THE GROVE THAT WATERS ITSELF: Hydraulic Lift, Fertility Islands, and the Desert Trees That Build Their Own Garden

Muse - research brief for Asherah - October 5, 2026

The GROVES shelf holds two films: Ashera's garden, and the Eden Kitchen Labs kitchen where experiments become dishes. It holds one scroll so far, the windbreaks brief, about what a grove does to the wind. This scroll is about what a grove does from below: how deep-rooted trees lift water at night and share it with shallow neighbors, how the soil under a desert canopy becomes an island of fertility in a sea of lean ground, why the people who fed themselves from the Sonoran Desert for eight thousand years planted their orchards out of mesquite, why sheep and geese may be the best orchard workers a desert garden can hire, and how a woodlot can be cut again and again for centuries without ever killing a single tree. The science below is others'. The readings for the Ark are the author's, labeled where they appear.

Evidence class: the hydraulic redistribution measurements, fertility-island soil studies, mesquite agronomy reporting, silvopasture field studies, and coppice history below are peer-reviewed research, federal agency publications, or reported field accounts; the indigenous-history passages draw on Nabhan's long-documented work and reported interviews; Ark-side readings and the closing synthesis are interpretation, labeled as such. Source links and provenance are in the foot of this brief.

THE PUMP THAT RUNS AT NIGHT

In the late 1980s, two ecologists in Utah did something clever with heavy water. M. M. Caldwell and J. H. Richards gave the deep roots of big sagebrush (Artemisia tridentata) a dose of deuterium-labeled water, then watched where it went. The labeled water showed up in the stem water of neighboring crested wheatgrass tussocks (Agropyron desertorum), plants whose roots never reached the deep source. The pathway was the sagebrush itself: water absorbed by deep roots in moist soil moved through the root system, was released from shallow roots into the dry upper soil at night when the plant's stomata were closed, and was stored there until the following day. They called it hydraulic lift. The effect was not trivial. When the researchers experimentally blocked the lift, the sagebrush's own transpiration dropped by 25 to 50 percent on the following days, meaning the tree-sized shrub was depending on its own nighttime redistribution to keep its deep roots effective (Oecologia, 1989). [1]

The mechanism is now documented in dozens of woody species and goes by the broader name hydraulic redistribution, because the flow runs both ways. A 2005 review in the Proceedings of the National Academy of Sciences laid out the picture: deep roots redistribute deep soil water into shallow layers mostly at night when leaf stomata are closed, and also on cloudy days, maintaining a shallow soil-water reservoir that is then used by the "lifter" itself and by neighboring vegetation. The same review records that downward movement also occurs, from shallow to deep soil after rain following drought, and that the effect is large enough to recharge soil-water reservoirs faster than percolation or capillary forces alone could manage. One of the review's striking speculations, credited to Caldwell's group: the quantity of deep roots commonly seen in most plants would be too small to matter for overall transpiration unless hydraulic redistribution were occurring, so the process effectively multiplies the usefulness of every deep root. [2]

How much water are we talking about? Todd Dawson measured a mature sugar maple in New York and came up with 102 plus or minus 54 liters of water lifted into the upper 35 centimeters of soil in a single night. That was enough to matter: stable-isotope analysis showed neighboring understory plants were getting 3 to 60 percent of their water from the maple's lifted supply. Some neighbors were effectively drinking from the maple's nighttime deliveries. [3]

The water does not only travel through roots. A USDA Forest Service synthesis of work in old-growth conifer forests notes that the hyphae of symbiotic mycorrhizal fungi associated with fine roots can also transport hydraulically redistributed water through the soil, and potentially move it directly between trees through common mycorrhizal networks. The amounts involved are small in absolute terms, under half a millimeter per day, but the impacts are not: redistributed water keeps root xylem conductivity alive in the nutrient-rich upper soil through the dry season, tempers the drought signals that would otherwise shut down stomata and photosynthesis, and modeling cited in the synthesis suggests hydraulic redistribution in Amazonian trees could increase total dry-season evapotranspiration by as much as 40 percent, a number large enough to touch regional climate dynamics. The grove that holds the land is also, every night, quietly irrigating itself. [4]

Evidence label: hydraulic lift is established, peer-reviewed plant physiology. The magnitudes (Caldwell and Richards' 25-50 percent, Dawson's 102 liters a night, the 3-60 percent neighbor share) are the measured values from those studies, not estimates. The Amazon 40 percent figure is a model result, labeled as such. The grove's "self-irrigation" phrasing is the author's summary, not a scientific term.

THE ISLAND UNDER THE TREE

Stand in the open desert and take a soil sample. Walk thirty feet to the canopy of a mature mesquite and take another. The second sample will be richer in nitrogen, phosphorus, potassium, and organic matter than the first. This is the fertility-island effect, and the soil-science literature has been documenting it for decades. A widely cited review of the spatial pattern of soil nutrients in desert ecosystems, drawing on the foundational work of Garner and Steinberger (1989), describes shrub islands as the locus of biotic activity and biogeochemical cycling in desert ecosystems on all continents, and high concentrations of nitrogen, phosphorus, and potassium under vegetation in creosote bush shrublands as evidence that biotic processes build these islands. Nitrogen limits net primary production in both the Chihuahuan and Mojave deserts, so the concentration of it under canopies is not a curiosity; it is the central fact of desert fertility. [5]

A review of Mojave and Sonoran desert plant succession puts the consequence in plain terms: perennial shrubs form "fertile islands," microsites with fertile soils and ameliorated microclimates, and these islands hold the greatest overall concentrations of annual wildflowers and forbs. The grove is not just a tree; it is a nursery for everything around it. [6]

The nurse-plant effect is how this plays out in plant generations. Desert seedlings face lethal sun, browsing animals, and starving soil. A mature plant lends all three remedies at once: shade that lowers leaf temperatures, spines or height that shelter the seedling from herbivores, and the enriched soil of the fertility island. A PBS SoCal natural-history account of the Sonoran Desert's nurse plants describes the final benefit, fertility, as perhaps the most decisive: desert soils are poor at holding nitrogen, the nutrient plants need in the greatest quantities to build the proteins of photosynthesis. The best place for a seedling to find it is under a mature plant, especially a nitrogen fixer, because mature trees drop nitrogen-rich leaves and fruit, attract animals whose droppings add nitrogen, and in the case of the legume family, pull nitrogen straight from the air. Even a seedling that will grow up to fix its own nitrogen gets a head start on the boost. [7]

None of this is news to the people who lived in these deserts. Gary Nabhan's long ethnobotanical work records that indigenous desert farmers in the Southwest deliberately transferred mesquite soils, the enriched earth from under mesquite canopies, into their gardens to improve fertility. They were mining fertility islands on purpose, eight millennia before soil science named them. [8]

MESQUITE, THE TREE THAT FED EIGHT MILLENNIA

Mesquite deserves its own section, because no tree on this continent makes a better case for the grove as infrastructure. Nabhan, who wrote a book called Mesquite: An Arboreal Love Affair and keeps a hammermill in his garage for grinding pods, describes the tree's record plainly: mesquite's branches have provided food, fuel, medicine, shade, and shelter to indigenous communities of the borderlands for more than eight thousand years. Velvet mesquite (Prosopis velutina) and honey mesquite (Prosopis glandulosa) are legumes that fix nitrogen and improve the soil, their seeds carry more protein than soybeans by Nabhan's account, and the flour milled from their pods has a low glycemic index. Mesquites cover nearly 200 million acres of arid and semi-arid lands in Mexico and the United States, and the scientific consensus Nabhan cites calls them among the most "under-managed" resources on the continent. Mesquite pod flour sells for $22 to $24 a pound in many states; mesquite hardwood sells for $5 to $10 a board foot to furniture makers and guitar makers; the mesquite barbecue industry values selectively harvested fuelwood at $200 to $400 million a year; and mesquite honey is already a multimillion-dollar industry along the border. [9]

The water story is the one that matters most for a desert garden. Reporting in The Counter, drawing on Nabhan and on Sonoran ethnobotanist Richard Felger of the University of Arizona Herbarium, lays it out: mesquite can bear pods unirrigated in Yuma, Arizona, on three inches of rain a year. Its roots reach deep into falling water tables, so it needs no irrigation drawn from aquifers, lakes, or rivers. It prospers on city streets with little more than stormwater runoff, and tolerates saline water, a point that matters because year after year of irrigation salts cropland until harvests fail, a pattern that has ended entire civilizations. Tim Crews, the Land Institute's research director, is quoted on the energetics: "Mesquite has a very high energy return on investment. You don't need to fertilize it. You don't need to add water. It pretty much grows itself." Felger's comparison is the one to remember: "Mesquite certainly makes more sense for that Southwest than growing alfalfa. The water requirements are reduced by orders of magnitude." [10]

The botany backs it up. Mesquite is a leguminous nitrogen fixer, one of about 45 Prosopis species in semi-arid climates on several continents, fast growing, hardy, and xerophytic. Its bean pods were an integral wild-plant staple of traditional diets across the Southwest and northern Mexico for centuries, milled into flour with stone manos and metates; the pod flour is gluten free, sweet, and aromatic. It fell out of use in the mid-20th century as processed foods took over and ranchers began treating the tree as a weed competing with rangeland grasses, but the flour is making a comeback in breads, pancakes, and baking. [11]

One more mesquite detail, because it connects to everything else in this brief: the tree's roots can run 150 feet deep in search of water, its taproot chasing moisture other plants cannot reach, and the same tree survives on less than six inches of rainfall and has been found growing in Death Valley. Cut it down and it resprouts denser; leave a root fragment and it grows again. It is a coppicing tree by nature, which brings us to the woodlot. [12]

Evidence label: mesquite's nitrogen fixation, drought tolerance, pod nutrition, and indigenous use are documented in the sources cited; the 3-inches-of-rain-in-Yuma claim, the flour prices, the 200-million-acre figure, and the Crews and Felger quotations are reported in those outlets (a field video, a science-food outlet, and Nabhan's own writing), attributed where used. "Eight millennia" is Nabhan's figure, cited to his account.

THE FLOCK UNDER THE TREES

The oldest agroforestry practice on record may be putting animals under trees. Silvopasture, the deliberate integration of livestock with trees, is being studied again as serious agriculture, and orchards are one of its best venues. A USDA agroforestry publication on integrating animals into California orchards lays out the approach: sheep and poultry are the most common grazers in traditional orchards because they are less likely to damage trees than larger livestock. Some breeds of goats and short-nosed pigs can work with closer supervision. The forage is free: weeds and existing grasses, cover crops seeded between rows, and dropped fruit and nuts after harvest. Young trees need protection from browsing, but once the bark is out of reach, the animals become workers. [13]

A French case makes the economics vivid. A farmer in France introduced sheep into his declining plum and apple orchard twenty years ago. His fruit production had been sliding because of soil compaction and disease pressure. A few years into grazing, productivity started climbing again, and replacement trees planted among the grazed ground showed unprecedented growth. The manure stimulated soil biota and structure; the sheep's cleanup of fallen fruit and debris cut tree pests; Shropshire sheep, the breed he used, do not strip bark, so they lived among the trees without damaging them; and the medicinal forage in the hedgerows he planted for them reduced his veterinary costs, the tannins acting as natural antiparasitics. The farm diversified its income, sold everything directly, and the soil regenerated. [14]

The peer-reviewed literature confirms the mechanism and adds the caveats. A 2021 study of orchard grazing in France lists the documented benefits: grass management that replaces herbicide and tillage, and sanitation of the orchard through the animals' consumption of fallen leaves and leftover fruit, which breaks the life cycles of plum curculio, codling moth, and voles. The same study lists the disservices honestly: overgrazing can compact soil and damage feeder roots, the pesticide residues of conventional orchards raise unknown animal-health questions, and the labor and fencing costs are real. [15]

In California, the numbers are being counted. A two-year University of California study with sheep in almond orchards found cost savings from less mowing, one fewer herbicide application, reduced winter sanitation work against navel orangeworm, and less compost needed, a potential savings of $7 an acre, with no revenue lost. And in Ontario, Schuyler Farms runs sheep and geese in its apple and cherry orchards, grazing only after harvest to keep food-safety questions clean, and reports that wind and soil erosion drop sharply in the pastured ground compared to cultivated ground. Their advice for starting: rent a flock first, watch carefully, and learn the animals before buying them. [16]

Dawn's own Ark Unit 1 keeps hens and is raising royal chickens in the kitchen garden. The old knowledge and the new studies agree on what those birds are: not just eggs, but sanitation crew, pest control, and fertility delivery, working the floor of whatever grove stands over them.

Evidence label: the silvopasture benefits and caveats above come from the USDA field publication, the European agroforestry case report, the French orchard-grazing study, and the UC and Ontario farm accounts cited; the $7-an-acre figure is the UC partial-budget result, not a universal claim. The Ark-side reading about Dawn's hens is the author's, labeled as such.

THE TREE YOU CUT WITHOUT KILLING

Here is a technology almost as old as agriculture that modern forestry mostly forgot: do not plant a woodlot. Coppice it. Cut a broadleaf tree to the stump, and the stump, the stool, sends up a new crop of straight poles. Cut those in seven to ten years, and it does it again. An indefinite succession of crops from the same root system, with no replanting, no nursery, no waiting a generation. A traditional-woodland account in the Rackham lineage describes sallow growing two inches a day after felling, reaching eleven feet in a single season, and even oak standing seven feet high and an inch thick after one summer. Where livestock could reach the regrowth and would browse it, the old practice was to pollard instead: cut the tree six to fifteen feet above the ground, above the reach of browsing stock and deer, leaving a permanent trunk, the bolling, that sprouts the same way. [17]

Britain's New Forest still shows the results. Pollarded beeches, oaks, and hollies there were cut repeatedly for centuries, the cuts kept above the reach of commoners' livestock and deer. The practice extended tree lifespans so dramatically that a 1698 Act of Parliament, the Act for the Increase and Preservation of Timber in the New Forest, had to forbid further pollarding of beech and oak because the navy wanted straight ship timber and pollard stems grew too gnarled. Trees cut before that act are still standing. [18]

Read that again: a wood-management system so productive and so long-lived that a parliament had to outlaw part of it to protect the shipbuilding supply chain. For a desert garden that needs fence posts, trellis wood, tool handles, charcoal, and fuel without importing lumber or planting a plantation, the coppice stool is the oldest renewable-energy device in the temperate and arid worlds. The desert candidates are the resprouters: mesquite, which coppices by nature, palo verde, desert willow. The arid analogue of the English hazel coppice is a mesquite coppice, cut on rotation, feeding the kitchen, the workshop, and the soil with its charcoal.

Evidence label: coppice and pollard mechanics and the New Forest history are from the traditional-woodland sources cited; the 1698 Act is recorded in the New Forest heritage account. The desert-candidate list and the Ark-side application are the author's proposal, labeled as such.

Not every fruit tree belongs in the desert, but the ones that do belong are old friends. Figs are among the most drought-tolerant fruit trees confirmed by university extension work; as Mediterranean natives they handle hot, dry ground, and once established they are remarkably self-sufficient. A forum of serious fig growers quotes the botanical literature on why: the fig is essentially a phreatophyte, a well-driller, whose deep roots search out groundwater in aquifers, rock cracks, and ravines. The surface can look bone dry while the tree drinks from depth. [19]

Pomegranates are native to the arid lands of Iran and Afghanistan and produce on as little as twelve inches of annual rainfall, thriving in hot, dry climates in USDA zones 7 through 11; once mature they need minimal irrigation, and they fruit on old wood with little pruning required. Date palms, Medjool and Deglet Noor, have grown in the Southwest for over a century; they love intense sun and dry air, handle alkaline and salty soils, and their deep roots work through tough ground once established. Only the female trees bear fruit, and growers hand-pollinate the flowers, a small piece of husbandry in exchange for a harvest from a tree that barely notices desert heat. Mulberry rounds out the menu: drought hardy once established, and its leaves are edible for livestock, which makes it a silvopasture tree as much as an orchard tree. [20]

The honest caveat, stated in every serious source: drought tolerance applies once the root system is fully developed. Every one of these trees needs regular watering for its first year or two. The desert's orchard menu is a promise to the patient, not the instant. Plant the grove, water the young trees, and then step back while the hydraulic lift, the fertility islands, the fungi, and the animals take over the watering and feeding.

Evidence label: the fruit-tree drought claims come from gardening and horticultural press and grower forums, labeled as such; the fig-as-phreatophyte passage is a forum post quoting Wikipedia, presented as the growers' working theory rather than primary research. The establishment-watering caveat is from the extension sources cited.

WHAT THIS MEANS FOR ARK UNIT 1

The following is the author's synthesis and speculation, not research findings. It is labeled as such because the commission requires it.

The grove is not a decoration at Asherah. It is the infrastructure that makes the rest of the growing possible. A grove of mesquite does five jobs at once: it fixes nitrogen into starving desert soil, it lifts water at night into the root zone of whatever is planted beneath it, its canopy builds a fertility island that accumulates the soil the whole garden will eventually be built from, its pods mill into a storable, low-glycemic flour that needs no irrigation and no fertilizer, and its stools coppice into a perpetual supply of wood and charcoal. Around and under the mesquite, the desert's orchard menu, figs, pomegranates, dates, mulberries, draws on the lifted water and the enriched soil. Under the trees, the hens and the royal chickens work the sanitation shift: fallen fruit eaten, pests interrupted, manure delivered exactly where the fertility island wants it. Sheep or geese could follow, rented before they are bought, on the Ontario model.

Wave-34's brief on the canyon taught that the desert keeps water in stone jars. This brief's lesson is the companion: the desert also keeps water in living wood, and keeps soil under living canopies, and keeps fertility in the bodies of the animals that work the grove floor. Plant the grove first. Everything else in the garden is downstream of the trees.

SOURCES

Caldwell, M. M. and Richards, J. H., "Hydraulic Lift: Water Efflux From Upper Roots Improves Effectiveness of Water Uptake by Deep Roots," Oecologia 79(1):1-5 (1989). Utah State University Digital Commons. https://digitalcommons.usu.edu/grcanyon/80/ - The original hydraulic lift paper: deuterated water from sagebrush deep roots to neighboring wheatgrass; 25-50% transpiration loss when lift was blocked.
"Root functioning modifies seasonal climate," Proceedings of the National Academy of Sciences (2005). https://www.pnas.org/doi/10.1073/pnas.0508785102 - Review: hydraulic lift mostly at night and cloudy days; bidirectional hydraulic redistribution after Burgess et al.; recharge faster than percolation alone.
Dawson, T. E., "Hydraulic lift and its influence on the water content of the rhizosphere: an example from sugar maple, Acer saccharum" (PubMed PMID 28307839). https://pubmed.ncbi.nlm.nih.gov/28307839/ - 102 plus/minus 54 liters lifted per night by one mature maple; neighbors used 3-60% hydraulically lifted water.
Warren, J. M. et al., "Hydraulic redistribution of soil water in two old-growth coniferous forests," New Phytologist, via USDA Forest Service Treesearch. https://research.fs.usda.gov/download/treesearch/29719.pdf - Mycorrhizal hyphae transport redistributed water; under 0.5 mm/day but ecologically significant; Amazon model of up to 40% dry-season evapotranspiration increase.
"On the spatial pattern of soil nutrients in desert ecosystems," citing Garner and Steinberger (1989), via The Free Library. https://www.thefreelibrary.com/On+the+spatial+pattern+of+soil+nutrients+in+desert+ecosystems-a018191322 - Shrub islands as the locus of biotic activity and nutrient cycling; N, P, K concentrated under canopies.
"Disturbance and Plant Succession in the Mojave and Sonoran Deserts of the American Southwest," Int. J. Environ. Res. Public Health 2010, via MDPI. https://Www.Mdpi.com/1660-4601/7/4/1248 - Perennial shrubs form fertile islands with the greatest concentrations of annual plants; nitrogen limits production in both deserts.
"Desert Nurses: How Seedlings Survive Harsh Conditions to Become Full-Fledged Plants," PBS SoCal / KCET Redefine. https://www.pbssocal.org/redefine/desert-nurses-how-seedlings-survive-harsh-conditions-to-become-full-fledged-plants - Nurse-plant shade, defense, and fertility; nitrogen fixers; animal contributions to island fertility.
USDA Forest Service, Tree Planters' Notes, "Inoculation response by irrigation system type for desert tree establishment" (mesquite section). https://rngr.net/publications/tpn/57-2/inoculation-response-by-irrigation-system-type-for-desert-tree-establishment/at_download/file - Mesquite as nitrogen source for desert ecosystems; Nabhan 1982: indigenous transfer of mesquite soils to gardens; rhizobial associations.
Nabhan, G. P., "Build a Border Wall? Here's an Idea That's Better for Communities and the Climate" (2019). https://www.garynabhan.com/news/2019/02/build-a-border-wall-heres-an-idea/ - Mesquite food, fuel, medicine, shade, shelter for eight millennia; nitrogen fixing; seeds higher protein than soybeans; 200 million under-managed acres; flour at $22-24/lb.
"Mesquite, crucial to Indigenous diets for centuries, works miracles with water and needs no fertilizer. Why don't we grow more of it?" The Counter. https://thecounter.org/mesquite-arizona-wild-foods-foraging-sustainable-crop/amp/ - Pods on 3 inches of rain in Yuma; deep roots; salt tolerance; Tim Crews on energy return; Richard Felger on mesquite vs. alfalfa water use.
"Indigenous Crop: Mesquite, Ancient Flour of the Future," Food Tank (2014). https://foodtank.com/news/2014/04/mesquite-ancient-flour-of-the-future - About 45 Prosopis species; leguminous nitrogen fixer; pods an integral staple for centuries; gluten-free flour comeback.
"Mesquite pods farming in Arizona - Mark Moody's farm," video description via YouTube. https://www.youtube.com/watch?v=FteM7aoXrqI - Velvet mesquite facts: roots to 150 feet, saline tolerance, centuries of life, high-protein low-glycemic flour. (Grower video description; figures as reported there.)
Mazaroli, D. N. and Carlisle, L. (2023), "Silvopasture Scenarios for California: Integrating Animals into Orchards," USDA Forest Service National Agroforestry Center. https://www.fs.usda.gov/nac/assets/documents/examples/srs-silvopasture-scenario-animals-in-orchards.pdf - Sheep and poultry as standard orchard grazers; forage from weeds, cover crops, dropped fruit; young-tree protection.
"PA61 - Sheep orchards," European Agroforestry Federation (AgroforestryNet). http://agroforestrynet.eu/wp-content/uploads/2019/12/PA61_Sheep_orchards_EN.pdf - French plum/apple orchard: 20 years of sheep grazing regenerated fertility; Shropshire sheep spare bark; tannin forage cuts veterinary costs.
"Orchard Grazing in France: Multiple Forms of Fruit Tree-Livestock Integration," Forests 2021, via MDPI. https://www.mdpi.com/1999-4907/12/10/1339 - Pest sanitation (plum curculio, codling moth, voles), grass management replacing herbicide; honest disservices: compaction, pesticide-residue unknowns, labor costs.
UC Agriculture and Natural Resources, "Partial Budget Analysis for Grazing in Orchards" (UCCE, Julie Finzel et al.). http://ucanr.edu/sites/default/files/2026-07/Partial%20Budget%20Analysis%20for%20Grazing%20in%20Orchards.pdf - Two-year almond study: $7/acre savings from less mowing, one fewer herbicide pass, navel orangeworm sanitation, less compost. Companion: Schuyler Farms, Ontario, sheep and geese in apple/cherry orchards, post-harvest grazing. https://www.fruitandveggie.com/silvopasture-potential/
"Coppice," culturalecology.info (traditional British woodland management, in the Oliver Rackham lineage). http://www.culturalecology.info/imagination/Coppice.html - Coppice stools and indefinite crop succession; pollarding 6-15 feet above ground, the bolling; suckering clones.
"Pollard Trees and Coppice Trees in the New Forest," New Forest Explorers Guide. http://www.newforestexplorersguide.co.uk/heritage/history-in-the-landscape/pollard-and-coppice-trees.html - Centuries of pollarding; the 1698 Act for the Increase and Preservation of Timber forbidding pollard beech and oak.
"Residents of Hot Dry Climates - Help with Watering Advice for in-ground PNW Fig Trees," OurFigs forum. https://www.ourfigs.com/forum/figs-home/1005217-residents-of-hot-dry-climates-help-with-watering-advice-for-in-ground-pnw-fig-trees - Grower quoting the botanical account of fig as phreatophyte: deep roots seeking groundwater in aquifers and rock cracks.
"Drought-tolerant fruit trees - 6 of the best for a backyard," Homes and Gardens; "8 Drought-Tolerant Fruit Trees," Gardening Know How; "10 Fruit Trees That Actually Thrive in Arizona's Caliche Soil," Positive Bloom. https://www.homesandgardens.com/gardens/drought-tolerant-fruit-trees https://www.gardeningknowhow.com/edible/fruits/drought-tolerant-fruit-trees https://positivebloom.com/10-fruit-trees-that-actually-thrive-in-arizona-s-caliche-soil-fad99951/ - Fig, pomegranate (12 inches of rain, zones 7-11), date palm (Medjool/Deglet Noor, hand pollination), mulberry (livestock-edible leaves); the establishment-watering caveat.