All techniques › Water harvesting & retention
Water Retention Landscapes & Farm Ponds
Storing runoff on the farm, from rainwater capture to dispersed check dams to open ponds. What is well measured here is how much water you lose, not how much you gain.
The short answer
Start with rainwater harvesting, because it is the legal path of least resistance. In England, storing purely harvested rainwater needs no abstraction licence, while taking more than 20 m³/day from ground or surface water does. Roof and yard capture is cheaper and simpler than anything else on this page.
An open pond in a Mediterranean summer loses a great deal of water. Measured evaporation at a Portuguese reservoir came to 450 mm between June and September, which on our arithmetic is about 45% of the depth of a one-metre pond in a single season. If you build open storage, build it deep and compact: evaporation scales with surface area.
Be careful what you believe about the benefits. Europe's flagship retention programme built roughly 100,000 structures and cannot fund its own monitoring, so its headline claims are projections rather than results. And exhaust soil water storage first — mulching and organic matter store water with no evaporation penalty and no permits.
What it is
Two related but genuinely different things, which are often blurred together.
Farm ponds and on-farm reservoirs are excavated or dammed stores of open water, usually for irrigation or stock. Concentrated, engineered, and permitted.
Water retention landscapes are many small, dispersed interventions across a catchment: check dams, weirs, bioretention basins, infiltration trenches, restored wetlands, beaver dam analogues, contour strips. The aim is to slow runoff everywhere rather than store it in one place.
The distinction matters more than it might seem, because the dispersed version largely avoids the evaporation problem that dominates the pond version. It puts water into soil and aquifers rather than under an open sky.
How it is meant to help
It stores winter and storm runoff for the dry season, when rainfall and demand are out of phase — the fundamental Mediterranean problem.
It recharges groundwater by holding water on the landscape long enough to infiltrate.
It reduces flood peaks and erosion by slowing runoff across many small structures.
Local cooling and humidity are also claimed as benefits, through evaporation. Note the tension honestly: you cannot count the same water as both stored and evaporatively cooling. It is one or the other.
The evidence is asymmetric, and that is the story
What is well measured here is the losses. What is claimed is mostly projected.
Evaporation: measured, European, and unambiguous
Measured at Alqueva Reservoir, Portugal by eddy covariance (HESS 24:5973, Copernicus):
| Period | Evaporation |
|---|---|
| June–September 2014, total | 450.1 mm |
| June, daily mean | 3.7 mm/day |
| July, daily mean | 4.0 mm/day |
| August, daily mean | 4.5 mm/day |
| September, daily mean | 2.5 mm/day |
What that means for a farm pond: an unshaded pond one metre deep in southern Portugal loses roughly 45% of its depth over a single summer to evaporation alone. A shallow pond can lose most of its contents in exactly the season it was built for. (That last figure is our arithmetic from the measured rates above, not a published result.)
In Andalusia, Spain, 76 reservoirs evaporated an average of 547 hm³/year — 8.0% of inflows — and the losses were driven chiefly by morphology, especially the ratio of flooded area to stored volume (Journal of Hydrology: Regional Studies).
That gives the one genuinely quantitative design rule on this page: minimise surface area per unit of stored volume. Deep and compact beats wide and shallow, every time.
Small agricultural reservoirs specifically are described as "an overlooked component of water accounting" in a warming climate (Earth's Future, Aminzadeh et al. 2024), and the "water storage paradox" — building more reservoirs while losing more to evaporation — is documented at regional scale (Scientific Reports 2025). For context, agricultural withdrawal accounts for 40–80% of annual water consumption in Spain, Portugal and Italy, and irrigation there depends heavily on local storage.
The practical conclusion: storing water in soil has no equivalent loss. On a Mediterranean farm, exhaust soil storage before building open-water storage.
Slovakia: a real programme at scale that measured almost nothing
The Landscape Revitalisation and Integrated River Basin Management Programme, adopted in October 2010, is the largest European example (EEA Climate-ADAPT; UN Decade on Restoration).
What was built (2010–2012): 488 villages and towns implemented roughly 100,000 retention elements in 18 months, creating or restoring 10 million m³ of retention capacity — about 4% of the intended ten-year plan — and providing 7,700 seasonal jobs.
What was claimed but not measured. The successor Košice programme (2021–2030, ~700,000 ha, €408 million) projects 12,000 l/s of aquifer recharge, regional cooling of 0.7 °C, 1.8 Mt of carbon sequestration and €32 million in annual crop-yield gains. These are expected results on full implementation. The case study states plainly that the programme "lacks sufficient funding for monitoring and gauging stations to evaluate the success of the catchment restorations." The 2010–2012 predecessor ended prematurely due to political changes, with funding suspended before monitoring was implemented.
So Europe's flagship water-retention programme built a great deal and measured very little. That does not mean it failed — the interventions are real, at scale, and government-adopted. It means nobody can currently tell you what they delivered, and you should treat the headline numbers as intentions rather than findings.
Much of the enthusiastic material also traces to the programme originator's own network, which is worth knowing when weighing it.
Permission comes first
Water you retain is water that does not reach downstream users or the groundwater-dependent ecosystems the EU Water Framework Directive explicitly protects (European Commission).
For England, the rules are reasonably clear:
| Situation | Requirement |
|---|---|
| Abstracting more than 20 m³/day (20,000 l) from one or several sources | Abstraction licence from the Environment Agency or NRW |
| Storing only harvested rainwater | No abstraction licence needed |
| Mixing harvested rainwater with ground or surface water, then abstracting | Licence may be required |
| Agricultural unit of 5 ha or more | Excavation and engineering works reasonably necessary for agriculture can fall under permitted development (Part 6 Class A) |
| Any such works | Prior notification always required; prior approval often required |
Sources: Environment Agency rainwater harvesting position statement; Farmers Weekly on planning an on-farm irrigation reservoir. Both read in September 2026; neither publication date was captured, and abstraction rules change, so confirm current guidance with the Environment Agency before you dig.
The most useful practical insight here: harvesting rainwater is legally far simpler than abstracting water. Capturing roof and yard runoff puts you in a materially easier position than taking from a watercourse or borehole.
England, Spain, Poland and Portugal are documented on the permits page. In Spain even a rainwater use is registered with a technical file and a meter, and any work within 100 m of a watercourse needs a six-month authorisation. In Poland a rain-fed pond up to 5,000 m² needs only a 132.33 zł notification (2026). The rest of eastern Europe remains open. The risk of getting this wrong is demonstrated rather than theoretical — in Slovakia, officials ruled retention structures illegal constructions and fined the people who built them. Check with your national water authority before you dig.
Where it fits
| Context | Fit | Why |
|---|---|---|
| Rainwater capture from roofs and yards, any zone | Best starting point | Legally simplest, cheapest, no abstraction licence in England |
| Continental catchments with flood-and-drought cycles | Good — dispersed version | Exactly what the Slovak model targets |
| Mediterranean, dispersed infiltration measures | Reasonable | Water goes into soil and aquifer, avoiding open-water loss |
| Wet temperate Europe, winter-fill storage | Good | Low evaporation; the seasonal offset is the real benefit |
| Mediterranean open ponds | Questionable | ~450 mm of summer evaporation. Build deep and compact, or not at all |
| Shallow, wide ponds anywhere hot | Avoid | The worst possible surface-area-to-volume ratio |
| Farms that have not yet maximised soil water storage | Not yet | Soil storage has no evaporation penalty. Do that first |
What it costs
The one European figure we have comes from Slovakia's programme, on a 2019 cost basis:
| Setting | Cost per m³ of retention capacity |
|---|---|
| Countryside / rural projects | €5/m³ |
| Urban construction | €28/m³ |
So roughly €5,000 for 1,000 m³ of dispersed rural retention capacity. Treat that as order-of-magnitude only — it is a programme-level cost for many small dispersed structures in a relatively low-construction-cost member state, and not a farm-pond construction quote. We have found no European excavator hire rates, pond lining costs or terrace restoration figures, and would rather flag the gap than invent numbers.
The binding constraints turn out to be funding and permitting, not construction. The Slovak programme's own stated obstacles were insufficient national funding, limited access to CAP money, land ownership disputes dating to 1950s nationalisation, and unresolved coordination between ministries.
Storage works from first fill. Aquifer and landscape-scale effects are multi-year and, on the Slovak evidence, rarely measured.
Commitment
Permanent, with an ongoing obligation — but less absolute than terraces or swales.
A pond is a lasting landform change carrying permanent maintenance: desilting, embankment integrity, safety fencing, liner replacement. Dam or embankment failure is a genuine safety hazard, in the same category as the landslide risk on swales and contour earthworks.
Small dispersed structures are the more reversible choice. A check dam or infiltration trench can be removed or allowed to revert far more easily than an excavated reservoir, and individual failures are small rather than catastrophic. That is a strong argument for the dispersed model over one large pond, quite apart from the evaporation argument.
There is also a regulatory dimension to reversibility: an unpermitted structure may have to be removed at your own cost, as the Slovak fines show. Permit first.
How to start
- Exhaust soil water storage first. Mulching, residue retention and organic matter store water with zero evaporation penalty and no permits. Building a pond before doing this is solving the harder problem first.
- Start with rainwater harvesting, not abstraction. Roof and yard capture is the legally simplest and cheapest route.
- Check permits before you dig — abstraction thresholds, planning and prior notification, and reservoir safety rules all apply.
- If you build open storage, build it deep and compact. Evaporation scales with surface area, and the Andalusian data show morphology dominates losses.
- Reduce the exposed surface where you can — shading, floating cover, or siting in shade.
- Prefer many small dispersed structures to one large pond where the goal is landscape water rather than irrigation supply. Less evaporation, more reversible, smaller consequences when something fails.
- Measure something. Europe's flagship programme failed to monitor itself. A staff gauge and a few dip readings put you ahead of a €408 million programme. The measure-something page sets out what to record and how to read it.
Risks and failure modes
Evaporation losses in hot climates — up to roughly 450 mm over a Mediterranean summer, which can be most of a shallow pond's contents. The central risk, and it worsens as the climate warms.
Building storage that a warming climate empties faster — the documented "water storage paradox".
Regulatory failure. Unpermitted structures declared illegal and fined, as happened in Slovakia.
Dam or embankment failure, a hazard to people and property downstream.
Downstream harm. Water retained upstream does not reach other users or the groundwater-dependent ecosystems the Water Framework Directive protects.
Siltation reduces capacity over time, and desilting is a recurring cost.
Believing unmonitored benefits. This is how a farmer spends real money on projected returns. Where a programme or supplier quotes outcomes, ask whether anyone measured them.
Claims checked. Two circulating claims are run down to their sources: a farm pond stores water for the summer and Slovakia's programme recharged aquifers and cooled the region.