Scholarlink

<p>Botswana covers diverse ecosystems from arid deserts to wetlands as such, vegetation in the area directly correlates with rainfall and moisture gradient. Whereas the spatiotemporal variability of vegetation phenology has been studied at regional and continental scales, the spatiotemporal variability in Botswana is still not well understood. This study, therefore, examines the spatio-temporal dynamics of phenological changes in Botswana, over 34 years (1982–2015). In this study, remotely sensed data, particularly the Global Inventory and Mapping Studies (GIMMS) Normalized Difference Vegetation Index 3rd generation (NDVI3g) (GIMMS NDVI) was used to extract phenology metrics. Data preprocessing involved aggregating bi-weekly datasets into monthly composites using the Maximum Composite Value (MCV) technique, filtered using the Savitzky-Golay. Spatial trends were investigated using the spatial linear regression analysis while the linear regression analysis was used to assess the temporal variation with a focus on Start of Season (SOS), Length of Season (LOS) and End of Season (EOS). There has been a significant increase (slope = 0.00101, p-value = 0.03, Mann–Kendall 0.03) in NDVI-based vegetation cover across the study area. Linear trend analysis showed that 81% of the area experienced positive trends in SOS, 59% in EOS, and 37% in LOS. However, statistically significant spatial trends (p &lt; 0.05) were limited, with 12.6% of SOS, 3.5% of LOS, and 7.7% of EOS. Temporal analysis of the seasonal cycle for 1982 indicated vegetation growth peaks in the summer months (January to April) and declines towards winter, with the lowest growth in late July and August. Temporal trend analysis across the entire study area indicated insignificant increases in SOS (0.44 days/year) and EOS (0.08 days/year), while LOS showed a non-significant shortening trend. The SOS extended by 4.4 days/decade, EOS by 0.8 days per decade, and LOS advanced by 1.5 days/decade. District-level analysis revealed diverse trends, with all districts showing extensions in SOS and most districts (80%) extending in EOS, except for the Southern and Southeast Districts. In contrast, LOS showed advancements in 90% of the districts. Kweneng District exhibited the highest magnitude of SOS extension (8.1 days/decade), while the Southeast, Southern, and Kweneng districts experienced the highest magnitude of LOS advancement. Overall, this study highlights the complexity and variability of phenological changes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Uncovering the shifts: land surface phenology in Botswana from satellite observations

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年2月15日

查看全文
  • Linganani Kombani,
  • Samuel Adewale Adelabu,
  • Olufemi Sunday Durowoju,
  • Colbert Mutiso Jackson

摘要

Botswana covers diverse ecosystems from arid deserts to wetlands as such, vegetation in the area directly correlates with rainfall and moisture gradient. Whereas the spatiotemporal variability of vegetation phenology has been studied at regional and continental scales, the spatiotemporal variability in Botswana is still not well understood. This study, therefore, examines the spatio-temporal dynamics of phenological changes in Botswana, over 34 years (1982–2015). In this study, remotely sensed data, particularly the Global Inventory and Mapping Studies (GIMMS) Normalized Difference Vegetation Index 3rd generation (NDVI3g) (GIMMS NDVI) was used to extract phenology metrics. Data preprocessing involved aggregating bi-weekly datasets into monthly composites using the Maximum Composite Value (MCV) technique, filtered using the Savitzky-Golay. Spatial trends were investigated using the spatial linear regression analysis while the linear regression analysis was used to assess the temporal variation with a focus on Start of Season (SOS), Length of Season (LOS) and End of Season (EOS). There has been a significant increase (slope = 0.00101, p-value = 0.03, Mann–Kendall 0.03) in NDVI-based vegetation cover across the study area. Linear trend analysis showed that 81% of the area experienced positive trends in SOS, 59% in EOS, and 37% in LOS. However, statistically significant spatial trends (p < 0.05) were limited, with 12.6% of SOS, 3.5% of LOS, and 7.7% of EOS. Temporal analysis of the seasonal cycle for 1982 indicated vegetation growth peaks in the summer months (January to April) and declines towards winter, with the lowest growth in late July and August. Temporal trend analysis across the entire study area indicated insignificant increases in SOS (0.44 days/year) and EOS (0.08 days/year), while LOS showed a non-significant shortening trend. The SOS extended by 4.4 days/decade, EOS by 0.8 days per decade, and LOS advanced by 1.5 days/decade. District-level analysis revealed diverse trends, with all districts showing extensions in SOS and most districts (80%) extending in EOS, except for the Southern and Southeast Districts. In contrast, LOS showed advancements in 90% of the districts. Kweneng District exhibited the highest magnitude of SOS extension (8.1 days/decade), while the Southeast, Southern, and Kweneng districts experienced the highest magnitude of LOS advancement. Overall, this study highlights the complexity and variability of phenological changes.