Browsing by Author "Adjorlolo, Clement."

Now showing 1 - 4 of 4

Assessing developmental footprint within an agricultural system using multi-temporal remotely sensed data.
(2014) Dlamini, Zibusiso Nelson.; Odindi, John Odhiambo.; Adjorlolo, Clement.
The advent of the new political dispensation in South Africa has seen an exponential growth in the rate of land transformation and encroachment by other land uses into agricultural land in the uMngeni Local Municipality. Accurate evaluation of the rate of transformation is necessary for effective monitoring and management of the natural agricultural resources. In this regard, the use of multi-temporal remote sensing data provides efficient and cost-effective method. The current research assesses the extent to which the development footprint in uMngeni Local Municipality has affected agricultural land categories or zones, using multi-temporal remote sensing data. The study endeavoured to map and quantify the magnitude of change in built-up land cover and other infrastructure by focusing on two time intervals: the periods from 1993 – 2003 and 2003 – 2013. Medium spatial resolution Landsat image data acquired for these periods were analysed to classify and extract the built-up features to appraise the level of change. Results revealed positive change in built-up infrastructure: ~13% increase between 1993 and 2003, ~38% increase from 2003 – 2013, with overall ~32% for the 20 years (1993 – 2013) period under consideration. Next, factors possibly contributing to the encroachment of other land uses into the agricultural landscape and the potential threats to the sustainability of the agricultural system are highlighted.
Estimating woody vegetation cover in an African Savanna using remote sensing and geostatistics.
(2008) Adjorlolo, Clement.; Mutanga, Onisimo.
A major challenge in savanna rangeland studies is estimating woody vegetation cover and densities over large areas where field based census alone is impractical. It is therefore crucial that the management and conservation oriented research in savannas identify data sources that provides quick, timely and economical means to obtain information on vegetation cover. Satellite remote sensing can provide such information. Remote sensing investigations, however, require establishing statistical relationships between field and remotely sensed data. Usually regression is the empirical method applied to field and remotely sensed data for the spatial estimation of woody vegetation variables. Geostatistical techniques, which take spatial autocorrelation of variables into consideration, have rarely been used for this purpose. We investigated the possibility of improving woody biomass predictions in tropical savannas using cokriging. Cokriging was used to evaluate the cross-correlated information between SPOT (Satellites Pour l’Observation de la Terre or Earth-observing Satellites)-derived vegetation variables and field sampled woody vegetation percentage canopy cover and density. The main focus was to estimate woody density and map the distribution of woody cover in an African savanna environment. In order to select the best SPOT-derived vegetation variable that best correlate with field sampled woody variables, several spectral vegetation and texture indices were evaluated. Next, variogram models were developed: one for woody canopy cover and density, one for the best SPOT-derived vegetation variable, and a crossvariogram between woody variables and best SPOT-derived data. These variograms were then used in cokriging to estimate woody density and map its spatial distribution. Results obtained indicate that through cokriging, the estimation accuracy can be improved compared to ordinary kriging and stepwise linear regression. Cokriging therefore provided a method to combine field and remotely sensed data to accurately estimate woody cover variables.
Integrating remote sensing and geostatistics in mapping Seriphium plumosum (bankrupt bush) invasion.
(2016) Mashalane, Morwapula Jurrian.; ; Odindi, John Odhiambo.; Adjorlolo, Clement.
The impacts of plant species invasion in natural ecosystems have attracted geo-scientific studies globally. Several studies have demonstrated that the effects of invasive species can permanently alter an ecosystem structure and affect its provision of goods and services, e.g. the provision of food and fibre, aesthetics, recreation and tourism, and regulating the spread of diseases. Plant invasion causes transformation of ecosystems including replacement of native vegetation. This study focuses on invasive plant impacting on grasslands called Seriphium plumosum. The plant is known to have allelopathic effects, killing grass species and turning grazing lands into degraded shrublands. The major challenge in grassland management is the eradication and management of S. plumosum. Central to this challenge is locating, mapping and estimating the invasion status/cover over large areas. Remote sensing based earth observation approaches offer a viable method for invasion plants mapping. Moreover, mapping of vegetation requires robust statistical analysis to determine relationships between field and remotely sensed data. Such relationships can be achieved using spatial autocorrelation. In this study, Getis statistics transformed images and geostatistical techniques, which involve modelling the spatial autocorrelation of canopy variables have been used in mapping S. plumosum. Getis statistics was used to transform SPOT (Satellites Pour l’Observation de la Terre)-6 image bands into spatially dependent Getis indices layer variables for mapping S. plumosum. Stepwise multiple Regression, ordinary kriging and cokriging were used to evaluate the cross-correlated information between SPOT6-derived Getis indices transformed layer variables and field sampled S. plumosum canopy density and percentage. To select the best SPOT6-derived Getis indices to map S. plumosum, 308 spectral Getis indices transformed layer variables were statistically evaluated. Results indicated that Rook, Positive and Horizontal Getis indices are most suitable for mapping S. plumosum with 0.83, 0.828 and 0.828 importance. The most accurate Getis index obtained using 5x5 (Lag 5) moving window yielded 0.83 mapping importance. Cokriging with the most important Getis index yielded the best in S. plumosum density prediction with root mean square error (RMSE) of 25.8 compared to ordinary kriging with RMSE of 26.1 and regression with RMSE of 35.6. This study demonstrated that Getis statistics and geostatistics were successful in mapping and predicting S. plumosum. The current study provides insights critical for developing sound framework for planning and management of S. plumosum in agro-ecological systems.
Remote sensing of the distribution and quality of subtropical C3 and C4 grasses.
(2013) Adjorlolo, Clement.; Mutanga, Onisimo.; Cho, Moses Azong.
Global climate change is expected to be accompanied by changes in the composition of plant functional types. Such changes are predicted to follow shifts in the percentage cover and abundance of grass species, following the C3 and C4 photosynthetic pathways. These two groups differ in a number of physiological, structural and biochemical aspects. It is important to measure these characteristic properties because they affect ecosystem processes, such as nutrient cycling. High spectral and spatial resolution remote sensing systems have been proven to offer data, which can be used to accurately detect, classify and map plant species. The major challenge, however, is that the spectral reflectance data obtained over many narrow contiguous channels (i.e. hyperspectral data) represent multiple classes that are often mixed for a limited training-sample size. This is commonly referred to as the Hughes phenomenon or “the curse of dimensionality”. In the context of hyperspectral data analysis, the Hughes phenomenon often introduces a high degree of multicollinearity, which is caused by the use of highly-correlated spectral predictors. Multicollinearity is a prominent problem in processing hyperspectral data for vegetation applications, due to similarities in the spectral reflectance properties of biophysical and biochemical attributes. This study explored an innovative method to solve the problems associated with spectral dimensionality and the related multicollinearity, by developing a user-defined inter-band correlation filter function to resample hyperspectral data. The proposed resampling technique convolves the spectral dependence information between a chosen band-centre and its shorter and longer wavelength neighbours. The utility of the new resampling technique was assessed for discriminating C3 (Festuca costata) and C4 (Themeda triandra and Rendlia altera) grasses and for predicting their nutrient content (nitrogen, protein, moisture, and fibre), using partial least squares and random forest regressions. In general, results obtained showed that the user-defined inter-band correlation filter technique can mitigate the problem of multicollinearity in both classification and regression analyses. Wavebands in the shortwave infrared region were found to be very important in regression and classification analyses, using field spectra-only datasets. Next, the analyses were up-scaled from field spectra to the new generation multispectral satellite, WorldView-2 imagery, which was acquired for the Cathedral Peak region of the Drakensberg Mountains. The results obtained, showed that the WV2 image data contain useful information for classifying the C3 and C4 grasses and for predicting variability in their nitrogen and fibre concentrations. This study makes a contribution by developing a user-defined inter-band correlation filter to resample hyperspectral data, and thereby mitigating the high dimensionality and multicollinearity problems, in remote sensing applications involving C3 and C4 grass species or communities.