Quantifying Surface Mining Expansion and Reclamation Using Deep Learning-based ConvoLSTM Model and Satellite Images: A Case Study in Lapland Region of Finland

Keywords: Keywords: Mining Mapping, ConvoLSTM, 2D-CNN, Remote Sensing, Deep Learning, CVA
Abstract Type: Paper Abstract

Abstract

Mining conflicts sustainable environment and causes disturbances for the livelihoods of people. Given the adverse impact on environment, indigenous community including Sami people and domesticated reindeer, it is of critical importance to peruse mining expansion and reclamation in Lapland, Finland. For the first time, this study employs a spatial-temporal deep learning architecture, the ConvoLSTM model, which enables accurate predictions of mining activities by capturing spectral, spatial, and temporal dependencies. Our custom ConvoLSTM model integrates a 2-Dimensional Convolutional Neural Network (2D-CNN) with a Long Short-Term Memory (LSTM) component. Using 10-meter Sentinel-2 imagery, we generated time-series land use/land cover (LULC) maps from 2015 to 2024 to track changes in mining extent. The performance of the ConvoLSTM model was carefully evaluated against a Random Forest (RF) and a standalone 2D-CNN model, where the ConvoLSTM achieved superior accuracy. In the post-analysis phase, the Change Vector Analysis (CVA) technique was applied to quantify the magnitude and direction of change in mining activities over the past decade. The unique contribution of this study lies in implementing a custom spatial-temporal deep learning model to map decade-long mining activities and detect changes using publicly available satellite data. The resulting time-series maps demonstrate significant conversion of forest land and bare soil into mining areas, highlighting the rapid expansion of mining activities in Lapland. These findings offer critical insights and a valuable resource for policymakers, researchers, and reindeer herders, facilitating informed decision-making for sustainable environmental management and natural resource conservation in Finland.