Growing with precision22/08/17Environment
PEN speaks to the president of the ISPA, Dr Nicolas Tremblay, about the benefits of precision farming and some of the barriers to a wider realisation
A term used to describe the use of technology to better measure and control crop production on a site-specific basis to improve efficiency, Precision Agriculture (or Precision Farming) involves areas such as improved crop and field measurements, more efficient application of inputs (seed, fertiliser), better farm management decisions, and the more effective utilisation of tillage equipment.
Such an innovative arena, though by no means new, continues to experience challenges as well as successes, and Pan European Networks spoke to the president of the International Society of Precision Agriculture (ISPA) about some of these issues, as well as about some of the benefits that continue to emerge from advances in precision agriculture.
What have been the biggest benefits bestowed on agriculture by GPS/GNSS since it was introduced some time ago? How important is it for signal availability to be enhanced?
Precision agriculture (PA) emerged as a direct result of GPS/GNSS being made available to the general public. This was truly a game changer. It led to a new era that included not only new technologies, but also new agronomic opportunities such as data collection and better tools to answer old agricultural science questions. Today, GPS/GNSS is available everywhere, but the signal to obtain centimetre accuracy needs to be made more available. The more spatially accurate we get in our collection of data, the lower the spatial uncertainty of our database gets. In the future, more data will come from ‘citizen science’ and mobile phones, and positioning accuracy will become an even more important limitation.
Aside from the use of positioning systems, what else could be counted as among the biggest advances in precision agriculture in recent years?
One of the biggest advances in PA in recent years has been the acknowledgement that it is not all about spatial considerations; it is also about temporal ones. If you get a perfect application rate map in a field for one particular year, you can be quite sure that it will be far from perfect the next. This is because seasons are not alike. Therefore, weather data and forecasts have been increasingly used in PA decision support systems. Temporal requirements are even more obvious when you consider livestock. It is not so much about where the animals are located, but rather about the current status and needs of each individual.
What are your thoughts on the evolution of the robotic automation of farm equipment – including tractors, sprayers, and harvesters with automated steering to work with sub-inch accuracy?
Robotics and the automation of farm equipment is perhaps the component of PA that is really improving productivity today. One clear example is auto steering. The ability to drive farm equipment in straight lines and to follow pre-defined paths has made product application significantly more effective by limiting overlaps and has improved soil quality by limiting the damage being caused to fields by wheels. Auto steering is a truly great engineering achievement that has had an impact on the efficiency of agricultural operations. I use the term ‘engineering achievement’ because there is no ‘biological system’ involved in the solution, which makes problems much harder to solve.
How important is it for data to be collected across farms, and what are the biggest barriers to achieving this?
This is perhaps the main element that is currently limiting our ability to make more significant progress in PA. Obtaining data from sensors and satellites is great, but the information collected is seldom directly related to the data needed to make effective decisions on farms. What we need is data to calibrate and then validate the relationships. But this data is nowhere to be found in great enough geographical and temporal diversity to appropriately finalise the necessary steps. This results in shaky relationships and measurements, of which the usefulness is questionable. PA will not be a resounding success until scientists are given access to on-farm data and learn how to use the information.
It is true that there are issues with format, vocabulary, compatibility and so on, but the barriers, to paraphrase a Japanese scientist I once heard in a Research Data Alliance (RDA) meeting, are “5% technical and 95% cultural”. Some farmers are afraid that ‘Big Brother’ could do something bad to them with their own data. There is also a feeling that this whole fuss about data is not very productive for them. Why would farmers share their data when even scientists have a hard time sharing and making their datasets available? The reality is that the data used to produce one typical scientific paper is not enough to support the decisions that have to be made on a farm, unless it is put together with many similar data to create a more comprehensive picture.
A good example of what can be achieved from data sharing is the Agriculture and Agri-Food Canada SCAN (Soil, Crop, Atmosphere and Nitrogen) decision support system (DSS) for nitrogen recommendation in corn. SCAN’s rules are based on meta-analyses that have been performed on hundreds of trials in North America, more specifically in Quebec and Ontario. Fortunately, there has been much development in the area of open data policies and organisations such as the RDA and ISPA are helping in the process.
The data is important, but less so than the related meta-data, which puts the results obtained into context. This is the cornerstone of our ability to personalise. The future relies on concepts that include open innovation, operational research, participatory research, and living labs. That is when the big data era will truly begin in agriculture, and artificial intelligence algorithms will help farmers make the best day-to-day decisions.
It has been argued that grower adoption of seed and fertiliser management continues to lag, which is an issue due to the fact that the use of fertiliser can harm waterways and underground water sources. What are your thoughts on this?
Growers want to see the value in the measures they adopt. They have to deal with uncertainties, and our job as scientists is to help them become more confident in their decision making. The harm being done to waterways and underground water sources is greatly owed to our inability, collectively, to get farmers to make the right management decisions anytime, anywhere, which results in farmers instinctively ‘buying insurance’ by applying fertilisers or pesticides (or applying more) just in case. This is how contamination happens. A hypothesis that tends to confirm this is that, in principle, profits and the environment are not necessarily in opposition. Currently, they are in opposition because there is still too much uncertainty surrounding our decision making.
Following on from that – how could precision agriculture tie into the idea of sustainable agriculture, which could broadly be defined as the production of food, fibre, or other plant or animal products using farming techniques that protect the environment, public health, human communities, and animal welfare?
PA is the perfect concept to achieve sustainability as it is based on the best scientific practices. As I have mentioned, it is about making the right decisions anytime, anywhere. By doing so, farmers will stop feeling like they need insurance, and we will become more sustainable – both financially and environmentally. The PA philosophy (using technology, data and evidence generation to produce useful information for decision making) is clearly the way to go, but we may have to review our strategy and our scientific reward system.
PA can help a great deal in achieving sustainability with more involvement from farmers throughout the research process, multidisciplinary teams, better access to on-farm data, better adapted data mining strategies, a focus on artificial intelligence algorithms in decision support systems, and delivery tools that are compliant with farmers’ realities. Finally, the careers of scientists should not be primarily based on the number of papers they publish, but on the beneficial practices they have effectively been able to get adopted, and on their impact.
How would you characterise political support for precision/sustainable agriculture in Europe?
Precision/sustainable agriculture has been given very good consideration so far by managers and policy makers in Europe. It is perceived as having benefits and making a positive impact, and it is clear that the future of agriculture relies on making factual information available to improve production practices, which is almost PA by definition.
However, PA is not limited by the availability of technologies today. It is instead limited by the ability of scientists to establish predictive models for prescription and impact. Given the complexity of the systems they have to deal with, and their inherent aspect of uncertainty, the limitation can only be circumvented by getting reliable and diversified data from farms.
I think that Europe is at the forefront when it comes to putting policies in place to help organise open access to data. There is a wealth of data coming in every day, but we have to put it in a context that we can understand, which is the on-farm ‘ground truth’. The support for precision/sustainable agriculture should not be diminished, but past strategies have not delivered fast enough, and with the new opportunities we have now we should be supporting the new ways of solving our ever-recurring agricultural problems, at the expense of the old ones.
Moving forwards, what role will the ISPA continue to play?
The ISPA has a direct link with the founders of the PA concept dating back to the early 90s, and our role is to foster the evolution of science in the PA sector. There are many platforms and meetings dealing with PA everywhere in the world. The ISPA, for example, has the legitimacy, legacy, and experience to handle the critical scientific issues that are truly at the heart of how PA can be improved for the benefit of all. This is primarily done through our series of International Conferences on Precision Agriculture (ICPA), of which the 14th edition will be held in Montreal, Canada, on 24-27 June 2018. We are at the stage of the call for abstracts and sponsorship establishment, and I would like to invite all interested parties to visit https://ispag.org/icpa.
A variety of PA aspects will be covered, but special attention will be given to data mining and artificial intelligence opportunities as Montreal is a well-recognised international hub for these new approaches, including big data and deep learning, which can greatly benefit agriculture in the future. The ISPA board is also busy giving access to information sources that are currently dispersed and hard to retrieve. Additionally, our specialised communities of interest are working on initiatives to address critical PA issues.
Precision livestock farming (PLF)
As farms continue to grow in scale, and as many automated processes become introduced to aid in things like feeding, farmers have come to work with average values per group, rather than at the individual level. As such, it seems variety has become an impediment to increasing economies of scale.
It is here that PLF has a role to play. Using modern information technology, farmers can now record numerous attributes of each animal, including reproduction, pedigree, age, growth, and meat quality. The availability of this information can deliver significant benefits, both for the farmer (with significantly higher reproduction outcomes, for instance, with each new animal also contributing to a higher meat value) and for society, with a low carbon footprint of livestock production for the environment and, of course, the delivery of high quality food products for the market.
The requirements of a farm animal are well known for each phase of its life, and they allow the precise preparation of an optimal feed to support the animal. They are also oriented on the required nutrition – providing more nutrition than is required makes no economic sense, but providing fewer nutrients can be negative to the health of the animal. The goal of precision livestock farming is thus also to provide nutrition that satisfies the animal’s requirements at the lowest possible cost.
New technologies for PLF are of use throughout the value chain, including in abattoirs, wherein the deployment of Slaughter Registration Systems – which read the unique tags given to each animal at the moment of slaughter, after which the carcasses are traced through the abattoir – adds the tag number and other slaughter data (such as weight, quality, fat and customer) to the carcass. The pertinent slaughter data (carcass weight, quality, fat) can then be fed back to the farmer, who can use this data to improve farming.
Dr Nicolas Tremblay
International Society of Precision Agriculture (ISPA)