"Feeding Cities - with Indoor Vertical Farms?" New blog post by Mike Hamm, CRFS Director

Mike Hamm compares three potential solutions to make urban agriculture more sustainable: fully indoor vertical farming, greenhouses and hoophouses.

The following was originally posted on the website of the Food Climate Research Network at the University of Oxford on March 30, 2015.


Back in May, 1970 I was a senior at St. Louis University High School in St. Louis, MO, USA.  On May 1 that year the first U.S. Earth Day was held and Dr. Barry Commoner spoke at nearby Washington University.  My friend, Keith, and I skipped out of school and attended his talk.  Part of his talk laid out the four basic ‘rules’ for human’s relationship to the environment – including the adage ‘there’s no such thing as a free lunch’.  The more I read the wondrous advantages that vertical, indoor farming has the more I flash back to that first Earth Day.  A simple Google search (link is external) brings forth wondrous conceptual drawings by architects and planners illustrating these marvelous skyscrapers for healthy eating – a veritable cornucopia of modern design.  However, a funny thing happened on the way to the TedX talk …  I couldn’t find evidence (i.e. peer-reviewed literature or web-available reports) that anyone promoting this strategy has done the research or analysis to understand the relative environmental or economic costs of these strategies and their potential impact on greenhouse gases.  


May I start with some disclaimers and a focus for this blog?  There is a wide range of indoor and protected agriculture as it is practiced globally.  On one end of the continuum are Nepalese farmers (link is external) using modified hoophouses made with indigenous bamboo and a plastic covering on the top and upper sides to grow tomatoes protected from rain; they thus get an ‘out-of-season’ crop that increases income and provides nutritious food. Or maybe passive solar hoop houses in Mongolia (link is external) as an example of usage in extreme environments.  Towards the other end, but using the sun for plant growth, are systems such as the A-Go-Grow technology (link is external) in Singapore.  On the other end of the spectrum are fully enclosed systems dependent on 100% electricity-generated light for plant growth.  I am focusing on the latter with comparative reference to other types of production systems.  I intend no criticism of those putting their time and money behind start-up businesses producing food for market in this way – I am very appreciative of those who follow their passions, put their money behind it, and try to make a difference in the world.  However, I do intend to look critically at the issue of fully indoor production systems as a matter of environmental unsustainability, public policy, publically supported research, and those who argue it being a primary means of food production for urban areas.  I also intend to look critically at some of the unexplored claims made for the advantages of these production systems. 


I take strong issue with those who go to great lengths trying to convince everyone this is the solution to world hunger, fresh water challenges, and agriculture’s contribution to climate change.  I have perused a number of websites focused on vertical indoor farming and have found very little beyond platitudes.  In the many ‘review’ articles on the topic I have found nothing that could be classified as data examining the veracity of the claims for these systems.  That tends to raise high my ‘skeptic’ antennae.  Several months ago a colleague alerted me to a recorded seminar (link is external)by Dr. Louis Albright, an emeritus professor of biological and environmental engineering at Cornell University[1].  Dr. Albright looks at the issue of producing crops in enclosed spaces with numbers – kilowatts of energy, moles of light, growth per square meter, kilograms of CO2 released.    


[1] Some of the numbers below are updated from what you will see in the seminar; received via personal communication with Dr. Albright.


He talks about three crops – wheat, lettuce, and tomatoes.  These are representative of the plant range promulgated by the vertical farming stewards.  Wheat represents a commodity crop that would be important if these systems are to truly feed a city.   Many vertical farming advocates imagine fully feeding cities from these structures and so examining a crop like wheat is an important surrogate for the range of grains.  Lettuce is probably the most commonly grown food crop indoors today and is one where we eat all of the plant biomass (except the roots).  Tomatoes are illustrative of a fruiting crop in which a limited amount of the plant is eaten – there are a lot of stems and leaves that have to be produced but not eaten and therefore ‘wasted’ in the food sense.  Can these crops be grown indoors with artificial light efficiently, at a cost somewhat competitive with current outdoor production or common indoor production, and in a manner that is more environmentally sustainable than the alternatives (with CO2 release as the surrogate – admittedly an incomplete accounting)?  In much of this I will use New York City, USA as a surrogate city to examine scaling requirements.  


We would need to produce wheat in great abundance – in 2012 the United States produced about 60 billion kilograms of wheat.  For reference - New York City has a population of about 8.6 million (2013 estimate) who each consumed (link is external)about 24 kg of bread (206 million kg. total).  Dr. Albright uses an estimate of 2 kg/sq.m./harvest and four harvests a year or 8 kg/sq.m./year total (these are generous estimates since the world record for outdoor wheat production is about 1.6 kg/sq.m.).  The wholesale market value of this, based on organic crop prices (link is external)of US$11.50/bushel is about US$3.37/sq.m./year.  


What about the cost of the light energy to grow this wheat? Assuming use of efficient red-blue LEDs (link is external) and $0.10 per kilowatt-hour the electricity for this production would cost about US$327/sq.m./year or about 100 times the wholesale value.  If turned into bread it would cost $11 per loaf of bread – just to cover the lighting cost.  Consider the scale required to feed New York City.  At this production level it would take about 26 million sq.m. of production space to grow New York City’s wheat.  Assuming these are stacked – say you can get 3 stacked in a 10 foot high room – it would mean 8.6 million sq.m. of floor space. This is equivalent to about three Empire State Buildings.  On the one hand, this seems very efficient – produce all the bread in three large buildings.  But what of the initial and recurring environmental footprint?  I am not sure what the environmental footprint would be of constructing these three buildings for wheat production or the annual operating costs beyond the lighting – but my guess is fairly high.  Also, there is no way to develop an agro-ecosystem as we can with land-based production – detrimental impacts will be locked in and degrees of freedom for the future very limited.  In summary, grain crops are a terrible choice for indoor production. 


Maybe more valuable crops – for e.g., lettuce and tomatoes – will give us an inkling of the potential in these indoor systems.  Dr. Albright analyzed two systems – one with 100% indoor light and one with 30% indoor light/70% solar capture (representative of a greenhouse with supplemental light).  Reported research from a greenhouse near Cornell University indicated it would take about 300 kWh/sq.m./yr with 30% artificial light or 0.40 kWh per head of lettuce (in the research he draws from the lettuce heads are 150-160 grams each).  The full indoor light system would use about 1000 kWh/sq.m./yr costing US$0.10 more per head (US$0.63 per kg) for light electricity. This puts a pretty high baseline on the head lettuce market price in markets where the profit margin is often slim.  It undoubtedly becomes too pricey for many people pretty quickly – those without a good amount of disposable income are pretty much shut out. 


One of the claims for these fully-lit, indoor systems is that they will reduce the carbon footprint of the food – that they are more environmentally sustainable.  True or not true?  With 100% supplemental light (using the highest efficiency lighting – 6.12 moles/kWh - and the lowest regional value for CO2/kWh in US electrical production – 460 g/kWh) his analysis shows about 3.95 kg. CO2 produced per kg of head lettuce from the lighting electricity.   A bit of perspective as I find these raw numbers fairly abstract.  The average US passenger car emits about 5,100 kg of CO2 per year – so the production of 8,300 heads of lettuce in this system is equivalent to one year’s car travel.  The average US resident consumes (link is external)about 5 kg of leaf lettuce per year – so those 8,300 heads will feed (leaf lettuce only - about 260 people).  Providing the leaf lettuce for New York City would entail the CO2 release (remember just for the lighting) equivalent to 33,000 U.S. passenger cars.  There are ways to make these systems more efficient, for e.g. increasing CO2 levels in the growth chambers increases the efficiency of plant’s light capture.  However, these improvements don’t alter the notion that by eliminating the sun from the plant the environmental cost of production is markedly elevated.  


Tomatoes are a different story.  Unlike lettuce, where we consume virtually the whole plant, we only consume the fruit of a tomato plant – a lot of solar/LED energy must go into producing stems and leaves – wasted from a human food perspective.  Modeling production at 67 kg/sq.m./yr of fruit the CO2 cost is 8.7 kg/kg tomatoes – more than twice that of lettuce.  The average U.S. resident consumes (link is external)about 8.2 kg of fresh tomatoes annually (and a lot more as processed in various ways).  Running the simple math – for New York City’s residents the total CO2 released in producing all the fresh market tomatoes (assuming zero waste prior to consumption) would equal that of 264,000 U.S. passenger cars.  Oops!


Certainly proposals have been suggested to use either solar energy or biogas to generate the electricity.  However, according to Dr. Albright, with the efficiencies and energy losses in the system it would take about 4.5 acres of solar cells for every acre of plant growth space.  In theory this is achievable yet the question should be asked – at what opportunity cost? Other aspects of urban dwellers’ lives need electricity.  Is it nonsensical to use photovoltaics to drive a system that could be driven directly by the sun when these same solar cells could be used to power other aspects of people’s lives?  Doing this in land-based greenhouses on the urban periphery (or maybe on flat roofs (link is external)/abandoned lots (link is external) inside the urban core) with direct sunlight and possibly supplementing with artificial light would markedly cut the artificial light requirement, maintain a high efficiency of production, and be in close proximity to consumption.


Some recommend coupling these 100% artificially lit production systems to biogas generators through recycling of urban food waste.  The basic notion is reasonable.  Recycling of human organic waste – both food and excreta – is something we need to grapple with globally for a number of reasons including nutrient and carbon cycling.  The question arises – are these systems efficient and cost effective enough that the operators can pay for the food waste they take, as some have suggested (link is external). Experience may tell other stories.  There is a 400 KW (combined heat and electricity) biogas generator (link is external)at Michigan State University that uses approximately 14,000 metric tons of animal and food waste per year. A large grocery chain in Michigan supplies approximately 4,500 metric tons of fresh food waste annually and pays a tipping fee to get rid of it – not the other way around.  This greatly helps cash flow to pay down the construction debt – things like the tipping fee balance the budget on the bio digester; they don’t pay to take it. 


In other words, incorporating renewable energy production/use into our food production systems is important for enhancing environmental sustainability, resilience development and climate change mitigation/adaptation.  But there are cost and trade-off issues that will mandate a high cost (environmental and economic) for the food - those with the fewest resources will likely not be able to afford these products.  The alternative is the need to massively subsidize these products, for extended periods of time to make them viable and access to them equitable.  These tradeoffs also beg the question of whether this renewable energy could better be used for other purposes – for example, powering our smart phones – when there is already a renewable resource in the sky for food production.


Maybe the real question is where should time, money and resources be put in thinking about how to feed city regions in more sustainable and resilient ways.  I agree wholeheartedly with the notion that city region food systems in the developed world need to be strengthened and food should be sourced, on average, closer to home.  It begs the question of why it makes sense to put a lot of intellectual activity and resources into something that negates the direct use of our one absolutely renewable resource – the sun – and totally replace it with artificial light.    From the above it is clear that the production cost and the cost in greenhouse gas release is going to be fairly steep.  Aren’t there other strategies for moving production closer to home in city regions around the world while still maintaining a diverse diet across the year?


Several lines of research and simple reasoning say yes!  Moving beyond 100% artificial lighting to consider other possibilities is illustrative.  Technologically, these range from field level tunnels that keep crops from frost kill in the early spring and late fall; unheated high tunnels that use only solar capture to produce crops year round in colder climates; and greenhouses with heat and supplemental light for full twelve month production of a wider range of crop possibilities (such as the 70/30 system related above). 


In a 2008 study (link is external)on crop product flow into New York State, the authors reported all imports of lettuce produced, on average, 0.70 kg CO2/kg lettuce over an average transport distance of about 2900 miles. Yet, according to Dr. Albright’s calculations, 100% artificially-lit systems would produce about 3.95 kg CO2/kg lettuce for the lighting – over five times greater.  


Production System

kg CO2/kg lettuce

A - Import average of 2900 miles

0.70 (for transportation only)

B - 100% Artificial lighting (High Pressure Sodium)

3.95 (for lighting only)

C - 70% Sun/ 30% Artificial Light with CO2 addition

0.71 (for lighting only)

D - Same as ‘C’ but on Long Island, NY

0.35 (for lighting only)


In Ithaca, NY the 70% solar/30% artificial system with added CO2 would produce about 0.71 kg CO2/kg lettuce and the same system on Long Island - with less cloud cover and therefore less artificial light required - is about 0.35 kg CO2/kg lettuce.  In other words, we can get in the ballpark of shipping lettuce 2900 miles with greenhouse production but not with 100% artificially lit indoor systems. 


We reported in 2014 on the lettuce production carbon footprint (link is external)in unheated hoop houses (January production in Michigan, USA –photos outside and inside Michigan State University Student Organic Farm (link is external) in winter) compared to getting it from about 2,200 miles away.  The unheated hoop house system generates about one-fifth the CO2 relative to an outdoor system - 0.198 vs. 0.857 kg CO2/kg lettuce.




See ‘Table 1 from paper for the breakdown in each system. The carbon cost for hoop house lighting is zero.  These unheated systems come with a caveat – since there is no heat beyond solar trapping and no supplemental light the productivity over a year is not as high as heated and partially lit systems.  Fruiting crops can’t be grown year-round in these systems; greens of various types and many root crops are feasible.  The predictability of harvest is more variable since weather changes do somewhat affect the inside environment.  Yet, for a number of crops there are advantages relative to carbon footprint and production cost. 


At the end of this analysis I am at a loss to understand why indoor vertical farms with 100% artificial light have seemed like such a good idea when there are so many other pressing issues to research if we are going to have a sustainable, resilient and secure food supply globally and within our respective nations.  I am unable to find evidence that 100% artificially lit systems are a priori a solution for feeding cities and their environs.  The environmental cost (judged only by carbon footprint in this case) is clearly higher than other strategies. It seems a shame that public policy or public research dollars would be diverted to 100% artificially lit systems.  A better use of time, resources, and capital with greater potential for long-term impact is continued development of controlled to semi-controlled environments with the sun as the basis of production.  Whether it is greenhouses with supplemental light and heat; hoop houses with neither supplementary heat nor light; as well as the range of variations between – these offer the more environmentally sustainable options.  They can be sited in urban areas on abandoned land (link is external) or rooftops (link is external)and in the surrounding peri-urban and near-rural areas.  In higher latitudes, these types of growing environments allow land-use over the full year instead of simply the normal, outdoor growing season (depending on your climate) – a form of sustainable intensification for city regions.  


There are still a range of complex issues – both with technology and human capital - to improve these greenhouse and hoop house systems.  These include small-scale equipment to improve efficiency of in-ground production, technology to help producers easily analyze their costs of production for various crops, continued improvement to the technological infrastructure across the range of sun-based indoor strategies, small-scale energy co-generation strategies for supplemental heat and light.   These greenhouse and hoophouse systems and strategies have a real potential to improve the sustainability and resilience of our food system.  They have a real potential to help supply a healthy diet in city regions across the globe.  They have a real potential to help insure sustainable livelihoods for a diversity of farmers in city regions.  This is clearly not the case for 100% artificially lit systems.  No, there is no such thing as a free lunch - no matter how much experts contend otherwise.


You may contact Mike at mhamm@msu.edu.

Did you find this article useful?