Image analysis
The satellite imagery used in this study was of Landsat 5, Landsat 7,
and Landsat 8 satellites (30x30 resolution) obtained from the Google
Earth Engine (GEE) website. Different satellites were necessary to
obtain images from different years, so Landsat 5 was used for 1996,
1999, and 2002 maps, Landsat 7 for 2005, 2008, and 2011 maps, and
Landsat 8 for 2013, 2016, 2019, and 2022 maps (Chandler et al., 2009).
In recent times Google Earth Engine has become an important database of
earth observation data and land cover classification in many
mining-related studies and other applications such as vegetation
mapping
and monitoring, agricultural application, disaster management, and earth
science (Gbedzi et al., 2022). In addition to these satellite images,
information obtained from the remnants of Pampean grassland identified
and delineated with Google Earth Imagery was added.
To analyze the different land uses, images from 27 October to 27 May in
each year of study were used. This date was chosen due the spring/summer
season is when the vegetation associated with the quarry is in bloom.
NDVI index (Normalized Difference Vegetation Index) in GEE was applied,
paying special attention to the threat posed by quarries in the Tandilia
Mountains. The application of the NDVI index has been used to monitor
characteristics of vegetation cover and health status but also for the
detection of changes in an ecosystem over time (Paruelo, 2008; Tong et
al., 2016). Calculation of NDVI for a given pixel always results in a
number that ranges from minus one (-1) to plus one (+1): bare soils
(quarries in that case), giving a value close to zero, and very dense
green vegetation have values close to +1 (Musa & Jiya, 2011). In
general, land covers are often mixtures of several types, so even
fine-resolution remote sensing data do not measure pure spectra, but
instead mixed reflectance of vegetation and non-vegetation, rendering it
difficult to clearly identify exposed bedrock (Yue et al., 2012). For
this reason, the satellite images were then processed with QGIS
software.
The image was processed using QGIS 3.24.2, a Geographic Information
Systems (GIS) software. QGIS allows the input, manipulation, analysis,
and presentation of data and information related to a place on the
earth’s surface and therefore works with geo-referenced points (Ershad,
2020). With this tool was possible to locate and characterize quarries
(dimensions and associated threats). To do this a visual interpretation
of land use was made: quarries with a lot of vegetation or water are
considered inactive, and quarries with bare soil are active. The
resulting maps show the evolution of the quarries during the years of
the study. Some characteristics of the quarries, such as the state of
their activity, only were reliable in the current year (2022). The state
of quarrying activity was confirmed from the information obtained in the
Mining Cadastre of the province of Buenos Aires(https://www.gba.gob.ar/produccion/areas_de_trabajo/mineria).
Moreover, the area of each remnant and the total amount of Pampean
grassland were calculated using the QGIS software.