How maths helps sub-Saharan farmers

A complex mathematical model derived by a team of analysts from the US and Europe is helping farmers and agronomists better understand – and potentially control – a major threat to food security in sub-Saharan Africa.

Maize lethal necrosis (MLN) is a relatively recent emergent phenomenon that is devastating to up to 90% of affected crops. This has severe implications because maize is a critical foodstuff in sub-Saharan regions, as well as an important source of income for farmers.

MLN has proved challenging to control because it is not in itself a plant disease, but a condition that arises from the interaction of two quite separate viruses – maize chlorotic mottle virus and sugarcane mosaic virus (or, occasionally, a closely related species).

To further complicate matters, both viruses are robust and adaptable. They can spread during growing seasons and when plots are fallow. They can be transmitted by insects ranging from beetles to aphids, as well as via seeds and infected soils.

Previous attempts to establish disease management strategies against MLN have proven to be ineffective because they have based their predictions only on modelling for one or the other contributing diseases: calculating the multiple variables arising from modelling both has thus far been beyond the reach of standard computing tools. {%recommended 3801%}

Crop rotation, a traditional farm management practice, has the potential to break the continued transmission of the two viruses, but, the researchers point out in the journal Phytopathology, to be effective, it must be coordinated across many farms simultaneously, to prevent viable pockets of disease persisting.

Other methods such as deploying pesticides and purchasing certified disease-free seed can also potentially reduce MLN effects, but these are options generally available only to the operators of large, corporate farms. Subsistence smallholders, in the absence of any government or non-government assistance, miss out.

A team, led by Nik Cunniffe of the University of Cambridge in the UK, set out to construct a detailed mathematical model that incorporates most possible variables for both contributing diseases.

To prevent the framework becoming unwieldy, however, the researchers had to make some choices. They decided, for instance, to include data on the region’s two annual rainy seasons, but opted not to include information describing erratic out-of-season rainfall.

They also opted to model only two types of farms – very big ones and very small ones – on the reasonable assumption that those in between the extremes would fall along predictable lines.

The modelling showed that, over all, crop rotation carried the best chance of controlling MLN, but only “when combined with other control methods and when done over large spatial extents”. The researchers suggest their model may also be useful in combating other plant diseases caused by two combined pathogens, notably sweet potato virus disease and rice tungro disease.