Pluvial flash floods in urban areas are becoming increasingly frequent due to climate change and human actions, negatively impacting the life, work, production and infrastructure of a population. Pluvial flooding occurs when intense rainfall overflows the limits of urban drainage and water accumulation causes hazardous flash floods. Although flash floods are hard to predict given their rapid formation, Early Warning Systems (EWS) are used to minimize casualties. We performed a systematic review to define the basic structure of an EWS for rain flash floods. The structure of the review is as follows: first, Section 2 describes the most important factors that affect the intensity of pluvial flash floods during rainfall events. Section 3 defines the key elements and actors involved in an effective EWS. Section 4 reviews different EWS architectures for pluvial flash floods implemented worldwide. It was identified that the reviewed projects did not follow guidelines to design early warning systems, neglecting important aspects that must be taken into account in their implementation. Therefore, this manuscript proposes a basic structure for an effective EWS for pluvial flash floods that guarantees the forecasting process and alerts dissemination during rainfall events.

Prolonged and high intensity rainfall often saturates urban drainage systems and generates urban pluvial flooding, resulting in hazardous flash floods. The city most affected by urban flooding in Colombia (South America) is Barranquilla since lacks a proper storm water drainage system. Heavy rainfall produces flash floods to quickly become torrential streams that flow down the streets endangering pedestrians. This research describes a low-cost early warning system (EWS) to detect in real time the hazard level of a stream in an ungauged basin. The EWS indicates whether it is safe or not for pedestrians to cross a flooded street, based on certain criteria used to assess the hazard level of the torrent. A hydrological and hydraulic model calculates the flow, velocity and water level in all cross sections along the stream. The model uses only real-time measurements of rain gauges and topographic survey data to determine the hazard level. Finally, a wireless sensor network sends the alert to a web platform and warns the community in real time.

No abstract

Contents 1 Introduction 1 2 Experimental setup 4 3 Method 5 4 Experimental results and discussion 6 5 Conclusions 12

No abstract