Exploring the Future Potential of AI-Enabled Smartphone Processors
Open Soil Profiles - testbed data portal for storing soil profile data
1. Open Soil Proles
Tomislav Hengl
ISRIC World Soil Information, Wageningen University
GSM2011.org, June 2024th 2011
2. GSIF components
1. Cyber infrastructure for input, analysis and visualization of
data.
2. Global databases (legacy data, gridded covariates) that
are main inputs to global soil mapping.
3. Software tools (modules and packages) and manuals for
creation of geoinformation, for instance, according to the
GlobalSoilMap.net specications.
4. Standards and protocols for data entry, map generation and
data sharing.
GSM2011.org, June 2024th 2011
3. Hypothesis
Probably the most valuable legacy soil survey data
are the eld records of soil properties
i.e. soil prole observations
GSM2011.org, June 2024th 2011
4. Denition
Soil eld record1: any georeferenced observation
or measurement of soil properties
(anything measured in soil medium)
1
Compare with e.g. http://www.geosamples.org
GSM2011.org, June 2024th 2011
6. Proposed implementation
1. Build an open (real time Public access) data portal that
can be used to collect and maintain world soil eld
observations
GSM2011.org, June 2024th 2011
7. Proposed implementation
1. Build an open (real time Public access) data portal that
can be used to collect and maintain world soil eld
observations
2. Use it to facilitate global soil mapping activities
GSM2011.org, June 2024th 2011
8. Proposed implementation
1. Build an open (real time Public access) data portal that
can be used to collect and maintain world soil eld
observations
2. Use it to facilitate global soil mapping activities
3. Collect millions of points. . .
GSM2011.org, June 2024th 2011
9. Basic principles
1. Soil surveyors / contributors remain owners of the data
(read/write rights)
GSM2011.org, June 2024th 2011
10. Basic principles
1. Soil surveyors / contributors remain owners of the data
(read/write rights)
2. Anyone can contribute
GSM2011.org, June 2024th 2011
11. Basic principles
1. Soil surveyors / contributors remain owners of the data
(read/write rights)
2. Anyone can contribute
3. Anyone can use
GSM2011.org, June 2024th 2011
12. Basic principles
1. Soil surveyors / contributors remain owners of the data
(read/write rights)
2. Anyone can contribute
3. Anyone can use
4. (ISRIC will maintain and improve it)
GSM2011.org, June 2024th 2011
15. There are probably millions of soil records out there
Figure: USDA NCSS Characterization Database, CSIRO National Soil
Archive, ISRIC WISE, SPADE, Iran National soil prole database,
Canadian Soil Information System, and African soil proles.
GSM2011.org, June 2024th 2011
16. Lower Saxony
Only in Lower Saxony
(federal state in northern Germany)
there are about 1.3 millions
of soil proles and cores. 2
2
Rainer et al. (2011): Soil carbon balance in sensitive ecosystems in Europe.
GSM2011.org, June 2024th 2011
17. Endangered species
Legacy soil eld records are being lost as we speak.
There is a considerable
danger of being lost to eective use forever (!)
GSM2011.org, June 2024th 2011
18. Types of soil observations (records)
1. Freeform Soil Records (FFSR)
2. Soil Prole Records in Original Form (SPROF)
3. Harmonized Soil Prole Records (HSPR)
GSM2011.org, June 2024th 2011
19. FFSR
Only a very limited number of basic site columns are
absolutely required.
Freeform soil records do not need to describe a whole soil
prole (0200 cm) nor do they require any pedological or
taxonomic data.
Variable types (columns) are extendible.
Entries are not restricted to standard measurement techniques.
Metadata is attached to each new record and is a part of the
record.
GSM2011.org, June 2024th 2011
20. SPROF
Soil Prole Records in Original Form (SPROF)
are soil prole records that describe complete prole
but are unharmonized
(do not comply with a single international standard)
GSM2011.org, June 2024th 2011
21. HSPR
Are complete (core columns).
Are ready for use (mapping and analysis).
Contain soil variables required to produce maps
according to GlobalSoilMap.net specications.
Do not carry any metadata (metadata is standard).
GSM2011.org, June 2024th 2011
22. Data entry scheme
New entry
Complete
NO YES
profile?
Soil Profile Record NO
in Original Form Harmonized YES
(SPROF) values?
Freeform soil Fill in the Harmonize Harmonized Soil
record missing to referent Profile Record
(FFSR) values method (HSPR)
SOIL
Open soil record REFERENCE Open soil profile
LIBRARY
OSRL0 OSPL0
Open Soil Profiles
(ISRIC hosted enterprise DB)
GSM2011.org, June 2024th 2011
24. Data display / sharing
For convenience of automated use and for consistency, we propose
the following attributes for HSPD:
One row = one record3
Column naming convention
Uncertainty columns
Access via URL or via ODBC (MySQL)
3
Although in the background we work with a relational DB.
GSM2011.org, June 2024th 2011
25. 3D variables
A proposed standard naming convention. E.g. pH estimated in 1:5
water solution for the initial soil depth layer becomes:
variable type estimation model depth sequence
P H I H O 5 _ A
variable type estimation model
T A X N U S D A GSM2011.org, June 2024th 2011
26. variable type estimation model depth sequence
Site observations
P H I H O 5 . A
Site information, such as soil classication by the USDA system can
be coded as:
variable type estimation model
T A X N U S D A
GSM2011.org, June 2024th 2011
27. Summary points
If meteorologists, ecologists have managed to build
global data sets of eld records why can't we?
GSM2011.org, June 2024th 2011
28. Summary points
If meteorologists, ecologists have managed to build
global data sets of eld records why can't we?
If a government agency does not intend to sell their soil
data, why not contribute to OSP (distributed hosting)?
GSM2011.org, June 2024th 2011
29. Summary points
If meteorologists, ecologists have managed to build
global data sets of eld records why can't we?
If a government agency does not intend to sell their soil
data, why not contribute to OSP (distributed hosting)?
Before we can use the OSP for soil mapping, we need to
convert FFSR and SPROF to HSPR
GSM2011.org, June 2024th 2011
30. Summary points
If meteorologists, ecologists have managed to build
global data sets of eld records why can't we?
If a government agency does not intend to sell their soil
data, why not contribute to OSP (distributed hosting)?
Before we can use the OSP for soil mapping, we need to
convert FFSR and SPROF to HSPR
There is an enormous value in legacy soil eld data
(forgotten goldmine)
GSM2011.org, June 2024th 2011
31. Next steps
Prepare entry templates for FFSR, SPROF and HSPR
Focus on National soil prole datasets
Invite people (IUSS, DSM, any soil surveyor) to contribute
Popularize it
GSM2011.org, June 2024th 2011