Ramani Huria and The World Bank are trying to understand the best way to calculate elevation in Dar es Salaam so that elevation measurements can be integrated into current flood models. Elevation will add a 3D dimension to flood maps which will improve understanding of water flow and help with flood projections. Measuring elevation requires a series of complicated measurements.
At the end of April 2018, three Civil Engineering students from Delft University of Technology in the Netherlands – Huck, Detmar and Martijn – arrived in Dar es Salaam to build cheap and practical devices to measure elevation. They will be spending two months in Dar es Salaam to work with Ramani Huria, a community mapping project. They believe the RH team is a very motivated team of young people that can make community mapping a real and usable source of open data, and they have realized how important and difficult it is to have an accurate and up to date map of such a fast growing city.
Their project objective is to generate elevation readings which are as accurate as possible using temperature and barometric pressure. They will do so using an arduino combined with a Bosch BMP280 inside a watertight box. The goal is to cancel out the noise and errors from the sensors to get an accurate and usable elevation reading to implement and integrate into OpenStreetMap.
This is the first time that they have worked with the BMP sensors so it is a new experience for them to trial how these sensors react within an urban environment. If they succeed in doing this, it will be possible to create an accurate elevation model of a big city using very cheap utensils.
The ultimate goal of Ramani Huria is to accurately be able to measure elevation so as to produce the city flood model using cheap methods and being able to replicate this process in other cities of Africa if they’re need it.
Workshop Breakout Session
From experiments so far, many of the environmental noises such as sun, wind and rain have been inspected and can now be taken into account. The absolute error of each sensor is now known and can be calibrated to one point, which has to happen every time a measurement is taken. More time will be needed to fully understand the relative error, but there are now even more filters that have been discovered to minimize this error. In the upcoming weeks the Ramani Huria team, led by the students from Delft, will begin field experiments using an android app which is able to read the arduino and sensor. The app already consists of a IIR filter (infinite impulse response) which filters the measurements to make the data more workable. The next task will be a server that collects all the data of those phones to a platform so these can be used easily and can be looked into at any time of the measurement. Every day, the potential for cheap, locally made devices to determine accurate elevation levels increases. These measurements will contribute to more accurate flood models and mitigation plans.