Infosheet WP 6 on Forecast Systems
Flood Early Warning - Why?
Flood early warning systems (FEWS) play a vital role in safeguarding lives, minimizing damage, and promoting resilience in flood-prone areas. Their main benefits include:
- Reducing Fatalities: These systems provide timely alerts, allowing people to evacuate to safer areas before a flood hits. Lives can be saved by ensuring that residents receive warnings well in advance.
- Mitigating Economic Losses: By alerting communities and businesses, early warning systems help prevent damage to infrastructure, homes, and crops. This reduces the economic impact of floods.
- Accelerating Post-Flood Recovery: With advance notice, emergency responders can prepare and respond more effectively. Recovery efforts can begin promptly, minimizing long-term consequences.
- Issuing Timely Warnings: Early warnings enable individuals to take protective measures, such as moving belongings to higher ground or securing property. This proactive approach helps prevent loss and damage.
Meteorological Inputs
Good knowledge about the current and future weather situation forms the basis for an FEWS.
Near real-time observations of rainfall, measured through an Automated Weather Station Network provide information of how much water reaches the ground locally. All rainfall information collected by such a weather station network must be automatically transmitted in near real-time by the stations to the local FEWS.
Weather Radars provide information about the spatial variability of rain along with rainfall rates. To do so they measure the reflectivity of water particles in the air, from which rainfall is estimated via so-called Z-R relationships that are provided by the radar manufacturer, but must be calibrated using historical records of rainfall data – ideally provided by the Automated Weather Station Network.
Because raw radar data estimates represent rain in the air, these data can be combined with on-the-ground rainfall observations to be more representative of the water hitting the ground.
Fig.1: Example of rainfall radar data.
Rainfall fields provided by radars, either with or without calibration, provide information about the recent weather. For warning it is important to also obtain information of the future weather. Nowcasting achieves this by projecting the movement of rainfall cells for several hours into the future. Realistically, nowcasts are usually only sufficiently accurate for the use by FEWS for 30 minutes to 1 hour into the future, and less so for thunderstorm (convective) rain.
Outputs from Numerical Weather Prediction Models (NWPM) can be used to gain projection of rainfall up to several days into the future. Rainfall forecasts from NWPMs that cover the entire globe can be available freely, but typically have a lower spatial resolution (often many kilometers) than needed for use in small areas. Regional NWPMs are typically used to adapt the outputs from global NWPMs to the local conditions and at a higher temporal resolution.
Flood Modelling
Flood modeling describes the processes that convert rainfall that hits the ground (or reaches it as snow melts) into simulations or forecasts of water inundating specific locations as result. This is important to understand water levels that may result from impeding heavy rainfall, especially if there is no good knowledge of where flood hotspots are located.
Hydrologic Models, often also referred to as rainfall-runoff models, simulate how much water infiltrates into the ground or runs into streams or canals when it rains. They can also estimate the amounts and timing of water running down streams and canals, albeit in a rather simple fashion. The results of hydrologic modelling can be improved if real-time observations of water levels in streams/canals are available for data assimilation.
Hydraulic Models simulate the movement of water through streams and canals using detailed local information and physical equations. They use as input runoff calculated by hydrologic models and produce as outputs simulations of water level and water velocity. FEWS often employ 1-dimensional hydraulic models, but two-dimensional hydraulic models can be used for flat terrain and urban situations.
Inundation Models calculate the spatial extent of inundation based on the outputs from the hydraulic modeling. This allows for an assessment of which infrastructure, such as buildings, streets, etc. are impacted.
Information Dissemination and Sharing
As the goal of flood early warning is to proactively inform first responders and warn the population it is paramount that the insights gained from the modelling and forecasting are analyzed and if necessary shared and disseminated with the right content to the right people at the right time.
The first step in dissemination FEWS results is to determine of a risk of imminent flooding exists. The needed calculations are typically done automatically based on pre-determined thresholds. Such Event Detection can apply to rainfall, runoff, flows, water levels, or impacted infrastructure and must be determined ahead of time.
Sharing live data, especially moving images from weather radars or forecast models, but also inundation maps, with the emergency management community provides them with better in-sights into the imminent risks.
Providing warnings to the impacted population, either directly or indirectly via collaborating agencies, is the most crucial step in flood early warning. This can be done by the institution operating a FEWS, or by other emergency management authorities. Warnings are often distributed via sirens, text-messages, or lately via social media with the goal of informing the impacted population and stakeholders of the imminent threat and providing guidance on how to react.
Fig.2: Screenshots of the Demonstrator used in CapTain Rain. Different Precipitation datasets are integrated, alarm thresholds can be set by the users.