You drive across Iowa or Illinois and see corn and soybean to the horizon. Two centuries ago you would have seen tallgrass prairie. A century before that, parts of it were oak savanna. The Midwest is not naturally a forest, and that history is exactly why prairie-to-forest succession is the right design pattern for a Midwest food forest, instead of copying a Pacific Northwest temperate rainforest model that has nothing to do with the climate you actually live in.
This guide is for the Permaculture-Curious gardener (or aspiring small farmer) in USDA Zones 3 to 6 across the US Midwest who has read about food forests and wants the regional version: which native edible species fit, how to use prairie strips, what Mark Shepard's Wisconsin farm actually proved, and what timeline a 1/4 acre suburban or 5 acre rural conversion realistically takes.
Before European settlement, the central US between the Appalachians and the Rocky Mountain rain shadow was a vast mosaic of tallgrass prairie, oak savanna, and bur oak woodland, maintained by indigenous fire management and bison grazing. National Park Service data on the Tallgrass Prairie Preserve documents an original prairie footprint of roughly 170 million acres. By 1900, less than 4 percent remained, converted to row crop agriculture.
That ecology matters for food forest design. East of roughly the 100th meridian (the line through central North Dakota down through western Kansas), annual rainfall exceeds 26 to 32 in (660 to 810 mm) per year, enough to support closed-canopy forest if fire is excluded. West of that line, the land is moisture-limited and stays prairie regardless. The eastern two-thirds of the Midwest is where prairie-to-forest succession will run on its own once you stop mowing and burning. The western third (Plains states) needs a different design pattern (windbreak agroforestry, oak savanna, dryland-adapted) which is its own article.
A food forest in the Midwest is not a temperate rainforest copy. It is accelerated, edited prairie-to-forest succession. Left alone, an abandoned Midwest field follows a predictable path: year 1 to 3 annual weeds (lambsquarters, pigweed), year 3 to 8 perennial grass and forbs (goldenrod, asters), year 8 to 20 shrub thicket (sumac, prickly ash, then dogwood), year 20 to 50 pioneer trees (eastern red cedar, black cherry, walnut, locust), year 50+ climax forest (oak, hickory, maple). The food forester replaces the weed species in each phase with edible analogues that perform the same ecological function. This is permaculture principle 5 (use and value renewable resources) applied to a regional ecology: you ride the existing succession curve rather than fighting it. The framework comes from Dave Jacke and Eric Toensmeier's Edible Forest Gardens and Mark Shepard's Restoration Agriculture.
The Midwest is unusually rich in native edible species, most of them undervalued by the supermarket but perfectly adapted to local soil, frost, and humidity. Build around these first, then add naturalised European fruit (apple, pear) as the productive overstory:
| Species | Layer | Hardiness | Native | Notes |
| Pawpaw (Asimina triloba) | Low tree | Zone 5-9 | Yes | America's largest native fruit, mango-banana flavour, full shade tolerant year 1-3 |
| American persimmon (Diospyros virginiana) | Low tree | Zone 4-9 | Yes | Cold-hardy compared to Asian persimmon, requires male + female |
| American hazelnut (Corylus americana) | Shrub | Zone 3-8 | Yes | Hybrid varieties from Badgersett yield 1-3 lb nuts/bush mature |
| Chestnut (Castanea hybrids) | Canopy | Zone 4-8 | Native-derived | Dunstan and Chinese-American hybrids replace blight-killed American chestnut |
| Serviceberry (Amelanchier alnifolia) | Shrub/low tree | Zone 3-8 | Yes | Blueberry flavour, ripe June, attracts pollinators in early spring |
| Elderberry (Sambucus canadensis) | Shrub | Zone 3-9 | Yes | Wind-resistant, wet-tolerant, 15-30 lb berries per mature bush |
| Wild plum (Prunus americana) | Low tree | Zone 3-8 | Yes | Suckering, makes thickets, excellent for jam and preserves |
| Groundnut (Apios americana) | Vine/root | Zone 3-7 | Yes | Nitrogen-fixing legume with edible tubers (24 percent protein), climbs hazel |
| Jerusalem artichoke (Helianthus tuberosus) | Herbaceous | Zone 3-9 | Yes | Edible tuber, 8 ft tall, prolific spreader, plant where you want it to stay |
| Apple (Malus domestica) | Canopy | Zone 3-8 cv | Naturalised | Cold-hardy varieties: Liberty, Honeycrisp, Empire, Wolf River |
Source: USDA NRCS PLANTS Database, University of Minnesota Extension agroforestry, Badgersett Research Farm hazelnut data, Mark Shepard New Forest Farm species notes.
Pawpaw deserves a paragraph because it is the highest-leverage species for Midwest food forests. It is the only North American member of the tropical Annonaceae family (cherimoya, soursop), native from southern Ontario to northern Florida, hardy to -25 deg F (-32 deg C) in dormancy, and produces the largest native edible fruit in North America (4 to 16 oz). It tolerates full shade as a young tree (years 1 to 3) then needs sun to fruit, which makes it a perfect understory-to-canopy transition species. Pair two genetically distinct cultivars (Sunflower, Shenandoah, NC-1) for cross pollination. The University of Kentucky maintains the national pawpaw cultivar repository.
The proven Midwest pattern is Mark Shepard's keyline-on-contour planting from New Forest Farm in Viola, Wisconsin: 106 acres, 25,000+ trees, established 1995 on degraded former corn ground. The pattern works at 1/4 acre suburban and 100+ acre commercial scale because the underlying physics is the same. Here is the design for a typical 5 acre rural parcel on rolling Midwest terrain (Iowa hills, Wisconsin driftless, Missouri Ozarks):
The prairie strips are the most important Midwest-specific element. Iowa State University's Prairie STRIPS research program documents that converting 10 percent of a cropped acre to prairie strips reduces soil loss by 95 percent, total nitrogen loss by 84 percent, and total phosphorus loss by 90 percent compared to all-crop fields. The strips also support 3 to 4 fold pollinator abundance and bird diversity. In a food forest context, prairie strips between tree rows do the same erosion control and biodiversity work, plus they generate seasonal forage and biomass for chop-and-drop mulch.
Mark Shepard's STUN method (Sheer Total Utter Neglect) is the most counterintuitive but pragmatically powerful contribution from his Wisconsin work. The argument: in a humid-continental Midwest climate with -25 deg F winters, summer drought, late frosts, derecho wind events, and high deer pressure, the only cultivars worth keeping are the ones that survive total neglect. Pamper a thousand trees and you produce a system that requires permanent pampering. Plant five thousand seedlings, never water, never spray, never fence, and the 30 to 40 percent that survive year 5 are the genuinely adapted population.
This works because the Midwest gene pool is deep. Chestnut, hazelnut, pawpaw, apple, and persimmon all have wide genetic variation across thousands of seedlings. The STUN method is essentially landscape-scale selection breeding compressed into a 7-year window. It is not appropriate for a 1/4 acre suburban yard where you want every tree to live, but at 5+ acres with cheap seedlings it is the dominant Midwest food forest paradigm now and what differentiates Restoration Agriculture from European-style food forest design.
The STUN caveats. STUN works on bare root seedling stock from cold-hardy parents (Badgersett hazel, Burnt Ridge chestnut, Z's Nutty Ridge, Permaculture Research Institute Midwest). It does not work on grafted dwarf apple stock from a generic nursery because grafted dwarfs are clones, not seedlings, and 90 percent of them will die without the input package that produced the original orchard. Start with seedlings if you commit to STUN; start with grafts if you want predictable cultivars and accept higher input cost.
Identify the keyline contour using a bunyip water level or a $200 builder's level. Mark tree rows on contour, 20-40 ft apart depending on canopy species. Plant 50-100 trees per acre. Suburban quarter-acre: 20-30 trees + 6-10 shrubs. Cost per acre: $400-1,200 in seedlings.
Sow native prairie mix between tree rows in autumn after first hard frost (cold-stratification dependent species germinate spring 2). Plant elderberry, serviceberry, hazelnut shrubs in clusters. Mulch heavily 4-6 in around each tree.
Serviceberry and hazelnut produce first significant harvests. Pawpaw starts flowering. Elderberry hits full production. Total yield per acre: 100-400 lb mixed. Suburban yield: 30-80 lb.
Apple and pear move from light to full production. Pawpaw and persimmon first significant fruit. Chestnut still establishing. Identify and mark the strongest survivors. Cull underperformers from next round.
Chestnut hits commercial yield (15-25 lb nuts per tree). Apple at peak. Pawpaw mature. Total per-acre yield 1,000-3,000 lb mixed. The system now feeds itself (mulch, nitrogen, pest control) with minimal input. Approximate 5-acre annual yield: 5,000-15,000 lb. Suburban quarter-acre: 200-400 lb.
If you have more than 1 acre, federal cost-share programs significantly reduce establishment cost. The relevant ones for Midwest agroforestry:
Most small suburban food forests (under 1 acre) are below the program threshold but worth being aware of if you scale up.
The full 7-Layer Backyard guide includes the cold-hardy cultivar source list, the keyline survey method, and the prairie strip seed mix for Zones 3 to 6.
Read the Free GuideThe USDA Plant Hardiness Zone Map updated 2023 shows the Midwest has shifted approximately half a zone warmer since the 1990 baseline. A site that was Zone 4 in 1990 (-25 to -30 deg F minimums) is now functionally Zone 4b to 5a (-15 to -25 deg F). Practical implication for food forest design: plant for one zone warmer than your historical assignment, but maintain windbreak and frost protection for the extreme-weather years that will continue to happen. A persimmon that was marginal in southern Iowa in 1990 is reliable in 2026. A peach that was zone-5 marginal in northern Illinois is now plausible. Hedge by planting a mix of zone-current and zone-warmer cultivars.
The Midwest food forest is a regional answer to the larger food-forest design framework. Pair this with our pillar food forest design guide for the universal 7-layer model, our food forest understory companion planting article for guild design, and our 7 layers of a food forest for the vertical structure. Cross-pillar context: what is permaculture covers the underlying design ethic, and carbon farming in your backyard covers the climate dimension Midwest prairie-to-forest restoration is particularly strong on.
Outside the Midwest? Start with the universal food forest understory strategies guide and adapt the species list to your bioregion.
A Midwest food forest is a designed multi-layer perennial planting in USDA Zones 3 to 6 across IA, IL, IN, OH, MI, WI, MN, MO, NE, SD, ND, KS that mimics the prairie-to-savanna-to-forest succession native to the region. It combines native edibles (pawpaw, hazelnut, serviceberry, elderberry) with naturalised fruit (apple, pear), integrates prairie strips for erosion control and pollinators, and uses Mark Shepard's keyline + STUN methodology developed at New Forest Farm in Viola, Wisconsin.
Yes, very effectively in the eastern two-thirds of the Midwest where annual rainfall exceeds 26 to 32 in. The western Plains states (western Kansas, western Nebraska, the Dakotas) are moisture-limited and require a different windbreak-agroforestry pattern. Mark Shepard's 106-acre New Forest Farm in Wisconsin, established 1995, is the standing proof at commercial scale; thousands of suburban and rural homesteads run smaller versions successfully.
The natural ecological succession in the eastern Midwest from open tallgrass prairie to oak savanna to closed forest, absent fire. Pre-1850 indigenous fire management kept most of the central Midwest in prairie or savanna phase. Absent fire, an abandoned field follows the sequence: annual weeds (year 1-3), perennial grass and forbs (3-8), shrub thicket (8-20), pioneer trees (20-50), climax forest (50+). Food forest designers accelerate and edit this succession by replacing weed species with edible analogues.
Native canopy: bur oak, hybrid chestnut. Native low tree: pawpaw, American persimmon, wild plum, serviceberry. Native shrubs: American hazelnut, elderberry. Naturalised canopy: cold-hardy apple cultivars (Liberty, Honeycrisp, Wolf River), European pear. Avoid generic Zone 6+ nursery stock; source from cold-hardy Midwest specialists like Badgersett (MN), Burnt Ridge (WI), or Z's Nutty Ridge (NY).
Sheer Total Utter Neglect, a planting and selection method developed by Mark Shepard at New Forest Farm in Wisconsin. Plant a large number of cheap seedlings (not grafted clones), provide no irrigation, no fencing, no spraying, no pampering. The 30-40 percent of trees that survive year 5 are the genuinely climate-adapted population. The remaining 60-70 percent die or underperform and are removed. The result is a system bred to local extreme weather over 7 years. Documented in Shepard's 2013 book Restoration Agriculture.
Prairie strips occupy the alleys between tree rows in a Midwest food forest, typically 20 to 40 ft wide. They replicate the Iowa State Prairie STRIPS research findings: 10 percent prairie reduces soil loss by 95 percent, nitrogen loss by 84 percent, phosphorus loss by 90 percent, and supports 3 to 4 fold pollinator and bird diversity. The strips also generate seasonal biomass that gets cut and dropped as mulch around the tree rows.
Year 1: keyline survey and 50-100 trees per acre planted. Year 2-3: prairie strips seeded, understory shrubs planted. Year 4-5: first meaningful yields from serviceberry, hazelnut, elderberry. Year 6-8: apple and pear bulk production, pawpaw and persimmon start. Year 10-15: chestnut maturity, total system function. A 5-acre site goes from bare ground to 5,000-15,000 lb annual yield over 15 years.
The Pacific Northwest food forest model (Geoff Lawton, Permaculture Research Institute) emerged from mediterranean and temperate-rainforest climates and emphasises citrus and persimmon canopies, dense polyculture year-round. The Midwest model adapts to cold-hardy climates with -25 deg F winters, late frosts, and clay soils. Species shift to chestnut, hazelnut, pawpaw, apple. Prairie strips replace year-round groundcover. STUN-style seedling planting replaces dense pampered polyculture. Mark Shepard's Restoration Agriculture book is the canonical Midwest reference.