1 of 8

Operational drought monitoring from satellites

Robin Wilson, Tudor Watson, Peter Foster, Ben Evans

​

http://rhoksoton.github.com/WaterMe/

2 of 8

The Problem

  • Droughts are a serious problem
  • They are likely to get worse with climate change

​

There is a need for timely, open data to allow assessment of drought conditions by:

  • Aid agencies
  • Farmers and local people
  • Governments
  • and more...

​

3 of 8

A solution?

The MODIS satellite produces daily data on light reflected from the Earth's surface

​

This can be processed to produce the Normalised Difference Water Index, which meaures water in plants

​

MODIS data is free, but NDWI is not produced operationally yet - we should change that!

4 of 8

What did we do?

Generating Data

Wrote Python code to acquire data and process to NDWI:

  • Download each day's data from FTP site
  • Reproject to latitude/longitude using NASA tool
  • Mask out areas with cloud, oceans and bad values
  • Add to a MySQL database

Data for West Africa acquired for the majority of May 2012

5 of 8

What did we do?

Visualising Data

Wrote an API and a map-based viewer using that API:

  • API written in node.js
    • latLng - gives data for a specific point
    • squareLatLng - gives data for a region
    • Both can take parameters for data to return, start and end time etc.
  • Map viewer gets data using API, plots on OpenStreetMap data

6 of 8

Demos

7 of 8

Screenshot

8 of 8

Further work

  • Importing can easily be done automatically to keep database up-to-date
  • Large storage space needed though...
  • Improvements to UI
    • Improve performance of webmap
    • SMS access
  • Scientific validation of data - is it correct, is it right location, is it useful?
  • Evangelise to aid agencies, users etc

​

Finally: release and test with real-world users