Second Global Workshop on Digital Soil Mapping

Rio de Janeiro, Brazil

Scientists from five continents converged 4-7 July 2006 in Rio de Janeiro, Brazil, to participate in the Second Global Workshop on Digital Soil Mapping, which focused on regions and countries with sparse soil data infrastructures. Following a stunning field trip that transected the range in soils, landscapes, and land-uses of the Rio De Janeiro state, the 3-day workshop consisted of five sessions on various aspects on DSM, each with a keynote address, several scientific presentations, and discussion.

The workshop began with Dr. Philippe Lagacherie, INRA-Montpellier, France, reviewing the state-of-the-art of digital soil mapping (DSM). Lagacherie proposed that a significant challenge for DSM is the effective integration of DSM into existing soil survey programs and data. The session addressing the DSM challenges of dealing with limited spatial data structures was launched by Dr. Andy Jarvis, CIAT, CGIAR, Colombia. Jarvis emphasized there is a great potential for global digital soil information, especially if DSM can provide soil property data at spatial resolutions useful for decision-making in less developed countries. The DSM protocol, quality, availability, and capacity building session included the keynote by Dr. Robert MacMillan, LandMapper Environmental Solutions, Inc., Canada, illustrating experiences with applied DSM, noting that accuracy assessment was becoming increasingly important. Dr. Budiman Minasny, The University of Sydney, Australia, presented the keynote lecture on new DSM methodologies, reviewing technological advances in hardware (e.g., remote sensing) and software (e.g., data models that infer mechanisms vs. black box approaches that mine data but do not infer mechanisms). Minasny suggested a hybrid approach to DSM, where soil knowledge drives both data mining and modeling approaches. The session on examples of DSM to predict soil properties began with the keynote by Dr. A-Xing Zhu, University of Wisconsin, USA, improving the efficiency of field mapping by using a similarity index to represent soil distribution in transitional areas. Dr. Thomas Mayr, Cranfield University, U.K., commenced the session on examples of DSM to predict soil classes with a keynote addressing the difficulties and potentials of using legacy soil survey data for modeling soil distribution in new areas. Mayr stressed the importance of data preparation, identifying missing data, and understanding the limitations of model feature space.

The diversity of thought-provoking presentations in each session sparked constructive debate. Final discussion elucidated specific priorities for the development and implementation of DSM, including 1) data (DSM should be data-driven); 2) training (train students as well as established soil surveyors); 3) standards and standardization (document DSM methods and measures of uncertainty); 4) international collaboration (IUSS working group can facilitate this, but should work towards funding international projects); and 5) marketing (work to develop interest and investment in DSM at home and abroad).

The organizing committee was co-chaired by Dr. Maria de Lourdes “Lou” Mendonça-Santos of EMBRAPA Solos, Brazil, and Dr. Alex McBratney of The University of Sydney, Australia. Lou and her colleagues at EMBRAPA and CPRM were gracious hosts of this stimulating meeting in the dramatic setting of Rio de Janeiro.

More details see www.digitalsoilmapping.org

Janis L. Boettinger

Utah State University

USA