Return to Recent projects

Assessment of algal community responses to multiple stressors over the last 150 years using hyperspectral imaging of lake sediment cores and machine learning algorithms

Aquatic sediments are one of the reliable archives of information that can be interpreted by paleolimnologists. Sediments accumulated much useful information archived from both autochthonous and allochthonous materials. Sediments contain a vast array of physical, biological, chemical, and visual information that follow the present-day environmental variability. Of most important data stored in sediments that can be used to reconstruct limnological conditions are indicators including particle size analysis, geochemical data, cyanobacteria (blue-green algae), and algae. There is a consensus among the paleolimnologists that complete knowledge of environmental change is possible using multi-proxy data. So, the parallel monitoring of multi-data can lead to a more robust assessment of environmental changes.

Most of the techniques used to extract specific information from sediments are destructive which restricts the number of samples to perform all the analysis required for multi-proxy data gathering. Moreover, in some cases because of the subsampling restrictions, reaching high-resolution data that satisfy today’s paleo studies requirements is not possible. Therefore, there is a need for non-destructive techniques capable of capturing different data from the sediment cores with high spatial resolutions.

Recently, due to the rapid advance of scanning technologies and associated algorithms, hyperspectral imaging (imaging spectroscopy) has emerged as a tool in paleolimnology, which is focused on tracking changes in a given lake as well as its watershed and airshed. This technique has gained attention in the geoscience community because of its detailed representation of sediment features. Hyperspectral imaging alongside remote sensing techniques can be used to determine the composition of sediment cores and restore the past environment’s physical, biological, and chemical variations.