Skip to main content
Have a personal or library account? Click to login
A Two-Stage XGBoost Approach for Mapping Arable and Non-Arable Soils under Salinity Stress in Southern Iraq Cover

A Two-Stage XGBoost Approach for Mapping Arable and Non-Arable Soils under Salinity Stress in Southern Iraq

Civil and Environmental Engineering
Volume 22 (2026): Issue 2 (June 2026)
By: ,   and    
Open Access
|Jun 2026
Figures & tables

Figures & Tables

Figure 1:

Study area

Figure 2:

Methodological framework for the research

Figure 3:

Histogram and summary statistics for soil samples

Figure 4:

Ground truth points

Figure 5:

Multi-temporal image fusion and cumulative data. a,b,c,f,g,h,I,j, and k multi-temporal image fusion for NBR2, NDVI, LST, NDMI (dry and wet season), NDSI, SI6, SI, and BI, respectively. d and e cumulative annual rainfall and aET data, respectively

Figure 6:

Results of Model A. a -Binary classification of SoilVR/NPV and other LULC classes. b-Spatial distribution of SoilVR/NPV

Figure 7:

Analysis of Gain and Weight Regarding the Importance of Variables

Figure 8:

Spatial distribution of arable/non-arable land

Model A performance

PhasesF1-Score [-]MCC [-]Accuracy [-]CategorySensitivity [-]
Training0.990.990.99SoilVR/NPV1
other classes0.99
Validation0.930.790.99SoilVR/NPV1
other classes0.78

Modeles Characteristics

Model CharacteristicsXGBoost-Model AXGBoost-Model B
Number of Trees [-]100180
Leaf Size [-]111
Tree Depth Range [-]2–62–8
Mean Tree Depth [-]33
Percentage of Training Available per Tree [-]100100
Number of Randomly Sampled Variables [-]13
Percentage of Training Data Excluded for Validation [-]1010
L2 Regularization (Lambda) [-]1.001.00
Minimum Loss Reduction for Splits (Gamma) [-]0.000.00
Learning Rate (Eta) [-]0.300.30

Model B performance

PhasesF1-Score [-]MCC [-]Accuracy [-]CategorySensitivity [-]
Training111arable1
non-arable1
Validation0.940.900.95arable1
non-arable0.88

The indicators and data used in the research

Data/IndexMathematical FormulaSourceTime
aET [mm]*****************************WaPOR(v2and v3)2013–2024
Rainfall [mm]*****************************CHIRPS2013–2024
NDVI [-] NDVI=NIR−RedNIR+Red {\rm{NDVI}} = {{{\rm{NIR}} - {\rm{Red}}} \over {{\rm{NIR}} + {\rm{Red}}}} Landsat8(B5,B4)2013–2024
NBR2 [-] NDR2=SWIR1−SWIR2SWIR1+SWIR2 {\rm{NDR}}2 = {{{\rm{SWIR}}1 - {\rm{SWIR}}2} \over {{\rm{SWIR}}1 + {\rm{SWIR}}2}} Landsat8(B6,B7)2013–2024
NDMI [-] NDMI=NIR−SWIR1NIR+SWIR1 {\rm{NDMI}} = {{{\rm{NIR}} - {\rm{SWIR}}1} \over {{\rm{NIR}} + {\rm{SWIR}}1}} Landsat8(B5,B6)2013–2024
LST [kelvin]*****************************Landsat8(B10)2013–2024
SI [-] SI=(Green*Red) SI = \sqrt {\left( {Green*Red} \right)} Landsat8(B4,B3)2013–2024
SI6 [-] SI6=Red*NIRGreen SI6 = {{Red*NIR} \over {Green}} Landsat8(B4,B5,B3)2013–2024
NDSI [-] NDSI=Red−NIRRed+NIR NDSI = {{Red - NIR} \over {Red + NIR}} Landsat8(B4,B5)2013–2024
BI [-] BI=(Green2+Red2+NIR2) BI = \sqrt {\left( {{Green^2} + {Red^2} + {NIR^2}} \right)} Landsat8(B3,B4,B5)2013–2024
River data [-]*****************************NCWRM2013–2024
DOI: https://doi.org/10.2478/cee-2026-0017 | Journal eISSN: 2199-6512 | Journal ISSN: 1336-5835
Journal RSS Feed
Language: English
Page range: 481 - 499
Submitted on: Jul 12, 2025
Accepted on: Aug 15, 2025
Published on: Jun 19, 2026
Published by: University of Žilina
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year
Keywords:
Soil salinization,
Remote sensing,
Binary classification,
Machine learning,
XGBoost
Related subjects:
Business and economics,
Political economics,
Economic theory, systems and structures,
Business management,
Industries,
Environmental management

© 2026 Mohammed Azeez, Hisham M. Jawad Al Sharaa, Abdul Razzak T. Zboon, published by University of Žilina
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 22 (2026): Issue 2 (June 2026)
Next article