Monitoring and assessment of the seasonal and inter-annual pan inundation dynamics in the Kgalagadi Transfrontier Park, Southern Africa

Abstract In arid regions, pans form a critical water source that supports local communities, livestock, and wildlife, with readily available water resources. However, these waterbodies are unevenly distributed across the landscape and highly ephemeral and this influences their ability to provide water, especially in remote and water scarce areas. In addition, it remains difficult, impractical, and costly to monitor their variability using traditional hydrometric networks. In this regard, their contribution as water sources remains uncertain. The availability of spatially explicit data, with improved sensor design characteristics such as Landsat-8 enables the monitoring of their surface water extent over space and time. This study therefore, for the first time investigated the potential of using Landsat-8 in detecting and monitoring the spatial and temporal dynamics of pan inundation in the water scarce region of Kgalagadi in Southern Africa between 2016 and 2018. This was achieved by testing the performance of multiple indices, namely; the Normalised Difference Water Index (NDWI), Modified Normalised Difference Water Index (MNDWI), Automated Water Extraction Index for Shadow (AWEIsh), Water Ratio Index (WRI) and Land Surface Water Index (LSWI). Overall, the results have shown the potential of remote sensing data to monitor pan inundation. The MNDWI produced the highest overall classification accuracy of 84.91% comparatively. The MNDWI was then used for monitoring and assessing pan inundation dynamics over different seasons between 2016 and 2018. During the study period, pan inundation varied significantly (α = 0.05) for different seasons. Nevertheless, 2017 had the largest surface water extent covering 23 195.8 m2, during the wet season and 17 913.3 m2 in the dry season. On the other hand, 2018 showed the smallest spatial coverage of 13 076 m2 for the wet season and 6032.587 m2 during the dry season. The observed spatial variability in pan inundation was attributed to rainfall and temperature variability. The study thus revealed the utility of remotely sensed data sets in providing a more robust approach for monitoring seasonal and inter-annual pan inundation variations in semi-arid environments of Southern Africa. This information is important for water management decision making, specifically for water-limited areas, to conserve these water sources to ensure their sustainability in supporting local communities, livestock and wildlife population.