UK United Kingdom

Get real: agronomists and breeders should review agricultural science

How will agriculture adapt to a changing climate? It’s an important question and, as more governments start worrying about future food security, one that’s worth trying to answer. Many well-known scientific…

To adapt crops to climate change, we need to know what’s working on the ground. bark/Flickr

How will agriculture adapt to a changing climate? It’s an important question and, as more governments start worrying about future food security, one that’s worth trying to answer.

Many well-known scientific journals frequently publish papers that claim to predict global crop yields or offer guidance on how farmers can adapt. But most of these papers are written in an academic bubble, and this seriously undermines their usefulness.

The most recent of these papers has just been published in Nature Climate Change. It claims its findings:

reveal that the use of seasonal climatic forecasts to predict crop failures will be useful for monitoring global food production and will encourage the adaptation of food systems to climatic extremes.

While the pathway to those findings is unfathomable to someone not deeply immersed in the arcane techniques of modelling chaotic systems such as the weather, its claims are wide open for debate.

It is a common characteristic of the authors of such papers that they do not ask potential users of their findings - agronomists, breeders, marketers - to comment on the claims. Reality therapy is not something they usually seek.

This comment applies not only in modelling, but also to most of the many papers in the experimental plant sciences that claim to be useful for improving food production.

What is to be gained from a new technique for predicting crop failures? The large financial industry based on grain futures relies on information about currently growing crops around the world. Good information is valuable to it. But does this recent paper improve this information?

To find out, the authors would have had to compare the usefulness of their model with currently used techniques of forecasting the yield. And one cannot tell from the paper if they have.

If there is no demonstrable improvement, why bother with the exercise? Is the paper a work in progress, an advance towards something that will eventually be demonstrably useful? If so, why is it not published in a more appropriate journal?

The second part of the paper’s claim to usefulness is that it “will encourage the adaptation of food systems to climatic extremes”. Who will do this adapting, exactly, and be encouraged by this paper?

First-rate breeders and agronomists are already hard at work solving problems of grain crops subject to volatile environments, not least in Australia. Is there anything that they would do differently if they read this paper? I would say not.

Our wheat breeders test their potential new cultivars in a wide range of environments across Australia, from the driest to the wettest parts of the wheat belt, sampling extremes of risk of frost damage and of heat damage.

Our agronomists keep making substantial improvements in farming systems that are substantially increasing yields per unit of rainfall.

There is really no serious basis for the claims of usefulness in this paper and the thousands of others that make similar vague claims of being agriculturally helpful. This is not a matter of fraud but of naivety, not only of the authors but of the institutions that support them. It highlights a serious problem with our system of peer review in agricultural research. It is a problem that can be readily solved.

Certainly reviews by peers in the given field remain essential. What is missing is additional review by people further along the value chain of agricultural R&D, the potential users. They are the ones who can check the promise of a given piece of work, because they are aware of the constraints and requirements of current practices and understanding in their own domains.

Involving them would be enormously beneficial. Their role would be much more than that of a filter. Just as importantly, they could provide guidance on what would make for substantial improvements in their own domains, including where the best opportunities may lie.

It is now common for our agronomists to work in tandem with designated farmer groups and their management consultants. Such collaborations are proving to be very effective.

The way is open for laboratory scientists to embrace similar close collaborations - not with farmer groups, for that would be a step too far, but with field scientists, the agronomists and the breeders.

Knowledge of what would potentially work in the field would then be available to them, thereby greatly enhancing the potential usefulness of their research.

Articles also by This Author

Sign in to Favourite

Join the conversation

8 Comments sorted by

  1. John Newlands

    tree changer

    My guess is that broadacre farming will continue to be a gamble. However horticulture will increasingly insure against erratic weather by use of microclimates, modified soil profiles and 'fertigation'. For example stone fruit can be decimated by both heatwaves and unseasonal frosts. The use of netting not only protects against hail and birds but traps air pockets and partly shades the trees. If it doesn't rain then a drip system will supply water, perhaps that has been recycled.

    My guess is that now most tomatoes are grown hydroponically not outdoors. The problem will be reliable grain production with crazy seasons. Perhaps we should rely less on grain and more on foods like corn and potatoes. They can be grown near the suburbs on treated effluent. The nation of Cuba could give us some pointers when broadacre farming gets too difficult.

  2. David Thompson

    Science Communications at Hawkesbury Institute for the Environment (UWS) at University of Western Sydney

    That's a fair point and of course it would be a missed opportunity if the results of what is often important and insightful work didn't flow down to end users and advisers.

    We held a new Soil Biology Masterclass in early August 2013 to achieve exactly this idea of taking new research out to industry and growers. The rates of knowledge discovery and new findings in most sciences now mean that researchers also need to find ways to take their results and shape them to be applicable to end users in order that the results aren't buried under a flow of new research.

    1. David Kemp

      Professor of Agricultural Systems at Charles Sturt University

      In reply to David Thompson

      John Passioura's long experience brings out a very critical issue. A useful example of how good Agronomy etc has delivered is in WA. Over the last thirty years, rainfall has been less than previously yet wheat yields have doubled. Its all been good applied science.
      Australian Scientists can deliver the goods, but the decline in support for R&D is associated with a plateauing in yield of cereals over the last 10-15 years (work by Mullen and associates). We are losing our edge at a time when we really need to up the effort in order to cope with what looks like an increasingly variable climate (supported by the models).

    2. Michael Hay


      In reply to David Thompson

      In my day, 1950's onwards, we called them 'field officers' or 'extension officers'. These were the persons trained in farming practices and adept at interpreting the scientific flow of words and implementing the ideas at farm level.
      We once had a system that worked - but like a lot of new ideas - the baby got thrown out with the bathwater. It would not be too hard to reinstate the practical training require to re-invent the 'field officer', but in this current environment he or she would need to be prepared to assess the marketing forces which accompany so many of the supposed marvellous new cures for incompetent farm practices.

  3. John Holmes

    Agronomist - semi retired consultant

    Several issues here. Part of the success of the of the State Department's of Agriculture before the ascent of the "small government is good" heresy and so called efficacy demands by placing all on short term contracts, was the interaction between a strong Adviser network and the dedicated Researchers.

    The Advisers kept the researchers feet on the ground. This included feedback as to the results of field trials and farmer experiences where preliminary results of new developments/insights were…

    Read more
  4. Comment removed by moderator.

  5. Carl Sudholz

    Proprietor at Fast Task Tools

    I couldn’t agree more. But this is not a new phenomenon. In the agriculture-software field especially, when evaluations of product have happened, they are highly questionable. I.e they tried it on a single farm; the sample size of users was less than 10 (most like friends of the scientists) and for only one repetition; their validation was by comparison to another model that has never been validated.

    Here are some quotes from my nearly completed literature review. These are just a tiny fraction…

    Read more
  6. Denise Cunningham

    farmer, teacher, writer

    Interesting article which applies skepticism, poses questions, and proposes solutions in just the right balance.
    I'm actually finding it difficult to get past the phrase "predict crop failures". Maybe I'm being overly literal, but to me that means the initial seasonal climatic forecasts were such that crops have already been sown (i.e., farmers have invested in fertiliser, seed, fuel and time), but the forecasters have subsequently changed their minds and they're now predicting that what has been sown will fail.

    What is needed is accurate long-range forecasting. We check rain / temp / wind forecasts almost daily and can verify that current methods can't often predict a week ahead, let alone a month, and especially not a season. But, we have to base decisions on the best available info...... you can never be SURE the forecast is wrong!