University of Twente Student Theses


Study of the effect of seasonal variation of Vegetation cover on soil erosion in the Ratamba Watershed, Banjanegara district - Central Java, Indonesia

Mboho, Uduak (2010) Study of the effect of seasonal variation of Vegetation cover on soil erosion in the Ratamba Watershed, Banjanegara district - Central Java, Indonesia.

[img] PDF
Abstract:Soil erosion is a serious global environmental problem. With an increasing human population pressure, there is an increasing expansion in agricultural practices to meet the increase on food demand. This has in turn led to increase in the risk posed by soil erosion. Therefore, it is very essential to assess soil erosion under various land use practices, so that conservation measures can be implemented in time. Firstly, it is a common practice to estimate annual soil loss using average input parameters such as the rainfall amount, rainfall interception, cover factor, ground surface, canopy cover and plant height for a year or at times series of years. This study investigates if the results obtained by this approach give the same as the sum of the monthly soil loss estimates which takes into account seasonal variation. To analyze this, the RMMF erosion model was used to predict annual soil loss using average parameters and the model was also adapted to run for monthly soil losses estimations and the results were summed up to derived the value for the total annual soil loss. The results show that the annual predicted soil loss rates using average factors were relatively higher than those using the sum of the monthly predicted soil losses. The average soil loss prediction for the study area using the annual average input factors was 23.43 t ha-1 yr-1and that of the sum of the monthly predicted soil loss was 12.01 t ha-1 yr-1. With regards to the different land uses, the prediction with the average values predicted the highest soil loss in the tobacco fields with an average soil loss of 43.56 t ha-1 yr-1 and for the sum of the monthly soil loss, the highest soil loss was in the dry land agricultural field with an average soil loss of 21.35 t ha-1 yr-1. The results also revealed that 18, 9122 tons in an annual basis of soil is lost from the study area annually using the average parameters. When soil loss was computed on a monthly basis, the results showed that 12, 0278 tons of soil is lost per year. From these results, there was a significant difference in the results derived from the two methods. This conclusion was drawn based on a statistical comparison to establish whether the differences in the different predictions are statistically significant. The quantile – quantile and box plots showed that the both had about 25% of the total predictions to be very low but the prediction from the averaged annual parameters gave extremely high values wish are probably not realistic. The predictions from the sum of monthly soil losses proved to give a realistic result. Secondly, soil loss assessment requires quantitative data. There are several approaches to acquire the data necessary for assessing soil erosion. And improved and enhanced modelling of this situation requires remotely sensed data which gives information of the surface and at times subsurface conditions of the area. And these are usually derived from satellite images and digital elevation models. The level of information derived this approach however depends on the availability, quality, resolution and also the cost of the image required. The cost implication though less frequently discussed, is also an essential important factor coupled with the resolution could limit the use of satellite images for recurrent monitoring of soil erosion especially in small areas like the study area of this work. To assess use coarse ii and readily available in small area, this study used the Moderate Resolution Imaging Spectroradiometer at a resolution of 500m for obtaining quantitative information erosion modelling parameters such as the cover management parameter based on NDVI for the 17km2 size study area. This apparently a small area and with such a resolution, it is believed that the obtained erosion parameter would be a generalized one without any true or detailed representation of the area. To evaluate the accuracy and reliability of the erosion parameter derived at this resolution, erosion modelling parameters such as cover factor was derived from NDVI values gotten from a LANDSAT image at a resolution of 30m. Aggregation functions were applied to the NDIV values from the LANDSAT to a coarser resolution as the MODIS image from which the cover factors were derived. Due to the unavailability of the LANDSAT images, this analysis carried only for one month (June, 2009), for which images were available for both the LANDSAT and MODIS images. A correlation was performed between the aggregated NDVI values of the LANDSAT image and that of the MODIS NDVI to establish a relationship. An r2 of 0.66 was obtained and that shows a good relationship and as such showed that the MODIS (500m) image which is free and cost effective can also be used for erosion modelling study in small watershed in the absence of high resolution images. But on applying it to soil loss assessment, it revealed that soil loss derived using the MODIS c-factor results overestimated the presence of vegetation cover since some fields in the area were just at the planting and vegetative stages. And when the soil loss result derived for the month of September was correlated with the soil loss prediction for the same month using the c- factor derived from the canopy cover, it gave a low value (r2=0.11) which showed that it’s extremely low prediction is not realistic. As a conclusion from this, the MODIS (500m) image which is free and cost effective can be used for Cfactor mapping but the results obtained when applied to soil loss assessments could be unrealistic and as such limits it’s usefulness for erosion assesses in small watershed. Finally, the management practice sensitivity analysis revealed that the management factor in the study area has little or no effect on the soil loss coverage but if the reverse slope terrace practice is implemented, there would be a reduction in the soil loss rate in the area.
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:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page