Dry Farming Vegetables Without Irrigation: What Holds Up

I lost most of a garden in 2011 by following a book written for a wetter country, and most of another in 2013 by following a different one. Dry farming vegetables is possible on this hillside, but almost none of the standard advice survives contact with four months of absolute drought and twenty-five centimetres of soil over limestone. This is what was left after I stopped copying and started measuring.

What this covers

• Why the sowing calendar matters more than any watering technique
• The eight-week window between last frost and dry ground, and how it constrains everything
• Mulch depth and material, and why five to eight centimetres and not more
• Where I still water, and where I stopped
• What I got wrong about deep initial watering

Start with the calendar, not the hose

The first thing to understand is that sowing here is not timed by the calendar. It is timed by the remaining soil water.

The question is never when do we sow tomatoes. It is: how many weeks of moist ground are left, and how many does this plant need to reach a size that can hold on?

On my ground, the top twenty-five centimetres hold usable moisture from about mid-October until roughly mid-June. Eight months. Outside that window, anything you sow depends entirely on watering.

That gives two correct sowing seasons and one wrong one.

Autumn, mid-October to late November. The best. The plant roots in with the rains, moves slowly through winter, and starts spring with a root system already in place. For anything that takes cold: onions, garlic, broad beans, most leafy greens, all the aromatics, every cover crop.

Early spring, mid-February to late March. For what will not overwinter. The aim is to get roots down before the surface dries.

April onwards is the wrong season. It is when most people sow, because that is what the books say and it is convenient. It is also when the plant cannot make it: up in a fortnight, six weeks of moisture ahead of it, and into June with ten centimetres of root.

The eight months of usable moisture are not a fixed quantity, either. How much of the winter rain actually enters the ground rather than running off a slope depends on what you have built to slow it — which is why swales and berms done in autumn change what is possible the following summer.

Anything I cannot sow by late March goes in as a transplant, not seed.

The eight-week problem

At five hundred metres, frosts run until mid-April and the first ones arrive in late October.

That creates the real constraint of this place: the gap between the last frost and the drying of the soil is about eight weeks. Mid-April to mid-June.

Eight weeks for anything cold-sensitive to establish. That, not heat and not poor soil, is the tightest limit I work under. Two ways round it — transplants started earlier under protection, or shorter-cycle varieties. I use both.

Spacing

Wide, and wider than a seed packet written for a rainy country will tell you.

The reasoning is straightforward: with a fixed and finite amount of soil water, each root system needs a larger volume to draw from. Plants do not share water in a dry summer. They compete for something that runs out and is not replaced until September.

It looks wasteful in April. It looks like the only sane arrangement by July.

The space between is not wasted if you use it. Perennials that need nothing once established will sit in it without competing for the shallow water the vegetables depend on — the caper bush is the one I have had most success with on the driest edges.

Mulch, and how deep

Five to eight centimetres, compacted. Not more.

I have read four to six inches recommended, which is ten to fifteen centimetres, and it caused me problems. Above about ten centimetres I get rodents in winter, and sowing becomes difficult. I lost two young trees to bark damage under deep cover before I worked out what was happening. Below three centimetres it does nothing measurable to soil temperature.

Timing is a deadline, not a preference. The material has to be down before the heat arrives — late April to mid-May here. Laid in June it does very little, because the ground is already warm and dry and all you are doing is keeping it that way.

Material: whatever the holding produces. I have no straw, because I do not grow cereals at any scale. What I have is chipped olive prunings, cut vegetation, leaves, spent aromatic stems after distilling, and vegetable stalks.

These are not substitutes for each other. The thing that matters is how fast each one breaks down. Chipped olive prunings last two or three seasons and belong under trees. Cut vegetation lasts one season, barely, and belongs in the garden. Aromatic residues are among the best material I have, but there is never much of it.

One warning that appears in no book I have read: dry loose material in July, in a hill village, is fuel. I keep boundaries and roadsides cut low and compacted instead, and accept the moisture I lose there.

Watering, and what I got wrong

I used to give transplants a deep initial soaking, on the theory that it drives roots down after the moisture.

The 2023 drought told me otherwise, and the evidence was not subtle. Across the whole holding, what I had watered generously in its first years died first. What I had watered sparingly, or not at all, survived. Three fig trees that had come up on their own from an old root, that I had never watered once, were in excellent condition when trees I had nursed were dead.

A watered plant has no reason to go deep. It finds water at ten centimetres. When the year comes that the top ten centimetres empty completely and the barrel is empty too, it has nowhere to go.

So: a modest soaking at transplant, and then nothing. And where I do water, it is infrequent and deep rather than frequent and light, for the same reason.

Seed

I keep my own, from open-pollinated traditional varieties, and I select for one thing.

I keep seed from the plants that did best in a bad year, not from the most productive ones. Since 2019 I have done this systematically. Six seasons on, my own lines do better on my ground than bought seed of the same variety. That is not a discovery — it is what everyone did before hybrids — but doing it deliberately, with a criterion, is different from saving whatever was to hand.

What I keep from this

• Sowing dates are set by remaining soil moisture, not by the calendar
• The eight weeks between last frost and dry ground is the real constraint here
• Five to eight centimetres of mulch, down before the heat, not after
• Match the material to the job: woody lasts and protects, soft breaks down and feeds
• Water once at transplant, sparingly, then stop
• Select seed for resilience in the bad year, not yield in the good one

Questions people ask

Can you really grow vegetables with no irrigation at all? Not entirely, in my case. I water at transplant and I keep a small reserve for the vegetable ground in an extreme year. What I have removed is routine summer irrigation. The garden survived 2023, when my water ran out on the twenty-second of August, but with reduced production.

How much rainfall do you get? Around 750 millimetres, almost all of it between November and March. The total is less important than the distribution — three to four months with none at all is what breaks most standard advice.

Does this work on shallow soil? It works differently. Depth of soil turned out to be more decisive than fertility on my ground. The places that held green vegetation into mid-June were worth planting intensively; the ones that browned off in mid-May were not worth attempting with vegetables at all.

What about weeds? Different answer in different seasons. Through winter I leave everything, because there is water to spare and bare ground erodes. From June the competition for water is real and I cut — cut at the base, never pull, so the root stays in the ground and the top becomes cover.

Related reading

• Building a Dry Stone Terrace — where the soil moisture this method depends on actually collects
• Rainwater Harvesting with Swales and Berms — the earthworks that build the reserve this approach draws on
• Caper Bush: A Drought-Tolerant Native — a species that asks for nothing once established

Before you go

I send one letter a month from this hillside — what was tried, what failed, and what the season is doing. No schedules, no promises about yields.