African and Portuguese researchers have created detailed maps of solar energy distribution in Mozambique with the use of measuring instruments and machine learning technologies. These data will make it possible to plan the positioning and capacity of solar power plants with more precision, especially in rural and remote areas which do not have a stable power supply yet.
Mozambique has one of the lowest electrification rates in Africa. More than 80% of the country’s rural population have no access to electricity. At the same time, the country is located in an area with high solar potential, the full use of which is hindered by the natural features of the region: unstable solar radiation during the day, cloudiness and the presence of dust and water vapor. This is why it is crucial for Mozambique to accurately predict the amount of incoming energy in a given place and at a given time to make sure solar farms operate efficiently.
To this end, the researchers from Mozambique’s Eduardo Mondlane University and Portugal’s University of Lisbon have developed a method for high-precision solar energy mapping with a time resolution of up to one minute and a full territorial coverage of the country. The method is based on the so-called clear sky index (K*ₜ), an indicator that reflects how close the actual level of solar radiation is to the theoretical maximum under a completely clear sky. The index is calculated as the ratio of the actually measured radiation to the calculated value for ideal conditions, showing the degree of atmospheric transparency at any given time.
To collect data, the scientists used 11 measuring stations installed in different provinces of Mozambique. Each station used state-of-the-art devices: pyranometers for measuring overall solar radiation, pyrheliometers for recording direct solar radiation and diffuse radiation sensors. In total, over 500 million measurement points with a time step of 1 to 10 minutes were recorded over the three-year period under study. In addition, satellite and meteorological data from international space agencies (NASA, NOAA, Meteonorm and AERONET) were included in the analysis, making it possible to expand coverage and improve the accuracy of the assessment of atmospheric and climate factors. A number of machine learning models were used to analyze and forecast solar radiation. The algorithms used included neural networks, random forests, support vector machines and other regression methods. These models allowed the researchers not only to accurately reconstruct the dynamics of solar flux based on incomplete or noisy data, but also to identify patterns in the spatial and temporal distribution of energy.
The analysis showed that solar energy distribution across Mozambique is extremely uneven. The south of the country has more clear days and a higher clear sky index, making this part of Mozambique particularly promising for solar generation. In the northern and central regions, significant fluctuations are observed due to cloudiness, atmospheric aerosols and the influence of seasonal cyclones.
It is planned to use the results of the study to improve the accuracy of energy production forecasts, choose the best locations for solar power plants based on regional specifics and reduce the risks of interruptions and fluctuations in power, especially in isolated and rural power grids.
Generally speaking, the study confirms the high potential of solar insolation in African countries. Based on the clear sky index, Mozambique surpasses most of the countries in Europe and the humid tropics of Asia, its climate conditions approaching Brazil and Mexico in terms of solar activity. The study shows that, although North African countries such as Egypt and Morocco have slightly higher indicators, Mozambique’s climate conditions make it a promising hub for the development of solar energy in the southern part of the continent.
