Presentation on theme: "Some learnings and open questions from the Montpellier DSM workshop First DSM working group meeting University of Miskolc, Hungary, 7-8 April, 2005 P."— Presentation transcript:
Some learnings and open questions from the Montpellier DSM workshop First DSM working group meeting University of Miskolc, Hungary, 7-8 April, 2005 P. Lagacherie Laboratoire détude des interactions Sol – Agrosystème - Hydrosystème
TOPIC 2 PROGRESS, EXAMPLES AND ECONOMICS OF DIGITAL SOIL MAPPING (12 papers) TOPIC 3 SAMPLING METHODS FOR CREATING DIGITAL SOIL MAPS (4 papers) TOPIC 4 REPRESENTATION OF DIGITAL SOIL MAPS (5 papers) TOPIC 5 QUALITY ASSESSMENT OF DIGITAL SOIL MAPS (4 papers) TOPIC 6 NEW ENVIRONMENTAL COVARIATES, & NEW SOURCES OF ENVIRONMENTAL COVARIATES, FOR DIGITAL SOIL MAPPING (8 papers) TOPIC 7 QUANTITATIVE MODELLING FOR PREDICTING SOIL CLASSES AND ATTRIBUTES (23 papers) TOPIC 1 INTRODUCTION & BACKGROUND TO DIGITAL SOIL MAPPING (1 paper) The Workshop topics
Outline DSM: some generic comments New environmental covariates and new sources of environmental covariates for DSM Collecting soil data Building DSM functions Evaluating the quality of digital soil maps Representation of Digital Soil Mapping outputs
DSM: some generic comments and questions DSM is expected to work better than conventional surveys DSM requires capital input and mixed skill teams Which DSM outputs? How to graft DSM on to existing soil database programs?
New environmental covariates and new sources of environmental covariates for DSM Great variety of environmental covariates tested. DEM remains the most common Need to express the mental models of pedogenesis to select the appropriate covariables Preprocessings of covariates (classifications, decomposition of scale factors) How to deal with newly available covariates in the future?
Collecting soil data for DSM Only a few contributions to this topic Sampling methods based on statistical and geostatistical techniques Which criteria for a DSM sampling? How to deal with them? From site sampling to area sampling: is this different ? How to use existing soil data?
Building DSM functions Topic 7 was the most populated Mostly non-spatial approaches Soil surveyor approaches vs pedometric ones How to better associate soil surveyors and pedometrician? How to deal with multiscale soil variations? How to deal with variations in time?
Evaluating the quality of digital soil maps A critical point Need of an unified framework of quality assessment How to validate area soil prediction? How to validate quality according to each user application? How to communicate quality to the users? The quality information must be meaningful to users
Representation of Digital Soil Mapping outputs The traditional 2D choropleth map is insufficient Modern GIS and visualization tools can be used to valorize DSM to the users…. …and to assist us in building DSM functions Which compromise between quantity of displayed information and complexity? How far must we go? Can visualisation tools and web be used to promoteDSM ?
To go further From the research phase to actual production of soil maps and soil data Need of a generic tool : Spatial Soil Inference System ? Lets organize us Structure a research project Select pilot areas that cover the various data configuration within Europe Build a Version 0 of a Digital Soil Mapper Create a common DSM toolbox with easy available functions
A Spatial Soil Inference System ( Lagacherie & McBratney, 2005) DTM RS image X Existing Soil map Scorpan layers Soil observations Spatial Soil Information System DSM Function library Scorpan F. Pedotransfer F Class Content F. Allocation F. User interface User data Predictor OUTPUT Function organiser