Content uploaded by Irena F Creed
Author content
All content in this area was uploaded by Irena F Creed on Jun 27, 2014
Content may be subject to copyright.
Observing Changes in Riparian Buffer Strip Soil Properties
Related to Land Use Activities in the River Njoro
Watershed, Kenya
E. M. Enanga &W. A. Shivoga &
C. Maina-Gichaba &I. F. Creed
Received: 19 February 2010 / Accepted: 14 October 2010
#Springer Science+Business Media B.V. 2010
Abstract Riparian buffer strip guidelines are under
scrutiny in the River Njoro Watershed in Kenya. This
study investigated soil properties (bulk density, carbon,
nitrogen, and phosphorus) in different land use types
(small scale agriculture in recent settlements, mixed
agriculture in established peri-urban settlements, large-
scale commercial agriculture, and the gazetted forest
reference condition) and their adjacent buffer strips.
Bulk density, carbon, nitrogen, and phosphorus within
30-m riparian buffer strips adjacent to recent settlement
land use areas were similar to those of the gazetted forest
reference condition, but only bulk density of the buffer
strips adjacent to peri-urban and commercial agriculture
land use areas were similar to the gazetted forest
reference condition. Phosphorus is a sensitive indicator
of the impacts of human activity, as increased concen-
trations were observed with increasing scale of land use
activity. For riparian buffers adjacent to recent settle-
ments, soil phosphorus was significantly higher in
buffers narrower than 30 m (5.01 mg P kg
−1
)than
gazetted forest (3.40 mg P kg
−1
)butnotsignificantly
different for riparian buffers wider than 30 m
(3.81 mg P kg
−1
) compared to gazetted forest. Based
on the research, it is recommended that policies
governing riparian buffer strips become (1) stricter,
with the current “maximum”of 30 m considered a
minimum; and (2) adaptive, with 30 m used in small-
scale agricultural areas, and wider riparian buffer strips
used in medium- and large-scale agricultural areas.
Keywords Riparian area .Buffer strip .Soil
properties .River Njoro .Kenya
1 Introduction
Water sustainability depends on ecosystem structure
and function; and in Kenya, water policy tends to
ignore water as an ecosystem service (Baldyga 2005).
In recent years, rural communities are increasingly
encroaching on forest reserves, riparian areas, and
commercial agricultural areas for their own pasture
and crop production. This trend has persisted in the
face of persistent dry conditions and decreasing water
flows in rivers serving Kenyan rangelands, despite
the recognition that these actions will impact the
communities that depend on the rivers’ecosystem
services (Chemilil 1995).
Riparian buffer strips have become an integral part
of watershed management in American and European
Water Air Soil Pollut
DOI 10.1007/s11270-010-0670-z
E. M. Enanga :W. A. Shivoga
Department of Environmental Science, Egerton University,
Nakuru, Kenya
C. Maina-Gichaba
Department of Geology, University of Nairobi,
Nairobi, Kenya
E. M. Enanga (*):I. F. Creed
Department of Biology, The University of Western Ontario,
1151 Richmond St. N,
London, ON N6A 5B7, Canada
e-mail: eenanga@uwo.ca
landscapes (Decker 2003; Sweeney and Blaine 2007).
Riparian buffer strips are vegetated areas adjacent to
streams, rivers, lakes, and other waterways that protect
aquatic environments from excessive sedimentation,
surface runoff pollutants, and contaminants from the
adjacent landscape. However, in many parts of Africa,
riparian buffer strips are not used and scientific support
for using riparian buffer strips to mitigate changes in
water resources is needed within African landscapes. In
traditional African culture, there is no demarcation or
separation ofpeople from nature since nature and people
are viewed to be the same (Lelo et al. 2005). People in
communal Africa have sustained their livelihoods for
many years, practicing cultivation within riparian areas
without posing a significant threat to the ecosystem
(Derman 1998). Consequently, people visit and use
riparian areas and streams on a daily basis (Mathooko
2001). However, with an increasing human population
and increased intensity of adjacent land use due to
increased commercial agricultural activities, there is
need to pay more attention to these areas to ensure that
they are not overburdened.
In the River Njoro Watershed, the riparian area
provides resources to local communities and supports
critical downstream watershed services. River Njoro
has a narrow strip of indigenous riparian vegetation
averaging about 15–20 m, and some sections have
been completely cleared to provide access to the
stream (Magana and Bretschko 2003; Mathooko
2001). The riparian areas are threatened by an
incompatibility between (1) tribal norms, (2) communal
regulatory mechanisms, and (3) government statutory
enforcement mechanisms. The “free access”mentality
has led to a decline in riparian services such as water
quantity and water quality. The increased prevalence of
downstream flooding during rainfall events due to
increased runoff during the rainy season and decreased
runoff during the dry season has been attributed to land
use change (Baldyga et al. 2007; Lelo et al. 2005;
Shivoga et al. 2007). Better quality stream water has
been observed adjacent to intact riparian buffer strips
compared to stream water adjacent to little or no
vegetated riparian buffer strips (Shivoga et al. 2007).
Additionally, a change from forested to agricultural
and grazing land uses has affected the physical–
chemical environment of River Njoro, reducing the
diversity and evenness of benthic macroinvertebrates,
an indicator of declining water quality (Kibichii et al.
2007).
Riparian buffer strips remove nutrients (carbon
(C), nitrogen (N), and phosphorus (P)) from water
flowing from adjacent lands to the river through
biological (e.g., nutrient uptake by riparian vegetation,
microbial assimilation, and microbial denitrification for
N-based nutrients and microbial respiration for C) and
physical–chemical (e.g., nutrient adsorption for phos-
phorus which binds to clay particles and sediments)
processes. Not all riparian buffer strips are effective in
mitigating changes to soil properties related to land use
types. Hydrology will influence the degree of nutrient
removal if (1) flows are dominated by surface/near
surface pathways, effectively bypassing the riparian
function and delivering nutrients straight to the river,
or (2) flows are dominated by subsurface pathways
where biological uptake, transformation, fixation,
and adsorption processes can occur (Buttle 2002;
Hill 2000).
Size impacts a buffer’s effectiveness because
buffers that are too narrow may not adequately protect
aquatic resources from adjacent land use activities,
while buffers that are too wide may deny landowners
productive use of their land (Castelle et al. 1994). The
optimal size of the riparian buffer strip continues to
generate debate, and no fixed width is universally
accepted, although Castelle et al. (1994) suggest that a
minimum of 15–30 m width is necessary to protect
streams and wetlands. Generally, recommendations
for riparian buffer strip widths vary from 10 to 100 m
on each side of the stream and are usually based on a
sound intuitive grasp of the processes that should be
protected (Allan et al. 1997). Appropriate riparian
buffer strip widths vary with stream size, stream
order, and ecosystem type as well as aquatic resource
functional value, intensity of adjacent land use, buffer
characteristics, and specific buffer functions required
(Castelle et al. 1994; Osborne and Kovacic 1993).
These recommendations are sensible, but the scientific
information for or against a specified riparian buffer
strip width is limited (Osborne and Kovacic 1993).
Generally, a specific riparian buffer width should
sustainably remove as much nutrient capacity within
the area as enters the area through upslope nutrient influx
(Castelle et al. 1994;Cooperetal.1995). Unfortunately,
buffer policies worldwide are significantly influenced
by political acceptability rather than scientific data
(Castelle et al., 1994), and the River Njoro Watershed
in Kenya is no exception. Within Kenya, the 2002
Water Act describes riparian zones as land lying within
Water Air Soil Pollut
a distance equal to the width of the watercourse, with a
minimum of 2 m and a maximum of 30 m (Republic of
Kenya 2002). This policy is being debated, as some
Kenyan stakeholders feel it is too wide, while
others express a need for a wider buffer strip. There
is no scientific basis to support a 30 m or any other
width of riparian buffer strip. Advocacy for a
properly functioning riparian buffer strip is strong
becauseRiverNjoroisstressedbynutrientsfrom
land use activities within its catchment (Mokaya et
al. 2004; Shivoga et al. 2007).
Previous studies in the River Njoro Watershed
have focused on indicators of river water quality
(Shivoga 2001; Shivoga et al. 2007); however, it is
difficult to disentangle local effects from the cumula-
tive effects of upstream activities. Comparing the
properties of soils within riparian buffer strips
adjacent to different land use types and to reference
condition soils in a natural landscape provides a more
robust picture of how buffer strips mitigate nutrient
influx. Surface soils interact directly with surface
runoff, and consequently, adjacent land use is
expected to have the greatest impact on surface soil
properties (Cooper et al. 1995) and the soil chemical
properties in the riparian buffer are a reliable indicator
of the quality of surface runoff discharged from the
adjacent lands to the stream. For example, P mostly
enters the stream adsorbed onto soil particles and
organic materials in surface runoff after storm events
(Pionke et al. 1996), influencing adjacent and down-
stream water quality. It may also adsorb to fine-grained
sediment and be deposited onto the riparian buffer zone
(Dillaha and Inamdar 1997), influencing the level of P
in the riparian buffer strip soils as a result of
sedimentation (Hoffman et al. 2009). Furthermore, a
study by Cooper et al. (1995) comparing riparian soils
under different land use areas in New Zealand revealed
large differences in several key physical, chemical, and
microbial properties that can influence the zone’srole
as a buffer of material transfer across the land–water
interface.
The purpose of this research was to determine if
Kenya’s policy of 30-m riparian buffer strip width
mitigates changes in soil properties related to land use
activities adjacent to riparian buffers in the River Njoro
Watershed. The soil properties of different land use
areas, including recent settlements, peri-urban settle-
ments, and commercial agriculture, were compared to a
gazetted forest riparian buffer strip used as a reference
condition. We evaluated the effectiveness of riparian
buffer strips adjacent to different land use types by
comparing bulk density and concentrations of C, N, and
P in the soils to those in a gazetted natural forest. The
objectives of this study were to (1) determine soil
properties in different land use types; (2) determine soil
properties within riparian buffer strips adjacent to
different land use types; and (3) determine if the
government regulated maximum riparian buffer strip
(30 m) results in minimal changes to soil physical and
chemical properties associated with land use activities in
the surrounding landscape.
2 Study Area
The River Njoro Watershed covers approximately
280 km
2
(Fig. 1). The river originates at an elevation
of about 3,000 m above sea level in the Eastern Mau
Escarpment, and descends in a northeast direction
before terminating at Lake Nakuru on the floor of the
Rift Valley at about 1,800 m above sea level. The
river provides 65% of the total freshwater in-flow to
Lake Nakuru (Gichuki et al. 1997) and is on average
10 m wide (ranging from about 1 m to 15 m). The
watershed has a population of over 300,000 people
(Ministry of Finance and Planning 2002) and includes
the urban center Njoro Town (30,000 people) and
much of the Nakuru municipality (240,000 people;
Lelo et al. 2005).
The climate is characterized by a trimodal precip-
itation pattern: long, intense rains from April to May;
short, intense rains in August; and shorter, less intense
rains from November to December. Total annual
precipitation is 956 mm, and the mean annual
temperature is 16.5°C, ranging from a minimum of
9°C (July) to a maximum of 24°C (January; Baldyga
2005). Geology is characterized by porous pumiceous
formations (McCall 1967). Soils include Humic
Acrisols (Ultisols), Phaeozems (Mollisols), Ando-
sols, Planosols (Aqualfs), Plinthosols, and Fluvisols
(Fluvents; Mainuri 2006). Soil textures range from
clay loams in the lower watershed to sandy clay in
the plantations and indigenous forest areas at higher
altitude, the focus of this study. Vegetation cover
ranges from 90% in upland indigenous forests that
are difficult to reach due to extreme topographic
relief on the eastern rift escarpment to 0% in areas
affected by anthropogenic practices (Baldyga 2005).
Water Air Soil Pollut
The study area was 30 m on each side of a 25-km
length of river with similar riparian forest and soil
compositions but different surrounding land use types
(Fig. 1). These land use types include (1) predomi-
nantly protected government forests (i.e., gazetted
forests; Fig. 2a); (2) small-scale agriculture associated
with new settlements on recently felled plantation
forests characterized by temporary structures (i.e.,
recent settlements; Fig. 2b); (3) medium-scale agri-
culture associated with older settlements surrounding
urban areas (i.e., peri-urban settlements; Fig. 2c); and
(4) large-scale agriculture with permanent structures
(i.e., commercial agriculture; Fig. 2d; Lelo et al.
2005).
3 Methods
The study was conducted in a stratified randomized
design. Four riparian areas with different neighboring
land use types (Fig. 2) were selected. For each land
use type, two sampling sites (eight in total) were
selected at random, and for each, a sampling cluster of
five transects positioned 50 m apart was established.
Each transect centered on a watering point (points
where human beings and livestock access the river for
water) running 30 m to the left and right side of the
river (Fig. 3), but the transects centered on the
watering sites were not included in the analysis to
minimize the effect of human disturbance within the
riparian area. The riparian buffer strip covered the
entire observed area for the gazetted government
forest (i.e., 30 m on each side of the river) and ranged
from no coverage to 30 m in width for the other land
use types.
One soil sample was collected at three points (10,
20, and 30 m from the streambank) along each of the
four transects on each side of the stream, giving a
maximum of 24 samples per sampling site. Soil
samples could not be obtained where there was
extensive underlying bedrock. For bulk density, soils
were sampled at surface (0–20 cm) depths using
cylindrical core rings (50 mm height and 50 mm
diameter), placed on aluminum trays where coarse
organic matter was removed manually, oven dried at
105°C, cooled in a desiccator, and then weighed. Bulk
Fig. 1 Map of the River
Njoro Watershed, Kenya
Water Air Soil Pollut
density (g/cm
3
) was calculated using the mass of soil
within the volume of the cylindrical core ring.
Previous studies observed little change in bulk density
10 cm below the soil surface (e.g., Augeard et al.
2007), so soil samples were not collected to determine
the bulk density of subsurface soils (20–50 cm).
For nutrient analyses, soils were sampled at surface
(0–20 cm) and subsurface (20–50 cm) depths using a
soil auger and placed on aluminum trays. Litter and
roots were removed, and the soil was air dried,
ground, passed through a 2-mm diameter sieve, then
analyzed for C, N, and P using near-infrared
technology. A Bruker MPA spectrometer was used to
obtain soil spectral reflectance signatures on air-dried
soil samples at the World Agroforestry Centre, formerly
the International Centre for Research in AgroForestry
laboratories, in Nairobi, Kenya. Air-dried and sieved
soil samples were placed on a Petri dish and positioned
on an optical window. Near-infrared light was then
broadcast onto the sample and the reflected light was
Fig. 2 Different land use
types in the River Njoro
Watershed agazetted forest,
brecent settlement,
cperi-urban settlement, and
dcommercial agriculture
Fig. 3 Experimental design
at each sampling station.
Dotted lines are a schematic
representation of the 30 m
strip of riparian buffer,
while the gray shaded area
is a representation of how
the actual strip of riparian
buffer varied. Transects are
perpendicular to the stream
Water Air Soil Pollut
collected as a reflection spectrum. The reflection
spectrum was converted to standard units with calibra-
tion models developed using standard methods on sub-
samples from 30 samples. The required soil physical
and chemical properties [C (%), N (%), and extractable
inorganic P (mg/kg)] were predicted with high repro-
ducibility (99%; Shepherd 2005).
One-way analysis of variance (ANOVA) was used
to test for significant differences among the bulk
density, C, N, and P levels in the soil of different land
use types and within the buffers adjacent to the
different land use types (McBride and Booth 2005).
Tukey’s pair-wise tests were performed where
ANOVAs yielded significant differences. To assess
the adequacy of the buffers, the soil properties within
buffers (“Inside”) were compared to both the soil
properties of the land use areas (“Outside”buffers) and
the reference condition, first using one-way ANOVAs
and then Tukey’s pair-wise comparisons. To determine
if the government regulated maximum of 30 m was
effective in mitigating changes to soil properties, the soil
properties of (1) the reference condition, (2) the riparian
buffer strips less than 30 m, and (3) buffer strips at least
30 m in width were compared using one-way ANOVAs
and then Tukey’s pair-wise comparisons. Data analysis
was performed using Sigma Plot 11.0. Significance was
assessed at the p<0.05 level.
Due to the study design, there were large differences
in sample sizes for the different land use types. The
gazetted forest had the largest sample size, because it
had the most intact riparian forest buffer, while the other
land use types were much more likely to have riparian
buffers of less than the full 30 m transect. To account for
these differences in sample sizes, comparable sample
sizes were randomly selected and analyzed. There were
minimal differences between these data sets and the full
data set, so the full data set is reported here.
4 Results
4.1 Slope in Different Land Use Types
The slope in commercial agriculture land use type was
significantly larger than that in the reference condition
represented by gazetted forest land use type (Table 1).
However, there were no significant differences in slope
between commercial agriculture land use type and peri-
urban and recent settlement land use types or among
gazetted forest land use type and recent settlement and
peri-urban land use types.
4.2 Soil Properties among Land Uses
Land use activities had a significant effect on all four
soil properties (Table 2). There was a general increase
in the bulk density of the surface soils (0–20 cm) with
increasing intensity of land use activity. Gazetted
forest land use type had significantly lower bulk
density than peri-urban land use type but not
significantly lower than recent settlement and com-
mercial agriculture land use types. There was a
general decrease in C and N with increasing intensity
of land use activity. In surface soils, there was
significantly more C in the gazetted forest than in
peri-urban and commercial agriculture land use types.
There was also significantly more C in the recent
settlements than in the peri-urban land use type.
Similarly, there was significantly more N in the
gazetted forest and recent settlements relative to the
peri-urban and commercial agriculture land use types.
The same general trends were seen in the subsurface
soils, but the differences were smaller. Concentration
of P was significantly lower in the surface soil in the
gazetted forest relative to recent settlement and peri-
urban land use types. The general trend of increasing
P with increased intensity of land use was similar in
both surface and subsurface soils.
4.3 Soil Properties within Riparian Buffer Strips
Adjacent to Different Land Use Types
Within the riparian buffer strips, all four soil
properties varied significantly (Table 3). There was
significantly less bulk density in the surface soils of
the buffer strips adjacent to commercial agriculture
Table 1 Slopes of the riparian buffer strips. Means are reported
with standard deviation in parentheses
Land use NSlope (degrees)
Gazetted forest 20 7.51 a (3.943)
Recent settlements 20 12.46 ab (4.209)
Peri-urban settlements 20 9.29 ab (8.839)
Commercial agriculture 20 13.39 b (6.942)
Means with the same letter are not significantly different from
each other (based on Tukey’s pair-wise comparisons, p<0.05)
Water Air Soil Pollut
land use types than the recent settlement buffers.
There was a general decrease in C and N with
increasing intensity of adjacent land use activity for
both surface and sub-surface soils. In surface soils, the
gazetted forest reference condition had significantly
more C than the buffer strips adjacent to the peri-urban
and commercial agriculture land use types. Recent
settlement buffers also had significantly more C than
Table 2 Surface (0–20 cm) and subsurface (20–50 cm) soil properties outside riparian buffer strips (within different land use areas)
and inside the gazetted forest reference condition in the River Njoro Watershed
Land use NBulk Density (g/cm
3
) Carbon (%) Nitrogen (%) Phosphorus (mg/kg)
Surface (0–20 cm)
F 4.238 27.95 25.15 4.235
p 0.008 <0.001 <0.001 0.008
Gazetted forest 48 0.85 a (0.25) 5.14 c (1.25) 0.46 b (0.11) 3.40 b (0.68)
Recent settlements 18 0.88 ab (0.21) 4.19 b (1.03) 0.42 b (0.11) 4.49 a (0.91)
Peri-urban settlements 17 1.05 b (0.16) 2.69 a (0.91) 0.23 a (0.09) 4.24 a (1.66)
Commercial agriculture 9 1.01 ab (0.22) 3.00 ab (1.01) 0.30 a (0.09) 3.99 ab (1.46)
Subsurface (20–50 cm)
F 6.579 3.929 2.650
P <0.001 0.011 0.054
Gazetted forest 48 NA 3.58 b (1.26) 0.31 b (0.12) 2.90 (0.48)
Recent settlements 18 NA 2.69 ab (0.81) 0.28 ab (0.79) 3.41 (1.39)
Peri-urban settlements 17 NA 2.57 a (0.52) 0.22 a (0.06) 3.59 (1.05)
Commercial agriculture 9 NA 2.56 a (1.00) 0.26 ab (0.09) 3.40 (1.62)
NA not available
Means are reported with standard deviation in parentheses. Means with the same letter are not significantly different from each other
(based on Tukey’s pair-wise comparisons, p<0.05)
Table 3 Surface (0–20 cm) and subsurface (20–50 cm) soil properties inside riparian buffer strips adjacent to different land use areas
and inside the gazetted forest reference condition in the River Njoro Watershed
Land use NBulk Density (g/cm
3
) Carbon (%) Nitrogen (%) Phosphorus (mg/kg)
Surface (0–20 cm)
F 3.993 8.629 7.95 13.745
p 0.01 <0.001 <0.001 <0.001
Gazetted forest 48 0.85 ab (0.25) 5.14 a (1.25) 0.46 a (0.11) 3.40 a (0.68)
Recent settlements 27 0.81 b (0.18) 4.80 ab (1.77) 0.47 a (0.16) 3.81 a (1.27)
Peri-urban settlements 15 0.99 ab (0.17) 3.53 bc (1.52) 0.31 b (0.15) 4.21 a (1.49)
Commercial agriculture 15 1.02 a (0.28) 3.36 c (1.43) 0.33 b (0.14) 5.64 b (1.86)
Subsurface (20–50 cm)
F 4.001 3.648 6.070
p 0.01 0.015 <0.001
Gazetted forest 48 NA 3.58 a (1.26) 0.31 ab (0.12) 2.90 b (0.48)
Recent settlements 27 NA 3.19 ab (1.56) 0.32 a (0.14) 4.09 a (1.74)
Peri-urban settlements 15 NA 2.54 b (0.93) 0.22 b (0.09) 3.87 ab (0.93)
Commercial agriculture 14 NA 2.51 b (1.18) 0.24 ab (0.11) 4.19 a (2.69)
NA not available
Means are reported with standard deviations in parentheses. Means with the same letter are not significantly different from each other
(based on Tukey’s pair-wise comparisons, p<0.05)
Water Air Soil Pollut
commercial agriculture buffers. Similarly, the concen-
tration of total N in surface soils from gazetted forest
and recent settlements were both significantly more than
the N concentration in peri-urban settlement and
commercial agriculture buffer strips. For subsurface
soils, there was significantly more C in the reference
condition soils than in peri-urban and commercial
agriculture buffer soils, and there were significant
differences among N amount in the subsurface soils
with recent settlement buffer soils having significantly
more N than peri-urban buffer soils.
The P of surface and subsurface soils within
riparian buffers generally increased with more inten-
sive land use in adjacent areas. The commercial
agriculture buffer soils had significantly more P in
their surface soils than the buffer soils in the reference
condition, peri-urban, and recent settlement land use
types. The subsurface soils within the reference
condition had significantly less P than the buffers
adjacent to commercial agriculture land use types but
were not significantly different than those of the peri-
urban areas and recent settlements, although the value
was still lower. All buffer soils (surface and subsurface)
adjacent to the three land use types had more P relative
to the reference condition gazetted forest.
4.4 Comparison of Soil Properties Outside vs. Inside
the Riparian Buffer Strips
There was only one case where the buffer had
significantly different soil properties than the adjacent
land: The commercial agriculture had higher P within
the buffer strip than in the adjacent land (Table 4).
The soil properties of the reference condition gazetted
forests were significantly different than the buffers
adjacent to the different land use types in most cases
(Table 4; Fig. 4). The P concentration in the reference
condition was significantly lower than the P within
recently settled land use type but not significantly
different than soils in the riparian buffer strips
adjacent to recently settled areas. Bulk density within
the peri-urban land use type was significantly higher
than that in gazetted forest, and differences between
riparian buffers adjacent to peri-urban bulk density
and gazetted forest were not significant. However, for
C and N, the reference condition had significantly
Table 4 Comparison of surface soil properties in the gazetted forest reference condition, riparian buffer strips that meet the
government regulation (≥30 m), and those that are narrower than the government regulation (<30 m) in different land use areas
Riparian Buffer Width NBulk Density (g/cm
3
) Carbon (%) Nitrogen (%) Phosphorus (mg/kg)
Recent settlement
F 1.345 2.481 1.160 10.089
p 0.266 0.09 0.319 <0.001
Gazetted forest 48 0.85 (0.25) 5.14 (1.25) 0.46 (0.11) 3.40 b (0.68)
At least 30 m 27 0.81 (0.18) 4.80 (1.77) 0.47 (0.16) 3.81 b (1.27)
Less than 30 m 9 0.95 (0.23) 3.96 (1.74) 0.39 (0.14) 5.01 a (1.42)
Peri-urban settlement
F 2.234 8.87 8.42 5.167
P 0.114 <0.001 <0.001 0.008
Gazetted forest 48 0.85 (0.25) 5.14 a (1.25) 0.46 a (0.11) 3.40 b (0.68)
At least 30 m 15 0.99 (0.17) 3.53 b (1.52) 0.31 b (0.15) 4.21 a (1.49)
Less than 30 m 14 0.94 (0.22) 3.87 b (2.07) 0.35 b (0.20) 3.47 ab (0.39)
Commercial agriculture
F 4.42 23.55 14.598 22.653
P <0.015 <0.001 <0.001 <0.001
Gazetted forest 48 0.85 b (0.25) 5.14 a (1.25) 0.46 a (0.11) 3.40 b (0.68)
At least 30 m 15 1.02 ab (0.27) 3.36 b (1.43) 0.33 b (0.14) 5.64 a (1.84)
Less than 30 m 14 1.04 a (0.21) 2.97 b (0.85) 0.29 b (0.10) 5.00 a (1.84)
Means are reported with standard deviations in parentheses. Means with the same letter are not significantly different from each other
(based on Tukey’s pair-wise comparisons p<0.05)
Water Air Soil Pollut
higher values than soils in riparian buffers adjacent to
peri-urban land use types, while the concentration of
P was significantly higher in riparian buffers adjacent
to peri-urban land use type than the reference
condition. Commercial agriculture land use type had
significantly lower C and N compared to the reference
condition, while P concentration was significantly
higher in commercial agriculture land use type than in
the reference condition. Concentration of P was
significantly lower in soil in commercial agriculture
land use type than the adjacent riparian buffer.
Comparison between inner (within riparian buffer)
and outer (outside riparian buffer, within the land use
type itself) using ttests revealed similar trends to those
observed with multiple pair-wise comparisons after
running the ANOVA (Table 5). There was significantly
more P inside (5.638 mg/kg) than outside (3.990 mg/kg)
riparian buffers adjacent to commercial agriculture land
use type (t=2.836, p=0.009). Differences in P in recent
settlement and peri-urban land use types were not
significantly different, although higher values were
observed outside than inside the buffer strip. However,
for bulk density, carbon, and nitrogen there were no
significant differences in soil properties between
inside and outside riparian buffers in all land use
types.
4.5 Evaluation of the Government Regulated 30 m
Maximum
There were no significant differences in soil proper-
ties between riparian buffers of at least 30 m wide and
buffers narrower than 30 m wide in all land use types,
except for P within recent settlements (Table 5;
Fig. 5). The concentration of soil P was significantly
higher in riparian buffer strips narrower than 30 m
wide compared to riparian buffer strips that were at
least 30 m wide. However, the concentration of P
within the soils of the reference condition and riparian
buffer strips that were at least 30 m wide adjacent to
recent settlement land use were not significantly
different. Therefore, within the recent settlement land
use type, riparian buffer strips that were at least 30 m
wide did not differ significantly in their soil properties
compared to gazetted forest land but other land use
types (commercial agriculture and peri-urban) had
significantly less C and N than gazetted forest and
significantly more P than the gazetted forest land use
type.
5 Discussion
Riparian areas are interfaces between terrestrial and
aquatic ecosystems. They represent an important filter
of sediments, nutrients, and contaminants in water
flowing from contributing hill slopes to streams
(Bilby 1988). Riparian buffer strip widths are based
on a sound intuitive grasp of the processes that should
be protected (Allan et al. 1997), although scientific
information for or against a specified riparian buffer
strip width is limited (Osborne and Kovacic 1993).
Government regulations, in Kenya, describe riparian
zones as land lying within a distance equal to the
-80
-60
-40
-20
0
20
40
60
80
Recent
Settlement
(inside)
Recent
Settlement
(outside)
Per-Urban
Settlement
(inside)
Peri-Urban
Settlement
(outside)
Commercial
Agriculture
(inside)
Commercial
Agriculture
(outside)
Percent Change from
Reference Condition (%)
C
N
P
Fig. 4 Percent change rela-
tive to reference condition
in concentration of carbon
(C, %), nitrogen (N, %), and
phosphorus (P, mg/kg)
in soils inside and outside
riparian buffer strips in
different land use areas
(recent settlement,
peri-urban settlement,
commercial agriculture)
Water Air Soil Pollut
width of a watercourse with a minimum of 2 m and a
maximum of 30 m (Republic of Kenya 2002).
However, there is no scientific basis to support a
riparian buffer strip width of between 2 m and 30 m
or any other width. Herein, we assess the suitability of
this regulation.
5.1 Defining a Reference Condition
One of the challenges of determining the effectiveness
of riparian buffer strips in mitigating changes in soil
properties related to adjacent land use activities is
identifying a reliable reference condition (White and
Walker 1997). The upper and mid sections of the
River Njoro Watershed have historically had similar
land use types before being converted to their current
land use types. Therefore, differences in the soil’s
physical and chemical properties are likely mainly
due to current land use activities. Natural factors like
underlying bedrock composition may be important in
less disturbed or more uniformly disturbed areas
(Allan 2004), but in the River Njoro Watershed the
Outside vs. Inside Riparian
Buffer Strip
NBulk Density
(g/cm
3
)
Carbon (%) Nitrogen (%) Phosphorus
(mg/kg)
Recent settlements
Inside buffer 27 0.807 (0.18) 4.797 (1.77) 0.469 (0.16) 3.809 (1.27)
Outside buffer 9 0.875 (0.21) 4.190 (1.03) 0.417 (0.11) 4.492 (0.91)
t−0.933 0.970 0.910 −1.481
p0.357 0.339 0.369 0.148
Peri-urban settlement
Inside buffer 15 0.989 (0.17) 3.531 (1.52) 0.307 (0.15) 4.213 (1.49)
Outside buffer 17 1.047 (0.16) 2.689 (0.91) 0.229 (0.09) 4.238 (1.66)
t−1.021 1.932 1.802 −0.044
p0.315 0.063 0.82 0.965
Commercial agriculture
Inside buffer 15 1.016 (0.28) 3.356 (1.43) 0.334 (0.14) 5.638 (1.86)
Outside buffer 18 1.013 (0.22) 3.003 (1.01) 0.290 (0.09) 3.990 (1.46)
t0.026 0.830 1.060 2.854
p0.981 0.413 0.270 0.008
Tab l e 5 Comparison of
surface soil properties
inside and outside riparian
buffer strips in different
land use areas and reference
condition gazetted forest
Means are reported with
standard deviations
-80
-60
-40
-20
0
20
40
60
80
Recent
Settlement
<30m
Recent
Settlement
>=30m
Peri-Urban
Settlement
<30m
Peri-Urban
Settlement
>=30m
Commercial
Agriculture
<30m
Commercial
Agriculture
>=30m
Percent change relative to
Reference Condition (%)
C
N
P
Fig. 5 Percent change rela-
tive to the reference condi-
tion in concentration of
carbon (C, %), nitrogen
(N, %), and phosphorus
(P, mg/kg) in soils within
riparian buffer strips <30 m
and ≥30 m adjacent to
different land use areas
(recent settlement,
peri-urban settlement, and
commercial agriculture)
Water Air Soil Pollut
different degrees of human activity (e.g., recent
settlements, peri-urban settlements, and commercial
agricultural land use) likely have had a much larger
impact than natural variation. There is a comparative-
ly intact and undisturbed section of riparian buffer
strip within the natural gazetted forest. It was used by
Kibichii et al. 2007 as an “unpolluted reference”in a
study examining macroinvertebrate assemblages
along a land use gradient, reinforcing the notion that
it provides a suitable reference condition for soil
properties in the River Njoro Watershed. Slope among
the different land use types was comparable (Table 1).
The only significant difference observed was between
the commercial agriculture land use and gazetted
forest. The rest of the pairs of the land use types were
similar which implies that slope may not have had a
pronounced effect on the other soil properties in the
comparable land use types.
5.2 Riparian Buffer Strip Widths
Riparian buffer strips of 30 m were only able to
minimize changes in soil properties related to adjacent
land use activity in some cases. Bulk density
generally increased with the scale of land use activity.
Bulk density provides an indication of erosion caused
by land use activities. Sediment eroding from con-
tributing areas can become trapped within the riparian
buffer strip thereby increasing bulk density (Cooper
and Gilliam 1987). In the River Njoro Watershed,
humans and livestock use the area surrounding the
river daily, as is common in other tropical countries
(Mathooko 2001). In addition, humans and livestock
spend a large portion of the day seeking shelter from
equatorial sunlight and heat in the shade of riparian
trees (Wyant and Ellis 1990). Intensive human
activities in riparian areas interrupt natural drainage
as riparian soils become compacted, sedimentation
rates increase, solar radiation increases, and stream
channels are altered (Klapproth and Johnson 2000).
This leads to increased bulk density, as witnessed here
in the peri-urban and commercial agricultural areas of
the River Njoro Watershed with a general increase in
soil bulk density with increased human activity.
Soil C and N concentrations decreased with scale
of land use activity. Soil C and N are closely related
because over 90% of the N in soils is associated with
organic matter (Oldham 2003a). The decrease in soil
C and N may be linked to reduced organic matter.
There are many natural sources of organic matter in
the gazetted forest. Recently settled lands (small-scale
agriculture) had the second highest C concentrations,
possibly because these areas have been only recently
converted from gazetted forest, and the soils may still
contain residues of organic matter left over from the
previous land use. In addition, recent settlement land
use activities restrict the harvest of organic material to
the crop itself, leaving the plant residues to be
incorporated into the soils, thereby maintaining the
soil C and N concentrations. In contrast, soil C and N
concentrations in peri-urban settlements and commer-
cial agriculture were much lower. Increased urbani-
zation creates more impermeable surfaces in the
watershed and changes the flow regime by increasing
runoff to streams and decreasing infiltration into the
ground water. This causes organic matter and associ-
ated nutrients to wash down-slope and into the
stream, bypassing the riparian buffer resulting in
lower C as observed in peri-urban land use types
compared to commercial agriculture land use type.
Dissolved organic C and nitrate-N are mobile and
therefore may be flushed out of the riparian buffer
strip to the stream (Creed and Band 1998; Hornberger
et al. 1994), or in the case of nitrate-nitrogen, become
denitrified (Vidon et al. 2010). Shivoga et al. (2007)
observed no net contribution of nitrate into surface
waters from sites near recent settlements, but they
observed a significant contribution of nitrate–N into
surface waters near land use areas with a higher
intensity of human activity. This suggests that N in
the peri-urban settlements and commercial agriculture
riparian buffer strips is being flushed to the adjacent
River Njoro; hence, the low concentration of soil C
and N in riparian buffer strips adjacent to this land use
types.
Soil P concentrations increased with scale of land
use activity. The movement of P within landscapes is
closely associated with the mobilization of soils with
sedimentation of particulate P during overland flow as
a major retention system in riparian buffers (Hoffman
et al. 2009; Oldham 2003b). Phosphorus adsorbs to
soil particles and is less likely to wash away like C
and N, thus P presence strongly indicates the effects
of human activities. Laundry detergent from individ-
uals washing clothes, along with waste from humans
and domesticated animals using the river, are the most
likely sources of inorganic P in peri-urbanland use areas
and adjacent riparian buffer strip soils. In commercial
Water Air Soil Pollut
agriculture areas, di-ammonium phosphate, the main
fertilizer used in the commercial agricultural farms
(Mokaya et al. 2004), is the most likely source of
phosphorus. These factors may explain the large levels
of extractable inorganic P in soils in riparian buffer
strips adjacent to peri-urban settlements and commer-
cial agriculture land use types. Hoffman et al. (2009)
notes that sedimentation which is the main physical
process in riparian buffers may account for high
retention P rates of up to 128 kg P ha
−1
year
−1
with
plant uptake temporarily immobilizing approximately
15 kg P ha
−1
year
−1
. In addition, riparian buffer strips,
because of their flatter slopes and high surface
roughness, effectively reduce the lateral movement of
suspended soil particulates which reduces phosphate
concentrations in streams (Amador et al. 1997;
Lowrance et al. 1984), causing increased loads in soils
in riparian buffer strip soils.
Based on the observed soil properties, we found
that a riparian buffer strip of at least 30 m (where the
policy would require 15 m) is needed to minimize the
effects of adjacent recent settlement land use activi-
ties. Soil P in riparian buffer strips adjacent to recent
settlement use that were wider than 30 m was similar
to soil P in the gazetted forest, implying that the
riparian buffer resets the soil P to what would be
expected naturally (reference condition). Additionally,
soil P in riparian buffers narrower than 30 m was
elevated, implying an overload in the narrow but
undisturbed riparian buffer strip. In peri-urban settle-
ment settings, a 30-m riparian buffer strip is not
adequate to prevent changes to soil properties,
because the soil P in riparian buffers narrower and
wider than 30 m was comparable. There was also no
significant difference between soil P in riparian
buffers narrower than 30 m and the gazetted forest
land use, suggesting that the buffer is too narrow.
Most of the sediment-bound P is washed off to the
stream, leaving very low levels of P in the narrow
buffer given that higher P levels were observed
outside the riparian buffer in the peri-urban land use
and more P in stream water as observed by Shivoga et
al. (2007). In the commercial agriculture land use
type, there was significantly more P in both the
riparian buffers narrower and wider than 30 m
compared to the gazetted forest, and there was no
observable difference between soil P in riparian
buffers narrower or wider than 30 m. This suggests
that the P entering even the 30 m buffer is much more
than the buffer can process. Therefore, a 30-m wide
riparian buffer is not adequate to mitigate changes in
soil properties related to commercial agriculture land
use type. For the commercial agriculture land use
type, soils within buffer strips of at least 30 m did not
have significantly different bulk density than the
reference condition, and the soils within buffer strips
of both less than and more than 30 m in width that
were adjacent to peri-urban and commercial agricul-
ture land use types had higher C, N, and P values than
the reference condition. In general, riparian buffers
narrower than 30 m were less effective in mitigating
changes to C, N, or P to reference condition levels
than buffers that were at least 30 m (Fig. 5).
The maximum 30 m wide riparian buffer strip
should become the standard for buffer strips adjacent
to relatively low intensity land use activities such as
the recent settlements. A larger buffer strip is needed
to mitigate changes in soil properties within buffer
strips adjacent to higher intensity land use activities
such as peri-urban settlements and commercial agri-
culture land use types.
5.3 Potential Links between Terrestrial and Aquatic
C/N/P Ratios
Changes to the soil properties of riparian buffers
could have far reaching effects. If the C/N/P ratio of
the soil properties reflect the C/N/P ratio of the water
that is discharged from the riparian buffer strip, as is
the case for nitrate concentrations (Ohrui and Mitchell
1998), these changes may have fundamental conse-
quences for not only the riparian vegetation but also
the downstream communities and Lake Nakuru, the
terminus of River Njoro. Although C (which is the
main component of organic matter) is not a pollutant,
riparian organic matter export to aquatic ecosystems
affects the rates of most biologically mediated
reactions. This regulates the fate of contaminants
such as N, P, Hg, and pesticides, thereby influencing
ecosystem metabolism (Vidon et al. 2010). Given
River Njoro’s high functional value (Castelle et al.
1994), elevated nutrient concentrations in the riparian
buffer strips would increase primary productivity;
however, if the nutrient load exceeds the rate of
vegetation uptake, nutrients will end up in the river.
Increased nutrient loads would likely lead to an
Water Air Soil Pollut
increase in primary productivity in the adjacent
aquatic ecosystem and the potential for harmful algal
blooms (Kronvang et al. 2001) as well as sediment
filling up Lake Nakuru. This would pose a threat to
flamingo populations and the tourism industry, upon
which Kenya heavily relies for foreign exchange.
5.4 Policy Implications
Recent initiatives within the management of the River
Njoro Watershed are affecting the government regu-
lation of riparian buffer widths. For example, Water
Resource Users Associations (WRUAs) have become
important in the management of rivers in Kenya. The
WRUAs are responsible for a given stretch of the
river and make recommendations on the best way to
manage the river. Initially, longitudinal tours were
conducted along the river in order to familiarize
upstream WRUAs with how their actions impact the
downstream communities. Similarly, downstream
WRUAs toured the upstream communities to under-
stand the sacrifices the upstream WRUAs make in
order to ensure that the downstream communities
continue to enjoy the services provided by River
Njoro. The WRUAs have been developing buffer
strip width recommendations within some portions of
the River Njoro Watershed, especially in the recent
settlements, which suggests better conservation prac-
tices within the country. Unfortunately, regulatory
mechanisms that enforce sustainable management of
communal property resources have failed and each
community is applying its own cultural values and
experiences to the use of riparian resources (Lelo et
al. 2005). This suggests that following traditional
communal use of riparian resources cannot continue
and more intensive, inclusive management strategies
must be employed using scientifically based criteria
for establishing buffer requirements and subsequent
utilization by policy makers and resource agencies
(Castelle et al. 1994).
6 Conclusions
This study assessed the suitability of the Kenyan
policy regulating a riparian buffer strip equal to the
width of the river, with a minimum of 2 m up to a
maximum of 30 m, for management of the River
Njoro Watershed. The research findings of this study
suggest that:
1. The current policy may be appropriate in mitigat-
ing changes in soil properties related to small-
scale human activities (i.e., recent settlements)
where settlers coexist with indigenous peoples,
and respect one another’s ways of life. However,
the current policy is not appropriate for mitigating
changes in soil properties related to larger-scale
human activities (i.e., peri-urban and commercial
agricultural land use areas).
2. Surface soils better indicate the effects of changes
in soil properties related to land use activities in
the River Njoro Watershed than subsurface soils.
3. Phosphorus is a more sensitive indicator of the
impacts of human activity, as increased concen-
trations were found at all scales of land use activity.
Carbon and nitrogen concentrations were reduced
only in the larger-scale land use activities of peri-
urban settlement and commercial agriculture.
4. The soil properties of riparian buffer strips could
have far-reaching effects, because if the C/N/P ratio
of the soil properties reflect the C/N/P ratio of the
water discharged from the riparian buffer strip, then
these changes may have consequences not only for
the riparian vegetation but also for the aquatic
ecosystems dependent on this riparian buffer strips.
Acknowledgments This study was funded by Sustainable
Management of Watersheds (SUMAWA): Biophysical, Livestock
and Human Interactions, a project of GL-CRSP (Global Livestock–
Collaborative Research Support Program). We thank Steve Huckett,
a graduate student working on the GL-CRSP project based at Utah
State University. We thank staff at the World Agroforestry Centre in
Nairobi and at Egerton University for access to facilities to complete
soil samples analyses. We acknowledge support from the
Department of Foreign Affairs and International Trade-Canada
(DFAIT-Canada) through the Canadian Bureau of International
Education (CBIE) for funding that allowed Enanga to complete
the study at the University of Western Ontario in Canada.
References
Allan, J. D. (2004). Landscapes and riverscapes: The influence
of land use on stream ecosystems. Annual Review of
Ecology and Evolution, 35, 257–284.
Allan, J. D., Erickson, D. L., & Fay, J. (1997). The influence of
catchment land use on stream integrity across multiple
spatial scales. Freshwater Biology, 37, 149–161.
Water Air Soil Pollut
Amador, J. A., Glucksman, A. M., Lyons, J. B., & Goress, J. H.
(1997). Spatial distribution of soil phosphatase activity
within a riparian forest. Soil Science, 162, 808–825.
Augeard, B., Assouline, S., Fonty, A., Kao, C., & Vauclin, M.
(2007). Estimating hydraulic properties of rainfall-induced
soil surface seals from infiltration exeperiments and X-ray
bulk density measurements. Journal of Hydrology, 341,
12–26.
Baldyga, T. J. (2005). Assessing land cover change impacts in
Kenya’s River Njoro watershed using remote sensing and
hydrologic modelling (Master’s thesis). Laramie: Univer-
sity of Wyoming.
Baldyga, T. J., Miller, S. N., Maina, G. C., & Shivoga, W.
(2007). Suitability of the Automated Geospatial Watershed
Assessment (AGWA) tool in assessing hydrologic response
and land cover change in River Njoro Watershed, Kenya.
Egerton Journal: Science and Technology, 7,56–71.
Bilby, R. E. (1988). Interactions between aquatic and terrestrial
systems. In K. J. Raedeke (Ed.), Streamside management:
Riparian wildlife and forestry interactions (pp. 13–29).
University of Washington: College of Forest Resources.
Contribution Number 59.
Buttle, J. M. (2002). Rethinking the donut: The case for
hydrologically relevant buffer zones. Hydrological Pro-
cesses, 16, 3093–3096.
Castelle, A. J., Johnson, A. W., & Conolly, C. (1994). Wetland
and stream buffer size requirements—A review. Journal of
Environmental Quality, 23, 878–882.
Chemilil, M. C. (1995). The effects of human-induced water-
shed changes on streamflows. PhD dissertation. London:
Loughborough University of Technology.
Cooper, J. R., & Gilliam, J. W. (1987). Phosphorus redistribu-
tion from cultivated fields into riparian areas. Soil Science
Society of America Journal, 51, 1600–1604.
Cooper, A. B., Smith, C. M., & Smith, M. J. (1995). Effects of
riparian set-aside on soil characteristics in an agricultural
landscape: Implications for nutrient transport and reten-
tion. Agriculture, Ecosystems & Environment, 55,61–67.
Creed, I. F., & Band, L. E. (1998). Export of nitrogen from
catchments within a temperate forest: Evidence for a
unifying mechanism regulated by variable source area
dynamics. Water Resources Research, 34, 3105–3120.
Decker, R. C. (2003). Current regulations, guidelines and best
management practices concerning forest harvesting and
riparian zone management. Buffer zone working group
literature review, fisheries and oceans Canada, science,
oceans and environment branch, environmental sciences
section, St. John’s, NL, Canada.
Derman, B. (1998). Balancing the waters: Development and
hydropolitics in contemporary Zimbabwe. In J. Donahue &
B. R. Johnston (Eds.), Water, Culture and Power (pp. 73–
94). New York: Island Press.
Dillaha, T. A., & Inamdar, S. P. (1997). Buffer zones as
sediment traps or sources. In N. E. Haycock,T. P. Burt, K. W.
T. Goulding, & G. Pinay (Eds.), Buffer zones: Their
processes and potential in water protection; proceedings
of the International Conference on Buffer Zones, September
1996 (pp. 33–42). Hertfordshire: Quest Environmental.
Gichuki, N., Wakanene, K., & Gatere, G. (1997). Biodiversity
profile of Nakuru district and its environs, Kenya. Nairobi:
Ministry of Environment and Natural Resources.
Hill, A. R. (2000). Stream chemistry and riparian zones. In J. B.
Jones & P. J. Mulholand (Eds.), Streams and ground-
waters (pp. 83–110). New York: Academic.
Hoffman, C. C., Kjaergaard, C., Uusi-Kamppa, J., Hansen, H.
C. B., & Kronvang, B. (2009). Phosphorus in riparian
buffers: Review of their efficiency. Journal of Environ-
mental Quality, 38, 1942–1955.
Hornberger, G. M., Bencala, K. E., & McKnight, D. M. (1994).
Hydrological controls on dissolved organic carbon during
snowmelt in the Snake River near Montezuma, Colorado.
Biogeochemistry, 25, 147–165.
Kibichii, S., Shivoga, W. A., Muchiri, M., & Miller, S. N.
(2007). Macroinvertebrate assemblages along a land use
gradient in the upper River Njoro Watershed of Lake
Nakuru Drainage Basin, Kenya. Lakes & Reservoirs:
Research and Management, 12, 107–117.
Klapproth, J. C., & Johnson, J. E. (2000). Understanding the
science behind riparian forest buffers: Effects on water
quality. Virginia Cooperative Extension (2000) Publica-
tion 420–151. Virginia State University, VA, U.S.A.
Kronvang, B., Jensen, J. P., Hoffman, C. C., & Boers, P.
(2001). Nitrogen transport and fate in European streams,
rivers, lakes and wetlands. In R. F. Follet & J. L. Hartfield
(Eds.), Nitrogen in the environment: Sources, problems
and management (pp. 183–206). Amsterdam: Elsevier.
Lelo, F. K., Chiuri, W., & Jenkins, M. W. (2005). Managing the
River Njoro Watershed, Kenya: Conflicting laws, policies
and community priorities. Paper presented at the Interna-
tional Workshop on African Water Laws: Plural Legisla-
tive Frameworks for Rural Water Management in Africa,
Johannesburg, South Africa, January 26–28, 2005.
Lowrance, R. R., Todd, R., Fail, J., Jr., et al. (1984). Riparian
forests as nutrient filters in agricultural watersheds.
Bioscience, 34, 374–377.
Magana, A. B., & Bretschko, G. (2003). Retention of coarse
particulate organic matter on the sediments of Njoro River,
Kenya. International Review of Hydrobiology, 88, 414–
426.
Mainuri, Z. G. (2006). Land use effects on the spatial
distribution of soil aggregate stability within the River
Njoro Watershed, Kenya (Master’s thesis). Kenya: Egerton
University.
Mathooko, J. M. (2001). Disturbance of a Kenyan Rift Valley
stream by the daily activities of local people and their
livestock. Hydrobiologia, 458, 131–139.
McBride, M., & Booth, D. E. (2005). Urban impacts on
physical stream condition: Effects of spatial scale, con-
nectivity, and longitudinal trends. Journal of the American
Water Resources Association Paper No. 04032.
McCall, G. J. (1967). Geology of the Nakuru Thomson’sFalls—
Lake Hannington area (Ministry of Natural Resources).
Nairobi: Geological Survey of Kenya.
Ministry of Finance and Planning. (2002). Kenya 1999
population and housing census. Nairobi: Central Bureau
of Statistics.
Mokaya, S. K., Mathooko, J. M., & Leitchtfried, M. (2004).
Influence of anthropogenic activities on water quality of a
tropical stream ecosystem. African Journal of Ecology, 42,
281–288.
Ohrui, K., & Mitchell, M. J. (1998). Spatial patterns of soil
nitrate in Japanese forested watersheds: Importance of the
Water Air Soil Pollut
near-stream zone as a source of nitrate in stream water.
Hydrological Processes, 12, 1433–1445.
Oldham, L. (2003a). Nitrogen in Mississippi Soils. Mississippi
State University Extension Service. Information sheet 767.
Oldham, L. (2003b). Phosphorus in Mississippi Soils. Mississippi
State University Extension Service. Information sheet 871.
Osborne, L. L., & Kovacic, D. A. (1993). The influence of
riparian vegetation on nutrient losses in a midwestern
stream watershed. Freshwater Biology, 29, 243–258.
Pionke, H. B., Gburek, W. J., Sharpley, A. N., & Shnabel, R. R.
(1996). Flow and nutrient export patterns for an agricul-
tural hill-land watershed. Water Resources Research, 32,
1795–1804.
Republic of Kenya. (2002). Kenya Water Act.Nairobi:
Government.
Shepherd, K. D. (2005). Infrared diagnostics: New technology
for rapid diagnosis of soil, plant and livestock health
constraints in farms and watersheds. Paper presented at the
3 rd World congress on conservation agriculture: Linking
production, livelihoods and conservation, Nairobi, Kenya,
October 3–7, 2005.
Shivoga, W. A. (2001). Influence of hydrology on the structure
of invertebrate communities in two streams flowing into
Lake Nakuru, Kenya. Hydrobiologia, 458, 121–130.
Shivoga, W. A., Muchiri, M., Kibichii, S., et al. (2007).
Influences of land use/cover on water quality in the
upper and middle reaches of River Njoro, Kenya.
Lakes & Reservoirs: Research and Management, 12, 97–
105.
Sweeney, B. W., & Blaine, J. G. (2007). Resurrecting the in-
stream side of riparian forests. Journal of Contemporary
Water Research and Education, 136,17–27.
Vidon, P., Allan, C., Burns, D., et al. (2010). Hot spots and hot
moments in riparian zones: Potential for improved water
quality management. Journal of the American Water
Resources Association, 46, 278–298.
White, P. S., & Walker, J. S. (1997). Approximating nature’s
variation: Selecting and using reference information in
restoration ecology. Restoration Ecology, 5, 338–349.
Wyant, J. G., & Ellis, J. E. (1990). Composition patterns of
riparian woodlands in the Rift Valley of Northern Kenya.
Vegetation, 89,23–37.
Water Air Soil Pollut
