S4ag Earthworks

Earthworks

Water that runs off your land takes soil, fertility, and resilience with it. Earthworks capture that water and put it to work — moving it laterally through the soil rather than down the slope and off the farm. Swales, berms, ponds, and keyline ripping are the physical infrastructure that transforms a runoff-and-erosion farm into a water-harvesting farm. Get this right once and the land pays you back for decades.

How this skill works: Each sub-tool pauses at a Checkpoint to confirm the assumptions it is about to build on before producing output. A recommendation built on a wrong assumption wastes time and money — confirm the checkpoint before acting. Each sub-tool ends with Next steps — the skills worth running once you have acted on this one.

---

Expert Lineage

The thinkers whose frameworks underpin this skill — and what they specifically discovered that changes how you farm.

P.A. Yeomans — Keyline Design Yeomans spent decades designing and building water systems across Australian farms and documented the methodology in The Keyline Plan (1954) and Water for Every Farm (1965). His specific discovery: the keyline — the inflection point on a valley slope where the land transitions from concave to convex — is the natural contour along which water will spread most efficiently across a hillside. Ripping along or slightly off-keyline moves water from valleys (where it concentrates and erodes) to ridges (where it is deficient). The practical implication: a single keyline ripping operation, properly located, can distribute rainfall across an entire paddock without a drop of concrete or pipe.

Bill Mollison — Swale Design in Permaculture Mollison formalised the on-contour swale — a level, non-draining ditch with a berm on the downslope side — as the primary earthworks tool in permaculture design. His key contribution: swales do not transport water, they hold it in place until it infiltrates. The difference between a contour drain (which carries water to a discharge point) and a swale (which retains it for infiltration) is the difference between managing a problem and solving it. Swales on contour turn every rainfall event into an irrigation and infiltration event.

Darren Doherty — Regrarians Platform Doherty refined Yeomans' keyline methodology into the Regrarians Platform, a practical design sequence that positions earthworks within a whole-farm context. His specific finding: earthworks sequenced without considering access, topography, and enterprise layout often create as many problems as they solve. The Regrarians Platform sequences keyline layout, access tracks, irrigation, and enterprise placement in a logical order so each layer is informed by the one before. The implication: design earthworks in context, never in isolation.

Mark Shepard — Farm-Scale Earthworks Integration Shepard documented large-scale earthworks integrated with enterprise design at New Forest Farm in Wisconsin in Restoration Agriculture. His specific contribution: on-contour earthworks sized and spaced for machinery access — swales wide enough to drive a tractor along, berms planted to productive species — deliver both water management and enterprise function. Earthworks are not an infrastructure cost; they are productive land if designed correctly.

Brad Lancaster — Rainwater Harvesting Principles Lancaster's Rainwater Harvesting for Drylands and Beyond established the decision hierarchy for water harvesting: slow the water, spread it, sink it, store it, use it. His most actionable finding: small, distributed earthworks on-contour outperform large storage tanks in drought resilience because they recharge the soil moisture profile across the whole farm rather than concentrating stored water in one place. Capture where it falls before you pump it anywhere.

Judith Schwartz — Water Cycle Restoration Schwartz's research (Water in Plain Sight) documented how land management changes infiltration rates and groundwater recharge at landscape scale. Her key finding: bare, compacted, and heavily grazed soils have infiltration rates 10–50x lower than biologically active, vegetated soils — meaning earthworks alone cannot solve a water problem caused by degraded soil biology. Earthworks and soil biology restoration work together; neither is sufficient alone.

---

Which tool fits

| You need to... | Tool | |---|---| | Find the contour lines and keyline before building anything | contour-reading | | Design a swale and berm system for your slope | swale-design | | Site and size a farm pond | pond-siting | | Apply Yeomans' keyline pattern across the farm | keyline-design | | Route tracks and roads so they manage water, not damage it | access-design |

Routing Decision

• Never built earthworks before, don't know where to start → contour-reading first; you cannot design anything without knowing where the contours are
• Have contours, want to build swales to catch runoff → swale-design
• Need water storage for stock, irrigation, or fire management → pond-siting
• Want to apply keyline ripping across a whole farm → keyline-design
• Planning new tracks, or existing tracks causing erosion → access-design
• Unsure if earthworks or soil improvement is the right answer → run contour-reading; it will reveal whether the problem is landscape water pattern or soil biology

---

Contour Reading

Locates the true contour and keyline on any slope before any design work begins.

Every earthworks error starts here — a swale built slightly off-contour drains rather than holds, undermines its berm, and eventually fails. Finding the true contour before any design work is the single most important earthworks skill. You cannot shortcut this.

Tools for finding contour:

| Tool | Cost | Accuracy | Best for | |---|---|---|---| | A-frame level | Under $20 in materials, self-built | ±2cm over 3m | Small-scale farms, no budget | | Bunyip level (water level) | Under $30 | ±1cm over long runs | Level transfer across obstacles | | Surveyor's level | Hire or buy ($200–2000) | ±5mm | Larger scale, consistent results | | Laser level | Hire $50–150/day | ±3mm | Fast over large areas | | GPS/RTK survey | Hire $300–500/day | ±2cm | Large farms, complex topography | | Contour map (1:5000 or better) | Free–$50 from survey authority | ±1m | Preliminary planning only |

A-frame method — step by step:

1. Build an A-frame from three lengths of timber or pipe: two legs of equal length (1.5–2m works well), one crossbar connecting them a third of the way down from the apex. Hang a plumb line from the apex.
2. Mark the centre of the crossbar when the frame stands on level ground — this is your level indicator.
3. Walk the slope. Place one leg on a known point. Move the other leg uphill or downhill until the plumb line hangs on the centre mark. Mark that point.
4. Continue across the slope, leapfrogging the legs and marking each new point. The line of marked points is a contour.

Understanding the keyline:

The keyline is not simply a contour line — it is a specific landform feature. On a valley slope, the land is concave (the valley) then transitions to convex (the ridge spur). The keyline is the inflection point between these two curvatures — the highest point where the valley floor meets the rising slope.

To identify it:

• Walk the valley from the lowest point upward. Notice where the slope begins to steepen and spread rather than narrow and concentrate. That transition zone is the keyline area.
• The keyline is typically where water would naturally start to diverge rather than converge.
• In low-rainfall environments, place earthworks slightly above keyline; in high-rainfall, slightly below.

Reading the slope for earthworks potential:

• 0–3% slope: water harvesting is easiest; pond sites viable with minimal earthmoving; swales can be large-spaced
• 3–8% slope: typical swale and keyline country; most earthworks tools apply here
• 8–15%: swales need more careful overflow design; berms must be well-compacted; access becomes a constraint
• Above 15%: earthworks risk failure if not expertly designed; biological solutions (deep-rooted perennials, mulch) often more appropriate; consult an engineer

Checkpoint — confirm before finalising:

• What is the approximate slope gradient across the proposed earthworks area?
• Do you have access to any existing contour maps (even topographic maps at 1:25000)?
• Is this preliminary design (maps are adequate) or are you about to dig (need ground-truthing with an A-frame or level)?

Designing from map contours without ground-truthing is common and often adequate for preliminary planning — but before earthmoving begins, every contour line must be verified on the ground. A map error costs nothing; an earthmoving error costs days of machine time.

Output:

CONTOUR READING ASSESSMENT
Site: [name or description]
Slope gradient (estimated): [%] across [distance]m
Keyline location: [description — distance from valley bottom, feature it relates to]

RECOMMENDED SURVEY METHOD
Method: [A-frame / bunyip / laser / GPS]
Reason: [why this suits the budget and scale]
Estimated time to survey: [hours]

CONTOUR LINES IDENTIFIED
[Contour 1]: [elevation or relative position] — [suitability for earthworks]
[Contour 2]: [elevation or relative position] — [suitability for earthworks]
[Contour 3 if applicable]

LANDFORM NOTES
Valley floor: [description]
Keyline zone: [approximate location]
Ridge/spur: [description]
Drainage patterns observed: [description]

EARTHWORKS POTENTIAL
[High / Moderate / Limited] — [reason]
Constraints: [any slope, access, or soil constraints noted]

RECOMMENDED NEXT STEP
[swale-design / pond-siting / keyline-design / access-design]

Next steps:

• Run swale-design (within this skill) — once you have the contours, design the swale and berm system.
• Run keyline-design (within this skill) — if multiple valleys are present, apply Yeomans' pattern across the farm.
• /s4ag-land-reading — combine contour reading with full landform and vegetation assessment before committing to design.

---

Swale Design

Sizes, spaces, and designs on-contour swales and berms for water harvesting and infiltration.

A swale is a level trench dug on contour. It does not drain — it fills and holds water until the soil absorbs it. The berm is the soil excavated from the trench, placed on the downslope side and compacted, then planted with trees, shrubs, or groundcover. Together, the swale and berm function as a water-harvesting and tree-planting system.

Design sequence:

1. Confirm the contour line is true level. Any gradient causes water to flow along the swale and concentrate at the low end — undermining the berm and eventually causing failure. Use an A-frame or laser level to walk the line before marking for excavation. A 20m swale needs to be level to within 2cm end-to-end.

2. Size the swale cross-section.

| Scale / Slope | Swale width (m) | Swale depth (m) | Berm height (m) | Notes | |---|---|---|---|---| | Hand-dug, garden | 0.3–0.6 | 0.2–0.3 | 0.2–0.3 | Use mattock or spade | | Small tractor, 3–8% slope | 0.8–1.2 | 0.4–0.6 | 0.5–0.6 | Backhoe bucket or grader blade | | Large tractor/excavator, 3–8% slope | 1.5–3.0 | 0.6–1.0 | 0.8–1.0 | Full earthmoving equipment | | High rainfall / flashy catchment | Add 30% volume | Add 30% volume | Increase berm compaction | Engineer check recommended |

3. Calculate catchment area and swale capacity.

The swale must hold all the water that falls on the slope above it, plus roof runoff if roadshed or building runoff is directed to it.

Approximate calculation:

• Catchment area (m²) × design rainfall event (m) × runoff coefficient = swale volume required
• Runoff coefficient: bare soil 0.6; grass 0.3; forest 0.1; sealed surface 0.9
• Add 20% safety margin to the calculated volume

Example: 500m² catchment, 50mm design storm, pasture (0.3 coefficient) 500 × 0.05 × 0.3 = 7.5m³ swale volume needed → design for 9m³

4. Design the overflow.

Every swale must have an overflow — a point where excess water exits safely rather than overtopping the berm unpredictably. Options:

• Armoured overflow (rock-lined low point at one end of the swale): simplest, most reliable
• Piped overflow to the next swale downslope: more controlled, more complex
• Natural overflow into established vegetation: suitable for gentle slopes and established berm

Never allow an unarmoured overflow. An overtopped berm erodes rapidly.

5. Space swales on the slope.

Swale spacing depends on slope, soil infiltration rate, and rainfall intensity:

| Slope | Suggested swale spacing (measured vertically) | |---|---| | 1–3% | 5–15m vertical interval | | 3–8% | 3–8m vertical interval | | 8–15% | 2–4m vertical interval — wider swales, steeper compacted berms |

Lower infiltration rate = closer spacing. Higher rainfall = closer spacing.

6. Berm planting.

An unplanted berm is a temporary berm. Plant it immediately after construction:

• Pioneer species first (fast-establishing groundcovers, grasses, legumes) to bind the surface within weeks
• Productive perennials (fruit trees, nitrogen-fixers, timber species) into the berm face
• Space productive trees at 3–6m intervals depending on species
• Berm top can support a walking path, fence line, or market garden bed if access is needed

Soil biology note: Swales dramatically increase soil infiltration and moisture over time — which feeds the soil food web. Dry soils have slow, bacteria-dominated biology; moist, aerated soils develop the fungal networks Ingham describes as the hallmark of healthy soil. Berm trees, with their mycorrhizal networks extending into the swale's moist zone, accelerate this transformation.

Checkpoint — confirm before finalising:

• Has the contour been verified as true level on the ground (not just from a map)?
• What is the estimated catchment area above the first swale, and has the swale volume been calculated against a design storm event?
• What machinery is available for construction, and will it fit the designed swale dimensions?

An undersized swale or an off-contour swale are the two most common earthworks failures. Confirm both before any ground is broken.

Output:

SWALE DESIGN
Site: [name]
Number of swales: [n]
Slope gradient: [%]

SWALE DIMENSIONS
Width: [m]
Depth: [m]
Berm height: [m]
Berm width at base: [m]
Cross-sectional area: [m²]
Volume per 10m length: [m³]

CATCHMENT AND SIZING
Catchment area per swale: [m²]
Design storm event: [mm]
Runoff coefficient: [value] ([surface type])
Required swale volume: [m³]
Designed swale volume (at proposed length): [m³] — [adequate / increase length]

SPACING
Vertical interval between swales: [m]
Number of swale lines required: [n]

OVERFLOW
Type: [armoured / piped / natural]
Location: [description]

BERM PLANTING
Pioneer groundcover: [species]
Productive trees: [species] at [spacing]m

CONSTRUCTION NOTES
Equipment required: [hand tools / tractor with blade / excavator]
Estimated construction time: [days]
Level verification: [tool and method]

Next steps:

• Run pond-siting (within this skill) — if a swale system will generate overflow, consider whether a pond can capture it.
• /s4ag-water — pair swale installation with irrigation scheduling to use the captured water productively.
• /s4ag-agroforestry — berm tree planting is agroforestry; design the species selection for enterprise value.

---

Pond Siting

Selects, sizes, and designs farm ponds for stock water, irrigation, fire management, or aquaculture.

A pond placed in the wrong location will either never fill, constantly leak, or fail catastrophically. Pond siting is the most consequential earthworks decision — get it wrong and you have a large expensive hole; get it right and you have a water asset that pays for itself every dry summer.

Pond types and primary uses:

| Type | Primary use | Si

…

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: human-avatar
• Source: human-avatar/skills-for-agriculture
• License: MIT
• Homepage: https://www.npmjs.com/package/@human-avatar/skills-for-agriculture

Install and usage instructions live in the source repository linked above.