Page 103 - Çevre Şehir ve İklim Dergisi İngilizce - Özel Sayı

Mustafa Sert - Abdullah Emin Akay
Ayhan Ateşoğlu

MOD13A1 vegetation index was applied to find NDVI; MOD11A2 land
surface temperature data was used to calculate the temperature state index;
MCD43A4 surface reflectance data was utilized to obtain the topsoil grain size
index (TGSI); and MCD43A3 albedo data was used to measure the reflectance
properties of the ground surface.
TGSI, which is particularly applied to monitor the land degradation and identify
the physical properties of soil, was found through the albedo measurements for
red, blue and green bands (Liu et al., 2018; Xiao et al., 2006).
TGSI = (ρred – ρblue) / (ρred + ρblue + ρgreen) (12)
CART-DT, RF, and CNN models were utilized to validate the model examples
in the study.
• CART-DT: The decision tree algorithm is a classification method formed
of nodes representing a feature and variable that models the decision-
making process by dividing the data into smaller parts (Quinlan, 1986).
This model was run using ’Python’ software (Lamrini, 2020).
• RF: It is a machine learning method which is commonly used for
processing remote sensing images and formed of a combination
of many decision trees, assigning a value to a response variable via
regression trees. In this study, the number of trees was taken as 100.
This model was run using ’Python’ software (Chen et al., 2020).
• CNN: It is a deep learning model designed for image analysis and
classification, similar to neurons in the brain. It discerns the features
in the image in layers. This model was run using ’Python’ software
(Guirado et al., 2020).
The study applied the landscape indices (fragmentation index (LFI) and
segregation index (LSI)), which are used in many fields such as biodiversity
analysis, habitat analysis, landscape pattern and monitoring of its change
over time (Sui and Zeng, 2001) with the aim of assessing and analysing
desertification consequences of the study.
FN1 = (Np − 1)/Nc (13)
• FN1: Landscape fragmentation index
• Np: Total number of landscape fragments
• Nc: Ratio of total area to minimum fragment area
Surface wetness index (SWI) was applied to measure the moisture content
of the soil surface, which is used to determine dry and wet times of the year in
relation to climate change (Zhuguo et al., 2004).

(14)

90 Journal of Environment, Urban and Climate