/ 10 March 2000

KwaZulu-Natal researcher who predicted

floods

David Le Page

A University of Zululand researcher publicly predicted the dramatic flooding in the Northern Province and Mozambique, but his warnings went unheeded by authorities, particularly dam managers.

According to a report published this week by British magazine New Scientist, Mark Jury made his predictions as long ago as September. Jury is currently on a conference tour in the United Kingdom, and will be discussing his forecasting methods at the University of Southampton.

Jury believes the impact of the flooding could have been reduced by better management of the dams on the Limpopo. He says that dam managers have kept their reservoirs too full, and have contributed to the flooding by panicking and emptying dams in a hurry as heavy rains swelled the rivers. The dams should have been partially emptied much earlier, says Jury.

Jury specifically predicted a 75% higher than usual rainfall in the Northern Province catchment areas for the Limpopo. The Limpopo has been responsible for most of the Mozambican flooding.

Jury has previously been involved in at least one conference called by the Mozambican Ministry for the Co-ordination of Environmental Affairs and the Mozambique National Institute of Meteorology, intended to allow action planning in response to long-range forecasts.

Jury’s predictions were at odds with those of international forecasters, who predicted above-average rains, but did not guess at their extent. Interestingly, the model used by Jury to generate six-month forecasts runs on a humble personal computer, while the international experts use supercomputers.

Jury agrees with other forecasters that the cause of the floods was the cold- water Pacific weather phenomenon called La Nia, when sea surface temperatures in the Pacific around the equator drop several degrees due to the strengthening of trade winds. La Nia usually brings above-average rains to south-east Africa. But Jury believes the effect of La Nia can be magnified by certain weather conditions over the Indian Ocean.

The crucial omission of many weather models, he says, is that they ignore a factor that normally diverts west-moving cyclones south and away from the continent. This is an east-moving jet stream, or high-altitude wind, which is unusually weak this year and so not playing its usual role.

One of the problems in accurately predicting these phenomena is a shortage of temperature data from the Indian Ocean. What there is is still supplied by passing ships.

A staff member at the South African Weather Bureau who deals with long-range forecasts was loathe to comment in detail on Jury’s predictions, but said “sometimes he gets it right, and sometimes we get it right”. The weather bureau uses statistical information, global climate modelling run on a Cray supercomputer, and data from the United States in deriving its own forecasts.

Bryan Davies of the University of Cape Town, who has been warning that poor management of the Cahora Bassa dam on the Zambezi could worsen the effects of future floods in Mozambique, says that dam management in South Africa may have played a role in the Mozambican flooding.

Unable to comment specifically on the quality of South African dam management, he nonetheless points out that dams should be managed to simulate seasonal flows. If this is done, people on flood plains are prepared for seasonal deluges, and plan their lives accordingly. Unfortunately, he says, engineers are prone to believing a myth – that dams stop all floods. When major floods prove them wrong, people living in river flood plains end up suffering.