Page 104 - Çevre Şehir ve İklim Dergisi İngilizce - Özel Sayı

The use of Geographic Information Systems and Remote Sensing Technologies in
Combating Desertification and Erosion

• Pi: Monthly rainfall
• PEi: Potential evapotranspiration
The outcomes of the study suggest that the random forest method
outperforms all other models in the splitting of desertified land based on
low and medium resolution remote sensing images. The land classification
results obtained by combining NDVI, LST, Albedo, and TGSI methods
have been highly accurate. The TGSI has proved to be a greatly effective
monitoring index. This study provides a database to monitor and analyse the
desertification process and to take necessary measures.
In their study “Land Degradation Neutrality Decision Support System-
Provincial Statistics and Sustainable Land Management Approaches and
Practices”, Erpul et al. (2023) aimed to identify the land use status and
changes, to determine priority areas, and to put forward nature-based solution
proposals against the impacts of climate change. LDN-DSS is an open source
system based on ‘Google Earth Engine’ (GEE) that enables decision makers to
make accurate and effective decisions with data and models. For the purpose
of monitoring LDN, three indicators were taken into account: land cover, land
productivity dynamics, soil organic carbon, apart from additional indicators
(erosion severity, desertification vulnerability). As a consequence, the study
presented the change of land cover in Türkiye between 1990-2018. According
to these results, an increase of 1.428,49 ha in areas covered with trees and an
increase of 29.892,27 ha in wetlands, a decrease of 99.843,60 ha in pasture/
meadow areas and a decrease of 3.830,74 ha in agricultural areas were found
at spatial scale. This study provides the opportunity to easily perform different
spatial data analyses in the desired project area using the GEE interface.
Since digital soil maps are characterised by coarse spatial resolution and
are outdated and therefore do not support physical process-based models
(mathematical and computer-based simulations) for improved estimation,
Samarinas et al. (2024) performed a study titled as ‘Soil Loss Estimation by Water
Erosion in Agricultural Areas Introducing Artificial Intelligence Geospatial
Layers into the RUSLE Model’. The aim of this study is to provide advanced
geographic layers to the RUSLE model by using Sentinel-2 satellite imagery
with a spatial resolution of 10 metres and 13 spectral bands to develop data
cube-based tools (Soil Data Cube) for various thematic maps of soil using
artificial intelligence and in-situ soil data (data from physical measurements).
This improves both the spatial resolution and reliability of the final map.
The study applied the European Digital Elevation Model (EU-DEM) with
a spatial resolution of 30 metres, produced by Copernicus, a programme to
study the Earth’s surface managed by the European Union. The ERA5-Land
with a temporal resolution of 1 hour produced by the European Centre for

91
Special Issue / 2024