| dc.description.abstract | The overall aim of this study was to assess sustainable strategies of wetland resources utilisation as well as the human impact on the Yala swamp ecosystem in West Kenya. The study involved different approaches, namely (i) socio-economic surveys measurements on macrophyte growth dynamics, (iii) plant/ soil/ water analysis, (iv) remote sensing and (v) geographical information systems. Land-cover changes and resources use interactions were analysed retrospectively to delineate principal drivers and trends of the trajectories of future land-use change on the local communities, (ii) field dynamics. This was done with a view focusing on “wise use” of the wetland. Data analysis was done using a combination of Excel spreadsheets, SPSS and Genstat, which involved descriptive statistics, chi-square, parametric statistics such as regression, ANOVA and post hoc LSD.
Wetland resources utilization in the whole of Yala swamp was investigated using a structured questionnaire. The socio-economic results indicated that Yala swamp provides a wide range of support and products to the local communities. These include mats, papyrus ropes, thatch material, fish, vegetables, forage and firewood. Other uses include small-scale farming, grazing, sand and soil harvesting for brick making, plus water withdrawal. Approximately 70% of the wetland products are used at the domestic level with the rest being used to generate modest incomes. Alternative sources of income are rare, especially during the extended dry season. Marketing of wetland products is ineffective, resulting in low profit margins, which again discourage sustainable wetland use. Nevertheless, farming is an important activity, which engages 90% of the farm holdings in the swamp and supplies about 70% of the domestic food requirements. It is no surprise that with respect to future development, most of those questioned preferred swamp conversion for farming.
The field measurements encompassed harvesting biomass and post-harvest growth rates measurements of Cyperus papyrus, Phragmites mauritianus, C. dives, C. distans, Echinchloa haploclada and Typha domingensis. In addition, plant materials were analyzed
iv
for nitrogen and phosphorous calorimetrically after Kjeldahl digestion. Macrophyte postharvest
growth characteristic results indicated a high mean growth rate in the first four weeks ranging from 5 to 300%. This was followed by a lower growth rate averaging 1-30% in the next ten weeks. The less disturbed sites recorded higher mean growth rate after the first four weeks of 10-30% compared to the highly disturbed sites, which on average was 1-15%. The growth rate after the 14th week was highly diminished in all the species. Comparative data on macrophyte that were not harvested showed insignificant gain in height after the 14th week. Even during the dry season, fast growth was restricted to the first 14 weeks, but with overall reduction in average height gain, growth rate and biomass in all the eco-types. This variability was attributed to seasonal ecological dynamics and not to the effect of repeated harvesting. The average biomass was about l,050g dry wt m2, which was within other tropical papyrus wetlands. In addition, there was a 50% change between the wet and dry season. Macrophyte nutrient content was high both in the flooded and swamp edge species with an N: P ratio ranging between 6 and 3.5, which was above ecological limiting levels. These results indicate that the macrophyte can be sustainably harvested at an interval of 14 weeks if the natural ecological set up is maintained.
Ecological dynamics in the wetland were assessed through analysis of selected soil and water parameters. Laboratory analysis encompassed the following techniques: total nitrogen (N) (Kjeldahl method), total phosphorus (P) (Mehlich’s double acid method), soil exchangeable potassium (K) (flame photometer after extraction), soil pH (extraction made to a ratio of 1:2.5 with water pH), water soluble carbon (Walkley-Black method), soil cation exchange capacity (measured total cations in the leacheates), electrical conductivity (through electrode) and soil bulk density (using standard disc on natural sites).
Growth patterns of macrophyte indicated pronounced temporal and spatial trends, which correlated well with the variability in ecological dynamics in both water and soil. Ecological conditions werf more favourable for macrophyte growth during the wet season (as
v
compared to the dry season) and in the less disturbed eco-types (as compared to the highly disturbed eco-types). While soil parameters were significantly influenced by the eco-type, but only varied marginally among seasons. In contrast to the soil, the water chemistry was influenced more by the seasons relative to eco-types. Both soil total N (0.25 -0.3%) and P (0.07- 0.06%) and water P (0.03 - 0.14 mg/1) and N (3.72- 2.01mg/l) were above ecological limiting levels.
Land-cover analysis was done using Landsat satellite images taken in the dry season (February 5, 1973, MSS and February 2, 2001 ETM). Ground data included inventory of plant species and land cover distribution using a hand-held global position system (Germini- type). A hybrid classification approach was used in change detection using ERDAS 8.6. Signatures were evaluated prior to classification using the transformed divergence (TD) with post-classification covering accuracy assessment, land use change matrix and normalized difference vegetation index (NDVI). In the identified seven land-cover classes, the most prominent change was noted in over three-fold increase in the agricultural area from 1,564 ha in 1973 (corr. to 7 % of the total wetland) to 5,939 ha in 2001 (corr. to 28 % of the total wetland). However, these changes excluded temporary land use during other seasons. Based on field survey, two vegetation communities were identified, namely the bushes-sesbania community (2359.96 ha) and sedge-dives community (3435.64 ha), which are periodically converted. Most of the land conversion was located along the swamp edges and in particular on the northern and eastern side of the swamp, where accessibility was good. The satellite images also allowed identification of siltation area, which had increased, along the Lake Victoria shoreline. The overall classification accuracy was high at 75% with Kappa statistics at 70%. The NDVI showed a high reduction of the positives values, i.e., from +0.909 in 1973 to +0.405 in 2001, which was mainly due to the reduction in the vegetation cover. These changes were attributed to anthropogenic activities, mainly farming, in the swamp.
vi
Land-use change assessment involved integration of household and census data with the processed images. The scale of integration was at the administrative level of ‘Location’. The main driving factors for land use changes in the Yala swamp were identified as (i) household numbers, (ii) household and population densities and (iii) the wetland accessibility (combining swamp coverage and terrain suitability). Areas with high swamp coverage and suitable terrain had higher conversion rates. These drivers act as proxy for a whole range of factors, in particular the demand for farming land. Another major underlying factor in land use change is the high dependence of the local community livelihood on the swamp resources. The statistically computed land use change using the conversion index (11,696 ha) offered a high co-validation with the land cover changes derived from the satellite images (11,735.44 ha). This coverage comprised the directly converted areas and the vegetation communities (bushes-sesbania, sedges-cfrves) that are periodically converted. Concluding, it can be stated that under the current utilization scenario, swamp conversion is expected to increase as a function of the household densities. However, the big challenge is to balance between increasing swamp farming and sustainable ecosystem utilization like macrophyte- based water filtering. Hence, there is need to reinvest in the resource-use strategies, wetland products marketing systems and ecosystem level management to stem resource degradation and increase benefits to the local communities. | |
