Training
Multiple-Point Statistics simulations (MPS) with Isatis.neo
Objectives
This course introduces you to Multiple-Point Statistics (MPS), a powerful simulation technique for modeling complex spatial variability using training images. Developed in collaboration with the University of Neuchâtel, the course combines theoretical foundations with hands-on practice using Isatis.neo and its integrated DeeSse engine.
You will learn how to select suitable training images, prepare conditioning data, and generate realistic categorical and continuous subsurface models. The course is particularly relevant to applications in mining, hydrogeology, remote sensing, and reservoir modeling. MPS enables you to assess uncertainty and reproduce complex geological features, such as channelized permeability, stratigraphic structures, and ore grades in vein deposits.
Target audience
Professionals and researchers involved in spatial modeling who want to enhance their ability to simulate complex geological structures and facies distributions using Multiple-Point Statistics.
- Geologists and geomodellers
Professionals working in mining, oil and gas, or hydrogeology who need to model complex geological patterns, such as channels, fractures, veins, or stratigraphic structures, that are difficult to reproduce using traditional variogram-based approaches. - Reservoir engineers
Professionals focused on building realistic facies or property models that improve reservoir characterization and flow simulations. - Environmental and hydrogeological scientists
Specialists who need to simulate spatial heterogeneity in aquifers and environmental systems while preserving geological realism. - Geostatisticians and data scientists
Professionals looking to deepen their knowledge of MPS and apply advanced simulation techniques using training images and geological analogs. - Consultants and technical advisors
Experts supporting clients with subsurface modeling projects who want to integrate innovative geostatistical techniques into their workflows. - Researchers and academics
Researchers involved in spatial data analysis, stochastic simulation, or geoscientific modeling who want to explore practical MPS workflows.
Content
General introduction
– Introduction to the geostatistical approach.
– Understanding conditioning data and training images.
– General principles of Multiple-Point Statistics.
– Introduction to the Direct Sampling algorithm.
Hands-on exercises:
– Introduction to Isatis.neo fundamentals.
– First application of DeeSse to stationary categorical and continuous cases.
From stationary to non-stationary simulations
– Understand the main DeeSse parameters.
– Learn why training images are needed and how to obtain or generate them.
– Identify the essential properties of a suitable training image.
– Handle non-stationarity within the simulation grid.
– Explore multivariate simulations.
Hands-on exercises:
– Complete a practical case study based on the Areuse delta.
– Generate a training image and an orientation trend to control simulations.
– Perform the joint simulation of two variables.
Applying MPS to real-world data
– Handle non-stationarity using analog data.
– Use secondary attributes to improve simulation results.
– Explore examples involving climate data, an Australian bauxite mine, bedrock topography, and geophysical data.
– Apply the Direct Sampling technique to time-series simulations.
Hands-on exercises:
– Complete a two-dimensional case study using secondary variables from the Herten fluvioglacial aquifer.
– Fill gaps in satellite images using multivariate and multi-temporal techniques.
Modeling with elementary training images
– Understand elementary training images and invariances.
– Explore an application example from a mining site in South Africa.
– Discover multi-scale simulations based on Gaussian pyramids.
Hands-on exercises
– Complete simple applications using elementary training images and invariances.
– Explore Gaussian pyramids.
– Build a first two-dimensional fluvioglacial facies model using the Herten aquifer case study.
Overview of advanced MPS methods
– Cross-validation.
– Multi-scale simulations on unstructured grids.
– Inequality and block conditioning.
– Connectivity conditioning.
Complementary hands-on exercises: modeling a fluvioglacial deposit
– Build elementary training images.
– Get introduced to Python programming for task automation.
– Build a stratigraphic model.
– Model a fluvioglacial aquifer from borehole data.
What the course includes
- Balanced learning approach. Theory paired with practical applications, so you understand and apply the concepts.
- Hands-on software training. Computer exercises in Isatis.neo, on real datasets.
- Personalized feedback. Individual guidance from experienced trainers throughout the online sessions.
- Comprehensive resources. A temporary software license, course documentation, journal files, and datasets to keep and reuse after the course.
- Certificate of completion issued at the end of the session.
Prerequisites
None. A theoretical understanding of geostatistical approaches is an advantage. This will help participants make the most of the course.
The trainer
Roberto Metzingen Rolo, PhD
Senior Consultant, Geostatistics & Data Science
An expert in turning geological data into reliable resource models, Roberto helps mining teams develop advanced estimation and simulation solutions that meet international standards. Specialized in Python and machine learning applied to geoscience, he designs tailored approaches for the complex challenges of mining projects.
Register, or ask us first
This form handles all three: registration, a quote for yourself or your team, and questions about the content or the prerequisites.
Address
44 Avenue de Valvins, 77210 Avon, France