Here’s the puzzle a lot of solar planning eventually runs into. A meaningful amount of solar power takes a meaningful amount of land, somewhere between four and five acres for every megawatt of ground-mounted panels. In a country with a huge population to feed and a limited amount of arable land to feed them from, that math turns into a real conflict fast: every acre under panels is an acre not growing food, and every acre spared for food is an acre not generating electricity.
So where do you put a solar array that isn’t already spoken for by farming, and isn’t just empty desert either?
Gujarat’s answer, tested first on the Sanand Branch Canal near the village of Chandrasan in Mehsana district on April 24, 2012, was to stop thinking of land as the only available surface. A canal is infrastructure, not farmland. It’s already claimed, already fenced off from crops, and it runs for hundreds of kilometers in a state crisscrossed by the irrigation network built to move water from the Sardar Sarovar dam.
Someone had to notice that a canal’s surface could hold a solar array the same way a roof holds one, and that a 1-megawatt pilot project could test the idea without waiting for a bigger, riskier bet to prove it first.
What the panels actually solved, beyond the electricity
The obvious win is power generation on land nobody had to give up. The less obvious win showed up in the water itself. A canal running through Gujarat’s heat loses a real volume of its water to evaporation before it ever reaches a farmer’s field, and shading even a portion of that surface slows the loss down.
A study led by researcher Sagarkumar Agravat at the Gujarat Energy Research and Management Institute found that canal-top installations save roughly 90 lakh litres of water per megawatt per year from evaporation, north of 9 million litres annually for a project the size of the original Chandrasan pilot. The same study found a second, unplanned benefit: panels mounted over flowing water ran about 10 degrees Celsius cooler than equivalent panels mounted on dry ground, which pushed their solar efficiency up by roughly 2.5 percent. Water that was already just sitting there, doing its regular job of getting crops watered, turned out to be a decent air conditioner for a solar panel too.
Confirmed by India’s own planning body
The pilot’s basic claim, first canal-top solar installation of its kind in the country, has held up well enough that NITI Aayog, the Indian government’s policy think tank, cites the Narmada branch canal system project as the template other Indian states have since copied, from New Town near Kolkata to canal networks well outside Gujarat. Rangan Banerjee, director of the Centre for Technology Alternatives for Rural Areas at IIT Bombay, frames the appeal in blunt planning terms: “Decentralized solar solutions, especially those integrated with existing infrastructure, can address local energy needs while minimizing environmental impact.” India’s canal network runs past 120,000 kilometers nationwide, which is a lot of unclaimed rooftop hiding in plain sight if a state actually wants to use it.
What came after the 1-megawatt proof of concept
Gujarat didn’t stop at the pilot. A later, considerably larger installation strung panels across 3.5 kilometers of canal between the Sama and Chhani areas of Baroda, a 10-megawatt project that, according to SS Rathore, chairman and managing director of Sardar Sarovar Narmada Nigam Ltd, generates 16.2 million units of power every year while sparing roughly 20 acres of land that an equivalent ground-mounted array would have needed. Scale that reasoning up further, to the more than 2,000 megawatts Gujarat has floated in planning documents, and the land spared runs into the tens of thousands of acres, none of it taken from a single working farm.
Why there aren’t more of these yet
None of this makes canal-top solar the obvious choice for every new project, and it’s worth being honest about why growth has slowed since the original pilot. Mounting a panel array over a moving canal costs real money that a flat field doesn’t. Kaushik Patel, of the same Gujarat Energy Research and Management Institute that ran the water-savings study, put a number on it: “Of the entire capital cost of a CETP, 40% is that of the mounting structure, more than the cost of panels,” which pushes a canal-top megawatt roughly 15 to 20 million rupees above a ground-mounted one. Utilities pay the same rate for electricity either way, so developers have little financial incentive to pick the harder, pricier build. Yogesh Sehgal, director of the solar firm SAM Solar Private Ltd, framed the incentive problem plainly: “The rate of solar paid by the discoms is the same for both ground-mounted and CTPV, so why will a developer invest more? Unless the government provides subsidies, CTPVs will not be economically viable.” Repairs are harder too. Jaideep Parmar, a deputy executive engineer at Sardar Sarovar Narmada Nigam Ltd, described what a simple wiring fault costs on a canal-top array: “2-3 hours are lost in rectifying it as the technician can’t go below the arrays without safety equipment. All this while, the entire unit has to be shut off.”
The 2012 pilot proved the physics and the water savings were real. The economics of building more of them, at least so far, have proved slower to catch up.
The part worth sitting with
I keep coming back to how unglamorous the actual insight was. Nobody invented a new kind of solar cell or a new kind of canal. Someone just looked at infrastructure that already existed, already had a job, and asked whether it could do a second one without asking anything new of the land around it. My own read on ambition has always leaned toward that same instinct: people who want something tend to go looking for the opportunity sitting in front of them, rather than waiting for permission to build something bigger and more complicated. A canal that was only ever asked to move water is now also making electricity and losing less of itself to the sun in the process. Real tricks like that tend to hide in plain sight, right up until someone actually looks.
