MIP+

Market uncertainty and risk aversion undermine farmers’ capacity to make informed decisions, which results in an overreaction to pests and an overuse of chemical inputs for food production. We propose a smartphone-based app that both assists in collecting relevant data from crops, and estimates the risk of yield losses, based on estimates of future produce prices and current pest damage, to decide the best course of action to manage one of the most limiting pests in greenhouse tomato crops in Colombia. Risk estimates and information on market prices and pest damage, rather than predefined or biased recommendations of what should be done, provide farmers the means to learn and adapt to uncertainty from calculated, small experiments over time. If our solution were scaled, farmers from developing countries will realize that zero-tolerance policies against pests harm their income and set a foot on the ongoing digital agricultural revolution.

Every year, more than four million tons of pesticides are applied in agricultural crops worldwide. Colombia is the fifth country with the highest pesticide use per ha in the world, which not only impacts public health and environment, but also increases vulnerability of production systems to the evolution of pesticide-resistant pests and reduces profit margins of small farmers. The overuse of pesticides impacts urban and rural populations, since residues are commonly found in produce, soils, and water. What is more, most pesticides applied are probably unnecessary, since decisions on what and when to apply are almost always made based on intuition, or suggestions from outsiders and rarely on data. The farmer community of Alto Ricaurte (Boyacá, Colombia), comprised of approx. 1500 people, primarily grows greenhouse tomatoes and their income is heavily affected by tomato pests and diseases. Around 17% of production costs are invested in pesticides, which are applied weekly under zero-tolerance-to-pest policies. Even though some farmers acknowledge that they tend to overreact to pest problems, they do not have the means to determine, based on evidence, the need for a pesticide application, which requires real-time predictions of produce prices and estimates of the current cumulative damage caused by pests.

Our solution is a smartphone app, named MIP+, that 1) Calculates a decision threshold to apply one of several control strategies, based on control costs and predictions of produce prices during harvest, 2) guides the user through a standardized procedure to collect data from a greenhouse tomato crop about the cumulative damage that has been caused by the pest, and 3) estimates the current risk of yield losses and presents the current pest population size relative to the estimated decision threshold. The first version of our app will include the most limiting tomato pest in the region, along with the three most popular and environmentally friendly control measures in the market, for which we have estimated efficacy. MIP+ will use online economic data repositories, along with weather forecasts, to estimate produce prices from three to 12 weeks ahead, which will be used to determine decision thresholds. The procedure to collect data from tomato crops and the calculation of risks of yield losses, according to the decision thresholds, will be included in an algorithm that will be implemented in the app.

Our target is a community of 350 farmers in the Alto Ricaurte region (Boyacá, Colombia), which comprises four municipalities (Santa Sofía, Villa de Leyva, Sutamarchan and Sachica) where the main source of income that feeds the economy is the cultivation of greenhouse tomatoes. Weather in Alto Ricarte is particular for the Colombian context. Temperature fluctuates around 15ºC, it is mostly dry all year, and soils are very poor in organic matter. Even though the tomato plant is native to south America, paradoxically, European varieties grow very well in the region, since they are adapted to such conditions. However, they also leave farmers with few choices, so diversity of produce is low and competition among neighbors is often strong.

Farmers in this region have very limited access to technical assistance, most of which comes from biased recommendations of pesticide salesmen. In a survey carried out in Alto Ricarte in 2017, we found that only 0.03% of the farmers use monitoring of their crops when it comes to managing pests. The rest of the respondents rely on outsiders, especially pesticide salesmen, or mere intuition. The main limitation is that deciding when or how to react to pest problems is not easy or intuitive. Ideally, the right moment to apply a measure is when the cost of applying it equals the yield losses caused by the problem. However, is it difficult for farmers to get estimates of tomato market prices, or to carry out the required calculations fast enough to make real-time decisions in the field.

We have been working with the community of farmers in Alto Ricarte since 2017. We started by carrying out a survey to get a better idea of their more general needs and limitations. Based on this information, we decided to work with one of the most limiting insect pests according to them, the tomato leafminer (TLM), Tuta absoluta. Over the next two years, we sampled crops regularly to collect the required data to develop sampling plans for decision-making. We made formal arrangements with several farmers to implement our pest management practices and decisions in their crops and to work alongside them and build trust. In exchange, we provided a small monetary compensation to cover any incidental losses and relieve aversion to tolerate pests. Throughout this period and thanks to our close relationship with the community, we organized meetings to discuss ideas and limitations to reduce pesticide use, and to further understand their mindset. In 2019, we organized a workshop to build a conceptual model that helped us understand the causes and effects of the TLM in the region. We found, once again, that the lack of tools for decision-making is a big obstacle, along with the poor communication among neighboring farmers.

Not only has our solution been developed alongside the community but can also be easily adjusted to be of use in other tomato growing regions in the country and abroad. The main advantage for an easy implementation and scalation is that our solution only relies on information and not additional and potentially expensive technology or inputs are required. We think that access to information and tools to use it will change the ways farmers make decisions. Better decisions imply optimization (and reduction of up to 50%) of chemical inputs to produce food, which translates in higher quality of produce, a healthier environment, and improved farmer and consumer quality of life.

Resilience is about building capital and capacities to adapt to uncertainty. Changing climate and markets increase uncertainty in the tomato production system. Our solution aims to promote empowerment, learning and memory building of farmers at the local scales to manage pest outbreaks, by relaxing the dependence in the capital and knowledge built at larger scales (i.e., private corporations). First, our solution reduces vulnerability to pest outbreaks caused by climate variations, while mitigating pressures against ecosystem services. Second, our solution increases the accuracy of decisions to optimize system inputs, considering market uncertainty, which allows for the release of financial capital.

We have already developed and tested the algorithms to guide the implementation of sampling plans for decision-making. We have also developed the mathematical functions to estimate dynamic decision thresholds based in the crop phenological stage. However, the component to estimate future tomato market prices (as a function of weather and supply-demand balance) is still lacking. Once this component is developed and adjusted, the whole systems must be integrated into one single platform (smartphone app). With the first version of the platform, we will be ready to deploy and validate our solution with the community of farmers in Alto Ricaurte (Boyacá, Colombia).

Conventionally, solutions to agricultural pest problems have been focused on the development of control inputs that could be either curative (pesticides) or prophylactic (transgenic varieties). Furthermore, other similar developments (apps) only provide assitance to identify the causal agents of ongoing problems. Much less attention has been paid to the development of procedures for a proper use of the existing technologies, when farmers already know the cause. Our approach is novel not only because it aims to optimize the use of exiting pesticides, but also because it takes advantage of the internet and data availability to both increase precision in deciding when and what to do, and favor learning by delivering risk estimates along with relevant information on economics and crop damage to pests. The usage of our solution will reduce the dependence of farmers in chemical inputs and build knowledge capital at the smallest scale of the production system (the farm). By making sounder management decisions, pest control costs could be reduced by up to 50% and stakeholders will move from a culture of collective risk-aversion to one of information and evidence use for crop management. Once farmers recognize what a given decision threshold looks like (for a given season, for example), they will figure out that most plants can tolerate more pests than they originally thought their crops did. Ultimately, our solution could change the way farmers manage their crops and businesses, from an emphasis on inputs to an emphasis on information.

Currently, our technology is not complete, so we are not serving anyone. After we socialize MIP+ with the region of Alto Ricaurte, we expect to directly serve at least 150 farmers in one year, which will impact about 600 people including their families and 2500 consumers. In five years, we expect to have reached the whole farmer community in Alto Ricaurte and started to introduce MIP+ to other producing regions in the country, serving directly about 500 farmers, which impacts approx. 2500 people including families and about 2 million tomato consumers.

We have already carried out a baseline study in 2017 in which we collected samples from water bodies, tomatoes, and soil. Furthermore, in 2019, we collected blood samples from selected farmers in Alto Ricaurte. We analyzed the samples for the presence of pesticide residues, considering the international maximum standards. In five years, after the implementation of MIP+, we will collect new samples to compare the impact of the decision-making system on the environment (soil and water) and public health (farmers’ blood). In 2017, we also carried out a survey that included specific questions about pesticide use and tools used for decision-making. We will measure pesticide use and preferences for decision-making in pest management through regular surveys constantly while we visit the area during MIP+ implementation.

Full-time staff: 7 researchers

Contractor: 1 (as part of the development of MIP+ we would hire a part time app developer)

Our team consists of three PhD-level researchers from different backgrounds and formations in different areas (ecological modelling, plant protection biology and phytopathology) and educational systems (USA, Spain, Sweden), and four M.Sc and B.Sc level agronomists formed in Colombia, with experience in Colombian agricultural systems. All of our team has 10+ years of experience working in the agricultural research sector along with farmers. Combined, this gives our team a wide range of disciplines from an academic perspective to face the technical challenges of developing MIP+ as well as the hands-on experience in the tomato production system to understand it fully. 

For the past three years this team has worked hand in hand with the farmers who we target with MIP+, gathering direct information about their needs, aspirations, and interpretations about pest problems at both the individual greenhouse scale, and that of the region. As such, as a team, we have gained the confidence and built the foundations of MIP+ with our target community.

At AGROSAVIA one of our main goals over the past decade has been to achieve gender equality among our researchers and we may proudly say that we currently sit at a 45:55 (female:male) ratio. We consider that every member of our team has earned their place through their work and dedication and currently have a team of three women and four men. Having women on our team is fundamental since a large proportion of the tomato farmers we work with are women which sometimes facilitates our team’s relationship with them.

We are applying to Solve because we truly believe that our solution will reduce the amount of chemical pesticides used by tomato farmers in Colombia, raising their profit margin, and contributing to environmental, farmer and consumer health. We have been developing our solution for the past four years and have reached a point where we require a small investment to transition from research to a viable product for farmers. We believe that a Solve grant would be enough to develop a first version of MIP+ which we could then field test and validate with growers to perfect it and continue to build farmer confidence. We also believe that Solve support could be fundamental to our development since, although we have IT support from AGROSAVIA for app development, it is not our area of expertise. Solve members would allow us to not only develop a nicer, smoother running app and web-based prompt, but to have a solid framework to build upon as we include other crop problems to MIP+.

As our first incursion into web and app-based developments, our focus is to develop the best and friendliest product for farmers. Since it is not our area of expertise, partners in the technological area and in how to best use the data acquired by the app to follow up on and feed our models will be fundamental for the adoption of MIP+.

We would like to interact with MIT media lab, specifically with the Lifelong Kindergarden group (LKG). They will be key to involve farmers in a learning process as they engage in using MIP+. We strongly believe that our solution requires a robust component that allow farmers to learn as they use the system for being sustainable. This includes providing the correct amount of information in the proper format so that it could be used by farmers to reduce risk in their future decisions. The LKG has great reputation and experience in developing technologies to engage people in learning experiences, just like the one we envision for the for MIP+ as a tool to promote a culture of information-based decision-making.

Although we do not work with refugees as such, we do work with farmers who are former members of the Colombian guerillas (FARC) who have disarmed and joined the peace accords signed in 2016. As part of the accords there is a crop substitution program for farmers to shift from illicit (coca, mainly) to licit crops along with programs to include former guerrilla members into the agricultural community. We strongly believe that the only way towards reconciliation is to grant former combatants opportunities and ways for them to include themselves into society. 

MIP+ aims at creating a sustainable tomato production system in Colombia, particularly in regions where the entire agricultural landscape and community is based around this single crop. We believe that reducing crop inputs (pesticides and others) through the use of knowledge is fundamental in reducing the carbon footprint of agriculture, increasing and protecting biodiversity in agricultural landscapes, contributing towards farmer and consumer health through more innocuous production, and achieving a more resilient system in the face of climate change and globalization. 

We also strongly believe that farmer learning, through easy to digest data and user-friendly solutions is fundamental in the adoption novel farming practices. Whereas technological developments in the agricultural sector usually come with a price tag associated with them, one that farmers in the developing world are often unable to pay, information, particularly that gathered by themselves only requires their own labor and the means to interpret it. Although this was once a difficult task, smartphones allow farmer to have all the processing power necessary to analyze complex information from varying sources, allowing them to make information driven decisions rather than following product recommendations which usually do not apply to their particular conditions.