Prioritization of Soil Erosion Prone Areas Based on Morphometric and Land Use / Cover Parameters in River Thiririka Watershed, Kiambu County Kenya
Loading...
Date
2024-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Kenyatta University
Abstract
Morphometric studies and land use / land cover analysis play a key role in integrated
watershed management. Sustainable resource utilization at a watershed level requires
an in-depth understanding of the vegetation characteristics, land surface features, land
use, drainage and hydrological patterns of the watershed. In developing countries,
poverty have led to unsuitable land management practices (e.g. deforestation,
continuous tillage), contributing to increased runoff causing land degradation and
increased soil erosion in watersheds. This inhibits the achievement of the Sustainable
Development Goals (SDGs) of zero hunger, access to clean water, and sanitation. To
reduce soil erosion at the watershed level, watershed managers need to make informed
decisions such as developing vegetative cover, agroforestry, and terracing. However,
this is limited in Kenya due to lack of readily available data to guide the process. This
study explored the potential use of basin and drainage network properties, land use /
land cover characteristics with Geographic Information Systems (GIS) and Remote
Sensing (RS) tools to identify sub watersheds susceptible to soil erosion in Thiririka
watershed in Kenya. Five sub watersheds were delineated and assigned a code from
SW1 to SW5 using the Shuttle Radar Topographic Mission (SRTM) 30 meter
resolution Digital Elevation Model (DEM) with Arc Hydro tools in ArcGIS 10.8
software. These was followed by the analysis of morphometric parameters of linear,
aerial, and relief characteristics of the watershed. Land use / land cover classes were
generated from an annual median composite of Sentinel-2 image for the year 2020,
collected using Google Earth Engine (GEE). The training polygons were systematically
sampled from the field using handheld GPS. A supervised classification scheme was
used to develop a random forest classifier to perform the classification. In addition, the
Normalized Difference Vegetation Index (NDVI) extracted from a median composite
of Sentinel-2 image for 2020 and the SRTM-DEM were incorporated to improve the
classification accuracy. The overall accuracy was 0.88, and Kappa statistics of the
classifications was 0.86. Further, to understand the spatial distribution of water in the
catchment, the Topographic Wetness Index (TWI) values were extracted from the
SRTM DEM. The effect of land use / land cover, vegetation cover and soil moisture to
soil erosion tested using a two way ANOVA showed that all the parameters have a
positive correlation with soil erosion. Finally, the effects of morphometric parameters,
land use/ land cover, vegetation characteristics and soil moisture on soil erosion were
assessed and assigned ranks 1 to 5. The ranks assigned for all the parameters were
averaged to get the compound priority value (CP). Results showed that sub watershed
5 (SW5) and sub watershed 1 (SW1) are highly susceptible to soil erosion needing
immediate management actions, while sub watershed 4 (SW4) and sub watershed 3
(SW3) show less susceptibility to soil erosion. This study provides information on sub
watersheds exposed to soil erosion, which is important for all the stakeholders in
watershed management such as agricultural officers, farmers, planners, and
policymakers to focus the appropriate sustainable watershed management practices.
Description
A Thesis Submitted in Partial Fulfillment of the Requirements for the Award of the Degree of Master of Science (Integrated Watershed Management) in the School of Pure and Applied Sciences, Kenyatta University May 2024
Supervisors:
1. Shadrack Murimi
2. Raphael Kweyu