Why Wheat Yields Fall After Paddy: The Hardpan Under India's Rice Fields

In British arable farming a plough pan — what Indian extension material calls a hardpan — is damage. It is what heavy machinery and untimely cultivation leave behind, and the remedy is an argument about how deep to loosen and whether to loosen at all. Professor Dick Godwin, who has spent a career on the engineering of it, puts the cost plainly: a compact soil restricts root development, impedes air and water movement and raises the risk of anoxic conditions — and if nothing is done, soil health deteriorates and yields follow it down.

Now cross to the Indo-Gangetic Plains, where the same layer is not an accident. It is the objective.

Puddling — working the soil wet until its structure collapses into slurry — is done deliberately before transplanting rice, precisely because the dense layer it leaves stops water draining away. It cuts percolation losses by roughly 14–16% and crop water demand by something like 10–25%. On about 10 million hectares of Indian farmland it is repeated every kharif, and has been for decades.

So the question a Punjab or eastern UP farmer faces is not the British one. It is not "how did this pan get here and how do I remove it". It is harder: the layer that makes flooded rice work is the same layer that penalises everything that follows it.

The pan sits just below the depth the implement reaches, and it is dense enough to measure easily. In a multi-year trial on sandy loam in the rice-wheat belt, soil in the 14–20 cm layer of normally puddled plots reached an average bulk density of 1.74 Mg/m³, against 1.57 in shallow-puddled plots after three years. Raising puddling intensity alone pushed the 16–18 cm layer from 1.63 to 1.67, and the 18–20 cm layer from 1.61 to 1.66.

Those look like small numbers until you consider what they describe. Bulk density is mass per unit volume of soil — as it climbs, the pore space that roots, air and water occupy is what disappears. A layer at 1.74 is not soil that a root pushes through slowly. In large part it is soil a root does not enter at all, and one that water crosses only very slowly, which is exactly why the technique works for a flooded crop.

This is also why the pan is durable. It was not made by one pass of one machine; it is the cumulative product of many seasons of working wet soil at roughly the same depth, and a single dry-season cultivation does not undo it.

The cost does not appear in the rice season. It appears in rabi, when wheat is drilled into the same field and its roots reach the layer the rice crop needed.

Published comparisons put wheat yields about 8% lower after puddled transplanted rice than after direct-seeded rice. The rice saved water in kharif; the wheat pays for it in rabi. Because the same field grows both, the honest accounting is the net of the two — and that netting is almost never done, because the two crops are planned, costed and subsidised as separate enterprises.

The scale makes this worth getting right. The rice-wheat system covers roughly 10 million hectares in India, producing on the order of 130 million tonnes of rice and 106 million tonnes of wheat — around three-quarters of the country's cereal output. An 8% penalty on the second crop of that system is not an agronomic footnote. It sits underneath the same rotation that ICAR now wants capped and diversified.

This is where the pan stops being a soil-science curiosity and becomes the practical question of the decade.

Direct-seeded rice asks the crop to do the opposite of what puddled rice does. Instead of sitting in standing water, the plant must send roots down to find moisture. The water case for it is strong — trial plots have used around 924 mm of irrigation under DSR against 1,441 mm under transplanting, a difference of over 500 mm — and that is before the labour and the puddling operation itself are counted.

But a field carrying a decade of puddling has a dense layer sitting exactly where those roots need to go. It is a large part of why DSR results vary so much between neighbouring fields with the same seed, the same drill and the same herbicide plan: what the previous ten years left at 15 cm is not the same in both. Our own map of which states are leading and lagging on DSR tracks the adoption; the pan is one of the reasons the map looks patchy even inside a single district, and it is worth reading alongside Chhattisgarh's DSR advisory.

A farmer switching to DSR on a long-puddled field is not merely changing an establishment method. He is asking a soil profile built for one system to serve the opposite one.

If the pan has to be broken, the instinct is to go deep. The engineering says otherwise, and the reason is worth understanding before anyone hires a subsoiler.

As a tine is pulled through soil, it lifts and fractures the soil forward and upward along failure planes spreading from the tip. Work deeper and the resistance to that upward movement grows, until a point where it becomes easier for soil to flow sideways around the foot than to be lifted at all. Past that point — the critical depth for that tine in that soil — the compacted zone is left essentially undisturbed and the machine has cut a channel and burned diesel. Going deeper does not achieve more; past a certain depth it achieves nothing.

Two related warnings from the same body of work travel well to India. Tillage where it is not required does more harm than good, disrupting root and water pathways that were working. And inadequate or badly adjusted equipment can re-compact the loosened soil on the very next pass, leaving the field worse than before it was treated.

That matters here more than in Britain, because in much of India the decision is not made by the person who owns the soil. Subsoiling is bought as a custom-hire service, priced by the hour or the acre, from an operator with one machine set one way. The incentive is to run deep and move on.

The first step costs nothing. Godwin's position, after decades of measuring this, is that nothing yet beats a visual assessment with the crop actively growing — as he puts it, we have yet to master the digital spade.

Walk the patches where the crop is visibly poorer in the growing season and dig. What you are looking for is restricted rooting, large slabs of soil that are hard to prize apart, and water that will not move down the profile. That tells you whether a pan exists, at what depth, and how thick — and those three answers determine the treatment. Surface problems from trafficking or slumping need gentle loosening. A discrete pan needs fissuring at its own depth, not below it. A deep, massive, structureless zone is a much harder problem and rarely fixed in one pass.

Then the decision that is specific to India, and it is a decision rather than a technique. If the field is staying in puddled transplanted rice, the pan is an asset and breaking it will cost irrigation water for no return. If the field is moving to DSR, to a diversified rotation, or to any crop that has to root deep, the pan is the first obstacle and should be broken once, properly, at the right depth — and then protected from the traffic that will rebuild it.

The layer itself is neither good nor bad. It is the physical record of a choice that was made every season, by people who mostly were not thinking about the crop that follows.