GSIF - Our Approach for Generating Open Soil Data

Share on:

The Global Soil Information Facilities (GSIF) is ISRIC's approach for production of open soil data The approach aims to collate soil data (both legacy/historic  data and new soil data) and upon their standardisation, use the standardised data for the production of maps at various spatial resolutions. The GIS layers provide the basis for modelling, spatial planning and decision making.

Key Principles

The following six key principles are the pillars of GSIF:

  1. Crowdsourcing: [RvdB1] Data collection: Everyone collecting soil data or working with soil information is invited to contribute point data, spatial data and / or modelling algorithms to the GSIF effort.  All point data sets are maintained ‘as is’ in the DATA repository together with their metadata, in particular the licence. Data licence: All provided soil data remain property of the original contributors, unless otherwise indicated by the contributor. The contributor  must indicate how the data they provided to ISRIC may be used processed and distributed. Datasets with at least a CC BY  or CC BY-NC Creative Commons licence will be standardised using the regular WoSIS workflow prior to their distribution (with the original licence). Alternatively, some datasets may only be used to make derivative predictions and visualisations ((SoilGrids-products); such products  will be made available with an open data licence in accordance with the ISRIC Data Policy.
  2. Open Source Software: GSIF products are mainly based on Free and Open Source Software[NB2]  (Linux, PHP, LaTeX, R and contributed R packages, RStudio, GDAL, GRASS, SAGA GIS, PostgreSQL, Geoserver, PostGIS, Python, Google Earth and similar), making them independent of commercial software packages.[RvdB3] 
  3. International standards: GSIF has been designed to serve global soil mapping initiatives and supports FAIR guidelines (Findable, Accessible, Interoperable, and Reusable) for data sharing. We closely follow internationally accepted standards e.g. International System of Units, international soil classifications systems, FAO soil field description guides, SoilML guidelines, Open Geospatial Consortium standards, World Geodetic System [RvdB4] 1984[NB5] , and similar.
  4. Reproducible research: The SoilGrids component of GSIF[NB6]  is based on automated procedures for mapping, pattern recognition and report/plots generation[RvdB7] . Derived maps can be updated by re-running the scripts with limited human intervention once new data sets become available.
  5. User-driven design: GSIF [NB8] is user-driven. Data processing services and databases (maps and reports), produced as a part of GSIF, are constantly adjusted based on usage statistics and web-traffic to user needs. [RvdB9] 

ISRIC runs an annual Spring School [RvdB10] to build a user network to stimulate use of the facilities, obtain feedback for improvements and further develop GSIF components through collaboration.

GSIF components

GSIF comprises a number of components that are interconnected, but can be considered to be independent procedures:

  • WoSIS (World Soil Information Service): a web service for serving a selection of standardised  soil profile data.
  • WorldGrids geo-data service
  • R-packages: GSIF (automated analysis and mapping functionality) and plotKML (visualization of soil field data and output maps);
  • SoilGrids: a collection of predicted global soil property and class maps (currently at resolution of 250 m) produced using automated mapping;
  • SoilInfo App: a smart phone App to access[NB11]  all global soil information on your mobile device.

General GSIF data generation scheme

Issues with this graph that need to be addressed before release:

  • Do not mention INSII on top, too sensitive. Of course we want them to be there, but this is not the case yet.
  • Left top box: i would mention: soil profile data contributed by third parties
  • Top middle boxes: is this reality[NB12] ?
  • Top middle right box: do we need to mention soter explicitely[NB13] ?
  • Middle left and middle middle: do the shaded boxes add information[NB14] ? If so, are these the right words?
  • Lower: why 2 boxes on this level and not just 1: GSIF toolbox[NB15] 
  • Where is the app?
  • Clarity: all earlier mentioned GSIF components should be in the graphs as one box, than it is in line with the text. So this picture needs to be discussed with colleagues and redone professionally, because it is a leading picture for many things we do[NB16] .


Who is it for?

GSIF aims at serving international agricultural development and research organizations, such as FAO, UNEP, UNCCD, CGIAR's, USAID, World Bank and similar. We also aim at serving National Environmental Agencies, National soil survey agencies, private entities, , soil scientists, farmers, agriculture engineers and other interested parties.

 [RvdB1]Is this the right terminology for data submissions by users? I associate this term with laypeople doing stuff, but up to date we mainly rely on specialist that have quality data, certainly not ‘ the crowd’ 

 [NB2]Check numbering.

 [RvdB3]Do not see the need of this statement. Who can convince me?

 [RvdB4]I do not know this. Important to mention?

 [NB5]Yes, GIS standard

 [NB6]Only refers to the SoilGridis part, not the WoSIS work.

 [RvdB7]Reality or future music? We should only talk about reality here, or explicate mention we are talking about future developments

 [NB8]SoilGrids products

 [RvdB9]Future music?

 [RvdB10]Links on the bottom of the text, as agreed for all text.

 [NB11]And upload?

 [NB12]No(t yet)

 [NB13]Probably not

 [NB14]Not really

 [NB15]Yes, better to reduce complexity in the figure

 [NB16]Agree; there have been many discussions about this.