A New Chip Tells Farmers When to Water | Hacker News Reader

A New Chip Tells Farmers When to Water

modernfarmer.com

I actually did an internship on a grape farm in the early 90's. Getting a degree in agonomy I was all hyped on modern technology and exited about tensiometers. The 20 year old me wanted to install them, but was shut down by the old Hispanic foreman during the meeting. He had been growing grapes for 30+ years. When I asked him how he decided when to water he said... simple. See these weeds growing under the grapes? (it was a type of pigweed I believe) When they start to droop it's time to water. Turns out, the grapes roots are deeper and the vines much larger than the weed, so this was in fact a free (and pretty accurate) tensiometer.

I love tech. And not to be a cynic, but I kind of doubt many serious "farmers" read this magazine. The old foreman maybe converted me to a degree of traditionalism. There doesn't need to be a tech solution for everything. Sometimes simple time tested solutions are better.

EDIT: One more point. Suppose one was able to figure out an individual plant's water needs. I don't know of any irrigation system that would let you meet it. The block, patch, field, whatever gets watered all together. Unless you send people to adjust individual emitters or something. And, if you are doing that you have undoubtedly spent far more money than you will ever make back in increased yields or water savings. This "modern farmer" magazine kind of drives me nuts and I see it posted here on HN frequently.

EDIT2: I guess I'm at my commenting limit so I'll leave the reply here. I agree with poster who says tech is necessary to feed more in the future. But this chip, in a production setting, is a non-solution looking for a problem. And there is way too much stuff like this that people think are "breakthroughs".

People have trouble comprehending scale when it comes to agriculture. A tray in a hydroponic gardens, even a roof full of hydroponic gardens... cool, but it isn't going to ever provide a significant portion of the worlds food. And it is way too expensive to set up and maintain for what it provides. Now... for specialty applications and a few off-season vegetables it's fine. But many thousands of acres and lots of water are required to produce the real bulk of our food... i.e. grains. The aquaponics systems pictured in the links below wouldn't even sufficiently feed the two guys in the picture. And... examining the moisture usage of individual plants per watering cycle... great for research or hobbyist or specialist crops, not practical day to day production agriculture. That's my point.

Unfortunately, simple, time-tested solutions likely won't get food production to the place it needs to be in 30 years.

Not to be deliberately morbid but there is a time tested solution to not enough food that re-appears again and again throughout human history. It's called famine and has been a tremendous regulator of human overpopulation.

Maybe we can beat famine back forever. But I'm not counting on it. The whole situation is actually pretty tenuous. You don't see a lot about it on the news, but I understand the Horn of Africa has been experiencing massive famine over the past several years and hundreds of thousands have died from lack of nourishment. From an outside (rather than personal) perspective, this is just part of the cycle. In the first world we have beat famine back over the past century and the third world has enjoyed much of the benefit of this as they import much of their grain staples. But, we have sucked the aquifers dry, damned up most of the rivers and overallocation the water, are running short of essential nutrients like phosphorous and the whole thing is oil based. So, I am guessing in spite of massive efforts, the cycle of famine re-appears around the world again. Maybe next time around we can make our political/social systems a bit more efficient and forward looking which is, in my opinion, the only way to eliminate famine forever. But I'm not necessarily counting on that either.

http://en.wikipedia.org/wiki/Peak_phosphorus

relatively speaking, does it matter if we run out in 100 years or 300 years? the point is... we are on an unsustainable path in food production.

You're answering the wrong question.

"How do you feed the number of people we expect to have on the planet in 30 years time?"

Wrong question.

"How many people can we expect to feed given the best possible conversion of sunlight into food?"

Right question.

Then you have to ask, "are we prepared to do that to our planet, or do we need to start seriously thinking about population control?"

I don't see why food production should be limited to sunlight: we're likely going to have fusion power before we hit any kind of real food cap (ie, limit on growing rather than hunger based on poverty or distribution problems).

Why not grow large tanks of GMO algae with fusion power?

GMO algae or yeast grown in tanks as a complete protein to replace the resource wasteful meat we eat so much of is a really good idea.

I expect it is the future and wish there were more work being done on it right now.

I can see using this in commercial aquaponics setups like the ones that Murray Hallam is perfecting: http://www.aquaponics.net.au/

Or the ones based on the of University of the Virgin Islands model: http://www.uvi.edu/research/agricultural-experiment-station/...

In those systems you could measure averages of water saturation in plants in each tank/division using your plant tensiometers and then use your pumps appropriately. It would just be another sensor in your installation along with those measuring PH, EC, temp, etc.

High pressure aeroponics with led multi spectrum lighting = great grow op....

Check out any of the commercial HP aeroponics, this is the future of cultivation.

http://aeroponics.com/aero43.htm

http://www.nasa.gov/vision/earth/technologies/aeroponic_plan...

From a personal hobby of mine, fast yield high bloom, small space requirements semi sterile and can be completely organic

I've been looking at this very tech and getting excited about it. Thanks for the links!

I'd argue that the weeds were a type of technology used in that role: they were using something that sent down probes in to the ground slightly shallower than the grape roots, and had a visual change when the levels dropped below a certain amount.

In fact, the weeds have some benefits: they average over a small patch, have good volume coverage, likely have a water loss correlated with the factors that cause grape water loss, etc.

If you view plants as evolved machines, then it makes sense that weed with certain properties (such as not strangling the vines (too much)) would make really good sensors.

Trying to replicate that, as you point out, is a really hard challenge, and only matters if you're doing something complicated with the data (such as integrating many sensors in to an automated system that locally adjusts the levels).

I don't know about grapes, but as a non-irrigating grain farmer, I could see attaching something like this to a few plants to keep records on general plant health in a given field. While there is little I can do if the plant is under stress, I expect the data can be correlated to final yields, which has the potential to improve my marketing abilities.

But as you point out, it comes down to dollars and cents.

I've recently imagined a spindly-legged robot that left a charging home base with water, poison, seeds and fertiliser. It would wander out into a home garden and handle whatever tasks were necessary - identifying and neutralising young weeds, checking soil moisture and watering at a per-plant level, applying fertiliser based on a log of past applications and knowing which plant it was servicing, etc.

Picture a daddy-longlegs spider, but with four jointed legs about one metre long (three always touching the ground while the fourth moves position). Then a light-weight body with storage and probes/applicators.

If the base station had water supplied, solar charging, etc - I'd find this useful in my vegetable garden. Swarms of them could potentially look after larger gardens.

Your robot would have to make a bunch of trips or else be really big for something like a grape vine or a citrus tree. Even for a tomato.

The proper way to water trees or grape vines is tens or hundreds of gallons at a time then let the soil dry down. Otherwise you don't get the root development that takes up nutrients and anchors the plant. Trees that are watered shallowly for their whole lives often blow right over on maturity. And, in watering in California, Arizona and probably most places in the western US you need what is called a leaching fraction. That means you need to water well beyond the root zone to push out and avoid the accumulation of salts. It is far more energy efficient to pipe or flow the water to the plant than carry it. And... we are talking about a significant amount of water.

This is what I mean by scale. People have trouble conceiving it in agriculture for some reason. They tend to believe a few computer controlled hydro gardens can feed the world. Nope. You know all the countless green circles you see on the ground while flying across the country? That's real food production in action and even with all that we may become strained. A change in global diet might help. Improved genetics might help. But farmers aren't stupid. Over the course of generations they have achieved some degree of efficiency. Gadgets might help get a bit more but the easy gains have been largely worked out.

This reminded me of 2008: http://www.businessweek.com/stories/2008-12-04/how-to-make-t...

[Spoiler alert] soil sensors + telemetry + spectrum = Internet of Things [/Spoiler alert]

Interesting-- tensiometers have been used in soil for a few years now, but never implanted into the plant. Tensiometers basically measure the pressure change caused by plant roots soaking up water.

so dumb question, would not some forms of spectral analysis or similar work? This chip solution seems a bit too work intensive for most crops, would not large fields be a bit much for installs?

for my dumb question I found this http://isda.ncsa.illinois.edu/Hyperspectral/hyperAgricultur....

Spectral analysis (via drone, satellite, airplane etc) is good but won't give you the resolution of 1 chip per plant. It does seem labor intensive to implant every plant, but not impossible. Even large farmers trim and take cuttings from individual plants.

You can get plenty of resolution off an airplane. I have seen aerial imagery of sufficient resolution to do stand count analysis, which is basically counting individual green shoots of corn in the spring.

As another poster mentioned, though, spectral analysis shows problems in the leaves of the plant, where it is too late: if your plant is yellowing due to lack of water or nutrients, you've already decreased the potential yield from that plant. It's better to detect deficiencies in the ground.

Assuming you could cost effectively do a spectral analysis it is still not optimal. By the time the plants are showing stress it is past the optimal time to water.

The true place for these chips is university and cooperative extension research. By knowing more about how plants use water they might be able to come up with a more easily measurable metric that farmers can use.

Reply on news.ycombinator.com