Browsing by Author "Hamad, Amina Amri"

Now showing 1 - 2 of 2

Distribution of lantana camara and its impacts on selected ecosystem services and livelihood in east Usambara mountains, Tanzania
(Sokoine University of Agriculture, 2021) Hamad, Amina Amri
This study aimed at assessing the distribution of L. camara and its fractional cover using very high-resolution satellite image (Worldview-3) complemented by field data from plots surveys. Since L. camara co-exist with several other invasive alien plants species in East Usambara it is not easy to detect it. Therefore, it was necessary to find ways to differentiate it from the other species using remote sensing. Field Spectrometer was used to obtain spectral data of eighteen invasive plants occurring in East Usambara and thereafter assessing suitable bands that can be used to detect L. camara from the available sensors currently but also future satellite missions. Results suggest that when using Worldview 2 the Near Infra-Red 2 (NIR2) (860-1040 nm) is the best band for L. camara detection while for Sentinel 2 the Short Wave Infra- Red (SWIR) Cirrus (1360-1390 nm) is more suitable. Moreover, hyperspectral EnMAP sensor can differentiate the species best in the Visible NIR (423 nm). Evidence from Random Forest model using all bands from 350-2500nm suggest that, the visible ranging 350-383nm is the best region for differentiation of the species with band importance of 0.8-1. With the absence of this range of bands in the visible, the SWIR bands ranging 1778-1790nm can be used to differentiate the species having band importance of 0.5-0.7. Gradient Boosted Machine model showed that SWIR band at 1790 nm was the best in the separation of the invasive species with band importance of 0.4. It is therefore important that while choosing images for detection of similar species it should contain these wavelength ranges. It is also important for future satellite missions to include these wavebands so as similar species as the 18 in this study can be detected and differentiated. Furthermore, the study used the Worldview 3 Satellite image to map L. camara in East Usambara utilizing the NIR 2 texture properties of the image as was found to be the bestiii band in differentiation of the species. Lantana camara in East Usambara was detected along the roads and forest edges with low abundance, while it was more abundant in agriculture lands and abandoned farms and tea plantations. Only 0.5% of the 11% invaded areas had L. camara abundance above 50% while 8.8% of invaded areas have L. camara abundance of less than 25%. Presence of L. camara affects ecosystem services and livelihood of the inhabitants. East Usambara inhabitants benefit from the ecosystem through provisioning of wild plant species for medicine, fuelwood and construction among others. They benefit also through crop cultivation and honey production which are among major economic activities in the area. Results show that 67% of the respondents perceive L. camara to cause a decline in varieties of wild plant species and 50% responded that it caused reduction in honey production which affects their livelihood. The study also found that L. camara reduced the growth of maize by 29%, while cassava was not affected. The reduction of maize growth is only caused when growing simultaneously with maize as there was no evidence of allelopathic effects of L. camara. Similarly, Inoculation of autoclaved soil with microorganisms from invaded soils increased biomass of cassava and reduced the growth of maize. Therefore, removal of L. camara will immediately mitigate its negative impacts on maize. The livelihood of the inhabitants depending on crop production is not affected directly by L. camara abundance rather it is affected significantly by increasing cost of farm maintenance which includes weeding. It is estimation that about 66.62 Million Tanzanian shillings is required to clear about 1277 hectares of L. camara invasion in East Usambara. It is recommended that L. camara should be managed while it is still not very dense and widely spread. The cost of managing will increase substantially with further invasion and mechanical means may not suffice.
The potential of hyper-temporal NDVI data to assess vegetation condition and grazing intensity
(International Institute for Geo-information Science and Earth Observation, 2010-02) Hamad, Amina Amri
Land degradation has been reported to be a major environmental problem in Crete for a long time and is largely caused by excessive grazing. Several attempts have been done using satellite images for the purpose of monitoring the effect of excessive grazing on vegetation as a whole and their distribution. These studies lacked the temporal aspects of monitoring grazed lands since they used satellite imagery of one date. This study used 10 years MODIS hyper temporal NDVI images of 16 days temporal resolution to assess vegetation condition and grazing intensity. Grazed areas were classified by vegetation types into 8 groups. Grass Index measured from field was used as a direct estimate of Grazing Index. Seasonal analysis was done where the pixel NDVI value at the peak and at the end of grazing season was compared separately to Grazing Index to assess vegetation conditions. Not only that but also trend analysis was used to assess vegetation conditions, where the slope of 10 years NDVI was compared to the Grazing Index by vegetation types. Assessment of Grazing Intensity was done by comparing the difference of NDVI at the peak and end of grazing season with the Grazing Index. Seasonal analysis showed that at a confidence interval of 0.05 four groups had significant positive relation between the NDVI at a peak of growing season and grazing index and one had a negative relation. Moreover trend analysis revealed that at a confidence level of 0.15 two groups had significant positive and negative relation between slope of 10 years NDVI and grazing index. These analysis indicate that different vegetation types have different response to high grazing intensities. NDVI difference was found to have a positive relation with Grazing Index indicating it is possible to use this method to estimate grazing intensity. Not only that but also R 2 ranged between 25%-87% which means the analysis explained sufficient variability. MODIS hyper temporal NDVI has shown a potential to assess vegetation conditions and grazing intensity however studies should be done more on how to capture the amount of brown biomass this could improve the estimates and monitoring of grazed lands.