Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Post-Classification Analysis of Land Use and Land Cover Changes in West Africa 2000-2013

Papadopoulou, Melina LU (2025) In Student thesis series INES NGEK01 20251
Dept of Physical Geography and Ecosystem Science
Abstract
This study assesses the effectiveness of post-classification change detection for monitoring land use and land cover (LULC) dynamics across West Africa between 2000 and 2013. Using publicly available maps from the USGS West Africa Land Cover dataset, the original 25-class scheme was reclassified into 8 generalized land cover categories to allow for a more interpretable change analysis. A pixel-wise comparison of the reclassified maps was conducted to detect transitions, supported by an accuracy assessment based on stratified random sampling and high-resolution satellite imagery interpretation. The methodological assessment revealed a moderate classification performance of the land cover transitions, with an overall accuracy of 46.5% and a... (More)
This study assesses the effectiveness of post-classification change detection for monitoring land use and land cover (LULC) dynamics across West Africa between 2000 and 2013. Using publicly available maps from the USGS West Africa Land Cover dataset, the original 25-class scheme was reclassified into 8 generalized land cover categories to allow for a more interpretable change analysis. A pixel-wise comparison of the reclassified maps was conducted to detect transitions, supported by an accuracy assessment based on stratified random sampling and high-resolution satellite imagery interpretation. The methodological assessment revealed a moderate classification performance of the land cover transitions, with an overall accuracy of 46.5% and a Cohen’s Kappa of 0.45. Classification accuracy was higher for the year 2000 than 2013, likely due to imagery quality differences and increased land cover heterogeneity over time. The results demonstrate that while the simplified classification system effectively captured dominant regional patterns, particularly agricultural expansion and deforestation, it also masked finer-scale variation and introduced thematic uncertainty in mixed land cover areas. Approximately 10.9% of the study area underwent detectable land cover change during the analysis period. Transitions were dominated by agricultural expansion (67.7% of changed area), followed by vegetation recovery, water-related change, and deforestation. Spatial patterns aligned with known regional dynamics, with cropland expansion concentrated in the Sahel and deforestation in the Upper Guinean forest zone. The study highlights both the strengths and limitations of post-classification analysis for large-scale LULC monitoring, particularly when working with coarsely resolved and thematically generalized data. It underscores the importance of consistent validation and appropriate classification design in ensuring the interpretability and reliability of long-term land change assessments. (Less)
Please use this url to cite or link to this publication:
author
Papadopoulou, Melina LU
supervisor
organization
alternative title
Post-Classification Analys av markanvändning och förändringar i marktäcke i Västafrika 2000-2013
course
NGEK01 20251
year
type
M2 - Bachelor Degree
subject
keywords
Land Use and Land Cover Change, Change Detection, Remote Sensing, West Africa, Post-Classification Change Detection, Accuracy Assessment, Transition Analysis
publication/series
Student thesis series INES
report number
706
language
English
id
9197174
date added to LUP
2025-06-11 16:06:49
date last changed
2025-06-11 16:06:49
@misc{9197174,
  abstract     = {{This study assesses the effectiveness of post-classification change detection for monitoring land use and land cover (LULC) dynamics across West Africa between 2000 and 2013. Using publicly available maps from the USGS West Africa Land Cover dataset, the original 25-class scheme was reclassified into 8 generalized land cover categories to allow for a more interpretable change analysis. A pixel-wise comparison of the reclassified maps was conducted to detect transitions, supported by an accuracy assessment based on stratified random sampling and high-resolution satellite imagery interpretation. The methodological assessment revealed a moderate classification performance of the land cover transitions, with an overall accuracy of 46.5% and a Cohen’s Kappa of 0.45. Classification accuracy was higher for the year 2000 than 2013, likely due to imagery quality differences and increased land cover heterogeneity over time. The results demonstrate that while the simplified classification system effectively captured dominant regional patterns, particularly agricultural expansion and deforestation, it also masked finer-scale variation and introduced thematic uncertainty in mixed land cover areas. Approximately 10.9% of the study area underwent detectable land cover change during the analysis period. Transitions were dominated by agricultural expansion (67.7% of changed area), followed by vegetation recovery, water-related change, and deforestation. Spatial patterns aligned with known regional dynamics, with cropland expansion concentrated in the Sahel and deforestation in the Upper Guinean forest zone. The study highlights both the strengths and limitations of post-classification analysis for large-scale LULC monitoring, particularly when working with coarsely resolved and thematically generalized data. It underscores the importance of consistent validation and appropriate classification design in ensuring the interpretability and reliability of long-term land change assessments.}},
  author       = {{Papadopoulou, Melina}},
  language     = {{eng}},
  note         = {{Student Paper}},
  series       = {{Student thesis series INES}},
  title        = {{Post-Classification Analysis of Land Use and Land Cover Changes in West Africa 2000-2013}},
  year         = {{2025}},
}