The amount of moisture in soil – and the way this moisture is distributed – combined with wind patterns in the lowest few kilometres of the atmosphere can influence where thunderstorms begin and how they develop. This new finding, from researchers at the UK Centre for Ecology and Hydrology (UKCEH) could help in the development of new early warning systems for such events, which are increasing worldwide and becoming more intense and dangerous as the climate warms.
Thunderstorms can develop quickly on hot afternoons, sometimes in less than half an hour of clouds building up, but predicting where they originate can be difficult.
A team of researchers led by meteorologist Christopher Taylor has now discovered that patches of dry soil 10–50 km across can combine with the wind field and affect how quickly convective storm clouds (cumulonimbus) form and grow.
“We already knew that differences in wind speed and direction with height (the ‘vertical wind shear’) in the atmosphere are critical ingredients for severe storm development, whilst gradients in land surface heating across the landscape can induce weak winds near the ground,” explains Taylor. “These two elements are usually studied separately, but we put them together and found that convective clouds grow very rapidly when the winds that steer them, some 3–4 km above the ground, oppose local surface-generated winds near the ground.”
This combination, he says, effectively increases the supply of moist, buoyant air into a cloud, accelerating the updraughts responsible for lightning and heavy rain.
“Storm initiations are clearly favoured in specific locations”
The result, he explains, challenges conventional thinking that over flat terrain, where cumulonimbus first develop, is essentially random. “In fact, under the conditions we studied – across sub-Saharan Africa – storm initiations are clearly favoured in specific locations, based on a combination of soil and wind conditions on that day.”
The work, which is detailed in Nature, could help in the development of more localized storm forecasting, he says, particularly in tropical areas where soil moisture gradients and wind shear are strong and can lead to flash flooding, lightning and strong winds.
The UKCEH team obtained its result by studying satellite images of 2.2 million afternoon storms in 2004–2024. They were able to obtain high-resolution data from the images and so observe fine-scale details of the wetness of soils.
The principle they have identified would be applicable to predicting thunderstorm formation in other parts of the world, such as Asia, the Americas, Australia and Europe – and not just the worst-hit tropical regions in Africa.
Ground-based measurement networks are scarce in Africa
Taylor and colleagues say they have been working with meteorological services in Africa for the last few years and contributing to international efforts to provide early warning systems for severe storms. Convective storms can be particularly damaging in built-up urban areas with intense rainfall damaging infrastructures such as roads and sanitation systems. “Unlike in the UK, where ground-based measurement networks are the backbone of weather forecasting, they are scarce in Africa and there are only a handful of meteorological radars here, explains Taylor. “We therefore had to rely on satellite data, which provide good quality information on some aspects of the coupled land-atmosphere system – notably the temperature (and therefore the height) of clouds and estimates of moisture in the top few centimetres of the soil.”
From this information, the researchers inferred how soil moisture affects evapotranspiration and atmospheric heating, how pressure gradients created by these heating patterns affect winds locally and, finally, how these inferred local winds interact with growing convective clouds.
The insights gleaned from this study could help improve the accuracy of short-term weather forecasts by providing a better indication of where storms are likely to appear within a region, Taylor says. “Just how much more skilful a forecast will be is an open question, but we have good reason to believe that in parts of Africa it could provide a big advance. In general, weather forecasting is a rapidly evolving field thanks to AI, and so the translation from research finding to application could be rapid.”
Pollutants and other aerosols trigger more intense thunderstorms
The researchers say they are now starting to look at how weather forecast models depict the processes described in their work. “Early indications suggest that models solving physical equations on a fine enough grid (of around 4 km) can capture the relationships between soil moisture, wind shear and cloud growth, but operational weather forecast models will require more accurate information on spatial variations of soil moisture to produce better forecasts,” says Taylor.
“We are also looking at how predictive models based on deep learning can exploit the new knowledge to provide forecasters with early indications of where storms may appear later in the day,” he reveals.
