Home ITCResearchPhD projectsNear-Real-Time Rainfall Detection and Estimation from Commercial Microwave Links and Meteosat Second Generation Data

Near-Real-Time Rainfall Detection and Estimation from Commercial Microwave Links and Meteosat Second Generation Data

What's the Master's Geo-information Science and Earth Observation about?
Student:Kwabena Kingsley Kumah
Timeline:March 2016 - 30 September 2020

Agriculture, meteorology, hydrology, climate studies etc., all presents a pressing need for accurate space-time information on rainfall intensities. For instance, in order to maximize yields in a rain-fed agricultural system, accurate rainfall forecasts (notably on the onset and severity of rainy seasons) are extremely important. Likewise, in flash flood prediction and forecast, near real time rainfall estimates are key input in hydrologic models.

Prevailing techniques for rainfall measurement are however limited due to drawbacks that can be  natural to each measurement technique or even specific to a geographic region. Practically, rain gauges presents a more physical method to collect and measure rainfall. From a point scale, rain gauge estimate can be very accurate especially for low to intermediate rainfall events. But to estimate accurate rainfall fields from rain gauges, it requires a dense network of rain gauges, which for most developing countries, can be a challenge. Remote techniques like weather radars and satellites can be an ideal way to observe rainfall over large spatial domain. However, whereas weather radars are mostly absent in most (developing) countries on the grounds of acquisition, operational and maintenance cost, satellite based rainfall estimates can be full of uncertainties and even still require ground based rainfall data for their evaluation.

A promising new alternative to observe rainfall, especially in gauge scarce regions, is to makes use of data from commercial telecommunication networks. Microwave (MW) signals from station-to-station microwave communication links are scattered and absorbed by rain drops, which leads to signal attenuation. More rainfall thus equals more signal attenuation, and the amount of signal loss can be used to calculate the average rainfall intensity along the microwave transect. Thus from individual MW links of varying lengths, a series of line measurements of space time rainfall intensities can be calculated, which can then be transformed into rainfall fields by interpolating between individual MW links. For developed countries where there is high densities of MW links, the standalone use of MW links to estimate rainfall fields has been shown to give good accuracy when compared with rainfall fields developed from dense network of rain gauges. But, for most developing countries like in Africa where MW link density is comparatively low and rainfall is typically convective, interpolating between individual MW links becomes difficult because of the localised nature of the rainfall phenomena. Meanwhile, geostationary meteorological satellites like Meteosat Second Generation (MSG) observes and records data on the earth’s atmosphere at a high spatiotemporal resolution (3km and 15 minutes). The data from MSG can be analysed to retrieve information on the rapid development and evolution of clouds and the rainfall they produce. With MSG satellite observing and recording data on the earth atmosphere at high spatial and temporal resolution, and ground based MW links of variable lengths and densities and with the possibility of retrieving accurate rain rates, the interest of this study to combine the two platforms for large scale rainfall observation is not trivial. 

The focus of this study is to develop a rainfall intensity estimation and monitoring system for gauge scare regions in Africa, specifically East Africa, Kenya (see Figure 1), using combined information from commercial telecommunication networks and MSG satellite data at spatial and temporal scale corresponding to those of convective rainfall prevalent in the region. It is however worth knowing that, both techniques are of entirely different physical basis and spatial extent for observing the same phenomena (rainfall). The scientific challenge then is; how to develop concepts and practical algorithms that combine information from the two platforms for large scale rainfall observation. To understand and answer specific research questions in this scientific challenge, it is proposed to; 1) study and infer the evolutionary process of convective clouds from MSG satellite observations 2) investigate and understand the local to regional scale dynamics of East Africa rainfall from the perspective of satellite observation and the corresponding rainfall induced MW link signal attenuation characteristics and 3) to investigate the extent or spatial scale at which clouds can be observed and characterized to match the observational scale of ground based MW links. It is anticipated that the combination of the two platforms (MW link and MSG satellite data) for rainfall monitoring in gauge scare regions like Africa can serve as a viable and sustainable alternative to weather radars, which has proven to be too costly to operate in most African countries.

Meet the team

K.K. Kumah (Kwabena Kingsley)
Graduate Student
prof.dr. Z. Su (Bob)
dr. B.H.P. Maathuis (Ben)
Dr. Joost Hoedjes
Research theme
Water Cycle and Climate

Water, food and energy security and environmental safety are key challenges to our societies. Better water resources management requires a fundamental understanding of the water cycle, water climate and water ecosystem interactions and impacts of human activities in the Earth’s climate system. Quantitative earth observation, hydrological modelling and data assimilation provide a powerful combination to quantify hydroclimatic variables for effectively addressing water management issues across the globe.

Show more