By Michelle Ang
The Mining, Resettlement and Livelihoods: Research and Practice Consortium is an initiative of The University of Queensland's Centre for Social Responsibility in Mining (CSRM). It is a multi-party, industry-university research collaboration on mining and resettlement in order to build knowledge for practical application. https://www.miningresettlement.org/
Just to recap on this research, the first phase of this three-part project is to categorize the types of landuse and infrastructure within the mine footprint using aerial imagery (LANDSAT) and remote sensing techniques to develop ratios of ‘productive elements’ to ‘waste elements’. To achieve this, the landuse classification was done using the manual digitizing method via the ArcGIS software combined with analysis outputs from the CLASlite software.
Step 1: Ground Truth Data from Experts
Prior to the classification process, we must first obtain assistance from the researchers and on the ground experts to identify the key land cover classes within the mining sites. Below are two videos formulated from the discussions.
a) Porgera
b) Sepon
Step 2: Finalized Land Cover Classes and Classification Schema
From the outputs above, we compiled the main mining land cover several other land cover classes from the mine sites' surrounding 3 km buffer; determined as the mine impact zone. Table 1 shows the final compilation of land cover classes to be used during the classification. We also developed a classification schema to help with the classification. Our schema represents surface features as a multi-level hierarchy from coarse thematic resolution land cover (i.e. ‘mine’, ‘non-mine’) to finer-scale land cover classes including sediment control and agricultural areas. The schema represents an idealized representation of land cover features.
Step 3: The Classification
Using a combination of automatic (CLASlite) and manual classification (ArcGIS) methods, we managed to map out the land covers from 1987 to 2017 for Porgera and 2002 to 2017 for Sepon. Below are two videos with the completed classification and area analysis.
a) Porgera
b) Sepon
Step 4: The Results and Conclusion
The mapped spatial patterns over time demonstrated the dynamic nature of mining landscapes. At both sites, mine waste (i.e. tailings dams, dumps) increased more rapidly than the mine productive area and progressive rehabilitation lagged behind. There was a notable increase in mining-related and community land covers such that by the 2017 these land cover types were in close proximity to one another at both sites. As land pressures build the relationship between the operator and local community intensifies. The results from this project represent one of the most detailed remote sensing studies in mining in terms of cross-disciplinary and high thematic, spatial and temporal resolution. Such data has the potential unite methods and perspectives from various social and environmental disciplines within a GIS platform to provide a more holistic perspective of a mining operation footprint that is required to solve the complex and diverse negative impacts associated resource developments.