With support from the STFC Food Network+ (SFN), this project is modelling a future of local, ‘closed-loop’ urban food production systems where nothing goes to waste.
​

Our food systems are under immense pressure to produce more whilst simultaneously reducing pesticide and fertiliser use, carbon emissions, water use, and food miles. At the same time, the way we purchase food is undergoing dramatic change, with goods being increasingly sourced online, rather than from high streets. The sight of empty shop units may appear depressing, but for Peter Ball they represent an opportunity: a chance to address fundamental issues of food production whilst building thriving local communities.
“I am a firm believer that many of these grand challenges we face with our current production systems can be addressed by returning manufacturing to a more local or regional level – whether it is food, drink, construction materials or other volume-low technology products” says Peter, a Professor of Operations Management at the University of York Management School.
“Besides reducing the climate and environmental impact of food production, this can support local economies, give greater understanding about where our products come from, put unused space to work, and generate employment. It also presents the opportunity to build local circular economies, with ‘waste’ resources being recaptured and fed back into the system as input materials for new products.”
For Peter, a particularly promising area is vertical farming, where plants are grown indoors in controlled environments. Besides reducing food miles by bringing agriculture into urban spaces, vertical farms are also more water- and space-efficient, and can remove the need for pesticides and herbicides. Additional benefits include the fact they can be set up practically anywhere (even underground), and that they allow year-round food production, sheltered from the elements.
“Here in York, for instance, the University teamed up with LettUs Grow, an indoor farming technology company, to set up a vertical farm demonstrator project right in the heart of the city ” says Peter. “Not only can crops be used locally at the peak of freshness and nutrients but we can engage consumers in where their food comes from.”
But isolated units don’t feed cities; for urban farms to have a significant impact on local consumption, these need to function as a network with their outputs matching consumer demands. But Peter believes this should go even further. His vision is for vertical farms to be integrated with local waste flows, so that the remaining value in food and food production waste is immediately put to use in nourishing the next crop. Up to now, however, there has been very little data to indicate whether this could be possible – a problem which inspired Peter to apply for a SFN Scoping Grant.

A model solution
“A key issue with using food waste flows as inputs for growing systems is that these are highly variable and dynamic, so ensuring a consistent supply of the right nutrients can be challenging” Peter explains. “Our aim was to provide a digital demonstrator to model waste flows in urban areas and explore how these could be optimised to create food production systems using circular economy principles.”
This drew upon Peter’s expertise in using simulation modelling (digital twins) to capture system behaviour and improve performance. To develop the model, Peter and his colleagues sourced live data on production from the vertical farm demonstrator in the centre of York. In addition, they used public data sources to generate information on potential locations and volumes of food waste outputs.
In the model, the waste flows were processed by IntelliDigest digestors, which use enzymes from bacteria and plants to turn inedible food waste into a broth-like nutrient liquid. The team analysed the simulated data to see how closely the output matched the nutrient requirements of growing crops.

From brewery waste to building materials
Beyond fertiliser, the team were also keen to explore whether additional waste streams could be harnessed to provide construction materials for vertical farms, or even an energy source to keep them running. “One of the barriers to scaling up vertical farms is the high initial capital costs, and the considerable energy demands” admits Peter. The team partnered with a local brewery, Brew York , to obtain samples of spent grain waste from beer production. Project partner WASWARE, a company that specialises in turning waste materials into novel biocomposite materials, then processed these into bioresin, and in turn bioleather and bioboard.
“These new materials showed strong potential for use in vertical farm construction” says Peter. “This has led to ideas not just around changing existing materials but changing how we design structures in the first place.”

The next steps
According to Peter, although the project’s findings are promising for vertical farm expansion, much work remains still to be done. “The next stage is to develop more sophisticated models with access to live data streams to populate them with real information. We also need to build more case studies that include other urban areas, including those in developing countries.”
“This collaboration funded by the SFN Scoping Grant enabled a diverse group of organisations to apply expertise into a common challenge and test out ideas” he added. “We are now pursuing a larger collaborative project on nutrient cycling, and have just started a project to develop new building materials and challenge existing farm design norms.”

The STFC code used for the model simulation can be found on GitHub.
The project was a collaborative team effort involving both academia and business. The team members were: Ehsan Badakhshan (University of York), Peter Ball (University of York), Joseph Bell (Wasware Ltd), Nicola Holden (SRUC), Jens Jensen (STFC), Ifeyinwa Kanu (IntelliDigest Ltd), Lydia Smith (NIAB), and Xiaobin Zhao (Wasware Ltd).
​
Find out more:
Using machine learning to unlock the secrets of plant productivity - STFC Food Network+
IntelliDigest: an innovative solution to put food waste to work - STFC Food Network+
 

With funding from an STFC Food Network+ (SFN) Scoping Grant, an exciting project has shown the potential for computational genetics to detect the most dangerous strains of a notorious food-borne bacteria, Shigatoxigenic E. coli

No matter how inviting a plate of food may look, it could be harbouring an invisible threat. According to the Food Standards Agency, 2.4 million people in the UK become ill from food-borne pathogens every year, at a cost of to society of around £9.1 billion. One of the most notorious of these is Shiga-toxin producing (Shigatoxigenic) Escherichia coli (STEC), which can cause symptoms ranging from gastroenteritis and kidney failure, to meningitis and even death. A public health priority risk, STEC infections are most commonly associated with undercooked minced meat products, underwashed salad vegetables, unpasteurised dairy products, and handling foods with soil residues. The World Health Organization estimate that in 2010 food-borne STEC caused more than one million illnesses, 128 deaths, and nearly 13,000 Disability Adjusted Life Years.

Controlling these outbreaks depends on being able to rapidly identify contaminated food products. But this is complicated by the fact that not all STEC strains are pathogenic, as Nicola Holden, Professor in Food Safety at SRUC, explains. “For an individual STEC strain to cause disease, it needs the right combination of several different factors. These include the anitgens on its surface, the toxins it produces, and the virulence factors that enable it to infect a cell. The exact combination of these will determine the ability to cause severe disease. It is similar to the different coronavirus variants, where small differences in, for instance, the spike protein, affect transmissibility and disease severity.”

To date, most cases of food poisoning from STEC have been caused by the O157 strain, so-called because it carries a distinct ‘O antigen’ that can be recognised in serology tests. But recent years have seen a worrying trend: a rise in STEC cases caused by non-O157 isolates.

“We currently don’t have any rapid means of identifying the STEC strains behind these new disease cases” says Nicola. “Diagnostic laboratories have to carry out additional tests to assess what combination of antigens, toxins and virulence factors a particular strain has, so they can work out the overall likelihood that it can cause severe disease. This takes time and introduces uncertainty.”

Unless we can detect these dangerous pathogens quickly, outbreaks can rapidly spiral out of control.  Consequently, Nicola believes it is time to overhaul these “historic and cumbersome” diagnostic methods and instead adopt a new approach that uses the power of genetics. “We saw during the COVID-19 pandemic how the introduction of lateral flow devices completely changed the game when it came to controlling the spread of the disease” she says. “Ultimately, that is what we need for STEC: a DNA-based method to enable rapid diagnostics using miniaturised point-of-care devices. This would help identify contaminated products and likely transmission routes quickly enough to control potentially dangerous STEC outbreaks.”

In December 2021, Nicola was awarded a STFC Scoping Grant to explore how feasible this would be by searching for genetic signatures that could distinguish between pathogenic and non-pathogenic STEC strains.

Step one: Assembling a genomic library
The first stage of the project was to compile as many genomes as possible from a diverse range of STEC isolates. Together with her co-investigators Dr Martynn Winn, a Computational Biologist at STFC Harwell, and Dr Tim Dallman, an expert in food-borne pathogens at the University of Utrecht, Nicola convened an online stakeholder workshop in December 2021. This brought together a wide range of research- and policy-related organisations who work on STEC, including the Scottish E. coli reference lab, the Food Standards Agency and Food Standards Scotland, and the Animal and Plant Health Agency.
​
“Working with these partners, we were able to access a good range of different STEC genomes held in reference databases, over 200 in total” says Nicola. “Crucially, these included 104 samples from human patients that we knew had been responsible for causing clinical disease.” The remaining samples had been collected during surveys on Scottish deer, cheese, and mince.

Step two: Comparing genes and genomes
Having acquired a diverse library of genomes, the team then used comparative genetic approaches to categorise genes as being ‘disease related’ and ‘non disease’ related.

First, they took an ‘informed’ approach, by searching for specific genes known to code for virulence factors in clinically pathological isolates. “Because these genes are more likely to be associated with pathological isolates, their presence indicates a likely clinical disease outcome” says Nicola. Mapping these virulence factor genes across the wider set of STEC genomes identified their presence in certain food and wildlife isolates, indicating that these may also be pathogenic.  

In the second stage, the team used an ‘unguided’ approach, that assumed no prior knowledge about the genes and their functions. Instead, the genomes as a whole were compared against each other to assess which regions were shared across the different samples. “This approach enables us to quickly assess which genetic regions are commonly seen across different pathological samples, and could therefore be associated with clinical disease” says Nicola.

Using these methods, the team successfully identified common genetic signatures that could distinguish different classes of STEC.

Step three: Explore the potential of Big Data
Besides these relatively straightforward comparative techniques, the team were also keen to investigate the potential of Big Data approaches, including machine learning- and artificial intelligence-based methods. With the SFN funding, they provided a three-month internship to PhD student Eddie Martin (University of Edinburgh) to explore whether these could help distinguish the most harmful pathogens from closely related ones that don’t cause disease.

“AI tools that use deep-learning offer exciting potential to investigate beyond the identification of genomic sequences” says Nicola. “For instance, they could help to determine if a similarity in sequence between two genes ultimately translates into functional similarity.”

Going forwardIn August 2022, Nicola and her colleagues convened a second stakeholder workshop at STFC Harwell to share their results so far, and to discuss the next stages for the project. The team are now seeking funding for a larger project to exploit the predictive power of Big Data to accurately classify pathogenic STEC.

“Further project development will fall into two main areas” says Nicola. “First, the basic bioscience – that is, refining the computational approaches so that these can sufficiently discriminate pathogenic bacteria. And secondly, applied bioscience to assess how to incorporate these methods into point-of-care devices for rapid diagnostics and surveillance.”

“Another challenge we hope to address is data accessibility” she adds. “There are currently many barriers to obtaining high-quality genomic data associated with clinical disease, and this is true for any human pathogen. This makes it crucial that this work is developed in partnership with our stakeholders. The SFN+ project provided an excellent opportunity to bring together a diverse team who would not have had the chance otherwise.”

According to Nicola, once DNA sequence-based diagnostic approaches have been refined, they would be an excellent route forward to discriminate between any set of pathogens and other organisms. We may never be able to eliminate food-borne bugs completely, but the near future shows promise for them no longer being such a scourge within our food systems.
​
You can keep up to date with Nicola’s work by following her on Twitter: @NicolaJHolden

Glossary:
Antigen: A toxin or other foreign substance which provokes an immune response in the body, particularly the production of antibodies.
​
Serology test: A laboratory test that assesses the presence of antibodies and other substances in a blood sample.

With support from the STFC Food Network+ (SFN), an interdisciplinary collaboration is breaking new ground in developing low-cost, accurate and autonomous technologies to measure soil properties. 
​​

“Climate change and soil degradation are twin threats to food security but we can address both at once by improving soil health” says Dr Marcelo Galdos, a soil carbon specialist at Rothamsted Research. “Improving soil carbon levels, for example, boosts productivity and yields by improving nutrient availability and making the system more resilient to climate extremes. At the same time, this can also contribute to bringing down carbon dioxide levels in the atmosphere.”

Our soils are in a perilous state, with a third of global agricultural soils thought to be moderately to highly degraded (FAO). But to restore our soils on a global scale, it is essential that we can quickly and accurately map properties related to soil health in order to understand which regenerative agricultural methods are most effective. A key challenge, however, is that soils vary immensely in terms of their physical, chemical, and biological properties and there is currently no single sensor capable of effectively monitoring all the relevant parameters. In addition, most soil measurement methods are highly time- and labour-intensive, requiring sampling from many different locations in a field.

Faced with this situation, Marcelo (then at the University of Leeds) applied for a STFC Food Network+ Scoping Grant to explore potential soil sensing technologies that could be both affordable and automated. This became the Multi-sensor Agricultural Robot for Soils (MARS) project.

Bringing down costs
The first objective of MARS was to investigate whether the cost and accuracy of soil measurements could be improved by using novel sensors. To do this, Marcelo joined forces with colleagues from the Schools of Computer Science, Electronic and Electrical Engineering, and Food Science at the University of Leeds, and with Dr Patrick Stowell from the School of Physics at the University of Durham. “Patrick provided us with a prototype of an affordable gamma ray sensor that he had designed” says Marcelo. “Professor Megan Povey from the University of Leeds meanwhile is an expert in using ultrasound in food-science applications, such as quality control. This project gave her an opportunity to apply this technology in a completely new area.”
The sensors were tested in tandem at the University of Leeds Research Farm site over the 2022 summer, measuring levels of potassium-40 using the gamma ray probe (as a proxy for soil moisture composition) and using the ultrasound probes to assess soil physical properties. “The gamma ray data was able to produce a working map of soil moisture concentration, and we even identified areas in the field with increased water content due to some blocked drainage pits” Marcelo says. “We found, however, that use of the ultrasound sensor would require more research under controlled conditions to calibrate the system.”

Going automatic
The second component of the project was to assess whether soil monitoring could ultimately become completely automated. “Robotics within UK agriculture is a very exciting sphere at the moment, and many start-ups have already developed promising protypes of automated systems” Marcelo says.

But although putting a sensor on a robot sounds simple, this is far from the case, as Marcelo explains: “We had to overcome numerous smaller challenges even before we could start trying to take any measurements. For instance, where should the sensor be installed? How would it be powered? How do we program the robot to move automatically? And how can we capture and record the data?”

A particularly difficult problem was developing a mechanism to raise and lower the sensor probe to touch the soil at exactly the right pressure. With support from students in mechanical engineering at the University of Leeds, the team designed and 3D-printed a custom robotic arm that could be mounted on top of the rover.  

Having equipped the robot with an ultrasound sensor in this way, Dr Syed Zaidi (School of Electronic and Electrical Engineering, University of Leeds) worked with postgraduate students to develop a web interface and self-driving algorithm so that the rover could be automatically programmed to move to a target coordinate and take measurements. The rover was tested in a field on the Leeds Farm, and successfully drove to specific points in the field were measurements would be collected. “Our next step will be to program the rover to take measurements based on covariates including topography, land cover and soil properties” says Marcelo.

Future plans
Having moved to Rothamsted in July 2023, Marcelo is excited by the prospect of developing this technology further at this centre for excellence in soil science. 

“Founded in 1843, Rothamsted has some of the longest-running continuous crop experiments in the world” says Marcelo. “This will give us the opportunity to test these technologies on fields for which we have soil data going back over many decades.” Besides developing the prototype rover further, he and his collaborators intend to investigate a range of other potential sensor/robot combinations, for instance integrating drones and satellite data into the analysis.

“Funding for feasibility studies such as the SFN Scoping Grants are immensely important for getting innovative ideas off the ground” Marcelo adds. “They allow you to develop a proof of concept to the point that you can attract further investment, besides giving you the freedom to seek out collaborators and build a team. This project ultimately led to me working with experts from a wide range of disciplines, including soil physicists, computer scientists, mechanical engineers, robotics researchers, and specialists in Internet of Things technologies. But this seems highly appropriate, because critical problems such as restoring our global soils will require innovation from many different areas.” 

Glossary:
Gamma spectroscopy: When used in soil sensing applications, gamma spectroscopy measures high-energy photons (gamma rays) that are continuously produced in soils due to the presence of radioactive elements, such as potassium-40. Water in the soil inhibits the flow of gamma rays from the soil to the surface, hence there is an inverse relationship between soil moisture content and the gamma signal recorded above ground.

Reducing food waste is an urgent challenge, both to achieve food security for the growing global population, and to help us reach net zero carbon emissions. Often, the focus is on consumers but in developing countries around 90% of food wastage occurs in the supply chain, before products even reach the market. A key reason for this is a lack of refrigerated storage and transport.

India, for example, has the sixth largest food and grocery market in the world, yet up to 40% of harvested crops are lost before they can be sold to consumers . With so much of the population dependent on agriculture, these loses keep many rural communities trapped in poverty.

‘Addressing post-harvest food losses in developing countries offers enormous potential to increase food security in these regions’ says Dr Bryan Shaughnessy, head of the Thermal Engineering Group at STFC RAL Space. ‘With such large volumes being lost daily, solutions that make even a small difference will translate into many tonnes of food being saved.’

Although his day job is to design thermal control systems for scientific spacecraft, Bryan became interested in the issue of post-harvest food loss when he attended an SFN meeting in 2017. ‘Talking to other attendees it became clear there was an enormous opportunity to reduce food waste in developing countries by providing cost-effective methods to keep food cool’ says Bryan. ‘It was also apparent that the technology and expertise I apply in developing instruments for use in space missions might bring a new and useful perspective to the problem.’ This led to Bryan being involved in an SFN pilot study  in 2018 to consider options to reduce food loss in India.

The transforming power of space technologies

In this latest work, funded by an SFN Scoping Project Grant, RAL Space partnered with Go4fresh, an Indian fresh produce agritech venture, to explore whether space thermal technologies could potentially help design cost-effective and efficient portable cold storage units for foods in transit. Although space rockets and fresh produce may seem worlds apart, they both face the same critical challenge, as Bryan explains:
‘One of the key factors that accelerates food spoilage is large variations in temperature, particularly in developing countries which may have long, highly fragmented supply chains combined with tropical climates. Similarly, spacecraft experience an extremely hostile thermal environment, which ranges from the extreme cold of deep space (around -270°C), to intense heating from the Sun.’

Without careful thermal control, spacecraft simply wouldn’t be able to survive these extreme temperature fluctuations. Consequently, engineers such as Bryan investigate how to exploit the physical properties of materials to keep these instruments within a safe temperature range. These methods include using highly-reflective surface finishes and insulation materials with low thermal conductivity. Because these technologies are typically lightweight and passive – not requiring a power source – they are particularly suitable for areas without a robust energy infrastructure.
For this study, the team focused on tomato production in Nashik, a district in Maharashtra state. With tomatoes having one of the highest percentage losses (around 12%) across vegetables in India, even a small reduction in waste would result in significant savings.

Capturing the temperature profile of a tomato in transit

The first step was to develop a simple thermal model to better understand the heat transfers that tomatoes experience during their journey from the farm to market, via various distribution centres and wholesalers. This was crucial to understand the most important factors that affect how quickly tomatoes spoil, and therefore the requirements that a portable cold chain container would have to meet.

‘Developing this model was actually quite a challenge, because we were adapting software that was designed to predict the temperatures of satellites in the vacuum of space’ says Bryan. ‘Some additional research was needed to better understand the weather conditions in Nashik and find a way to represent this in our model.’

A simulation was then run to represent a ten-day period - a typical transit time for tomatoes destined for non-local markets. The target at this stage was to keep the produce below 25°C throughout the journey.

Simple changes, significant effects

For the baseline scenario, where the tomatoes were transported in open crates, the model indicated that the tomatoes experienced wide temperature fluctuations, reaching a maximum temperature of around 50°C. Even before technical solutions were considered, the analysis demonstrated that simple modifications could make a dramatic difference.
Painting the outer surfaces in white paint, for instance, reduced the temperature fluctuations and the maximum temperature to more like 35°C. ‘This might sound like a simple solution, but white paint is often used as a surface treatment for spacecraft thermal control because of its low solar absorptivity and high infrared emissivity. This means it radiates away more heat than it absorbs from the Sun,’ says Bryan.

The team next investigated the impact of coating the exterior of the box with a finish called a Second Surface Mirror. Like white paint, these thin, transparent materials reflect a high proportion of sunlight but can radiate heat in the infrared, and are often used on spacecraft as part of a passive thermal control system. This performed a little better than the white paint, causing a slight further reduction in temperature. ‘Although Second Surface Mirrors are routinely used on spacecraft, this is interesting because it might be possible to make a similar ‘solar reflective’ sheet using locally-available materials’ Bryan says.

Adding a coolant material, such as an ice pack, also had a significant effect, and kept the tomatoes within safe temperature limits for four days, doubling the time they could be transported. However, the benefit only lasted until the ice pack thawed, and relied on power being available to freeze the ice pack in the first place. The packs also presented the risk of causing chill damage to the tomatoes if they came in direct contact.

Another idea borrowed from space technologies was to add a layer of insulation to effectively decouple the interior of the box from the external environment. Virtually all spacecraft are insulated using blankets made of multiple layers of reflective sheets interleaved with netting. But since these materials aren’t readily available in India, the team instead modelled a scenario where the tomatoes were packed with straw. Incredibly, this achieved a 10°C temperature difference between the inside and outside of the crate, and reduced the maximum temperature experienced by the tomatoes to under 25°C.

The greatest benefits, however, were achieved when these approaches were applied in combination: the outer layer of white paint or a Second Surface Mirror, an inner insulation layer, and an ice pack. In this scenario, the tomatoes experienced no temperature fluctuations and stayed underneath 22°C throughout the ten-day period.

​COVID-19 shook global food supply chains to the core – but did this result in any long-term changes in societal attitudes towards food? A project funded by the STFC Food Network+ (SFN) combined the power of machine learning approaches with large-scale Twitter data to find out.
 
The COVID-19 pandemic impacted every aspect of our lives, with food being no exception. For instance, as countries went into lockdown and panic buying swept supermarket shelves bare, many people experienced food insecurity for the first time. And whilst communal meals and eating out became impossible, online shopping and home meal preparation kits soared in popularity. Meanwhile, the sudden increase in home cooking led to greater interest in the health and nutritional attributes of what we eat.
 
But now that most countries are returning to ‘near normality’, it remains unclear what the long-term repercussions of the pandemic will be. Did these sudden societal changes result in any long-term fundamental shifts in our purchasing behaviours and attitudes towards food? Understanding this could help policy makers to identify the most important food supply concerns of consumers during a crisis situation, which could inform strategies to respond to future emergencies.
 
Twitter: an ideal tool for research
 
To investigate whether consumer attitudes towards food were affected by the pandemic, a project backed by the SFN used machine learning methods to mine food-related information contained within social media posts on Twitter. With more than 500 million users worldwide, Twitter has become a global tool for sharing news, expressing opinions, and interacting with others in real time. This makes it highly suitable for capturing large amounts of organic data to explore sentiment and attitudes surrounding food concerns both before and during the COVID-19 pandemic.
 
“Public attitudes on Twitter have already been used to evaluate consumer perceptions towards brands, predict stock market fluctuations, and assess the success of political campaigns” said project lead, Mohammad Delgosha, Associate Professor in Business Analytics at the University of Birmingham. “But up to now, little attention has been given to the feelings and perceptions of consumers regarding the unique and sheer scale of the impact of the COVID-19 pandemic on food supply chains.”
 
Mohammad’s research focuses on data mining, specifically analysing Big Data by employing machine learning techniques to large amounts of unstructured data to find solutions for social, environmental, cultural and practical challenges. In the past, for instance, Mohammad has used textual Big Data to investigate the positives and negatives of automated algorithms being used to manage gig workers in digital platforms such as Uber or Deliveroo.
 
According to him, Twitter data is highly useful for exploring consumer opinions in an unbiased way. “When we use surveys to try and assess consumer sentiment, these are inevitably constrained to the researcher’s existing knowledge or speculation. Twitter, on the other hand, offers significant amounts of open-ended data to explore open-ended questions” he said.
 
Capturing meaning from Big Data
 
The first stage of the project involved capturing tweets generated during two 16-week periods: from 7 August to 8 December 2019, and then the 16 weeks immediately after the World Health Organization declared COVID-19 as a global pandemic on 11 March 2020. By applying search algorithms to Twitter’s Application Programming Interface, posts were extracted that contained keywords related to food (for instance, food, crop, fruits, vegetables, meat, milk, and groceries) or supply-related terms (such as supply, chain, logistics, systems). This resulted in a dataset of around 182,000 tweets from 109,600 different people before COVID-19, and 427,000 tweets from 183,000 people after the lockdowns started.
 
The sheer size of this dataset meant it would have been impossible to manually analyse it using human coders. So Mohammad used two natural language processing techniques, called topic modelling and sentiment analysis.
 
“Topic modelling is an unsupervised statistical method for discovering abstract ‘topics’ within textual data. Its principal function is to combine main concepts within a text into a single, understandable structure” said Mohammad. This method scans the collection of text to detect frequently used words or phrases, and groups them to provide a summary that best represents the information in the document.
 
Sentiment analysis, on the other hand, is a tool used to understand the emotion, opinion, or judgment behind written or spoken language, and to evaluate if this communicates a favourable, unfavourable, or neutral message. Businesses, for example, often apply sentiment analysis to customer comments or online reviews to assess how their customers feel about their product, services, or brand.
 
The impact of COVID-19 on consumer concerns about food
 
“Our results indicate that the public sentiment and topics associated with food systems were markedly transformed by the COVID-19 pandemic” said Mohammad. The sentiment analysis, for instance, showed that overall public attitudes towards food had a more negative tone and became more pessimistic during the initial COVID-19 wave, compared with before the pandemic. For instance, before the pandemic, around 47% of tweets had an average positive sentiment, with 36% being negative, and 17% neutral. During the first phase of the pandemic, however, the proportion of negative tweets jumped to 54%, with 24% being positive and 22% being neutral.

A project supported by the SFN aims to empower smallholder farmers in India to become citizen scientists and champions for healthy soils, using a simple tool developed for mobile phones.

​‘Across the world, it is the millions of smallholder farmers who are the ones working the soil. Unless we involve them and develop tools that they can easily understand and use, we won’t be able to help soils recover.’ Rajneesh Dwevedi.
​
​When it comes to food security, the focus tends to be on what is happening above ground rather than below our feet. But this needs to change – and fast. Worldwide, a third of agricultural soils are thought to be moderately to highly degraded (FAO), limiting their productivity. Besides putting millions of livelihoods at risk, global soil degradation also has significant impacts on climate change by releasing vast quantities of carbon to the atmosphere.

A key characteristic of healthy soils is that they are highly biodiverse: teeming with life that ranges from microscopic organisms, invertebrates such as nematodes, insect larvae and earthworms, and mammals, reptiles, and amphibians. All of these play a fundamental role in maintaining the benefits of soils that we depend on. For instance, as part of their metabolism many microorganisms transform essential organic and inorganic compounds into forms that plants can use. Larger species, meanwhile, such as earthworms, ants and termites, engineer soil structure through their movements and open up pores for water and gas to flow.

But modern intensive farming methods can have devastating impacts on soil biodiversity. Widespread use of pesticides and fertilizers, compaction from heavy farm machinery, and disruption to soil structure from ploughing devastate the complex webs of life below ground. This results in less efficient nutrient cycling, poorer soil structure and ultimately smaller harvests. It is a vicious cycle: as soil biodiversity decreases, agricultural productivity decreases – causing farmers to resort to even more intensive farming to try and maintain yields.

In contrast, lower impact and regenerative farming methods – such as organic farming and ‘no-till’ farming – can maintain and even restore soil biodiversity. Wider adoption of these could help us start to reverse the perilous conditions of our soils, but a major barrier to this is a lack of ready tools and technologies to easily measure soil health. ‘Maintaining soil ecosystem services is a key challenge for sustainable food production, and one that depends on soil biodiversity’ says Mr Rajneesh Dwevedi (Lady Irwin College, Delhi). 'Achieving this will ultimately depend on the knowledge and actions of farmers. Though farmers often understand the importance of soil health, they are not able to monitor and infer the ecosystem condition accurately.’

As part of a SFN-backed collaboration with the STFC, Rajneesh is addressing this by developing an easy to use, intuitive tool for assessing soil health, designed to be suitable for farmers worldwide regardless of their level of education. To start with, he is focusing on India, where a large proportion of soils (particularly in the Gangetic plains) are severely degraded. Since the presence or absence of certain organisms can be a direct indicator of soil health, Rajneesh’s specific aim is to develop an accessible tool that can quickly identify soil species. Currently, existing guides for assessing soil biodiversity are often highly technical keys and charts, requiring expert knowledge to decipher. But Rajneesh’s vision is for a mobile phone application that uses the power of artificial intelligence to accurately identify soil species from photographs taken by the user.

Having no previous experience of machine-learning approaches before, Rajneesh is working with Dr Melina Zempila from STFC RAL Space to refine the identification algorithms specific for soil organisms that will form the basis of the tool. In particular, Melina’s expert knowledge is helping to refine the method of classification using advanced image analysis based on information from the visible spectrum of light.

But to train a program, you first need a labelled dataset it can learn from. Consequently, the first stage of the project saw Rajneesh travelling across North India between March and August 2022, collecting more than sixty soil samples from as many different farms as possible. This included both farms using intensive, pesticide and fertilizer-heavy practices and those based on organic methods. ‘It was a fascinating new experience to closely study so many different soils across India and to see how the biodiversity varied’ says Rajneesh.

Back in the laboratory at Lady Irwin College, each sample was carefully analyzed and high- resolution images taken of the species to curate a database of labelled images. ‘Not surprisingly, our preliminarily observations found that soil biodiversity can vary significantly with soil type, with organic farms being richer than the conventional farms’ Rajneesh says.

In July, Rajneesh visited the STFC RAL Space facility, based at Harwell, Oxfordshire, to meet Melina and discuss the next stage of the project: developing the identification algorithms. ‘It was a great experience to see the computing capabilities at STFC, and discuss the next stages of the project together. The initial results so far have been promising, with our prototype model being effective in identifying soil fauna present on the soil surface. Below ground species, however, remain difficult as they look similar in the visible spectrum.’

Another challenge will be to tweak the algorithms so they still work effectively on simpler, mobile-phone images rather than the high resolution photographs. ‘As a first goal, we hope this tool will enable farmers to become citizen scientists, capable of mapping soil biodiversity and collecting information simply by taking a photograph. A longer-term aim is that this information can then be used policy makers to identify priority areas for restoration. It could also help farmers to select the best crops for their fields, based on the soil health’ says Rajneesh. If successful in India, the method could be adapted to the soils of countries worldwide.

Reflecting on his involvement so far, Rajneesh says: ‘I’ve really enjoyed working on this, particularly as it has motivated me to learn new things. My background is in biology, so artificial intelligence was completely foreign to me. Even so, I’ve found it a magical experience to get deep into the mathematics behind it. The SFN serves a great purpose in bringing people with different expertise together, to apply science and technology into making new solutions that ultimately help people.’

‘People often assume that grater agricultural productivity always comes at the expense of nature, but I hope we can help show that these do not have to be mutually exclusive. There can be a middle way’ he concludes.

The past few years have demonstrated just how vulnerable our food systems are to large-scale disruptions, from pandemic lockdowns and social distancing to labour issues and extreme weather events. For the consumer, these can result in product shortages and empty supermarket shelves. Such uncertainty looks set to continue, as food producers start to feel the effects of climate change, and global political systems remain fragile. Consequently, a key part of the STFC Food Network (SFN)’s work is to research ways we can build resilience into food supply chains, so that they can keep functioning even when situations (and people’s shopping behaviour) change quickly.
 
Potentially, this could include harnessing the power of information contained in social media posts. One of the winning projects in the 2020 SFN Sandpit competition is investigating whether it could be possible to apply deep learning techniques to build a computer model that can forecast specific food-system disruptions (such as flour running out of stock) on the basis of the text and images that individuals post online (e.g. comments about flour).

Project lead Dr Laura Wilkinson (Senior Lecturer in Psychology at Swansea University) explains the rationale: “The COVID-19 pandemic saw an unprecedented level of conversation around food on social media, with many people experiencing shortages of key staples such as flour for the first time. We suggest that understanding how people reacted during the pandemic may help us to understand what might happen to the food system if other events occur that lead to uncertain times.”

The project’s ultimate aim is to train a machine learning programme to automatically detect patterns in people’s social media activity that indicate changes to food systems, and how consumers react to these. Training such a model, however, requires a detailed, annotated dataset, so that it can ‘learn’ which information is relevant to spot trends. To compile this, Laura and her colleagues created a database of just over a million anonymised Twitter posts, using searches with food-related keywords.
 
The trouble is that some of these posts may not actually be about food at all, as Laura explains: “Many food words can also refer to other things. For example, the word ‘coconut’ could have been used to describe someone wanting to be able to eat a coconut or could have been used to describe their favourite shampoo scent. On the other hand, a tweet might be referring to food but not actually mention a food word like 'banana'. For instance, it might be referring to home deliveries and mentioning supermarkets. Unfortunately, computers are not very good at interpreting these situations.”
 
This means that, in order to advance the project to the next stage, each tweet in the database first needs to be manually checked and classified as either being about food or not. As a small project team, this would have taken them an unfeasibly long amount of time… so they decided to call in reinforcements, by launching the project on Zooniverse!
 
Founded in 2009, Zooniverse is perhaps the best-known citizen science platform, and gives any member of the public the opportunity to contribute valuable data for research projects. These typically involve categorising data that computers would struggle to do automatically, such as picking out animals in camera trap footage, transcribing historical documents, or identifying cell features from microscope images. For this project, participants were asked to read through the Twitter posts and decide whether they were about food or not.
 
Since launching on Zooniverse in July this year, the response has been phenomenal. By the beginning of August, over 209,000 Tweets had been classified by 1,200 volunteers, reaching a high of over 36,500 classifications in a single day. But besides the raw data, Laura has been impressed by how the platform encourages dialogue between contributors and the research teams. “What makes citizen science projects distinct from many other research projects is that there is a high degree of knowledge sharing: the Zooniverse contributors are not passive participants” she says. “On the discussion boards for the project, participants have often reflected on what particular tweets mean to them and how they relate to their own food experiences during the pandemic. This really adds value to the study, giving us a much greater breadth of lived experiences.”
 
The discussion boards also revealed how social and cultural differences can impact the understanding of food-related content, which could help inform the design of future studies. As Laura explains: “Our database is restricted to UK-located Tweets, but Zooniverse contributors are based all around the world. This has caused some questions about certain terms, for instance about the names of some UK supermarkets that aren’t widely known abroad, and whether Easter eggs in the UK are edible. The two-way discussion boards are a great feature, as they enable us to respond rapidly to these issues, and share information with the whole community of contributors.’
 
The team hope to start training the first machine learning model in September 2022. Once the workflow is established, Laura hopes it can be adapted to process information on an international scale and from different social media platforms.
 
“I’m talking about Zooniverse to anyone who will listen – I think it’s an absolutely brilliant platform, which I definitely hope to use more of in the future” she says. “For instance, it would be very interesting to train an algorithm to classify photographs of meals that people post on social media for specific properties, such as the number of food components and portion size.”
 
“As recent events have shown, we live in uncertain times and we can’t take our food systems for granted, presuming that we can always get exactly what we want from the supermarket, whenever we want” she added. “It is imperative that when global disruptive events occur, we have the information and the means to ensure that the most vulnerable people in our societies who may be food insecure aren’t hit the hardest.”

The COVID-19 pandemic brought to light just how fragile food supply chains can be: with traditional trade disrupted, many producers struggled to find alternative routes to sell their products. Consequently, across the world shoppers were faced with empty shelves at supermarkets, even though farmers were having to throw away mountains of unsold produce.

Some of the worst hit were the many thousands of smallholder farmers and urban microenterprises trading fresh, perishable produce in low- and middle-income countries. Typically, these rely on in-person interactions at marketplaces, trade hubs and bazars, meaning that the COVID-19 lockdowns effectively shut down their trade.
Sarang Vaidya, co-founder of fresh produce agritech venture Go4fresh, describes the situation in his native India: “Disruptions to traditional marketplaces during the COVID-19 pandemic resulted in extensive on-farm food wastages, unfulfilled demand, and major setbacks to the livelihoods of millions of small grocery shops, street vendors, and hawkers. If we want to help communities to become more resilient to future shocks, we need to provide access to alternative markets.”

A marketplace on a mobile phone
With mobile phones being almost ubiquitous across India, Sarang believes that one solution could be to create ‘virtual’ marketplaces that directly connect smallholder farmers to buyers (such as retailers, wholesalers, restaurants, and canteens). In 2019, this vision inspired him and his team at Go4fresh to develop a collaborative, digital platform to connect producers and businesses in food supply chains.

However, currently very little is known about how such a platform should be designed to best meet the needs of smallholder farmers and microenterprises. To address this, Sarang is leading a SFN-backed project which is engaging end-users to identify the key data sources and features required to build a scalable, user-friendly virtual marketplace for affordable smartphone devices. The UK-India partnership involves experts in data science from STFC, besides a variety of Indian-based food-sector partners, including farmer groups and fresh produce buyers.

The central objective was to build a prototype virtual marketplace for four vegetables (tomato, cabbage, okra, and green chilli) along an established trade route in India. This ran from the village of Otur in the Pune district of Maharashtra, to a market in the Kandivali neighbourhood of the state capital Mumbai (a distance of 200 km, or roughly five hours of driving).

Through the Indian team’s links, the project consulted over 60 smallholder famers, microenterprise owners, and transporters. Due to ongoing COVID-19 lockdowns and social distancing restrictions, many of the interviews were conducted using phone/video calls or online chat functions. This qualitative research was supplemented with quantitative data, including live market prices, delivery dates, crop production data, and consumer prices on ecommerce platforms.

Mapping the challenges against widespread adoption of digital technologies
“We found that both the supply and market ends of the chain faced distinct challenges hindering the adoption of digital technologies” says Sarang. For those supplying fresh produce, a key issue was low current use of digital technologies, with less than a third using digital sources of information, such as market data and weather alerts. Another concern was that current data sources were often updated only infrequently, limiting their accuracy over time. In combination, this meant that smallholder farmers were often unaware of the current market prices for their products, resulting in them selling their goods for below-average price.

Meanwhile, for the microenterprises purchasing fresh produce, the most prominent issue was the large variation seen among different lots of the same food product (including quality, weight, and packaging) and the lack of standardisation. This inconsistency meant that most traders visited the market every day to ensure they could access the best-quality products available.

From their findings, the team mapped out the different variables currently influencing transactions between buyers and sellers, and used these to build a prototype digital marketplace. This also sought to address the low access smallholders currently have to pricing information by linking to benchmark markets with live data feeds. STFC co-PIs in the project Dr Jens Jenson and Dr Tom Kirkham, contributed their expertise in data architecture, selecting a platform and user interface, and effective use of data science & sensor technology.

A promising prototype
The result was an ecommerce platform designed to be resilient to network failures, provide strong data security, and be compatible with mobile devices. “The buyer interface has an end-to-end order management, with an integrated payment gateway. Meanwhile, the seller interface is a simpler version of a multivendor marketplace with features to add products, quantities, and target prices” says Sarang. “Based on over 300 transactions, we found that when farmers used the app to check the forecasted price for their product, this could increase their income by 15-22%, by helping them to find the best place to sell their products” he says. “This also enabled them to decide whether it was worth transporting their produce to the market in Mumbai to get a better price, or if they would receive the same amount at a local market.”

In the future, Go4fresh envisages that the apps could also allow smallholders to capitalise on growing food trends, including the rising demand in India for healthy and nutritious food, besides food that is produced more sustainably. “Digital marketplaces can allow smallholder farmers to directly connect with the buyers who are most interested in their products, resulting in the best price for them and higher-quality products for consumers” Sarang says.

For the next stage, the team will refine the template, based on input from a wider range of stakeholders covering different food products and supply routes. A particular aim is to simplify the interface and introduce an Indian language version for those with limited English. Ultimately, however, Sarang hopes this approach will have an impact far beyond India.

“In many countries across South-East Asia and Africa, we see a similar scenario. If rolled-out at scale, digital marketplaces could help provide a more sustainable and resilient livelihood for millions of marginal communities” he concludes.

Farming already accounts for around 38% of the global land surface (Food and Agriculture Organization), but this will need to expand even further if current methods have to scale to feed the rapidly growing population. And with cultivated soils degrading at a frightening rate, agriculture is increasingly encroaching into pristine, biodiverse habitats. With ‘business as usual’ clearly being an unsustainable option, we urgently need to rethink the way we farm.

This has sparked huge interest in vertical farming, where plants are grown stacked in layers within a protected environment (typically indoors). In these soil-free systems, plant roots are either immersed in a nutrient solution (hydroponics) or exposed to the air and irrigated with a nutrient-rich mist (aeroponics). Vertical farming offers many benefits, including high space-efficiency, reduced water usage, shorter growing times, reduced need for pesticides/herbicides, and shelter from extreme weather. In addition, since vertical farms can be set up practically anywhere (even underground), they could enable hyper-localised production, thus shortening supply chains and providing fresh, nutritious food all year round. 

But there are currently many challenges with vertical farming, including high capital costs, considerable energy consumption and inefficient resource use. According to Chungui Lu, Professor of Sustainable Agriculture at Nottingham Trent University, a crucial first step to address these issues would be facilitating greater precision control of plant growing conditions. “Young crop plants are incredibly sensitive to a large range of different environmental factors, including air and water temperature, light, relative humidity, and CO2 levels” he says. “Even small differences can have a great effect on growth. However we currently have very little data on the optimum conditions for indoor farming systems.”

With so many different factors affecting plant growth, it would be virtually impossible to manually work out the best combination in real-time to maximise yields. So as part of an SFN scoping project, Chungui joined forces with Dr Gadelhag Mohmed and Dr Steven Grundy from Nottingham Trent University, and Professor Wantao Yu (University of Roehampton) to explore how big data approaches could be used to derive ‘optimum growing recipes’.
​
“Essentially we had two main questions” says Chungui. “Can we monitor in real-time all the factors that affect plant growth? And can we then use artificial intelligence, big data and internet of things methods to predict the conditions that give the best yield and quality for different crops?”

The team’s approach centred on using neural networks, so-called because they mimic the way the human brain processes information. Structurally, neural networks are made up of a number of layers of nodes (neurones), each connected to nodes in the next layer. The connections between the nodes are weighted, and these weights are adjusted during a training process using a labelled dataset. The resulting network of connected neurones can then automatically (and accurately) process novel data to achieve the correct output.

Traditionally, developing a neural network relies on prescribing important features within the training dataset, which requires expertise and can be time-consuming. But what sets deep learning apart is that the algorithm doesn’t need important features to be specified in advance: given enough labelled samples, it will discover the most relevant attributes on its own. “It would take forever to experimentally test each individual combination of the factors that affect plant growth, so being able to automatically infer this would be a tremendous advantage” says Chungui.

But this intuition comes at the cost of considerable processing power: a single neural network can be made up of tens to hundreds of layers of artificial neurones connected by millions of weights, and in a single experiment go through over a billion computational operations. Since this is far beyond the capabilities of most computers, the team collaborated with STFC’s Tom Kirkham to access the Hartree Center at STFC Daresbury Laboratory in Cheshire. This enabled them to construct a cloud computing and data processing platform to develop the model. “A real strength of being part of the SFN network is being able to access experts and facilities to support really cutting-edge techniques. It makes a really good platform for research” says Chungui.

To train the model, the team collected data from sensors which measured the temperature, day length, relative humidity, CO2 levels, and light levels. The output data included measurements of shoot fresh/dry weight, root dry weight, and leaf area, and photosynthetic efficiency. “The algorithms in the resulting trained model had a strong predictive power, demonstrating that we can effectively extract hidden patterns in these huge datasets to identify the best growth conditions” says Chungui.

For their next steps, the team are looking to combine this approach with high-quality, real-time data captured using a state-of-the-art plant phenotyping scanner: the PlantEye multispectral 3D scanner. “These impressive machines can measure a whole range of factors related to plant crop and productivity, including biomass, plant height, 3D leaf area, and greenness– all without having to destroy the plant” says Chungui.

Chungui is particularly keen to extend this work to explore tailored ‘light recipes’ for different crops. Plants mainly use the red and blue parts of the visible light spectrum to power photosynthesis, and differences in the relative proportions of these can have a big impact on growth. Since blue light uses considerably more energy than red, working out the optimum recipe can result in great cost savings. Besides productivity, the team intend to investigate how different light combinations also affect flavour and nutritional value.

But ultimately, Chungui’s research group wants to see this technology in the hands of farmers, empowering them to make evidence-based decisions. “We are now exploring ‘Smart Green Grow Vertical Farming’ models, where the growing systems have built-in data transfer nodes and communicate with remote data bots to monitor and control environmental and plant performance factors. Ultimately, this would link farmers and their crops together with instant food production enhancement data.”

Manure and other organic materials offer a more sustainable alternative to conventional, mineral-based crop fertilisers that carry a high carbon footprint. But our lack of understanding about how these materials interact with soils currently limits their widespread use. With support from the STFC Food Network (SFN), Dr Ruben Sakrabani (Cranfield University) is addressing this knowledge gap so that more farmers can transition to a greener way of crop production.

“Producing mineral fertilisers for agriculture requires high amounts of energy, causing it to be a major source of greenhouse gas emissions, and resulting in an expensive final product” says Ruben. In contrast, producing ‘natural fertilisers’ such as compost, manure or slurry, causes fewer greenhouse gas emissions and is generally much cheaper, since many are waste products from farming practices. These materials can also enhance soil structure, increase soil carbon levels, and encourage beneficial microorganisms that support crop health. But there is a key problem against their widespread adoption, as Ruben explains: 

“A fundamental issue with organic fertilisers is that their composition can vary considerably, making it difficult for farmers to consistently apply the right level of nutrients, such as nitrogen and phosphorus. Whereas mineral fertilisers with similar formulations will essentially be the same, whether you produce them in London, New Delhi or anywhere else in the world” he says.

As part of a project funded by Innovate UK, Ruben is helping to develop organo-mineral fertilisers which combine the ‘best of both worlds’: the environmental benefits of organic materials with the consistency of inorganic fertilisers. The research team from Cranfield University have partnered with CCm Technologies, a specialist in carbon capture technology, to trap CO2 from industrial sources (such as a factory chimney) into organic materials. These are dried into pellets that farmers can apply directly to the soil.

Nevertheless, incorporating the organic element does introduce some variability between batches. “As long as there is any uncertainty over what exactly the pellets contain, farmers will tend to choose the ‘safer’ option of mineral fertilisers” says Ruben. “But we realised that if this variability can be easily and accurately quantified, it will no longer be an issue as farmers can adjust the amount they apply to the soil. The novel aspect of our work is to investigate whether this can be done with techniques which have never been used for this purpose before.”
​
Through a scoping project grant from the SFN, Ruben launched a collaboration with two researchers based at STFC Rutherford Appleton Laboratories, in Harwell Oxfordshire. These were Dr Genoveva Burca, a neutron imaging and diffraction scientist at the ISIS Neutron and Muon Source, and Dr Sara Mosca, a Raman spectroscopy scientist at Central Laser Facility. Their experiments combined different non-destructive techniques such as neutron imaging and Raman spectroscopy (Box 1) onto the same individual pellets.

“This initial feasibility work demonstrated that these two techniques are entirely workable with the fertiliser pellets, and give an unprecedentedly detailed map of both the physical and chemical characteristics” said Ruben. Whilst Raman spectroscopy defines the chemical bonds present in the pellets, neutron imaging gives information on the humidity distribution and how different particles are arranged.
​
The team have now progressed to testing their organo-mineral fertilisers in field trials at the Luton Hoo Estate in Bedfordshire, and have just harvested their first crop of winter wheat and winter barley. Promisingly, for both crops using the pellets as a fertiliser resulted in the same yield as crops that were treated with conventional mineral fertilisers. To verify this over the long-term, and to investigate how carbon sequestration is affected, the Cranfield University team have secured funding from Cranfield University and CCm Technologies to launch a three-year monitoring study using oil-seed rape and spring barley.

Ultimately, Ruben hopes this research will be applied to develop a simple handheld device that can quickly assess different nutrient levels in batches of fertiliser. “Besides reducing our reliance on mineral fertilisers, our second aim is for this technology to boost yields by helping farmers apply the right nutrients at the optimal time. Plants are like growing children: their development occurs in phases and their needs vary depending on the stage they are at” says Ruben. He is keen to start engaging end-users now, however, and organised a demonstration event at the field study site ahead of the COP26 UN Climate Conference. This was attended by a wide range of interested stakeholders, including researchers, fertiliser companies, farmers, agronomists, water companies and policy makers.

“Innovation is all about pushing the frontiers – and that is exactly what we are doing here” Ruben concludes.