Reading a live river dashboard during a flood
Five checks that turn a live river water level dashboard from a screen of numbers into a decision aid when the river is rising.
It is 2 a.m., the rain has not stopped, and you are looking at a live river water level dashboard with thirty stations on it. Which numbers matter, and which can wait? A dashboard is built to be read quickly, but in a flood the reader is tired, the pressure is high and one wrong reading can mean a missed warning or a needless evacuation. This article gives a short routine for reading a water level dashboard at river level monitoring stations, whatever software it runs on.
What is a live river level dashboard for?
A dashboard is, in Stephen Few's definition, "a visual display of the most important information needed to achieve one or more objectives; consolidated and arranged on a single screen so the information can be monitored at a glance" [1]. He adds that it tells you what is happening, not why: diagnosis needs further investigation, and a dashboard is the starting point for it [1].
For a river, the objective is usually narrow: is the water rising towards a level at which people must act, and how long is there to do it? A typical river dashboard therefore shows a map or list of stations, the latest level at each, a short trend chart, and a status against preset marks. Treat it as the place where you decide where to look next, not as the final answer.
How do you read a river water level dashboard? Five checks, in order
1. How old is the reading?
Before trusting any number, find its time stamp and the station's online status. A level shown in bold type that was recorded forty minutes ago is history, not news. During a flood, networks fail first at the stations you most need. A dashboard that shows when each station last reported, and whether it is online, lets you separate "the river is steady" from "the station has gone quiet". Never read a missing update as good news.
2. Where is the level against its marks?
Compare the level with the station's warning level and danger level, the stage values set for that site (our guide to warning level vs danger level explains how they differ). WMO's Manual on flood forecasting and warning says that triggers must be "linked to local conditions and features of risk", such as the level at which water leaves the channel or a low road floods, rather than standard figures applied everywhere [2]. So ask which mark the number is near, and what happens at that mark at this site.
Check also that the level is on the same datum as the marks. Two records that disagree by a constant often differ only in their zero; see gauge zero, datum and benchmarks.
3. How fast is it rising?
The level says where the river is. The rate of rise, the change in level per hour, says how soon it will get to the next mark. WMO lists it as a trigger of its own, with the example of a rise faster than 25 cm per hour [2]. You can work it out from the trend chart by dividing the change in level by the time between two points.
Illustrative example. Suppose a station reports every 15 minutes, with a warning level of 4.50 m:
| Time | Level (m) | Change in 15 min (m) | Rate (m/h) |
|---|---|---|---|
| 14:00 | 3.20 | – | – |
| 14:15 | 3.28 | 0.08 | 0.32 |
| 14:30 | 3.41 | 0.13 | 0.52 |
| 14:45 | 3.57 | 0.16 | 0.64 |
| 15:00 | 3.78 | 0.21 | 0.84 |
The river is 0.72 m below the warning level, and the rate has risen from 0.32 to 0.84 m/h. At the latest rate, the warning level would be reached in 0.72 / 0.84 ≈ 0.86 h, about 50 minutes. That is a rough indication, not a forecast: the rate may keep rising or fall away. But it shows why a shrinking gap to the warning level, together with an accelerating rate, deserves attention long before the level itself looks alarming.
4. What are the stations upstream doing?
A single station cannot tell you what is coming. Look upstream on the same river and on the tributaries, and at rain gauges in the catchment. If the upstream level is rising faster than yours, the peak has not arrived. In small, steep catchments the gap between rain and river response can be short (see flash floods vs river floods). Note that the time a flood takes to travel between two stations is not simply distance divided by water velocity, because the flood wave moves differently from the water. The most reliable guide is the lag between peaks at the same two stations in earlier floods.
5. Can you trust the data?
A flat line for hours, a single spike, or a sudden jump can each be a real event or a sensor or link fault. WMO notes that rapid changes in a value or its trend are often pointers to erroneous data, and that comparison with plots from adjacent stations is a simple and effective check [2]. It also says that estimated values should be flagged so their presence is apparent to the user [2]. If your dashboard marks estimated or suspect readings, read them as such. Our article on quality control for hydromet time series shows the common patterns.
Common ways to misread a dashboard
- Reading the colour, not the number. A green tile can hide a steep rise that has not yet crossed a mark.
- Trusting the latest value alone. A single point has no direction; use the trend.
- Mixing clocks. Confirm that all stations show the same time basis.
- Treating a dashboard as the warning. Warning decisions follow the agreed procedures and the responsible agencies; the dashboard informs them. How warnings are then issued is covered in how a flood early warning system works.
Key points
- A dashboard shows what is happening at a glance; it does not explain why [1]. Use it to decide where to look next.
- Check the age of each reading and the station's status before the value.
- Read the level against site-specific marks, and the rate of rise against the gap remaining.
- Use upstream stations and rainfall to judge whether the peak has arrived.
- Treat flat lines, spikes and flagged estimates with caution, and compare with neighbouring stations.
Where ClimMaTech fits
Visiflow (Device Cloud) is the cloud platform for ClimMaTech devices. According to its page, it provides a device list and map showing which devices are online, live readings such as water level and surface velocity with signal strength and battery status, charts of past readings, alerts when readings cross a threshold, and shareable reports; see Visiflow (Device Cloud). For teams that issue community warnings, the Rakshak Control Centre is available for demonstration.
References
- Stephen Few, Dashboard Confusion, Perceptual Edge, 20 March 2004. perceptualedge.com
- World Meteorological Organization, Manual on Flood Forecasting and Warning, WMO-No. 1072, 2011 edition: sections 6.2 (data validation, missing data) and 8.4 (selection of warning stages). old.wmo.int PDF