Latest News Archive

Please select Category, Year, and then Month to display items
Previous Archive
12 October 2020 | Story Leonie Bolleurs | Photo Supplied
Adriaan van der Walt
Although several international studies have used temperature metrics to statistically classify their seasonal divisions, a study in which Adriaan van der Walt was involved, would be the first known publication in a South African context using temperature as classification metric.

Gone are the days when we as South Africans would experience a three-month spring season, easing into summer, and then cooling off for three months before we hit winter.

Adriaan van der Walt, Lecturer in the Department of Geography at the University of the Free State (UFS), focuses his research on biometeorology (a specialist discipline exploring the role and climate change in physical and human environments) as well as climatology and geographic information systems.

He recently published an article: ‘Statistical classification of South African seasonal divisions on the basis of daily temperature data’ in the South African Journal of Science.

In this study, which Van der Walt undertook with Jennifer Fitchett, a colleague from the University of the Witwatersrand, data on daily maximum and minimum temperatures was collected from 35 meteorological stations of the South African Weather Service, covering the period between 1980 and 2015.

They went to great lengths to ensure that they had a complete set of data before presenting it to demonstrate seasonal brackets.

First for South Africa

Their statistical seasonal brackets indicate that South Africans now experience longer summers (from October to March), autumn in April and May, winter from June to August, and spring in September.

Although considerable work has been done using rainfall to determine seasonality in Southern Africa, Van der Walt believes that these methods did not work well as there are too many inconsistencies in this approach, as identified by Roffe et al. (2019, South African Geographical Journal). To make matters more complicated – as a semi-arid region, and with desert conditions along the west coast – some regions do not have enough rainfall to use as a classifier.

Temperature, on the other hand, worked well in this study. “Temperature, by contrast, is a continuous variable, and in Southern Africa has sufficient seasonal variation to allow for successful classification,” says Van der Walt.

He continues: “Although several international studies used temperature metrics to statistically classify their seasonal divisions, this study would be the first known publication in a South African context using temperature as classification metric.”

Van der Walt says what we understand as seasons largely relates to phenology – the appearance of blossoms in spring, the colouration and fall of leaves in autumn, and the migration of birds as a few examples. “These phenological shifts are more sensitive to temperature than other climatic variables.”

Seasonal brackets

According to Van der Walt, they believe that a clearly defined and communicated method should be used in defining seasons, rather than just assigning months to seasons.

“One of the most important arguments of our work is that one needs to critically consider breaks in seasons, rather than arbitrarily placing months into seasons, and so we welcome any alternate approaches,” he says.

A number of sectors apply the temperature-based division to their benefit. “For example, in the tourism sector it is becoming increasingly important to align advertising with the season most climatically suitable for tourism,” says Van der Walt.

Temperature-based division is also used to develop adaptive strategies to monitor seasonal changes in temperature under climate change. However, Van der Walt points out that each sector will have its own way of defining seasons. “Seasonal boundaries should nevertheless be clearly communicated with the logic behind them,” he says.

News Archive

Mathematical methods used to detect and classify breast cancer masses
2016-08-10

Description: Breast lesions Tags: Breast lesions

Examples of Acho’s breast mass
segmentation identification

Breast cancer is the leading cause of female mortality in developing countries. According to the World Health Organization (WHO), the low survival rates in developing countries are mainly due to the lack of early detection and adequate diagnosis programs.

Seeing the picture more clearly

Susan Acho from the University of the Free State’s Department of Medical Physics, breast cancer research focuses on using mathematical methods to delineate and classify breast masses. Advancements in medical research have led to remarkable progress in breast cancer detection, however, according to Acho, the methods of diagnosis currently available commercially, lack a detailed finesse in accurately identifying the boundaries of breast mass lesions.

Inspiration drawn from pioneer

Drawing inspiration from the Mammography Computer Aided Diagnosis Development and Implementation (CAADI) project, which was the brainchild Prof William Rae, Head of the department of Medical Physics, Acho’s MMedSc thesis titled ‘Segmentation and Quantitative Characterisation of Breast Masses Imaged using Digital Mammography’ investigates classical segmentation algorithms, texture features and classification of breast masses in mammography. It is a rare research topic in South Africa.

 Characterisation of breast masses, involves delineating and analysing the breast mass region on a mammogram in order to determine its shape, margin and texture composition. Computer-aided diagnosis (CAD) program detects the outline of the mass lesion, and uses this information together with its texture features to determine the clinical traits of the mass. CAD programs mark suspicious areas for second look or areas on a mammogram that the radiologist might have overlooked. It can act as an independent double reader of a mammogram in institutions where there is a shortage of trained mammogram readers. 

Light at the end of the tunnel

Breast cancer is one of the most common malignancies among females in South Africa. “The challenge is being able to apply these mathematical methods in the medical field to help find solutions to specific medical problems, and that’s what I hope my research will do,” she says.

By using mathematics, physics and digital imaging to understand breast masses on mammograms, her research bridges the gap between these fields to provide algorithms which are applicable in medical image interpretation.

We use cookies to make interactions with our websites and services easy and meaningful. To better understand how they are used, read more about the UFS cookie policy. By continuing to use this site you are giving us your consent to do this.

Accept