Flood watch is a tool which aims to provide up-to-date information concerning flooding hazard to the general public. It takes in a diverse source of data, including terrain relief, river height sensor, and historic flood risks, and congregates these sources into a single flood danger metric which is then overlaid onto the map. The purpose of this tool is to provide any interested party with the information to make decisions in the event of an emergency.
This repository contains both the front-end android application,
located in the android-dev folder, and the back-end tools for
analysis, which includes web-crawler inside the StationDataCrawler
folder, and the image analysis tools inside the
AnalysisBackend.zip file. Some preliminary results are provided
in the PhotoResultsSubmission folder.
The source code for the Android application is located in this folder. It can be built using the usual tools, and details are provided in a separate readme. Some screencaps of the application are also provided inside the folder. Eventually, the purpose of the app would be to fetch our calculated flood risk metric as a contour map for the user’s geographical location.
This is a set of Python tools which obtains the list of height
sensors for in rivers in Queensland, and proceeds down the list,
fetching the most recent readings from these sensors and storing
the results in a csv file. Further usage details are located
inside the read_me file in the folder.
This is the main tool used to analyse information from the separate sources and provide the output for the Android application. Right now it exists as several functions which must be chained together by hand, but eventually the process will be entirely automated.
The basic idea can be outlined below:
- We fetch the elevation data of a local region from
ELVIS. - The river sensors are placed over the elevation data. For each sensor, if the water height rises above the danger height we can calculated the flooded regions or potentially flooded regions (FAZ) using the elevation data near the sensor.
- The second step is fetching the Water observation from space data, which gives the number of occasions water was detected for regions. This data is smeared over the entire map to provide a probability for each region to be flooded.
- The FAZ and historical flood probability data is then combined to give a general risk metric. To do this, the flood probability data is binned into three categories of high, medium and low risk. Then combined with the binary FAZ data for each point on the map we have a 6 degree scale describing the current flooding risk.
- This map of the general flooding risk can updated as frequently as the river sensor data are to provide a fully dynamic picture of flooding risk. The elevation and Wofs data can be updated at less regular intervals when necessary.
- The Android application simply provides a user friendly feature to access this map, in addition to other useful functions.