The river Thames in London at Vauxhall Bridge; one of the locations visited in recent field tests of our EnviroTracker™ Smart City Dashboard, as part of our project exploring the utility of using social media reports of flooding to enhance model validation and situational awareness when flooding strikes.
We at Ambiental are developing innovative flood forecasting and flood monitoring technologies in order to tackle urban challenges. We are currently conducting a project for London, building decision support systems which combine flood risk insight and Artificial Intelligence (AI) to enable greater resilience in the face of growing environmental challenges.
The bulk of the project involves deploying our proprietary FloodWatch® flood forecasting technology, to give advanced warning of floods in Greater London. Sudden, high intensity storms, resulting in flash flooding can cause major, unexpected impacts on the nation’s capital city in the form of risk to life, damage to property, and severe impacts for transport and critical infrastructure.
Our modelling software can predict flood events hours in advance through continuously processing Met Office rainfall forecast data feeds. This model simulates the entire hydrological system by calculating hydrological flows and then hydraulically modelling flood evolution.
The objective of the 9-month project is to develop a flood data delivery system dashboard. Through the integration of multiple data sources, including real time sensor telemetry and big data, the aim is to provide improved, actionable intelligence to our government stakeholders. Serving this data via a common application interface unlocks the information’s true power. The new EnviroTracker™ smart city dashboard front end application concept is designed to enable city authorities to manage and visualise high resolution flood forecasting footprints.
A screenshot from Ambiental’s FloodWatch Dashboard for Greater London.
The initiative is funded under the UK Space Agency’s (UKSA) Space for Smarter Government Programme (SSGP). Our data driven software application aims to deliver geospatial insights, which in turn, lead to positive benefits for UK Government. As part of its activities, SSGP provides funding for research and development of applications, with the intention to increase uptake of satellite and Earth Observation (EO) data within the public-sector.
During our field tesing we tweeted ‘Is Government prepared for flooding in London?’ With innovative technologies like FloodWatch™ and EnviroTracker™ from Ambiental it is hoped that UK flood risk management capabilities will continue to improve.
Our FloodWatch product is a newly developed flash flood forecasting system, which is currently being used in Malaysia and is now being deployed in London. Our work aims to mitigate the effects of sudden flash flooding, which in the past has severely impacted cities like London with little or no warning. Key stakeholders for the project include Transport for London and the Greater London Authority.
A map showing some of the many Twitter posts made by the @EnviroTrackerPD testing account involved in assessing spatial accuracy, timeliness and natural language processing to extract flood intelligence from social media.
The Environment Agency has estimated that 140,000 people in London are at high risk of flooding, whilst 230,000 are at medium risk. Studies have shown that almost a third of all London Underground stations have significant or high flood risk. As a result, any alerting of imminent flood dangers will improve emergency response, which in turn can reduce danger and potentially lower financial impacts.
Not all mentions of flooding are reports of disaster. This tweet was made at the Green Business Leaders Awards where Ambiental was Highly Commended for our FloodFutures™ project. The EnviroTracker™ dashboard uses AI to differentiate whether social media posts are reports of flooding or just mentions of the word flood in a different context.
This exciting initiative also explores the value in using big data, specifically social media, to inform decision-making around flood incident management. It is notoriously difficult to predict rainfall generated (pluvial) flooding accurately because it can occur anywhere. Artificial Intelligence (AI) can deliver improvements by providing ordered reports of flooding which serve as model validation and support actionable intelligence workflows. A key component of the project will be to assess techniques for interpreting Twitter posts which mention floods. This will provide a form of validation for improving the skill of subsequent flood model iterations. Furthermore, the project tracks land-use alterations, specifically changes in green space. This data can be fed back into the dynamic modelling environment and used to maintain the accuracy and currency of the flood forecasts produced.
Ambiental’s article in Septembers GeoConnexion Magazine discussed how we used our flood modelling tools for ‘Analysing footprints in the flood’.
About the Author
Paul Drury is the GIS Data Manager at Ambiental. His role includes project management of production operations and reporting back to stakeholders. He also oversees the preparation, integration and quality assurance checking of data assets. Paul is an expert in GIS and data analysis with a developed understanding of the environmental data industry and underlying technical concepts. He has a BSc (hons) in Environmental Sciences from the University of Brighton.