Wednesday, October 28, 2015

Earthside - Indoor farms

This is more of a movement than a single example. In many former industrial centers, abandoned warehouses and factories are being converted to grow food. The approaches vary but generally focus on water and energy efficiency and year-round productivity. Keywords are urban farms, vertical farms, etc. I'll list two examples and then discuss pros and cons, plus the (fairly obvious) synergies with space-based agriculture after the break.

An excellent example is The Plant in Chicago. I recommend checking out their site; they have a very interesting business model that encourages multiple companies to operate farm areas in their facility. One key element is that they intend to be net zero energy consumers; this has relevance for agriculture in space since both environments would be using only sunlight for production.

Another example is the Mirai group, a systems integrator that handles this type of farm in Japan. Here's an article  from Inhabitat showing their core facility (the Green Room), a 25,000ft² (2323m²) facility with 15-tier grow racks producing about 10,000 head of lettuce a day. They don't specify variety so it's hard to make a direct productivity calculation; let's assume it is a leafy type harvested at around 500 grams. That would mean 5,000kg per day, or 2.15kg/m² per day (143g/m² per day for each layer). Their site states that seven tiers fit within 2.7m, so the facility must be roughly 6m high. Just under 39cm total height for a layer is much more aggressive than my planned 46cm, but clearly it can work for lettuce. My reference 4m system would fit 10 of these layers or about 1.4kg/m² per day (compared to my projection of 8 layers and 1.1kg/m² per day).

This movement has generated controversy and criticism. Not all is as rosy as the press releases make it sound.
Let's review the positives:
No pesticides
No herbicides
Reduced or eliminated microbial contamination
Reduced water use
Potentially reduced electricity use
Potentially reduced fossil fuel use for shipping
Year-round production
High yield production, particularly on a land-area basis

Some negatives:
Potential for chemical leaching from plastics
High capital costs for structure and equipment
High technology requirements
Specialized training required
Soluble nutrients required, some of which are petroleum byproducts
Hazardous wastes are generated in the manufacturing processes for lighting
Most approaches focus on leafy greens rather than balanced fruit and vegetable output

Some arguments that do not seem to be supported by evidence:

Different nutrient content of hydro/aero food
 - This certainly can happen, but hydro veggies are more likely to have better nutrient profiles due to reduced stress. This is something that should be sampled and measured on an ongoing basis to provide feedback for nutrient and environment controls.

Higher carbon footprint
 - One can build a vertical farm using very wasteful methods and materials and probably manage to pump out more carbon than some traditional farms. That's not very likely, and it would be more representative to look at a modern intensive farm with multiple field operations a year consuming diesel and with product shipments running hundreds of miles by truck. One could also look at off-season produce shipped from other hemispheres, traveling thousands of miles by boat, train or truck and burning fossil fuels the whole way. In general, a project like this given even a little bit of thought to environmental impact will have an overwhelmingly positive effect on resource consumption.

Unnatural monoculture methods are bad
 - In dirt, yes, but the word unnatural is a curious word to use in this case. Many species grow in monoculture stands or fields in nature; in fact, the ability to crowd out other species is often a big part of what makes a food crop desirable since it can outcompete weeds and needs less or no herbicide. In dirt, companion crops are often a very good idea in theory. In practice they tend to complicate mechanical harvesting and our society is not willing to go back to all hand-harvested food.

Unnatural cloning techniques are bad
 - No. Just, no. Entirely natural process. Just because it works best in a sterile laboratory doesn't mean it's some hideous techno-scourge that will kill us all in our sleep. Of cancer. Because labs, right?

Unnatural light spectra are bad
 - See above. Plenty of plants are adapted to a variety of light conditions including under water (reduced reds) and understory rainforest (reduced reds and blues).

Makes GMOs easier to grow
 -  This is true, but growing them in a carefully controlled environment with little risk of accidental release is a positive. Hating all GMOs universally to the point that anything that might make growing them possible becomes an evil enemy is irrational.

Doesn't help poor people, third-world countries, far-north residents, etc., etc.
 - Not every advance in agriculture has to solve all possible problems with food security worldwide. Sometimes when a technology is first invented it is very complex and fragile. Scientists deal with this because they handle one-off tech all the time. Engineers convert these lab projects into reliable (though generally still very complex) processes. A murky cycle follows that trades reliability and simplicity back and forth, with extra innovation helping improve the sum of the two over time and market forces pushing things all over the map. Eventually we get something that a student can build for a science fair or an aid organization can pack into a shipping container and send into some war-torn or disaster-stricken area. Some of them will even work as intended.
 We're at the optimization stage and have been for decades. Given a set of constraints (for example, must work in Nigeria with nothing more advanced than plastic water bottles) and enough money, a workable solution can be developed. If you see a project that doesn't do what you think it should, ask yourself what the designers of that project are trying to achieve; the results should be informative.

 Anyway, enough with the rantish material. Agriculture in space has no choice but to use high-intensity methods like hydro and aero production. Floor space, volume, power, water, air, heat and light are all very expensive. Vertical farming faces cost challenges as well, and this pressure drives innovation that improves performance on Earth and could improve performance in space.
 I think that organizations investigating space agriculture could partner with groups (researchers as well as companies) investigating terrestrial agriculture. This surely already occurs, but much of it appears to be in the form of technology transfer from NASA's suppliers to hydrofarm companies. If NASA and similar groups were to broaden their criteria and targets in order to explicitly produce improvements that will benefit Earth-based agriculture, that could induce companies to contribute and help fund that research. Perhaps an industry group for growers and their suppliers would make sense as a central clearinghouse for research funding and results. Suppliers providing lighting, instrumentation, etc. to researchers would be well-positioned to commercialize these developments and sell them to private companies. Actual end users could provide yield data which could be aggregated and made available to researchers so they can analyze the real-world effects of various environmental parameters and specific plant varieties. Space programs would benefit from additional research funding, but also from real-world production data on a much larger sample set and variety of species and conditions. That could help steer breeding programs and further research that would specifically benefit space food production.

 The ideal situation in my opinion would be an industry organization with voluntary, paid membership; this would include sellers, integrators, suppliers, job trainers, breeders and possibly philanthropic groups. Members would gain access to an expert panel of biologists and engineers as well as to the results of research programs funded by the organization's fees. Members would also go through a quality review to ensure their practices and products are safe and effective. The organization would cooperate with universities and research organizations around the world, helping to fund research that will benefit their members while helping researchers tie their work into active production facilities. The body of knowledge about plant behavior under soilless cultivation would be greatly improved and the productivity of member facilities would grow. A lot of food would be grown for less money and less resources, and the improvements would continue to accumulate. People or companies looking to start a vertical farm would be able to trust that organization members are providing safe, effective processes and products using the latest information available. Space programs would be able to select from a wide pool of expert suppliers who already have ties to their own research programs, reducing cost and risk for a critical part of long-term manned space exploration.

 Most of the grocery stores I've seen are enormous single-level buildings. They have more than the average amount of HVAC equipment up top, but otherwise there's a huge expanse of flat space. Imagine if each of these stores were to build vertical farms in that space. Create local jobs, provide fresh and local produce year-round without shipping costs, cut down on food waste and boost the profitability of their property at the same time. Some might argue that they could be installing solar panels instead; that's true, but it is cheaper to move electricity than to move tomatoes. Put solar farms on cheap scrubland and grow crops on the roof instead.
 Right now the up-front costs would be too high and the operation would be too complex for most stores. What we need is a tipping point where the pool of trained labor is available, the equipment is ready to be mass-produced cheaply and the complex parts of the operation can be automated or contracted to experts. Concerted effort and investment in the field could get us there by the end of the decade.


  1. Given that light saturation occurs at around 100 W/m^2 for most plants, what are the prospects for using light pipes to grow multi-level crops using just sunlight?

  2. Plants care about PAR, photosynthetically-active radiation. In reality we are concerned with units of nanomoles of useful photons per m² per second (µmol/m²s). Sunlight is about 37% useful and each watt/m² converts to about 4.57 µmol/m²s.

    Light saturation is a question of intensity. Edible crops vary from 500 to 2000 µmol/m²s (110 to 440 W/m²) depending on species and many other variables. For AM1.5 sunlight at peak intensity of 930 W/m² we can collect 4250 µmol/m²s. That means each square meter of solar collector can illuminate anywhere from 2 to 8 m² of crops at full noon.

    Saturation tells us whether we are wasting extra light on a crop, but that's only part of the story. Reaching saturation for an hour or two a day isn't enough; most food crops require 6-8 hours of good light and many can make use of 16+ hours. This is measured as the DLI or daily light integral. DLI is measured in units of mol/m² per day.
    At the low end of the scale is lettuce, requiring around 17 mol/m². As a densely leafy plant lettuce can handle fairly high intensities, but even at 500 µmol/m²s (1.8 mol/m² per hour) the plant does not benefit from more than about 9.5 hours of light. On a 16-hour light cycle, lettuce needs an intensity of 295 µmol/m²s or a bit under 65 watts per m² of sunlight. (Wattage of good spectrum LEDs could be as low as half that value.)
    Most commercial crops are in the neighborhood of 26 mol/m² per day. Some higher-intensity crops like tomatoes can handle around 38 and potatoes can take 42.
    At the high end is brutally intensive wheat, which can be driven at 140 mol/m² per day on a 20-hour 425-watt light cycle. If there are no resource bottlenecks or sources of stress, the edible yield of a plant is based almost entirely on the amount of photosynthesis performed. This is measured by the DLI, the sum of all useful photons striking the plant over the course of the day. Wheat driven at this level can produce many times the yield of field wheat.

  3. So, how much light is available? Depends on where the farm or facility is located. NREL has excellent maps for the USA; I'm not sure about good sources for that kind of data elsewhere. Use the PV solar radiation maps / models; these take into account weather (sunny days), seasons, latitudes, etc., to arrive at a figure for available solar energy in kWh/m² per day. 1 kWh/m² per day is equal to 16.45 mol/day, or about enough for a lettuce crop. The majority of the US gets at least 4 kWh/m², enough to grow 1.6m² of vegetable crops for every 1m² of collector area. The desert southwest can see 6-7 kWh/m² per day, potentially growing up to 2.8m² of vegetables or 4.5m² of leafy greens per m² of collector area. An offshore farm in the Gulf of Mexico could see similar insolation with much higher humidity and more even temperatures.

    A third factor I haven't mentioned is CO2 concentration. C4 plants (maize, sorghum, sugarcane) have an innate ability to concentrate CO2, making photosynthesis much more efficient. These plants generally see little benefit to extra CO2 and can even handle very low CO2 concentrations, perhaps as a finishing step in larger life support systems. C3 plants (wheat, rice, pretty much everything else) lack this ability; their key enzymes are only at about 25% efficiency at ambient CO2 levels (380-400 ppm). Boosting CO2 to 1000-1200 ppm can double or triple their efficiency and enable the use of much higher light saturation levels.

    All of this applies only if the first-order problems are solved. Plants need proper nutrients, water, temperature and humidity in addition to proper lighting. Some plants require a well-defined day-night cycle, so even in ideal conditions a year-round lightpipe farm will use a combination of seasonal and all-year crops for best productivity.

    If one were to capture sunlight using PV panels, convert to electricity and drive LEDs with it then that 4kWh/m² per day available power only yields perhaps 0.8kWh/m² per day, not even enough for underground lettuce on a 1:1 basis. Of course, a space-based facility would see the full 1.3kW/m² intensity 24/7, or 31.2kWh/m² per day; this would allow for 6.2kWh/m² of collector area in the form of LED lights. Clearly light pipes or fiber optic concentrators are the better solution as long as harmful UV can be properly filtered.

  4. The reason we count the number of useful photons is because photosynthesis occurs as a series of discrete steps. Blue light has more energy for the same number of photons than red light does, yet both colors result in about the same amount of activity. The extra energy in the blue photon is wasted because it is used for one step and anything left over becomes heat. This is why LED grow lights are mostly red; you get more PAR for the same wattage. Some blue is included because some plant species will grow deformed or fail to flower without it.

    Plants use only a tiny fraction of the photons available to them, but using PAR and DLI as measurements allows one to calculate how efficient a plant variety is at using light. Researchers investigating this will typically grow a batch of plants under carefully controlled conditions to get accurate PAR and DLI data, then look at the dry mass and composition of the plants at harvest. In most cases nearly all the carbon in a plant is captured CO2. Much of the CO2 captured by a plant is later burned for energy, so what we see at harvest is the plant's energy reserve. This efficiency value is a little like the feed efficiency for an animal; it can be used to predict the yield of a well-understood crop based on the amount of energy provided. There is also an edible yield ratio for crops which is similar to the slaughter yield for animals; for example, wheat typically has an edible yield of about 50% of dry matter.

    This is exactly the kind of research I'd like to see done in parallel with directed plant breeding programs. Commercial programs usually target yield, simultaneous ripening and resistance to transportation damage as their endpoints. I'd like to target human-useful nutrient content, taste, photosynthetic efficiency and compact growth habits in addition to yield. Improvements to these would be useful to everyone, not just to a space colony program.