Source: Coastweek, Article: David Musyoka
Researchers at the Kenya-based institution have developed a new remote sensing-based methodology to map flowering plants in Africa.
Using this methodology, which uniquely combines two hyperspectral mapping formulas, the researchers at the International Centre of Insect Physiology and Ecology (ICIPE), have produced the world’s first ever floral map.
“Based on its successful testing, we now intend to upscale and adapt the methodology we have developed so that it can also be used on satellite data,” said Tobias Landmann, head of Icipe’s Geo-Information Unit.
“We expect that its wider application will contribute towards filling critical gaps that exist, especially in Africa, on floral diversity,” he added.
The researchers present the suitability and accuracy of their new methodology, using a test study site of the spatial distribution, abundance and temporal cycle of flowering melliferous plants (those that produce nectar and pollen, which is collected by honey bees and converted into honey) in Mwingi County, eastern Kenya.
“In our test study, we deployed a hyperspectral sensor on an aircraft, which enabled us to collect data on all flowering species over an area measuring 100 square kilometres in Mwingi County,” Landmann said.
“We were able to attain an overall 83 per cent mapping accuracy. As a result, we produced a map showing the distribution, abundance, season and duration of bloom of melliferous plants, otherwise known as a floral cycle.
These findings show that it is possible to use remote sensing to map flowers precisely,” he explains.
Hyperspectral technology is one of the most advanced remote sensing strategies, as it combines powerful digital imaging and spectroscopy, thereby increasing the ability to detect discrete materials of interest, and to gather key information towards identifying and classifying them.
In a paper published in the Remote Sensing of Environment journal, Landmann observes, most studies conducted globally on flowering species previously have used ground-based methods.
However, such approaches are not cost effective or comprehensive, and they are also prone to too many errors.
Remote sensing has several advantages – it saves time, it is relatively economical, and it allows collection of data in areas that would be inaccessible on the ground.
He further noted that remote sensing can either be conducted via space or airborne platforms.
Space platforms, however, do not always provide the appropriate data resolution.
Airborne platforms are more ideal because they operate from a much lower level than space platforms, thereby providing higher resolution data.
“Our hope is that better understanding of floral diversity will enable decision makers and beekeepers to discern the interaction between bee colonies and the floral environment, leading to optimized beekeeping,” Landmann said.
Knowledge of floral diversity is critical, as it provides understanding of how environmental factors in a given area affect the productivity of bees, or their vulnerability to pests and diseases.
As an example, a decrease in the availability of melliferous plants within a landscape, for instance due to deforestation, would indicate that beekeeping and crop pollination have been compromised.