University of Twente Student Theses

Login

Assessment of the herbaceous above ground biomass gain on the fringe of Lake Naivasha, Kenya

Mudereri, Bester Tawona (2012) Assessment of the herbaceous above ground biomass gain on the fringe of Lake Naivasha, Kenya.

[img] PDF
2MB
Abstract:The herbaceous vegetation in the fringe zone of Lake Naivasha is vital for providing ecosystem services to both humans and animals inhabiting the fringe. Various fauna depend on the herbaceous vegetation for forage and shelter. The fringe zone area has declined significantly in the last 30 years mainly attributed to increases in agriculture, declines in the ground and lake water levels. These water level declines have been attributed to changes in the rainfall patterns in the upper catchment and increased abstraction for irrigation of farms surrounding the lake. The increase in the number of herbivores and loss of land to agriculture has increased susceptibility of the remaining herbaceous patches to overgrazing thus causing a reduction in productivity. Though there has been studies conducted on the fringe zone vegetation, most of them have focused on papyrus productivity. Measurement of productivity is critical as it indicates the net carbon assimilation from the atmosphere. Productivity can also explain land surface area conditions like change in water and nutrient availability among many other ecological processes. This research aimed to determine how the biomass gains varied spatially as influenced by changes in the fringe zone including water levels. To achieve this, an ASTER image was classified using a combination of automated isodata clustering and expert knowledge to separate the herbaceous classes from other natural and semi natural vegetation classes. The disk pasture meter was used to estimate the standing biomass for both June and September sampling periods. Biomass gain was the difference between the total biomass accumulated between the successive periods. This was compared between different herbaceous communities and inherent characteristics of the dominant species. The effect of lake level was analysed by establishing the effects of frequency of inundation and the ground water level on the biomass gained. A least squares regression analysis was conducted to determine if there were interaction effects in the explanatory variables for biomass gain. We found that 17 classes comprising three separate herbaceous classes could be discriminated successfully at 90 % overall accuracy and a kappa coefficient of 0.89. Expert knowledge and use of ancillary data in classification improves classification of satellite images as compared to using only pixel based approaches. The overall standing biomass in the fringe zone was 381 g/m2 and an average decrease in biomass gain of 3.7g/m2 . The average grass height in the fringe zone of Lake Naivasha was 5.5cm with Cynodon dactylon and Kikuyu grass (Pennicetum clandestinum) dominating in most of the area especially in the north. There was no significant difference (p>0.05; CI=0.95) in the medians of the standing biomass in June and September using a Wilcoxon rank test, which could have been a result of season position of the study period. Although descriptive statistics showed variations among the different herbaceous groups, a one way analysis of variance established that there was no significant difference (p>0.05; CI=0.95) in biomass gain among the groups. Frequency of inundation has influence on biomass gain as evidenced by the one way ANOVA significant differences (p<0.05 among the frequency of inundation groups). A Tukey Kramer post hoc test established that patches inundated for 20-60% times a year and those less than 20% have significantly different productivity levels. The lake level and ground water are highly correlated for an estimated distance of 9.5 km from the lake centre. Ground water has no significant influence (p>0.05) on the biomass gain between the period June to September. Results of the test for interaction effects using regression analysis established that the grassland type, frequency of inundation bulk density and elevation are the best explanatory variables for biomass gain yielding an average R2 of 0.4. The model can be improved in successive studies by increasing the accuracy of productivity estimation and include other variables like precipitation and grazing that seem to have very profound effects on herbaceous biomass gain. We conclude that lake water level has influence on the biomass gain of grasslands directly on the margin of the lake influenced by frequency of inundation. The fluctuations in water levels affected the herbaceous biomass gain however, further study on how level and period of inundation influence biomass gain could be further explored.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:https://purl.utwente.nl/essays/93691
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page