**5. Conclusion**

Wide use of chlorinated pesticides such as chlordecone has led to severe contamination of the environment in the French West Indies. Chlordecone is an organochlorine insecticide that was regularly applied to banana crops more than 20 years ago and is now on the list of POPs prohibited by the Stockholm Convention. Chlordecone is a hydrophobic, non-biode‐ gradable pesticide that is strongly bound to the organic soil matrix which is why soils are still contaminated today. The level of soil pollution and the fate of the pollutant in the eco‐ systems depend on two main determinants at two different scales: agricultural practices and the soil physical properties, i.e. the clay microstructure.

Obviously, the history of agricultural practices (land-use changes and the intensity of pesti‐ cide use) partly explains soil contamination. The stocks of pesticide still present in soil are important indicators, which calls for the reinforcement of databases on agricultural practi‐ ces, especially on pesticide use. However, soil type mainly explains the concentration of pol‐ lutant in the soil. Our experimental data showed that clay microstructure controls the accumulation of pollutants in the soil and their release into the environment. In the case of andosols, this study revealed the importance of the fractal microstructure and the associated tortuous porosity in trapping the pollutant. Moreover, as chlordecone tends to bind with soil organic matter, the high organic content of tropical volcanic soils also favors its reten‐ tion.

More generally, it is worth noting that since pesticide retention depends on soil type, the concentration of pesticide in the soil cannot provide reliable information about the actual risk of contamination from the soil to the ecosystem. This makes mapping contamination hazards problematic because there is no univocal relation between soil pollution and its ca‐ pacity to contaminate crops. Additional studies are thus necessary to characterize the behav‐ ior of pesticides in soils and to identify the link with plant absorption. This will increase the relevance of mapping pesticide contamination and availability

Concerning possible soil decontamination procedures, the efficiency of chemical or biologi‐ cal species able to degrade the pesticide will be controlled by the microstructure of the clays. Future research into pesticide degrading microorganisms and other remediation tools that could be used to clean up polluted soils should take the accessibility of the pesticide in the soil microstructure into account.
