling ponds, if the skeleton riverbed never fleshed out, just lay in the un, year in, year out, would it at some point cease to be a river? We impound the Colorado River behind a dam and call it a lake. We siphon the lake (reservoir? river?) water into an uncovered canal (aqueduct? river?) which snakes across 336 miles of arid desert. The water is pumped 2,900 feet uphill to Tucson, where it is not only injected underground to help recharge the depleted aquifer but mixed with treated sewage effluent and released into the downtown sections of the dry Santa Cruz River in an urban revitalization scheme known as “Rio Nuevo.” When is a river a river?

A River City: Structure
Tucson is a grid. Straight-arrow Speedway Boulevard, thirteen miles, east to west. Straight-arrow Kolb Boulevard, nine miles, south to north. Tucson is laid like a tapestry across the alluvial plain, the warp and weft of streets strung between mountains. The city’s mechanics, its cybernetic geographies, seem immutable but for the dendrite patterns of arroyos and washes cutting through the circuitry. These urban arroyos have concrete channels and banks; they have natural beds and banks; they are sterile flumes; they are tangled with riparian vegetation, and they lend a certain wonderful, organic disorder to the scene. They lend curves. Sinuosity. Some twine through the city in ribbons of green, then funnel into culverts under perfect square miles of concrete, under clustered mall parking lots, and reemerge farther downstream as though gasping for air. Where the washes still see sky, the roads and blocks of houses press against their contours and mirror their courses as though the city, before the culverts, before the malls, sought to dance, or spoon. Even Kolb, after nine rigid miles, runs against Pantano Wash for a final mile and softens into curves.

A City River: Loss and Adjustment
The Rillito is a narrative of loss. Fish, amphibians, crayfish, gone. Thousands of species of plants and animals gone and soon to be gone. To travel to a local, still-extant riparian zone and know an oasis is to know the death of an oasis. The Rillito is a narrative of adaptation. Scraggly bushes attempt riverbed lives between big flows. Raccoons, squirrels, coyotes thrive on the edges the river cuts through the city. Occasionally a mountain lion will slink through, between the mountain massifs, using the dry riverbed as path. Tens of thousands of Mexican free-tailed bats, their colonies the largest congregations of mammals in the world, live beneath the Rillito’s bridges. Between the rattle and clank of passing cars, the trailing snippets of radio, one can hear the high-pitched squeaking of bats. At sunset the bat chatter rises to a clittering, clackering song. Then, in individual coagulations and communal patches, bats collapse from the cracks in the bridge. They spread their long, translucent wings into flight, not the antic hyper-flittering of some bats, but powerful, swooping loops round the pillars then out into the evening, whirring away towards the Catalina Mountains in great streams, all on the same frequency, all twitching down then up at once, all out into a great bat joygyre that lasts for five, ten minutes and is gone.

Anatomy of a Flood: Confluences
The Rillito is a world where single rocks or root systems of buried trees tell the story of the river. Shopping carts obstruct flows for just the few seconds necessary for water to shift between the convoluted braids. Channels ramble and swerve between the slightest variations in slope. Often there is no clue as to the cause of the anastomotic points—the moments of convergence and divergence—as though the channel clove at the resistance of a single grain. Floods are chaotic, surprisingly sensitive, and bristling with variables—velocity, viscosity, volume, pressure, density, temperature—that affect a flood’s turbulence, recirculation, eddies, waves. Take sediment loading. A flood interacts with different materials of the streambed differently, and different tributaries supply different materials. Because of its length, Pantano Wash carries relatively fine sediment as it reaches the Rillito, whereas Tanque Verde Wash, flowing right out of the Rincon and Santa Catalina, delivers a relatively coarse load. The mixing of the two flows is like cigarette smoke rising in thin laminae, then suddenly disrupting into turbulence and dissipation. Say a flood has reached a sediment equilibrium—that, fully saturated, it is transporting as much sediment as its variables allow—and in comes a side stream, running clear along bare rock, a different temperature, velocity, and boom, the dynamics skew into chaos, into a different flood, perhaps reaching equilibrium again downstream, where bam, another tributary disrupts the dynamics. Depending on the variables, different types of erosion will occur simultaneously in varying proportions, overlapping and accentuating each other. Different channels within the main channel will alternately be scoured or deposited. But chaos doesn’t signify complete disorder—floods can be chaotic yet exhibit certain general characteristics. Some roll along as semiconsolidated slurries of water, 60 to 80 percent mud and debris.They’ll slide along for miles until some juncture, some critical loss of gradient or speed, when they’ll lose their power, and in this exact waning moment the flood loses its peculiar homogeneous cohesiveness and slackens, dumping the heaviest loads. Over the heavy-rock pattern the irregularly pulsing, dying flood deposits lateral terraces, asymmetric ridges and swales, dunes and antidunes, sand patches and scroll patterns.

A City River: Runoff
Tucson sprawls. A city of parking lots. A city upon which rainwater hits, shards, and runs off. The streets sluice the rain not into storm drains and treatment plants but into the washes, which run into the rivers, which run in flood. The water running off the pavement is different than the water that would run off the desert floor beneath the pavement. What it has gained in car oil, in industrial chemicals, in plastic bags, in dog shit, it has lost in natural sediment. This is important. There was once equilibrium between the sediment scoured from the riverbed and the sediment deposited on a riverbed in runoff flood. No longer. Runoff is free of sediment and running fast and the less sediment it has and the faster it runs, the more it can scour, suspend, and remove without replacing.

A River City: Stabilizing Instability
The Rillito River system is now the story of Tucson. It is a story of control. Take the Broadway Bridge over Pantano Wash. Constructed in 1965, it took but ten years for the Rillito and Pantano’s riverbed to lower, or downcut, twelve feet. At that rate the bridge’s pilings would have been undermined in two decades. This would not do. In 1979, Pima County enlisted the Army Corps of Engineers to build a $750,000 grade-control structure—essentially a giant concrete ramp—directly downstream of the bridge. The structure is intended to stop headward erosion—any knickpoint migrating upstream won’t be able to ascend the ramp to eat at the riverbed around the bridge pilings. Thus the upstream riverbed is maintained at its 1979 level, which, due to continued downstream downcutting, is now some twenty-feet above the downstream riverbed. Or take the Southern Pacific Railroad Bridge over the Rillito, constructed in 1925. The Rillito maintained its perpendicular course under the bridge until a huge flood in December of 1978 ate into 50 acres of land on a western meander bend. This shifted the river channel to the extent that, rather than flowing west, under the bridge, the river now flowed north, adjacent to the bridge. Again, the city sought stability over an instable system. The river still courses to the north, but hits a rip-rapped and soil-cemented bank that pivots the flow into an awkward 90-degree turn, redirecting it back under the bridge. Chaos, controlled.

Knowing a River: Language
If our recent arrival underlies our inchoate understanding of an arid world, our language overlies our ignorance. We betray a linguistic unsteadiness, as though there are objects or processes the English language lacks words for—as it did for saguaro, raccoon, butte, arroyo. “We didn’t think it could do this,” Tucsonans said of the 1983 flood. “It.” The signs along some of the bridges say Rillito River. Other signs say Rillito Creek. Pima County’s Regional Flood Control Department website alternates between creek and river. A stream can be a brook, beck, burn, bourne, nant, creek, crick, kill, lick, rill, river syke, bayou, rivulet, or run. A tributary can be a branch, fork, or prong. A dry or ephemeral river can be an arroyo, wash, draw, wadi. Yet the Rillito, because it is dry, or short, or ignored, is baffling as creek or river or “it.” The same when in flood: “I didn’t realize this could happen.” This. A flood, generally, is when a river exceeds its channel’s capacity and water inundates the surrounding floodplain. Flood is inundation, and if 20,000 cubic feet of water per second (c.f.s.) is roaring down the Rillito, gouging out hundreds of feet of bank and destroying millions of dollars worth of property, yet not breaching the banks, not inundating, many will still hesitate to call it a flood. To hydrologists, engineers, it is a “significant flow event” with perhaps some “corresponding lateral erosion.” Some call it a flood. On one level, this is mere semantics, needless quibbling; on another, our linguistic constructions define our meanings. We name the world as we know it, or, perhaps, as we don’t know it. Our understanding remains in suspension; our vocabulary has yet to put down roots.

Knowing a River: Disassociations
We don’t know the Rillito as a river because we don’t need to know the Rillito. It has little or no bearing on our day-to-day existence—it flows but a few times a year and destructively maybe once in a lifetime. We turn the tap and out flows clear water. We are equally removed from water’s origins if it comes out of a plastic bottle or from hundreds of miles away or thousands of feet underground or from hundreds of miles away to be injected thousands of feet underground. We don’t know the Rillito because it’s not water—it’s dust, and rock, and scrabbly vegetation, and trash, and when it’s water it’s more sludge than water. The connections are not made. This is not a judgment; this is the world as we know it—where disassociations reveal more about ourselves than associations.

A River City: Flood
Tucson was reminded of its rivers in 1983. It had been the second wettest September on record when Tropical Storm Octave delivered five days of steady rain, culminating in a twenty-fourhour, 3.58-inch deluge on the third of October. The Rillito roared at 29,700 c.f.s., the Santa Cruz a staggering 52,700 c.f.s. For three days, Tucson’s rivers dwarfed the Snake or the Hudson or the Colorado. Trees, telephone poles, cars, sections of bridges, mobile homes, and culverts swept along in the flood. Landfills became lakes. Natural gas lines exploded. Sewage lines ruptured, adding thousands of gallons of raw sewage to the burgeoning flow. People stuck in river-crossings were swept away. People rescuing people stuck in river-crossings were swept away. People watching the rising floodwaters were swept away when the undermined banks beneath them collapsed. The flood killed twelve people, destroyed all but one of the bridges across the Santa Cruz, and caused 250 million dollars in damages. Out of the receding floodwaters emerged floodplain ordinances. Future riverbank development was required to be set 500 feet from the existing banks of the Rillito and Pantano unless approved by the county’s Chief Engineer. Developers were allowed to ignore setbacks if they stabilized the banks. Consequently, stabilization became the primary solution to flood management. We fixed the meanders in place or straightened the river’s sinuosity. We widened the channel. We sprayed the majority of the Rillito’s banks with soil-cement or overlaid them in rip-rap or fencing. We kept building.

A City River: Acceptance
The Rillito is a narrative of change. We can’t go back to the days when Tucson didn’t edge against the Rillito. We can’t go back to when the Rillito’s mouth stretched a mile wide or its course shifted by miles in single floods. The Rillito is strung between golf courses and cottonwoods, development and open space, bats and bridges, and ideas about how it is and how it was. The Rillito will not flow again but in flood. The aquifer will not rebound. We can drop Tucson’s per-capita water consumption from its current 177 gallons a day to 100, we can xeriscape every yard, use reclaimed water for all nonpotable necessities, cut agricultural and industrial uses of groundwater—and, with current population growth, still overdraw the aquifer. No amount of conservation or effluent use or injection of canal water can match our population growth. Based on current weather patterns—without even takinginto account that it’s projected only to get hotter—Tucson’s water use is projected to double in the next fifty years.

A City River: Future
The problem is, piecemeal flood protection doesn’t work. A smooth concrete riverbank only amplifies a flood’s velocity, providing it that much more power to bore into the next bend’s unprotected bank. If upstream protection begets downstream protection, at some point the entire river’s banks will have to be systematically stabilized. Not to mention the riverbed itself. The huge flood of 2006, while not overtopping banks or gouging out meanders, lowered Pantano’s bed beneath Speedway Boulevard eight feet. This type of downcutting threatens to undermine even where stabilized banks or grade structures toe-in with the natural channel—a knickpoint cannot migrate up and over the concrete check-dams, but it can corrode the riverbed from their bases. We armored the confluence with the Santa Cruz, so any downcutting the Santa Cruz incurs is not transferred upward. But sediment-deficient runoff will still scour the Rillito’s riverbed. The future of the Rillito is that of a continuous, tightly stepped series of grade control structures, similar to rice terraces, or we’ll simply line the riverbed with concrete and the Rillito River will become a flume.

Confluences: City and River
The story of the Rillito isn’t just the overdrafting of the aquifer, the artificial headcuts, the impermeable streets and runoff, the inadvertent yet profound morphological changes, but every point, every confluence, every influence. The seemingly unrelated factors that tie in everywhere—buffelgrass replacing native vegetation so that the Sonoran desert is primed for a massive, unprecedented burn and subsequent unprecedented runoff. A cow browsing a desert shrub, the withered roots no longer trapping sediment, no longer absorbing water, the next storm’s rain thus running onto the path the cows trod, making it a rill, eventually an arroyo. Gravel and sand is mined out of the riverbed to build roads that block natural sediment flows. The mine pits act as giant sediment receptacles, allowing flood waters to momentarily pause, deposit their sediment loads, and then scour more from the banks and riverbed downstream. Or the pits act as scour holes, with headward erosion gnawing upstream out of them. The Rillito is all these stories. But the bats, too, tell the story. Bats and the biomimicry of streets to washes. Bats and the rare, sudden, and human awe of a dry river in flood.