Resource information
The study was carried out on twelve runoff
plots installed at Kabete Campus Field Station,
University of Nairobi, on a 10% natural slope of
eutric Nitisol to assess the effect of grass strips
in controlling soil loss and runoff on sloping land.
Four treatments, a control plot without grass strip
and three different widths (0.5, 1.0 and 1.5 m.) of
Nandi setaria (Setaria anceps), were tested under
natural rainfall during 1982 and 1983, simulated
runoff, and simulated rainfall of 80 mm/hr.
Results showed that the treatment effect was
highly significant in reducing both soil loss and
runoff under natural rainfall. Annual soil loss
for the control plot was 97.7 t/ha. while for the
0.5, 1.0 and 1.5 m. wide grass strips annual soil
losses were 35.4, 35.6 and 17.8 t/ha. (36, 36 and
18%) respectively. Annual runoff from the control
plot was 100~. (20%) and percentage runoff for
the 0.5, 1.0 and 1.5 m. wide grass strips were
56, 44 and 24% of the control piot, respectively.
The difference between the three different grass
strip widths was not stat~stically significant
although annual soil loss from the 0.5 m. wide
grass strip was double that from the 1.5 m. wide
grass strip.
The simulated runoff test showed that soil
losses from the plot with 0.5 m. wide grass strip
were only 4.5, 7.8 and 41.1% of the soil losses
from the control plot under the three application
. -4 -4 -4
rates of 1.5 x 10 ,2.5 x 10 and 3.6 x 10
3 -1 m s respectively •• Runoff from the 0.5 m. wide
grass strip were 1.9, 3.5 and 50.2% of the runoff
from the control plot for the same application
rates.
Under heavy simulated rainfall, soil losses
from the plot with 1.5 m. wide grass strip were
17.1, 25.1 and 36.7% of the soil losses from the
control plot for dry, wet and very wet antecedent
soil moisture conditions, respectively. Runoff from
the plot with 1.5 m. wide grass strip were 32.7, 41.0
and 46.0% of the runoff from the control plot for the
same three applications. Efficiency of grass
strips in controlling soil loss and runoff
increased from the 0.5m. wide grass strip to
1.5 m. wide grass strip and decreased from the dry
to very wet antecedent soil moisture conditions.
Strip width was exponentially related to soil
loss and runoff with coefficient of determination
(r2) of 0.96 for soil loss and 0.88 for runoff
under dry antecedent soil moisture condition.
Sediment was deposited on the plots with grass
strips, with most of the deposition taking place
on the uphill edge of the grass strip, extending two
metres in one and half years. There was little
deposition in the grass strips. The maximum mean
depth of deposition was 7.6 cm. at the 0.5 m. wide
grass strip. Much of scour took place on the upper
section of the plots. In almost all the plots
the ground slope was reduced by about 2%
Two thirds of sediment transported was in sand
size aggregates although the soil contained more
than 50% clay. Grass strips had no influence in
the trapping of different sized aggregates.
Of the six erosivity indices tested, the amount
of storm (A), in mm., was highly correlated
(r = 0.88) with soil loss followed by the El1S(r=0.86)
and the EI30 (r=O.73) indices. The KE> 25 index gave
the least correlation coefficient (r=O.41). The
~nnu~l r~inf~ll ~rosivi~y f~e~or (R) for R~b~~~ w~s
est.imated to be 246 m~t-cm/ha-hr and the soi l,
erodibility factor (K) was 0.21 t-ha-hr/ha-m-t-cm.
which wa s higher than that from previous findings.
