2 research outputs found
Factors influencing impala distribution patterns in Nairobi National Park, Kenya
No full textMonitoring the distribution of wild animals using appropriate methods and survey techniques is essential not only for sustainable management but also to avoid wastage of resources. This study applied remote sensing to investigate the factors influencing the distribution of herbivores in Nairobi National Park, Kenya. Impala was selected as indicator specie for the herbivores within the park, because the population of impala had drastically reduced over time. The influence of food availability, water and disturbance on herbivore presence was investigated. A positive significant statistical relationship between impala population density and feed availability was observed. However, the correlation between impala population density and water distance was negative, indicating less impalas as the distance from water sources increase. An interesting finding was the expectation of greater impala population presence next to roads. The study demonstrated a rapid method for gaining information useful for conservation and land use planning practices, such as in the determination of the carrying capacity or even for redistributing animals within the park
Crop Residue Cover Assessment Using Remotely Sensed Data
No full textThe goal of reducing dependence on fossil fuels combined with the ongoing need to increase food production, and the concern about climate change, have created a demand to better quantify the fate of crop residues. Quantifying the amounts of crop residues returned to the soil are important input requirements for many carbon models. This research addresses the critical need for corn residue estimation using ground and space based remote sensing platforms. The questions asked were: (i) can residue be discriminated from soil, and other vegetation within 450 to 1750 nanometer wavelength? and (ii) can surface residue be accurately mapped using remote sensing? This research investigated different mathematical and statistical models to address these questions. Mathematical based approaches utilized 2, and 4 component mixing models, while statistical-based approaches relied on ANOVA, and regression analysis. A regression model developed directly from principal component analysis though complex, explained 47% of the variability across sites. Simpler models such as the Normalized Difference Water Index (NDWI), and the Normalized Difference Vegetation Index-wide band (NDVlw), though with variable consistency, could significantly (p \u3c 0.05) detect surface residue cover from soil. In a subsequent analysis, the accuracy of NDVlw derived from satellite sensors to predict surface residue improved when more information from spectral bands within the green, red, and near infra-red wavelengths were incorporated in the model. Finally, this research mapped com residue cover referred to as non- photosynthetic vegetation (NPV), remaining on the soil surface at the end of themonth of November in the year 2009. The AWiFS satellite derived residue map used in this study covered an area of about 837,000 ha, which was 40 % of the 2,000,000 ha total com fields within South Dakota. A 4 component mixing model was used to characterize surface residue cover. The coefficient of determination between NPV and NDVlw from the AWiFS derived map was 0.69. An independent check confirmed that areas practicing no-tillage adoption had high NPV. These findings suggest that NDVIw can be. used to assess the spatial variability of residue cover. However, these results need additional validation from independent field sites to confirm precision. Future research should investigate the most critical parameters for designing a general automated model for residue estimation. Mapping crop residues can assist in the assessment of policies and programs meant to support agricultural sustainability
