Accurately measuring meteorological visibility is an important factor in road, sea, rail, and air transportation safety, especially under visibility-reducing weather events. This paper deals with the application of Machine Learning methods to estimate meteorological visibility in dusty conditions, from the power levels of commercial microwave links and weather data including temperature, dew point, wind speed, wind direction, and atmospheric pressure. Three well-known Machine Learning methods are investigated: Decision Trees, Random Forest, and Support Vector Machines. The correlation coefficient and the mean square error, between the visibility distances estimated by Machine Learning methods and those provided by Burkina Faso weather services are computed. Except for the SVM method, all the other methods give a correlation coefficient greater than 0.90. The Random Forest method presents the best result both in terms of correlation coefficient (0.97) and means square error (0.60). For this last method, the best variables that explain the model are selected by evaluating the weight of each variable in the model. The best performance is obtained by considering the attenuation of the microwave signal and the dew point.
<p>Commercial Microwave Link (CML) data can provide important rainfall information, in particular in regions with low density of rain gauges and with no radar coverage. We have set up and operate a CML data acquisition (DAQ) system for Burkina Faso and report on the first larger scale analysis of the derived rainfall information.</p><p>Our real-time DAQ system started as a pilot project covering only eight CMLs and was gradually extended. For the monsoon season 2020 and 2021 it collected data for more than 1000 CMLs in Burkina Faso with a temporal resolution of one minute. Our first analysis is focusing on the 300 CMLs which operate in the frequency range between 11 GHz and 13 GHz in and around the city of Ouagadougou, the capital of Burkina Faso. We carry out a comparison with official daily rain gauge data, both for individual CMLs as well as for CML-derived rainfall maps. Our results for the period of the 2019, 2020 and 2021 rainy season indicate good performance of the CML rainfall information, with a Pearson correlation coefficient of 0.8 and higher.&#160;</p><p>The processing of the longer CMLs in the frequency range between 7 GHz and 9 GHz, which connect the urban centers in Burkina Faso, currently is in progress. To tackle the challenge of noisy dry periods we are investigating the use of cloud cover and cloud type information from MSG SEVIRI data.</p>
Several factors can attenuate radio signal between transmitting and receiving antenna. One can cite: vegetation, atmospheric gases, fog, water vapor, transmission instruments, rain, temperature, etc... The sources of attenuation differ according to the climate and the relief of each continent or even each country. In this work we aim to show that there is link between microwave signal attenuation and weather visibility in the presence of dust. Weather visibility is a very important factor for the safety of road, sea, rail and air transportation. In the presence of dust, the visibility is strongly reduced and there is also a strong attenuation of the microwave signal propagating between two antennas. By performing a linear regression on the attenuation-visibility scatter plot, we propose a method for real-time estimation of the visibility knowing the microwave signals attenuation. A correlation measurement between the visibility estimated by our method from the real attenuation data of the mobile phone operator Telecel Faso SA (Burkina Faso) and the visibility measured by the National Meteorological Agency of Burkina Faso (ANAM) gave a correlation coefficient of 0.86.
<p>Many studies have already shown that attenuation data from commercial microwave link (CML) networks can be used to derive rainfall information, also on a country-wide scale. Particularly in regions with coarse station networks and without radar coverage, CMLs provide an attractive solution to increase the spatial and temporal coverage of rainfall observations. There are, however, several challenges that we face when transferring the successful applications from Europe to developing countries. In this contribution we present recent results from dense CMLs networks in two African cities, discuss the challenges that we are facing when trying to expand CML rainfall estimation, and present potential solutions to tackle these challenges.</p> <p>We show rainfall maps with temporal resolution of 15-minutes derived from CML networks in the city of Ouagadougou (Burkina Faso) and the city of Lusaka (Zambia). There is only one rain gauge for comparison in each city, which limits the options for validation. However, comparison of the CML-derived rainfall maps with the gauges shows good agreement. These results clearly show the large potential of the dense CML networks in African cities for rainfall observation.&#160;</p> <p>Country-wide rainfall estimation based on CML data in developing countries can not always be done in the same manner, as e.g. in Germany. Based on our experience, a large number of CMLs in developing countries are long 7-GHz CMLs. At these frequencies the path attenuation is less sensitive to rainfall and the long CMLs seem more prone to fluctuations during dry periods. This makes the data processing more challenging. We suggest that a combination of CML data processing with data from geostationary satellites is considered a basic requirement and not only an option for further improvement. While this combination is methodologically feasible, it implies large organizational efforts. Either large amounts of satellite data have to be moved to the individual institutions that do CML data processing, or CML data, which is hard to get access to, has to be transferred to an institution that has direct access to the satellite data.</p> <p>To be able to bring rainfall estimation from a combination of CML and geostationary satellite data to an operational level, simplified access to CML data and concerted processing is required. We do not suggest a final solution, but we present ideas to initiate a discussion that should pave the way towards making operational usage of CML data in developing countries a reality.</p>
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.