Open Access

Time Series Analysis of Landsat Images for Monitoring Flooded Areas in the Inner Niger Delta, Mali

Polina Lemenkova
Polina Lemenkova
 and 
Olivier Debeir
Olivier Debeir
   | Jan 19, 2024
Artificial Satellites's Cover Image
About this article
Previous Article
Next Article
Cite
Published Online: Jan 19, 2024
Page range: 278 - 313
Received: Feb 24, 2023
Accepted: Nov 29, 2023
DOI: https://doi.org/10.2478/arsa-2023-0011
Keywords
© 2023 Polina Lemenkova et al., published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.

Figure 1.

Topographic map of Mali. Mapping software: Generic Mapping Tools (GMT) scripting toolset. The area of the Inner Niger Delta is indicated by green rotated square. Data source: GEBCO/SRTM. Cartography source: authors.
Topographic map of Mali. Mapping software: Generic Mapping Tools (GMT) scripting toolset. The area of the Inner Niger Delta is indicated by green rotated square. Data source: GEBCO/SRTM. Cartography source: authors.

Figure 2.

Flowchart summarising general steps of data processing. Diagram source: authors (R library ‘DiagrammeR’).
Flowchart summarising general steps of data processing. Diagram source: authors (R library ‘DiagrammeR’).

Figure 3.

The location of the Landsat 8–9 satellite image in Mali, Mopti region of the Inner Niger Delta. The images were downloaded from the EarthExplorer repository, USGS. Background image: ESRI World imagery.
The location of the Landsat 8–9 satellite image in Mali, Mopti region of the Inner Niger Delta. The images were downloaded from the EarthExplorer repository, USGS. Background image: ESRI World imagery.

Figure 4.

Landsat 8–9 images of Inner Niger Delta in natural colour RGB values showing floodplain for six years (always November): (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022.
Landsat 8–9 images of Inner Niger Delta in natural colour RGB values showing floodplain for six years (always November): (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022.

Figure 5.

Concept flowchart of the remote sensing data processing and analysis. Flowchart is prepared using R library ‘DiagrammeR’. Source: authors.
Concept flowchart of the remote sensing data processing and analysis. Flowchart is prepared using R library ‘DiagrammeR’. Source: authors.

Figure 6.

Processing Landsat satellite image in RStudio for extracting NDVI.
Processing Landsat satellite image in RStudio for extracting NDVI.

Figure 7.

NDVI based on Landsat 8–9 images of Inner Niger Delta: (a) 2013, (b) 2015,(c) 2018, (d) 2020, (e) 2021, (f) 2022. Mapping: RStudio. Source: authors.
NDVI based on Landsat 8–9 images of Inner Niger Delta: (a) 2013, (b) 2015,(c) 2018, (d) 2020, (e) 2021, (f) 2022. Mapping: RStudio. Source: authors.

Figure 8.

Histograms of the NDVI of the Landsat 8–9 images on Inner Niger Delta: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Plots: RStudio. Source: authors.
Histograms of the NDVI of the Landsat 8–9 images on Inner Niger Delta: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Plots: RStudio. Source: authors.

Figure 9.

Soil Adjusted Vegetation Index (SAVI) based on the Landsat 8–9 images of Inner Niger Delta for November: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Mapping visualisation: RStudio. Source: authors.
Soil Adjusted Vegetation Index (SAVI) based on the Landsat 8–9 images of Inner Niger Delta for November: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Mapping visualisation: RStudio. Source: authors.

Figure 10.

Histograms of SAVI based on the Landsat 8–9 images of Inner Niger Delta: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Plots: RStudio. Source: authors.
Histograms of SAVI based on the Landsat 8–9 images of Inner Niger Delta: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Plots: RStudio. Source: authors.

Figure 11.

Enhanced Vegetation Index (EVI) based on the Landsat 8–9 images of Inner Niger Delta for November: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Mapping visualisation: RStudio. Source: authors.
Enhanced Vegetation Index (EVI) based on the Landsat 8–9 images of Inner Niger Delta for November: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Mapping visualisation: RStudio. Source: authors.

Figure 12.

Histograms of EVI based on the Landsat 8–9 images of Inner Niger Delta for November: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Plots: RStudio. Source: authors.
Histograms of EVI based on the Landsat 8–9 images of Inner Niger Delta for November: (a) 2013, (b) 2015, (c) 2018, (d) 2020, (e) 2021, (f) 2022. Plots: RStudio. Source: authors.

Figure 13.

Clustering based on the Landsat 8–9 images (bands 4-3-2): (a) 2013 RGB in true colour composites (TCC), (b) 2013 clusters, (c) 2015 RGB in TCC, (d) 2015 clusters, (e) 2018 RGB in TCC, (f) 2018 clusters. Mapping: RStudio. Source: authors.
Clustering based on the Landsat 8–9 images (bands 4-3-2): (a) 2013 RGB in true colour composites (TCC), (b) 2013 clusters, (c) 2015 RGB in TCC, (d) 2015 clusters, (e) 2018 RGB in TCC, (f) 2018 clusters. Mapping: RStudio. Source: authors.

Figure 14.

Clustering based on the Landsat 8–9 images of Inner Niger Delta: (a) 2020 RGB in TCC (4-3-2), (b) 2020 classification, (c) 2021 RGB in TCC, (d) 2021 classification, (e) 2022 RGB in TCC, (f) 2022 classification. Mapping: RStudio. Source: authors.
Clustering based on the Landsat 8–9 images of Inner Niger Delta: (a) 2020 RGB in TCC (4-3-2), (b) 2020 classification, (c) 2021 RGB in TCC, (d) 2021 classification, (e) 2022 RGB in TCC, (f) 2022 classification. Mapping: RStudio. Source: authors.

Figure 15.

Correlation plot showing probability cases for land cover classes in the Inner Niger Delta between 2013 and 2022 based on the results of the clustering of the Landsat 8–9 images using Kendall correlation method. Mapping: Python. Source: authors.
Correlation plot showing probability cases for land cover classes in the Inner Niger Delta between 2013 and 2022 based on the results of the clustering of the Landsat 8–9 images using Kendall correlation method. Mapping: Python. Source: authors.

Metadata of the satellite images used in this study: Landsat 8–9 USGS1.

Date Spacecraft / ID Path/Row Entity Product ID Scene ID Cloud/Coverage
2013/11/10 Landsat 8 197/50 LC08_L2SP_197050_20131110_20200912_02_T1 LC81970502013314LGN01 0.12
2015/11/16 Landsat 8 197/50 LC08_L2SP_197050_20151116_20200908_02_T1 LC81970502015320LGN01 1.12
2018/11/24 Landsat 8 197/50 LC08_L2SP_197050_20181124_20200830_02_T1 LC81970502018328LGN00 0.00
2020/11/29 Landsat 8 197/50 LC08_L2SP_197050_20201129_20210316_02_T1 LC81970502020334LGN00 0.00
2021/11/16 Landsat 8 197/50 LC08_L2SP_197050_20211116_20211125_02_T1 LC81970502021320LGN00 0.00
2022/11/11 Landsat 9 197/50 LC09_L2SP_197050_20221111_20221113_02_T1 LC91970502022315LGN00 0.00

Results of the NDVI, SAVI and EVI computations of the Landsat 8–9 images.

Time NDVI Extreme Values SAVI Extreme Values EVI Extreme Values
minimal maximal minimal maximal minimal maximal
10 November 2013 −0.2311377 0.5879165 −0.2235105 0.4281315 −0.3576084 0.6849960
16 November 2015 −0.2619293 0.5245128 −0.1681281 0.2896889 −0.2689985 0.4634939
24 November 2018 −0.2603942 0.5258653 −0.1830571 0.2938004 −0.2928849 0.4700728
29 November 2020 −0.2496693 0.4980780 −0.1384092 0.2833834 −0.2214491 0.4534074
16 November 2021 −0.2522458 0.5417392 −0.1564469 0.2790245 −0.2503089 0.4464335
11 November 2022 −0.3072812 0.5237642 −0.1810541 0.2929085 −0.2896799 0.4686475
eISSN:
2083-6104
Language:
English
Publication timeframe:
4 times per year
Journal Subjects:
Geosciences, other
Journal RSS Feed