LAND USE AND LAND COVER CHANGES ON A TROPICAL LAKE BASIN: THE SOCIOECONOMIC DRIVERS AND CASCADING IMPACTS ON HYDROLOGY AND ECOLOGICAL SYSTEM
Odongo V.O1†., Dawit W. Mulatu1, Muthoni F.K1., Ndung’u J.,456 Meins F4., Mudereri, B. T1., van der Tol1, Becht R1., Su B1., van Oel P.R1., van der Veen A1., Onyando J. O2., Augustijn D4., Hulscher S. J. M. H. 4, Kitaka N5 ., Mathooko J5., Otiang’a-Owiti, G. E3.,Groen, T. A1, Skidmore, A. K1
1
University of Twente, Faculty of Geoinformation Science & Earth Observation (ITC), P. O. Box 217, 7500AE Enschede, The Netherlands.
2
Egerton University, Department of Agricultural Engineering, P. O. Box 536-20115, Egerton, Kenya.
3
Kenya Wildlife Service Training Institute (KWSTI), P. O. Box 842-20117, Naivasha, Kenya. 4
University of Twente, Faculty of Engineering Technology Water Engineering and Management P. O. Box 217, 7500 AE Enschede, The Netherlands.
5
Egerton University, Department of Biological sciences, P. O. Box 536 -20115, Njoro, Kenya. 6
Kenya Marine and Fisheries Research Institute, P. O Box 81651-80100, Mombasa, Kenya
Abstract
Lake Naivasha experiences frequent lake level fluctuations despite being on the decline in the last three decades. The main possible cause has been postulated to be increased abstraction around the lake. However, land use land cover changes (LULC) in the basin may be impacting on the level fluctuations and decline. The LULC interfere with runoff, evapotranspiration, and infiltration conditions of a catchment. The frequent lake levels fluctuations resulting from LULC impact significantly on riparian vegetation productivity and species composition. In addition, it hasimpact on the aquatic ecosystem water quality especially in terms of turbidity. Lake Naivasha basin has experienced significant LULC transformations predominantly caused by socio-economic drivers. Implications of past, present and future patterns of socio-economic drivers of LULC is vital to the understanding of social, ecological and limnological functioning of the basin. These factors are proposed to impact on the entire hydrological regime of the lake. In this study we present first results of the Earth Observation Integrated Assessment (EOIA) project for the governance of Lake Naivasha basin using Interdisciplinary approach that applies GIS and RS towards the understanding of the dynamics of Lake Naivasha ecosystem.
1. Introduction
Land use and land cover (LULC) are considered as one of the most important components of the terrestrial environment system (Lin et al., 2009). Changes in LULC mirror the impacts of human activities on the global environment (e.g. Houghton et al., 1999; Schneider and Eugster, 2005). At global scales these changes have been known to have impacts on continental and global atmospheric circulation leading to even larger impacts on regional and continental climate (Lambin and Geist, 2006). Numerous studies have investigated the complex relationships between land surface and other components of the climate at the local to global scales, detailing the differences in magnitude of land surface changes in different geographic localities over the Earth (e.g. Betts et al., 1996; Pielke et al., 1998; Pielke et al., 2002). Based on these studies, there is evidence that large-scale LULC changes, particularly in the tropics, generate remote climatic effects of global extent far from where the surface has been directly affected by land-cover changes (Franchito and Rao, 1992; McGuffie et al., 1995; Zhang et al., 1996; Pielke et al., 2002). In particular, these changes impact greatly on the hydrology, forestry, agriculture, ecology and socio-economic conditions of a region. Most significant is the growing human population that exerts increasing pressure on the LULC, as demand multiplies for resources such as food, water, shelter, and fuel. These socioeconomic factors often dictate how land is used regionally as well as locally (Muttitanon and Tripathi, 2005). RS and GIS have been widely applied and recognized as powerful and effective tools in detecting the spatio-temporal dynamics of LULC changes. Moreover, RS can provide valuable multi-spatio-temporal data for monitoring LULC patterns and process (Mertens et al., 2000; Reid et al., 2000; Campbell et al., 2005; Xie et al., 2005). The application of RS and GIS has been a new trend of the international research on LULC to explore the driving forces and to establish an approach of driving forces of LULC change using earth observation (EO) and RS techniques. The two major categories of driving forces of LULC change are: bio-physical and socioeconomic drivers. Socioeconomic driving factors include population change, economic development level, technological progress, political and economic structure and value concepts (Briassoulis, 2000). Therefore, it is vital and timely to assess major socioeconomic drivers of LULC changes in developing countries and their impact on sustainable development of their economy and environment. The present study attempts to investigate these impacts for the Lake Naivasha Basin, Kenya.
Lake Naivasha has been subject to wide fluctuations in water levels over time and is said to have almost dried in the past years (Gaudet, 1977; Abiya, 1996; Verschuren et al., 2000). These natural fluctuations, coupled with consumption by humans, changes in land use/cover over time and climate variability have led to decrease of the water levels which have led to shrinking of the lake (Becht and Harper, 2002a; Ondimu and Murase, 2007; Otiang'a-Owiti and Oswe, 2007; Olaka et al., 2010; Trauth et al., 2010). Consequently, the shrinking of the lake has made the lake ecosystem vulnerable and its fragility is a challenge to conservationists and scientists. The lake is a RAMSAR1 wetland and supports significant economic activities. These include fishing, irrigation, agriculture, geothermal power generation, domestic water supply, sewage effluent disposal and tourism. However, these hydrological benefits could be threatened by land use/cover changes exacerbated by socioeconomic driving forces. Hence, the objective of this study was to quantify the LULC in Lake Naivasha Basin and elucidate its impact and linkage to hydrology, ecology and socio-economic conditions of the basin.
2. Study area
The Lake Naivasha Basin is situated in the Kenyan Rift Valley approximately 70 km from Nairobi at a latitude of 0o 09′ to 0o 55′S and longitude of 36o 09′ to 36o 24′E. The maximum altitude is about 3990 m above mean sea level (a.m.s.l) on the eastern side of the Aberdare Ranges to a minimum altitude of about 1980 m (a.m.s.l). The catchment area is approximately 3500 km2 and is part of the Lake Naivasha basin (Figure 2.0.1).
The major soils in the study area are of volcanic origin. The soils found on the mountain and major escarpments of the catchment are developed from olivine basalts and ashes of major older volcanoes. They are generally well drained, very deep (1.2-1.8 m) and vary from dark reddish brown to dark brown, clay loam to loamy soils with thick acid humic topsoil in shallow to moderately deep and rocky places (Rachilo, 1978; Nyandat, 1984). Climatic conditions in the study area are quite diverse due to considerable differences in altitude and relief. The annual mean temperature ranges from 8 oC to 30 oC. The rainfall regime within the basin is influenced by local relief with the catchment being in the rain shadow of the Aberdare ranges to the East and the Mau Escarpment to the West. There are two rainy seasons experienced in this catchment. Long rains occurring in the months of March to May and the short rains experienced between October and November. The Lake Naivasha basin experiences an average annual rainfall of 610 mm, and the wettest slopes of the Aberdare ranges receive as much as 1525 mm. Figure 2.0.2 summaries the monthly average precipitation and temperature variations in the Lake Naivasha basin.
Figure 2.0.2: Monthly climate average distribution of temperature and rainfall for Lake Naivasha 3. Methodology
3.1 Analysis of Land Use/Cover Data
Data used for image classification were Landsat MSS of 31st January 1973, Landsat TM 17th January 2011, ASTER of 14th March 2011, Worldview 2 of 17th December 2010. These images were chosen on the basis of availability and less cloud cover. A stratified random sample of 302 ground reference data of major LULC spaced at a minimum distance of 1 km from each other were collected with the aid of a GPS supported with ground photo evidence for species identification later. Aerial photos of August 2010 acquired from Department of Remote Sensing and Resource Survey (DRSRS) of Kenya were also used to support interpretation and extraction of extra ground reference data.
Image classification and accuracy assessment
Unsupervised classification was conducted on all the images using ISODATA algorithm with an initial set of 50 classes. The Jeffries-Matusita (J-M) class separability test was also performed to distinguish different classes based on their spectral profiles.
Ground reference data collected during the study (1st January 2012 to 30th March 2012) and aerial photos of August 2010 were adopted to distinguish classes for supervised classification of Landsat TM, ASTER and World View 2 images representative of the year 2011. Half of the reference data together with the results from unsupervised classification were used to develop regions of interest (ROIs) representing different land use/cover classes. The ROIs were then used in training the maximum likelihood classifier to come up with 12 main dominant land use/cover classes of Lake Naivasha Basin for the year 2011 (Figure 4.1.1). For the 1976 Landsat MSS image, ROIs were developed by unsupervised classification using ISODATA algorithm together with vegetation map of 1976 published by the British Ordnance Survey. Forthwith, a maximum likelihood
classification was conducted using the ROIs that led to 8 dominant land use/cover classes of Lake Naivasha Basin representative for the year 1973 (Figure 4.1.1).
Half the reference data and the vegetation map of 1976 were then used to conduct the accuracy assessment of the classified images of 2011 and 1973 respectively.
Change detection
To detect in detail the process of LULC in the Lake Naivasha Basin during 1973-2011, an analysis of land use/cover change matrix was conducted through spatial overlay of the two land use/cover maps. The spatial overlay was based on a pixel by pixel comparison following the map algebraic formula shown by Equation 1.
Cij=Aij*1000+Bij (1)
Where Cij is the change detection map between time A and B at spatial distribution position i,j. Each pixel for Land use/cover at time A was multiplied by 1000 to warrant comparison during addition such that when LULC at time A at pixel i,j is 1000 and the corresponding LULC at time B at the same pixel position is 1 then, the Cij for that position is unchanged (1001) between the two times. If the change map pixel reads 1002 it means there has been a change from land use/cover type 1 to 2 between the two periods. Table 4.1.1 shows the change detection matrix between the year 1973 and 2011.
3.2 Hydrology
Assessment of annual runoff
LULC changes affect the surface conditions which inherently affect the conversion of precipitation to runoff. To evaluate this, runoff coefficient (Equation 2) was used to assess the evolution of surface conditions in transforming precipitation to runoff.
(2)
Where C is the annual runoff coefficient, Q is the total runoff volume (mm), and P is the total annual precipitation (mm) received in the basin. Discharge data of the last outlet gauging station (2GB01) entering the Lake and 27 precipitation gauging stations within the basin from 1960 to 2010 were used for this analysis.
3.3 Socio Economic
Time-series socioeconomic data related to drivers of LULC changes were elaborately reviewed through both statistically published and unpublished annual reports in order to set the socioeconomic factors for analysis. Because of the limited published annual time-series socioeconomic data and differences between administrative boundaries, there was a challenge to collect data on basin level. As a result, to compile a time-series socioeconomic data at basin level for selected socioeconomic variables required visiting of various local and national offices. These data were collected from the Kenya National Bureau of Statistics (KNBS), Kenya Horticultural Crops Development Authority (HCDA), Kenya Ministry of Agriculture, Districts Ministry of Agriculture offices covering the Lake Naivasha basin (i.e. Naivasha, North Nyandarua, South Nyandarua and Olkalou) and Naivasha Municipality Council Bureau. Finally, the following time-series socioeconomic data were collected for the Lake Naivasha Basin: population, urban population, rural population, economically active agricultural population, major cereal production (wheat, maize and beans), flower export volume and Naivasha town municipality total expenditure and income from different sectors of the economy.
The LULC change dynamic degree can describe the speed of the local LULC change. The LULC dynamic falls into single LULC dynamic degree and integrative one (Liu et al., 2003; He et al., 2009). In this paper, a single land use/cover dynamic degree was employed to determine the change rate of certain land use/cover type in a fixed study period and computed using Equation 3.
%
100
1
T
U
U
U
K
ai ai bi i (3)Where K is the change rate of a certain land use/cover type i in a fixed study period; Uai and
Ubi is the area of the land use/cover type i at the beginning and the end of the study period, respectively; and T is the study period. The result of this equation is the annual change rate of land use/cover type i. Hence, the time-series land use/cover data for each land use/cover type can easily be extracted. The major socioeconomic forces triggering LULC changes were explored using statistical analysis. Pearson correlation matrix was carried out to explore the correlation between these socioeconomic drivers of LULC.
3.4 Ecology
Eleven line transects measuring 450m long were established in 5 ranches along the fringe of the Lake based on land cover, accessibility and evolutionary history of grazing. Twenty plots were located on formerly fringing swamp that had dried due to decline of lake levels and channelization of river Malewa. The other 24 plots were located on traditional grazing lands fringing the lake. Each transect had 4 plots measuring 30 m2 at an intervalof 150m. In each plot herbaceous biomass was estimated on June and September 2011 using a calibrated Disc Pasture Meter (DPM) in 34 plots after excluding 10 which could not be measured in September due to flooding. The vegetation communities along the fringe zone were delineated after a semi-automated classification of ASTER satellite imagery that was taken on 22nd September 2011. After delineation of main vegetation communities, four classes of herbaceous vegetation types were derived. The 34 sampling plots were overlaid on classified map to determine their category. The distribution of sampling plots in different herbaceous vegetation types was; the temporary flooded grasslands (n=6), dense grass sward of medium height (n=9), herbaceous vegetation on shrub-land (n=13) and herbaceous vegetation under forest canopy (6). The herbaceous productivity between the two time periods was measured as the biomass gain following “method 3” of Scurlock et al. (2002): (
) where is biomass gain and is herbaceous aboveground herbaceous
biomass at time period A one way ANOVA test was used to test for statistical differences on biomass gain between the four herbaceous vegetation types.
Moreover vascular plant species within each plot were identified together with a visual estimate of their percentage coverage. Trees and shrub species were inventoried within the whole 30m2 plot while five replicates of 1m2 quadrants were used to inventory the herbaceous species in each plot.. To explore the factors that drive the vegetation composition along the fringe zone four classes of explanatory variables were collected namely: (1) disturbances regimes: grazing intensity measured as herbaceous aboveground biomass (AGB) (Aarrestad et al., 2011), the frequency of inundation and fractional bare soil cover (2) soil resources: N, P, K and Ph, (3) light resources: fractional tree cover (FTC), fractional bush cover (FBC) and water resourcesmeasured as Topographical index (TPI) (Weiss, 2001). The sites were classified into two; the occasionally inundated former swamp and never inundated. The canonical analysis of principle coordinates (CAP) (Legendre and Anderson, 1999) with Bray-Curtis ecological distance was used to explore variation in species composition along the environmental gradients. The best CAP model was selected as the
one with the lowest deviance after sequentially omitting one variable with least explained variance and low contribution to the loadings of the first two CAP axes.
3.5 Water quality
The spatio-temporal variation in turbidity in Lake Naivasha was estimated using secchi depth measurements. A 20 cm diameter Secchi disk was lowered into the water while unwinding a waterproof tape attached to it to the point where the disk just disappears. The disk was then raised until it reappeared. The average depth of the water where the disk disappears and reappears was recorded as the secchi depth reading. This was done in six different locations in the lake Figure 3.5.1. The measurements were taken weekly from 20th January 2011 to 26th May 2011. A time series graph was plotted to show the temporal variation of the secchi depth in the studied sampling sites.
4. Results and Discussion
4.1 Analysis of Land Use/Cover
LULC analysis indicated that the grassland was the dominant cover type occupying about 27% and 45% in 1973 and 2011 respectively (Figure 4.1.2). This is typical of a savannah ecosystem which is known to be dominated by grasses. The 18% increase of grassland in this period was attributed to decline in bushland as shown by the conversion matrix in Table 4.1.1. Between 1973 and 2011 bushland cover reduced by 24%. Of this amount 18% of bushland has been converted to grassland. The rest have been converted to farmland (1.7%), shrubland (2%), woodland (3%), fallowland (2%) with builtup, horticulture irrigation and forest each taking up <1%. Farmland has more than doubled (3.5%) over this period while forest cover has declined by 3%. However, woodland has seen a growth of 2% in this period. This may be attributed to the increased tree planting campaigns in the basin mainly by non-governmental organisations (NGOs) e.g. World Wide Fund for Nature (WWF), Green Belt Movement and lately Imarisha Naivasha. Moreover the increased economic value of timber may have motivated residents to plant trees as a source of income for their livelihood. Aquatic vegetation has declined by 0.6% in this period. This is evidenced by recent observations of relative decline of papyrus vegetation in the Northern swamp of Lake Naivasha (Harper and Mavuti, 2004). The aquatic vegetation decline could also be explained by the expansion of informal settlements in the riparian area causing degradation of the Lake ecosystem. Over the years beginning early 1980s there has also been an upsurge of built up areas, horticulture farms around the lake and increased irrigation activities which were minimal in the 1970s. Overall, 65% of LULC in Lake Naivasha Basin has transformed to a different LULC class over the period of 38 years.
4.2 Assessment of runoff
The results of runoff generation of the basin indicate that the runoff conditions have undulated over time with minima and maxima runoff coefficient cycles occurring between 2 to 4 years (Figure 4.2.3). Interestingly, the minima and maxima runoff coefficients show a downward and upward trend respectively. This tapered-like pattern is suggestive that during wet years there has been an upward trend in runoff volumes whereas during drier years the basin has experienced downward trend in runoff volumes than in the 1960s and 1970s. Figure 4.2.4 shows the relationship between runoff coefficient, annual total runoff and annual total precipitation. The results suggest that the relationship between annual runoff and runoff coefficient was stronger (r2=0.87) compared to that of precipitation (r2=0.26). This is suggestive that the precipitation conditions in the basin have remained fairly the same over the last four decades. However, the situation is different with runoff conditions which exhibit high correlation with runoff coefficient indicative of surface changes exacerbated by LULC. The above evidence is collaborated statistically by runoff and precipitation results between 1961-1985 (when there were minimal LULC changes) and 1986-2010 (when there were significant LULC changes). The mean seasonal runoff volumes during the long-rainy period (March, April, May, June) between these two periods have increased significantly (p<0.01) by up to 13% (See also Figure 4.2.5) even though the precipitation amount between the two periods has remained significantly unchanged. However, caution should be taken in interpretations of these results since further analysis using
remote sensing and hydrological models are much better placed to confirm such an interpretation. It would be feasible to examine the evolution of evapotranspiration for the different LULC classes over the period of the changes. These evapotranspiration results, together with precipitation and runoff results reported here, are proposed to give insightful hydrological evidence that goes a long way to providing important information for stakeholders in the basin towards improved water governance of Lake Naivasha Basin. This is still an ongoing part of this research.
Figure 4.1.2: Lake Naivasha Basin per cent LULC cover for year 1973 and 2011 0 5 10 15 20 25 30 35 40 45 50 aq u atic v e ge ta tio n b u sh lan d far m lan d fo re st gras slan d sh ru b lan d w at e r w o o d lan d b u iltu p h o rticu ltu re irrig ati o n fall o w lan d
%
co
ve
r
LULC 1973 LULC 2011Table 4.1.1: Land use/cover change matrix
LULC 2011
aquatic vegetation
bushland farmland forest grassland shrubland water woodland builtup horticulture irrigation fallowland TOTAL
(%) LU LC 1 9 7 3 aquatic vegetation 0.23 0.10 0.05 0.46 0.03 0.05 0.22 0.13 1.26 bushland 0.00 2.32 1.69 0.20 18.11 1.92 0.01 3.06 0.26 0.19 0.36 1.98 30.08 farmland 0.05 0.83 0.05 1.92 0.23 0.00 0.49 0.00 0.00 0.06 0.05 3.69 forest 0.26 0.54 10.32 2.27 2.09 0.00 0.68 0.00 0.00 0.01 16.18 grassland 0.00 1.96 2.26 1.15 14.91 1.68 0.01 2.90 0.11 0.19 0.28 1.47 26.92 shrubland 0.00 0.41 0.51 1.10 2.44 0.46 0.00 1.97 0.00 0.01 0.07 0.08 7.04 water 0.42 0.02 0.04 0.00 0.20 0.01 3.56 0.08 0.00 0.02 4.35 woodland 0.00 0.58 1.24 0.23 4.49 0.51 0.00 3.15 0.01 0.01 0.04 0.21 10.48 TOTAL (%) 0.65 5.69 7.14 13.05 44.80 6.93 3.63 12.56 0.38 0.40 0.80 3.94 100.00
R² = 0.8667 1000 1500 2000 2500 3000 3500 4000 4500 0 0.05 0.1 0.15 0.2 Ann u al t o ta l r u n o ff ( mm ) Runoff coefficient R² = 0.2609 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 0 0.1 0.2 A n n u al t o tal p re ci p itat io n (m m ) Runoff coefficient
Figure 4.2.5: Monthly average runoff for periods 1 (1961-1985) and 2 (1985-2010) for Lake Naivasha
Basin consistent with LULC changes.
4.3 Socio Economic
The socioeconomic driving forces for changes in built-up areas (rural and urban settlement areas) were analyzed using Pearson correlation matrix. The correlation matrix exhibited nine variables that were entered into the bivariate correlation analysis. These variables were forest cover/km2, farmland/ km2, grassland/ km2, total population, total output volume of major cereals/Ton (totalmainc~n), total output volume of cut flowers/Ton (flowerexpo~e), total income of the Naivasha town municipal/Million KSH (totalearni~n), and total expenses of the Naivasha town's municipal /Million KSH (naivashamu~n).
The growth of cities following urban economic development and population concentration, and rural urbanization based on the growth of smaller towns in rural areas. The first trend is typical in Naivasha town due to the proximity of Naivasha to Nairobi, i.e. the capital city of Kenya. The town is a popular destination for local and international tourists, and the presence of large-scale horticultural farms around the lake. Naivasha town is the main corridor for national and international road transport communications and cargo transport (import and export) from and to port of Mombasa. The continued growth of the town also requires improvement in infrastructure and other public services, as a result the town expenditure increased by 134% within the last three decades. The Naivasha municipality expenditure has a strong and positive relationship with total population of the basin, and the municipal income from different sectors of the economy (Table 4.3.1). The settlement areas in the Lake Naivasha basin could not be identified by remotely-sensed image due to
0 50 100 150 200 250 300 350 400
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
A ve rag e m o n th ly r u n o ff (m m ) Month Period 1 (1961-1985) Period 2 (1986-2010)
the numerous but small rural towns. However human settlement area was 4.91 km2 in 1973 and 12.51km2 in 2011. This trend showed 0.02% annual increment with direct impact on the surrounding land use/cover while also increasing pressure on the Lake Naivasha ecosystem. These can be confirmed by mushrooming of the three new informal settlements within the riparian boundary that were converted from shrub land, woodland and grassland respectively. The built-up areas, forest cover, grassland and farmland use/cover trend for the Lake Naivasha basin from 1973 to 2011 are presented in Figure 4.3.4.
Table 4.3.1: Pearson correlation matrix for socioeconomic and remotely-sensed data for the
Lake Naivasha basin
The second trend is rural urbanization based on the growth of smaller towns in rural areas mainly in the upper catchment area of the basin. Currently, small rural centres are developed in rural areas and the newly established district offices would advance the urbanization process of the basin area. The new district offices supported by the national government to improve their facility, as a result these areas have a tendency to be a town centre within a short period of time and aggravate the pressure on land resource. The number of rural settlement area had increased. The rural settlement expansion and/or town development certainly explained in the Lake Naivasha basin and it is revealed by a continued decline in forest cover by 2.31% per year and the cultivated land area increased annually by the rate of 1.29%. The result supported by the negative correlation between built-up areas and forest cover and with the total output volume of major cereals in the basin. The positive correlation with grassland (correlation coefficient = 0.9986) and with farmland (correlation coefficient = -0.9980) has also supported the argument of rural settlement expansion in the Lake Naivasha basin. Moreover, the result indicated that built-up area expands positively with Naivasha municipal income and expenditure with the correlation coefficient 0.8754 and 0.8202, respectively. The result illustrated in the correlation matrix (Table 4.3.1) and the scatterplot matrix (Figure 4.3.2). The remotely-sensed data on the LULC change analysis revealed the decline in forest cover. The forest cover declined from 531.96 km2 in 1970’s to 427.39 km2 in 2011while settlement areas increased from 4.66 km2 in 1970’s to 12.51 km2 in 2011. The result also supported by an inverse relationship between built-up areas and forest cover (correlation coefficient = -0.9816). Thus, the Pearson correlation result confirmed that
grassland 0.9980 -0.9916 0.9556 -0.2832 0.9866 0.8491 0.7950 0.9999 1.0000
farmland 0.9986 -0.9902 0.9582 -0.2805 0.9855 0.8537 0.7994 1.0000
totalearni~n 0.8202 -0.7341 0.8839 -0.1340 0.7336 0.8822 1.0000
naivashamu~e 0.8754 -0.7857 0.8981 -0.2119 0.7889 1.0000
totalpopul~n 0.9782 -0.9896 0.9151 -0.2977 1.0000
totalmainc~n -0.2670 0.3138 -0.2035 1.0000
flowerexpo~e 0.9697 -0.9145 1.0000
forest -0.9816 1.0000
builtup 1.0000
builtup forest flower~e totalm~n totalp~n naivas~e totale~n farmland grassl~d
the five socioeconomic variables are significantly determined the driving forces for changes in built-up area conversion in the Lake Naivasha basin.
Figure 4.3.2: The scatter plot matrix for socioeconomic and remotely-sensed data for the
Lake Naivasha basin
As population increases, the demand for food increases. This demand can most easily be satisfied by increasing the cultivated area into previously uncultivated areas or by migration into unsettled areas. This is conceivable in a very low population density area, where large areas are still available for agriculture expansion (Mertens et al., 2000). However, in the Lake Naivasha basin a presence of high population and competition to possess arable land between large-scale commercial farms around the lake and small-scale agricultural practice in the upper catchment of the basin results a pressure on land resource. The lake Naivasha population trend for the last three decades presented in Figure 4.3.3.
built up forest Flow er export volume Total Main Ceral production/ton Total Population Naivasha Municipality Expenditure Total Earning of Naivasha Tow n farmland grassland 6 8 10 12 6 8 10 12 200 300 400 500 200 300 400 500 0 5.0e+07 1.0e+08 0 5.0e+07 1.0e+08 0 50000 100000 0 50000 100000 200000 400000 600000 200000 400000 600000 0 1.0e+08 2.0e+08 0 1.0e+08 2.0e+08 0 1.0e+09 2.0e+09 3.0e+09 0 1.0e+092.0e+093.0e+09 150 200 250 150 200 250 1000 1200 1400 1000 1200 1400
200 400 600 800 1000 1200 Are a i n Sq .K m 1970 1980 1990 2000 2010 Year forest farmland grassland
Figure 4.3.3: Population in the Lake Naivasha basin (KNBS, 2010)
The population had a positive relationship with built-up area (correlation coefficient = 0.9782). The negative relationship of population with forest cover revealed that an increase in population would have an impact on the forest resource in the Lake Naivasha basin (Correlation coefficient = -0.9896). Except for the total output volume of major cereals, the forest cover has a high inverse relationship with the other variable in the Pearson correlation matrix. The forest covers around the Aberdare area was fenced to protect the resource spearheaded by conservation groups especially the Rhino Ark project where the settlement expands up to the edge of the forest area. This shows that smaller towns in the upper catchment rural areas are spread out and would exert pressure on the protected forest cover due to an increase in population. In addition, the major cereal production volume in the Lake Naivasha basin contends on the land resource in order to maintain the production level and it has a negative correlation coefficient, except for flower production volume. As agricultural output increase, the demand for land increase will also increase and it has an impact on LULC. Thus, the result indicated that demographic factors have significant impact on LULC in the Lake Naivasha basin.
Figure 4.3.4: Built-up areas, forest cover, grassland and farmland trend in the Lake Naivasha
basin from 1973 to 2011. 0 200000 400000 600000 800000 Po pl at io n C ou nt 1980 1990 2000 2010 YEAR
Total Urban Population In LN Basin Rural Population in LN basin Total Population 4 6 8 10 12 Are a i n Sq .K m 1970 1980 1990 2000 2010 Year built up
4.4 Ecology
The mean biomass gain (productivity) was -37.0, -23.5, 28.4 and 66.1 gm-2 for herbaceous vegetation on shrub-land, dense grass sward of medium height, temporarily flooded grasslands and herbaceous vegetation under forest canopy respectively. However the one way ANOVA revealed no statistical difference in biomass gains between the herbaceous communities. This may be due to the short period between measurements, limited samples and the fact that the biomass gain did not account for total productivity since significant biomass fluxes such as litterfall, decomposition and herbivory were not considered. Nevertheless the temporarily flooded grasslands had the highest biomass gain despite the fact that they are highly preferred by grazers due to their abundant and nutritious forage. Thus the higher productivity in this community maybe attributed to frequent flooding that increase soil moisture and nutrients in alluvial deposits (Kitaka et al., 2002). The LULC conversions in the upper catchment of Lake Naivasha has been linked to increase in river discharge and consequent frequency of inundation (Becht and Harper, 2002b), inflow of sediments and nutrients (Everard et al., 2002; Kitaka et al., 2002).
The negative gains in herbaceous vegetation on shrub-land may be attributed mainly to intense competition for resources (water, nutrients and light) with the shrubs while that in the dense grass sward of medium height was mainly attributed to prevalent intense grazing. The positive gain on the forested herbaceous communities may be linked to the dominance by two unpalatable and shade tolerant herbs; Achyranthus aspera L. and Hypoestes forkeolii
(Vahl.) R.Br. Thus their shade tolerant and non-palatability character coupled with limited
grazing enabled them to exploit the moist and nutrient rich soils under forest canopies. Overall, the frequency of inundation (FoI), grazing intensity and topographical index (TPI) contributed significantly in CAP model (Table 4.4.1). Moreover the frequency of inundation (FoI) and grazing intensity explained the highest variance on species composition for sites located on former North swamp (Figure 4.4.1). In contrast, availability of light resources and soil PH explained the highest variance on species composition on sites located on traditional grazing lawns (Figure 4.4.1). The significant contribution of the of the frequency of inundation on the species composition on the recently formed grazing lawns after drying of the fringing swamp along the fringe zone further confirms the importance of the processes that regulates the fluctuations of the lake levels in driving biodiversity patterns along the fringe zone. If hypothesis that the fluctuations in lake levels are largely driven by LULC changes in the upper catchment is confirmed, the results therefore demonstrate the potential of the processes in the upper catchment impacting significantly on biodiversity of the fringe zone despite the two ecosystems being spatially separated. This offers support to the on-going payment of the ecosystem services (PES) programme that encourage the lower catchment resource users to extend incentives to the upper catchment resource users to institute appropriate environmental conservation programmes eventually aimed at conserving biodiversity on the lower catchment.
Table 4.4.1: Significance of environmental variables on species composition in a CAPscale model Variables Df Var. F Pr(>F) PH 1 0.2689 0.9959 0.434 K 1 0.2870 1.0628 0.354 TPI_300 1 0.542 2.0073 0.003* Log_FTC 1 0.6833 2.5307 0.013* Log_FBC 1 0.2835 1.0501 0.366 Log_FBS 1 0.4521 1.6743 0.081. Log_AGB 1 0.6263 1.3194 0.014* Log_FoI 1 0.7521 2.7855 0.005** Residual 32 8.6405 Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Figure 4.4.1: Ordiplot for species composition using Bray-curtis distance for sites with short
(left hull) and long (right hull) grazing histories.
-1.5 -1.0 -0.5 0.0 0.5 1.0 -1 .0 -0 .5 0 .0 0 .5 1 .0 CAP1 C A P 2 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + PH K TPI_300 Log_FTC Log_FBC Log_FBS Log_AGB Log_FoI -1 0
4.5 Water quality
Figure 4.5.1 shows a time series plot of the secchi depth in Lake Naivasha from January to May 2011. Mouth of Malewa sampling site generally showed the lowest values throughout the study period. This site was situated at the lake’s inlet and therefore gives an indication of the state of the discharge from the input rivers specifically River Malewa. This may be attributed to the changes in landuse and landcover which may result in an increase in runoff which carries particulate matter that end up in Lake Naivasha as suspended solids. These eventually leads to high sedimentation rate which is attributed to human activities on the lake Naivasha catchment (Stoof-Leichsenring et al., 2011, Mergeay, 2004). Crescent sampling site had the highest values meaning that it's the clearest part of the lake. This may be due to the limited connection of Crescent Lake to the main lake especially during low lake levels which may prevent the suspended material from reaching the Crescent Lake. The rest of the sampled sites in the lake were moderately turbid and somewhat similar. In reference to the previous studies, it suffices to mention that secchi depth readings have been on the decline from 250 cm in 1979, 50–75 cm in 1997–98, to not more than about 20 cm in 2001 and 2003 (Mavuti & Litterick, 1981; Kitaka et al., 2002; Ballot et al, 2009). In 2011, this study revealed much lower readings of up-to to 10 cm especially at the Mouth of Malewa site.
Figure 4.5.1: Time series graph of secchi depth from sampling stations in Lake Naivasha
0 20 40 60 80 100 120 140 0 9 /0 1 /1 1 2 9 /0 1 /1 1 1 8 /0 2 /1 1 1 0 /0 3 /1 1 3 0 /0 3 /1 1 1 9 /0 4 /1 1 0 9 /0 5 /1 1 2 9 /0 5 /1 1 1 8 /0 6 /1 1 Se cc h i D e p th (c m ) Date 1. Crescent 2. Hippo 3. Kamere 4. South East 5. Mid lake 6. Mouth of Malewa
Conclusions
Approximately 65% of LULC in Lake Naivasha Basin has been transformed to different LULC class over the last four decades. Statistical analysis suggests that rapid settlement expansion (built-up area), agricultural land expansion and population growth are the major driving forces of the LULC in the Lake Naivasha basin over the period between 1973 and 2011. As a consequence of LULC changes, the surface conditions have been transformed thus impacting on runoff generation from the basin. Accordingly, such impact does affect the frequency of floods downstream which is essential in maintaining the highly productive temporarily flooded grasslands that act as a key forage resource especially in dry season when large herbivores aggregate along the fringe due to scarcity of forage in their wet seasons range. Moreover, the frequency of inundation had significant impact on plant species composition in recently formed grazing lawns after drying of the former North swamp. Thus processes in one ecosystem can influence biodiversity patterns in another spatially separated ecosystem through a mobile linkage mechanism such a river discharge. This offers empirical basis for the implemented PES programme between the lower and upper catchment resource users in a bid to conserve the fringe biodiversity. The use of remotely sensed data, GIS, and regression analysis can allow policy makers to have a better understanding of the causes and effects of the changes in LULC, and provide insights in making appropriate adjustments for long-term utilization and management of land resources.
References
Aarrestad, P. A.,Masunga, G. S.,Hytteborn, H.,Pitlagano, M. L.,Marokane, W., and Skarpe, C. 2011, Influence of soil, tree cover and large herbivores on field layer vegetation along a savanna landscape gradient in northern Botswana. Journal of Arid Environments, Vol. 75, No. 3, pp. 290-297.
Abiya, I. O. 1996, Towards sustainable utilization of Lake Naivasha, Kenya. Lakes & Reservoirs: Research & Management, Vol. 2, No. 3-4, pp. 231-242.
Becht, R., and Harper, D. M. 2002a, Towards an understanding of human impact upon the hydrology of Lake Naivasha, Kenya. Hydrobiologia, Vol. 488, No. 1-3, pp. 1-11.
Becht, R., and Harper, D. M. 2002b, Towards an understanding of human impact upon the hydrology of Lake Naivasha, Kenya. Hydrobiologia Vol. 488, No., pp. 1-11.
Betts, A. K.,Ball, J. H.,Beljaars, A. C. M.,Miller, M. J., and Viterbo, P. A. 1996, The land
surface-atmosphere interaction: A review based on observational and global modeling perspectives. Journal of Geophysical Research-Atmospheres, Vol. 101, No. D3, pp. 7209-7225.
Briassoulis, H. 2000, Analysis of Land use change: Theorethical and Modeling Approaches. The Web Book of Regional Science. Regional Research Institute, West Verginia University.
Campbell, D.,Lusch, D.,Smucker, T., and Wangui, E. 2005, Multiple Methods in the Study of Driving Forces of Land Use and Land Cover Change: A Case Study of SE Kajiado District, Kenya. Human Ecology, Vol. 33, No. 6, pp. 763-794.
Everard, M.,Vale, J. A.,Harper, D. M., and Tarras-Wahlberg, H. 2002, The physical attributes of the lake Naivasha catchment rivers. Hydrobiologia, Vol. 488, No., pp. 13-25.
Franchito, S. H., and Rao, V. B. 1992, Climatic-change due to land surface alterations. Climatic Change, Vol. 22, No. 1, pp. 1-34.
Gaudet, J. J. 1977, Natural drawdown on Lake Naivasha, Kenya, and the formation of papyrus swamps. Aquatic Botany, Vol. 3, No. 0, pp. 1-47.
Harper, D., and Mavuti, K. 2004, Lake Naivasha, Kenya: Ecohydrology to guide the management of a tropical protected area. Ecohydrology and Hydrobiology, Vol. 4, No. GEOBASE, pp. 287-305. He, D.,Zhou, J.,Cai, J., and Qiao, Z. 2009, The analysis of land use dynamic change and driving force in
the Olympic village periphery area, Urban Remote Sensing Joint Event pp. 1-7.
Houghton, R. A.,Hackler, J. L., and Lawrence, K. T. 1999, The US carbon budget: Contributions from land-use change. Science, Vol. 285, No. 5427, pp. 574-578.
Kitaka, N.,Harper, D. M., and Mavuti, K. M. 2002, Phosphorus inputs to Lake Naivasha, Kenya, from its catchment and the trophic state of the lake: Kluwer Academic Publishers, pp. 73-80. KNBS. 2010, The 2009 Kenya Population and Housing Census: Population Distribution by
Administrative Units.Population Distribution by Age, Sex and Administrative Units,VOLUME IC, Kenya National Bureau of Statstics (KNBS), Kenya.
Legendre, P., and Anderson, M. 1999, Distance-based redundancy analysis:testing multispecies responses in multifactorial ecological experiments. Ecological Monographs, Vol. 69, No., pp. 1-24.
Lin, W. S.,Zhang, L.,Du, D. S.,Yang, L. M.,Lin, H.,Zhang, Y. Z., and Li, J. N. 2009, Quantification of land use/land cover changes in Pearl River Delta and its impact on regional climate in summer using numerical modeling. Regional Environmental Change, Vol. 9, No. 2, pp. 75-82.
Liu, J.,Liu, M.,Zhuang, D.,Zhang, Z., and Deng, X. 2003, Study on spatial pattern of land-use change in China during 1995–2000. Science in China Series D: Earth Sciences, Vol. 46, No. 4, pp. 373-384.
McGuffie, K.,Henderson-Sellers, A.,Zhang, H.,Durbridge, T. B., and Pitman, A. J. 1995, Global climate sensitivity to tropical deforestation. Global and Planetary Change, Vol. 10, No. 1-4, pp. 97-128.
Mertens, B.,Sunderlin, W. D.,Ndoye, O., and Lambin, E. F. 2000, Impact of Macroeconomic Change on Deforestation in South Cameroon: Integration of Household Survey and Remotely-Sensed Data. World Development, Vol. 28, No. 6, pp. 983-999.
Muttitanon, W., and Tripathi, N. K. 2005, Land use/land cover changes in the coastal zone of Ban Don Bay, Thailand using Landsat 5 TM data. International Journal of Remote Sensing, Vol. 26, No. 11, pp. 2311 - 2323.
Nyandat, N. N. 1984. Climate and Soils of the South Kinangop Plateau of Kenya. Their Limitations on Land Use. (Thesis).
Olaka, L. A.,Odada, E. O.,Trauth, M. H., and Olago, D. O. 2010, The sensitivity of East African rift lakes to climate fluctuations. Journal of Paleolimnology, Vol. 44, No. 2, pp. 629-644.
Ondimu, S., and Murase, H. 2007, Reservoir level forecasting using neural networks: Lake Naivasha. Biosystems Engineering, Vol. 96, No. 1, pp. 135-138.
Otiang'a-Owiti, G. E., and Oswe, I. A. 2007, Human impact on lake ecosystems: the case of Lake Naivasha, Kenya. African Journal of Aquatic Science, Vol. 32, No. 1, pp. 79-88.
Pielke, R. A.,Avissar, R.,Raupach, M.,Dolman, A. J.,Zeng, X. B., and Denning, A. S. 1998, Interactions between the atmosphere and terrestrial ecosystems: influence on weather and climate. Global Change Biology, Vol. 4, No. 5, pp. 461-475.
Pielke, R. A.,Marland, G.,Betts, R. A.,Chase, T. N.,Eastman, J. L.,Niles, J. O.,Niyogi, D. D. S., and Running, S. W. 2002, The influence of land-use change and landscape dynamics on the climate system: relevance to climate-change policy beyond the radiative effect of greenhouse gases. Philosophical Transactions of the Royal Society of London Series a-Mathematical Physical and Engineering Sciences, Vol. 360, No. 1797, pp. 1705-1719. Rachilo, J. R. 1978, Soil Conditions in the Kinangop Area. A Preliminary Investigation. Kenya Soil
Survey Project, Nairobi.
Reid, R. S.,Kruska, R. L.,Muthui, N.,Taye, A.,Wotton, S.,Wilson, C. J., and Mulatu, W. 2000, Land-use and land-cover dynamics in response to changes in climatic, biological and socio-political forces: the case of southwestern Ethiopia. Landscape Ecology, Vol. 15, No. 4, pp. 339-355. Schneider, N., and Eugster, W. 2005, Historical land use changes and mesoscale summer climate on
the Swiss Plateau. Journal of Geophysical Research-Atmospheres, Vol. 110, No. D19, pp. Scurlock, J. M. O.,Johnson, K., and Olson, R. J. 2002, Estimating net primary productivity from
grassland biomass dynamics measurements. Global change biology, Vol. 8, No., pp. 736-753. Trauth, M. H.,Maslin, M. A.,Deino, A. L.,Junginger, A.,Lesoloyia, M.,Odada, E. O.,Olago, D. O.,Olaka, L.
A.,Strecker, M. R., and Tiedemann, R. 2010, Human evolution in a variable environment: the amplifier lakes of Eastern Africa. Quaternary Science Reviews, Vol. 29, No. 23-24, pp. 2981-2988.
Verschuren, D.,Laird, K. R., and Cumming, B. F. 2000, Rainfall and drought in equatorial east Africa during the past 1,100 years. Nature, Vol. 403, No. 6768, pp. 410-414.
Weiss, A. D. 2001, Topographic Position Index and landforms Analysis. ESRI UC 2001, 2001, San Diego California, pp. 4.
Xie, Y.,Mei, Y.,Guangjin, T., and Xuerong, X. 2005, Socio-economic driving forces of arable land conversion: A case study of Wuxian City, China. Global Environmental Change Part A, Vol. 15, No. 3, pp. 238-252.
Zhang, H.,Henderson-Sellers, A., and McGuffie, K. 1996, Impacts of tropical deforestation .1. Process analysis of local climatic change. Journal of Climate, Vol. 9, No. 7, pp. 1497-1517.
