As published in Landbou Weekblad April 2014
In my last post I illustrated that more than 40% of our crop-loss is due to pests, diseases and weeds, in spite of annual applications of 2.5 million tons of pesticides worldwide (Paoletti & Pimentel 2000). Furthermore, adding to the challenge of feeding growing populations and rising demand, is climate change, which will reduce agricultural output by 10-15% in the next 60 years (Nelson et al. 2009), and the end of the green revolution. As agricultural output is levelling off (World Bank), we are starting to pay for the methods of the green revolution, where excessive use of chemicals has depleted our soils and made them tired.
That is not where the problems stop. Only ½ a degree change in temperature has an enormous effect on the life cycles of microbes, as research shows. This in turn means that a small advance in climate change has a fundamental shift in how diseases and pests will behave and influence our output. A change in behaviour of diseases means they will surprise us, and being surprised by a disease usually means we lose crop. This effect will have a faster pace as we progress into climate change. Nelson et al. did not include this effect into his calculations, when he concluded, that agricultural output would decrease by 10-15%. Chances are good, if we do not adopt, that projected output could even be considerably less.
The most difficult question, of how we will increase agricultural output by 100% in the next 40 years, to satisfy rising demand for meat and biofuels, compounded by growing populations, needs answering. There are two tools, which can clearly help: Agroecology with the ability to increase yield and produce plants with considerably more resilience against pests, diseases and droughts and Information / Communication Technology (ICT), which is the next step up from plain software tools.
In the case of geospatial software, farmers already have access to yield maps and variable rate seed and fertilizer applications. The full potential of management zones in the geospatial field, in form of cropping techniques and proper soil management, is still open to utilization. This is even more the case, with the exploitation of a complete new technology based on ICT, high throughput computing, machine intelligence and Apps. MySmartFarm has such an offering, with its imminent release in months.
The potential of big data and ICT, replacing plain desktop applications, is endless. One of them is a new disease modelling application, developed by MySmartFarm and making use of completely new scientific methods and analytics technology. These models, predicting diseases, are based on individual small regions and are therefore very accurate and superior to current existing models. This type of development will make use of phenomenological science, in contrast to standard practice reductionist science. The difference being that phenomenological science studies the entire phenomena of a diseases, rather than individual aspects of it, such as only temperature. For a farmer, in laymen terms, this means, that new disease prediction models are directly applicable to his region, based on many monitored environmental aspects, which influence the life cycles of diseases and pests.
This research, by MySmartFarm, much relies on data of farmers, data that is time stamped and geo-tagged and over many years accumulated. The more data exists, the more accurately the analytics engines and scientists’ interpretations will be for these models. Whereas 10-20% of prediction models can be confidently used in the very near future, the rest will only produce high correlation values in 3-5 years from now. Everything else will be inaccurate. The most promising aspect of this methodology of modelling diseases is that each disease model can be easily upgraded every year for every region, in that compensating for the effects of climate change.
Wolfgang J. von Loeper
