Monitoring available water resources accurately and consistently is a key information requirement for successful water resource management, says geographical information services company GeoTerra Image (GTI) director Mark Thompson.
"With the Western Cape feeling the pressure of the drought, we are confident that we will be able to provide insights into the current drought as well as contribute positively towards understanding and monitoring surface-water resources across South Africa using Mzansi Amanzi."
Mzansi Amanzi is a Web-based monthly water monitoring tool that enables users to visualise and determine the extent of surface water resources across the entire country, using earth observation technologies.
National monthly monitoring across the country at high spatial detail will provide a better overall assessment of the catchment areas that have water resource potential, which can provide a better perspective for managers on overall water resources availability, says water management specialists DHI South Africa director Jason Hallowes.
"This could enable authorities to identify problem areas immediately prior to complications occurring, and to better understand the necessity of delaying or implementing water restrictions in specific areas."
Hallowes adds that the information generated through the system can also support other activities in certain areas from a health and safety perspective. "Stagnant water can lead to increased disease occurrence in malaria areas, and contribute to increases in enteric disease in poorly serviced communities."
Explaining how the water monitoring Web service works, Thompson says "GTI places the key in the ignition and starts the motor" that processes the satellite imagery through its unique algorithms in its cloud-based models, and produces the geospatial data product.
He explains that the cloud-based processing pulls in all the Sentinel 2 satellite imagery that has been recorded over South Africa during the previous month, after which the company developed modelling algorithms first identify any cloud-obscured areas for exclusion, and then classifies the remaining image pixels into those that contain water and those that do not.
"The derived water datasets are stored in our Google storage bucket and then synced from the cloud to our Web application."
When determining the accuracy of the modelled derivative, he says it is important to understand the resolution and minimum mapping unit of the product because the models can accurately determine surface water if the surface water covers an area larger than 0.5 ha. Each pixel in the Sentinel satellite imagery is resampled to a resolution of 20 m, allowing us to repeatedly and accurately identify and map water features down to 0.5 ha in size.
"Owing to the resolution of the input data, the models will classify a farm dam, but may miss the small pond in the back of your garden."
Thompson says the beta version of the online Mzansi Amanzi Web portal was launched in November last year and the company is planning to release several updates this year, which will have various user and information content enhancements of the platform and its features.
"The first release was primarily a visual interpretative indication of water in South Africa, with the next release focusing on making the quantitative statistics available to the user through tabulated spreadsheets and digital data . . . to quantify the surface-water changes that occur," he explains.
Hallowes says in South Africa, only a fraction of the water resources is physically monitored by the Department of Water and Sanitation which measures about 205 dams across the country on a weekly basis.
Hallowes indicates that there is a significant number of larger dam and reservoir assets held by municipalities and private organisations, including farmers, natural lakes and industry, which are not monitored.
". . . to gain a good perspective on the actual availability of water resources in an area, an understanding of the overall storage provides a much better assessment, which can lead to better decision-making. The Mzanzi Amanzi web tool provides such an assessment on a regular basis."
The surface water monitoring system is the first of its kind in South Africa. The company recognises that, while water management authorities have existing systems to help manage water resources, GTI is positive that this service can add value to help and improve current monitoring systems.
GTI is collaborating with the DHI South Africa/ EkoSource partnership for testing the operational capability to derive volumetric estimates from South Africa's surface-water output across numerous dams in the country, using the surface area water data provided by Mzanzi Amanzi.
"Once the volumetric estimates have been completed, the product will be improved to provide dam volume estimates on a catchment basis for key individual larger and medium-sized dams, as well as an overview of water availability in all the smaller dams in an area. The final product will also assess water use and flow predictions to provide an assessment of how these systems will look six months to a year in the future," Hallowes adds.
Thompson confirms that the company wants to compile a quantitative accuracy figure to its monthly surface-water maps and it is engaging with several entities that could assist in validating the water classification output against field verified surface water.