Desertification indicators service

This service will support monitoring required by the United Nations Convention to Combat Desertification

The following presentation on this service was given at the MELODIES: Exploiting Open Data conference in October 2016. Slides are available here.

Service Q&A

Who are the target users of this service?

Entities involved in reporting to the United Nations Convention to Combat Desertification on a periodical basis. Also, land managers monitoring land restoration projects, and organisations dealing with biodiversity (e.g. IUCN - International Union for Conservation of Nature).

How will these users access this service?

Via a webGIS portal for readily available applications. Else, re-directed from the portal for bespoke applications.

What products does this service provide?

  1. A Land Use/Land cover map with 30m resolution and an 11 class nomenclature
  2. A Land Use/Land cover changes map showing the differences between user-defined years
  3. A Land Condition maps - states of ecological maturity, including land degradation; temporal trends of Net Primary Productivity corrected for inter-annual climate variability; interpretation and validation, statistically-based, report. Full climate gridded archive if not provided by the user.
  4. A Desertification Susceptibility Indicator map - this dataset classifies areas according to their estimated Susceptibility to Desertification

How will these products benefit users?

These products will provide the users with the information required to act on management and political decisions as well as for reporting obligations to the United Nations Convention to Combat Desertification. Also for detection of hot spots for land degradation and green spots for ecological restoration.

Which Open Data sources drive this service?

  1. Landsat imagery
  2. ECMWF climate data
  3. SPOT VEGETATION S10 archive
  5. Global Surface Summary of Day
  6. LADA Global Land Use Maps
  7. ESA GlobCover Version 2.3
  8. PROBA-V 300 m syntheses (older than 1 month) - S10 TOC - 300 m 

What processing is performed on this data?

For the Land Cover subservice, the imagery is processed by a set of preparative algorithms for cloud removal, radiometric and geometric corrections, value normalisations and vegetation index generation. The prepared images are then run through a classification algorithm in order to attain a per-pixel classification of Land Use/Land Cover. Lastly, the results go through generalisation and cleaning procedures in order to retrieve the final products.

For the Susceptibility to Desertification subservice, drought indicators are derived from long term climatic reanalysis data over time. The mean and variance time trends reveal the level and variability of the drought in each location, which is then downscaled using vegetation and soil radiometric indexes by using a geostatistical co-estimation method.

For the Land Condition subservice, vegetation density products (archived time-series) are downloaded from the sources. Corresponding climate archives are interpolated at the same spatial and temporal resolution from georeferenced station summaries. Cloud computing includes: Net Primary Productivity normalized by precipitation and corrected for aridity zones, and stepwise regression to determine temporal trends. Validation is based on independent measures of Soil Organic Carbon, and on statistical tests, respectively. Visualization tools are then deployed to facilitate ingestion by the user.

How does this service use Linked Open Data?

One of the possible ways we plan to use Linked Open Data is for querying the Land Use/Land Cover product for different years in order to retrieve the Land Use/Land Cover changes map in a much more dynamic, flexible and faster fashion. Similarly, to retrieve land condition states and trends for relevant spatial delimitations such as administrative units or, as done recently for IUCN, eco-region terrestrial ecosystems.

How Open Data has improved this service

The open data sources listed above are at the heart of this service. The service uses Open Data because users are often in developing countries with low economical resources for ad-hoc projects. Also, OD favours standardization of results.

How the Shared Platform has improved this service

Having out services on the cloud platform has helped to improve the architecture of the services as well as their usability. The increased efficiency of processing due to the compute power of the cloud platform is expected to bring great improvements in the time taken to generate each product. In addition, cloud computing is required to upgrade the service from regional to global applications.

Our biggest challenges so far...

We have faced numerous challenges in constructing these services so far. The top three are:

  1. Understanding the new technologies available to us - particularly in the Linked Open Data field - and how our service can benefit from them.
  2. The technical challenge of converting our raster datasets into Linked Open Data.
  3. Automating our workflow to achieve some bespoke user interaction capabilities.