AquaConSoil 2023

A significant effort to redevelop brownfield sites has been underway in France for several years to fight against soil artificialisation and scarcity of building surfaces. It is estimated that between 90 000 and 150 000 hectares are occupied by industrial brownfields in France, which constitutes a considerable reservoir of land that can be used for construction without destroying natural areas. The regeneration of brownfields is thus of great economic, social, and environmental interest to the country.

Redevelopment of former industrial sites requires, nevertheless, to account for soil pollution to check that it is compatible with the intended use and to take the necessary measures to ensure compliance with regulations. Soil can be evacuated and/or remediated to reduce the risks of exposure of future users or to implement building foundations. But this has a cost: the higher the number and concentration levels of pollutants, the more expensive it will be to recycle, reuse or treat the soil.

The question raised is thus: how can we optimize a site development plan by reducing both the health risks and the costs of cleaning up polluted soil? The Cartorisk method and code were developed for this purpose. The concepts and the computation tool were created by four partners – eOde, Element-Terre, Geovariances, and Ginger Burgeap, thanks to the financial support of Ademe, from 2016 to mid-2023, in 5 major steps:
1. Design of the method
2. Computer development (R package)
3. Tests on two real case studies, with the following preliminary work:
– Analysis of health risks (exposure routes, main pollutants, etc.)
– Identification of risk indicator substances
– Geostatistical modeling of concentrations of risk indicator substances
4. Consultation of stakeholders of brownfields regeneration and soil remediation on the results produced:
– Through a one-day workshop
– Through a questionnaire on the way the health risk maps should be represented for clarity
5. Adjustment of the method and the code according to the feedback of these professionals.

The Cartorisk approach allows producing two types of maps and results depending on the progress of the conversion project:
− Opportunity maps to define the possible land use with the minimum of health risks and soil excavation and/or clean-up,
− Risk maps for a specific development plan and associated excavation and/or soil remediation costs.

The code is used to apply analytical equations of soil-gas transfer and exposition through soil ingestion, ingestion of plants, and inhalation of indoor air to the cell units of the 3D geostatistical soil contamination model and compute probabilities that the hazard quotient and individual excess risk exceed the regulatory thresholds. Health risks can be mapped in 2D at different levels of the future buildings (underground, ground floor, floors). Unit costs are applied to the geostatistical model of soil pollution in the areas to be excavated and/or remediated to estimate the total costs while meeting desired risk levels.

Input data are composed of a set of conditional simulations of pollutant grades, a table of the parameter values used in the transfer and exposition equations, and the unit costs applied according to the soil pollution levels (several reuse and clean-up options).

The results produced by the code are:
− Health risk maps (image and csv files),
− Excavation and/or soil remediation costs, represented as graphics and tables of statistics, exportable as image and csv files.

To conclude, Cartorisk is a flexible approach that may help decision-making to define a regeneration plan for brownfields. It allows testing and comparing different redevelopment scenarios, assessing their impact on soil excavation costs, and compliance with health risks regulation. Uncertainty on health risks and cost estimates due to the imperfect knowledge of soil contamination is quantified thanks to geostatistics.

Operating from 1975 to 1989, the principal activity of the site was the distillation of products from degreasing taweries operations, including chlorinated compounds. The site comprises 400 m² of buildings on less than 3700 m². The main characteristics are the presence of a river 15 m below a steep embankment and residential occupancy in the vicinity. Many products waiting for treatment and treatment residues were stored in barrels, ending with more than 5000 piled up on the site when the activity ceased.

Since the early 90s, the site has been in the hands of the French Agency for Ecological Transition (Ademe). An initial soil, water, and groundwater diagnostic revealed multiple contaminations by greases, petroleum solvents, organohalogen solvents, and BTEX. A detailed risk assessment was then performed. At this stage, uncertainties remained, and several remediation works options were considered.

In 2007-2008, new studies were undertaken that rigorously followed the MACAOH protocol, dedicated to volatile compounds (on-site methanol extraction and small vertical sampling lag of 50 cm as main specificities). Several source zones were identified, calling into question the remediation work based on an accelerated process of natural degradation of pollutants, completed by pumping of released degradation products and treatment of the water table.

A new campaign has been decided and undertaken in 2022 regarding the soils and 2022- 2023 for water. Objectives are a better understanding of the location and levels of the contaminants in soils, water, and groundwater, the natural pathways, the risks for residents to the east and south, etc. They cover an assessment of the state of media quality and the implementation of a remediation plan.

A geostatistical methodology has been undertaken with the specificity of relying on an iterative campaign. Initial modeling based mainly on 2008 data has been conducted to assess uncertainties and find the most hazardous areas. The optimal locations of the first 16 boreholes were determined to:
– Confirm or infirm 2008 data,
– Decrease the absolute uncertainty,
– Better delineate impacted areas.

Sampling was not preferential and considered the possible future locations of boreholes of the second part of the 2022 campaign. Using the first results, a detailed exploratory spatial data analysis has shown a poor correlation with old data. Iterative sampling then proved its advantage by giving the flexibility necessary to set up an optimized sampling plan for the second part of the campaign.

Finally, a geostatistical study was carried out to calculate the mass balance to develop scenarios for cost-benefit analyses and to describe better contaminant transfer behavior on site and to the river.