iii ASSESSMENT OF HYDROLOGICAL CONTROLS ON GULLY FORMATION NEAR LAKE TANA, NORTHERN HIGHLANDS OF ETHIOPIA A Thesis Presented to the Faculty of the Graduate School of Cornell University In Partial Fulfillment of the Requirements for the Degree of Master of Integrated watershed management and Hydrology By Tigist Yazie Tebebu May 2009
62
Embed
ASSESSMENT OF HYDROLOGICAL CONTROLS ON ...soilandwater.bee.cornell.edu/Research/international/docs/...iii ASSESSMENT OF HYDROLOGICAL CONTROLS ON GULLY FORMATION NEAR LAKE TANA, NORTHERN
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
iii
ASSESSMENT OF HYDROLOGICAL CONTROLS ON GULLY FORMATION
NEAR LAKE TANA, NORTHERN HIGHLANDS OF ETHIOPIA
A Thesis
Presented to the Faculty of the Graduate School
of Cornell University
In Partial Fulfillment of the Requirements for the Degree of
Master of Integrated watershed management and Hydrology
∗ A Piezometer which did not reach the bed rock due to deep soil layer
Runoff depths of September 2008 were measured manually at 10
minutes interval during a storm and on average every 6 hours to measure non-
storm flows. The velocity of runoff (m s-1) was determined by float method in
which a float was released on the water surface a distant of 4.2 m above the
weir outlet and the time required for the float to reach the weir was recorded.
The depth records and runoff velocity were subsequently converted to the
11
corresponding discharge and runoff volume using a rating equation developed
for the gully and weir.
VAQ = Equation 1
Where Q is the discharge (m3 s-1), v is the velocity of runoff (m s-1) determined
by float method, and A is the cross sectional area of the weir or the flow (m2)
calculated by multiplying the base width of the weir, W, (m) by the runoff
depth, D, at the outlet (m).
Figure 2: The dimensions of a rectangular weir constructed at the outlet of the gully
Several 70 cm long metal rods marked at 10 cm intervals were driven
into the bed of the gully at the beginning of the study period rainy season at
the head cut gully segments and selected gully bases with 10 cm of the rods
exposed above the gully surface (Figure 3). The length of the exposed ends of
the metal rods were measured using the 10 cm intervals at the beginning and
end of the monitoring period.
12
Precipitation in 2008 rainy season was measured daily using a
calibrated rain gauge at a standard runoff plot (Adet Research Center) about 1
km north of the sub ‘catchment in the monitoring dates. It was crosschecked
with five simple rain gauges that were installed 60 m north of the weir.
Infiltration tests were done with double ring infiltrometer at the north and south
gully branch upslope recharge areas separately during August and September
2008. The diameter of the inner ring and the outer ring measured 32 cm 53
cm.
Figure 3: The location of piezometers, the constructed weir, and metal rods installed in the head cut areas.
Monitoring Gully through Community Interviews
The evolution of the gully was studied in detail using the assessment of
gully erosion rates through interviews and measurements (AGERTIM) method
(Nyssen et al., 2006). The AGERTIM method has been used for the
assessment of gully erosion rates over the last decades in the absence of
historically written or photographic information (Nyssen et al., 2006). The
0 8 0 160 240 32 040Me ters
13
selected gully was visited with five key informants in four age groups (farmers
of the age of 20, 30, 40, and 50 years) and the age of the various gully
segments was estimated through different questions. The key informants
located different segments of the gully head and the major changes that
occurred during the past three decades. Based on these different time
horizons an event-oriented calendar (Table 2) was prepared.
Table 2: Event oriented calendar used in the questions
The extent and location of the gully in its early stage was first
reconstructed with the oldest informants. Key events in the informants’ lives,
including the necessity to cross the road to fetch water from the rural-
developed well, appearance of a log over the gully and used as a bridge,
flooding of the road during the initial settlement of the nearby town, served as
time references for the progression of the gully.
The commencing points of the gully located at different ephemeral
springs were investigated in more detail. Information from the oldest key
informant (age around 50) were discussed and crosschecked with information
provided by younger informants. According to these interviews, major incisions
started in 1981 at the location of three ephemeral springs near the current
14
junction of the north and south branch. The main branch started around 2004
and almost has joined the other gully system.
Gully Evolution and Erosion Rate
The current surface area of the gully system was analyzed from a 0.58
m resolution Quick Bird image from 2005 and compared with the area of the
current gully system. The extent of the gully was delineated with a Garmin
GPS (2 m overall accuracy) and compared with the gully area of the 2005
satellite image (Figure 4).
Figure 4: The location of springs, dyke and major active gullying segments in the Debre- Mewi gully system. Gully areas marked with red triangles expanded in width and depth during the 2008 rainy season.
The gully volume was determined from the cross sectional area of the
gully and the length of the considered gully segments. A hand held Garmin
Etrex global positions system (GPS) receiver (Garmin International, Inc.,
Olathe, Kansas) was also used to measure and determine the spatial
15
coordinates for each gully segment. The gully segments were numbered
based on their location in the study area. The length of each segment between
two cross profiles, where the top width, bottom width and depth of the gully
segment were measured, using a 50m long surveyor’s tape. For each
segment, two or three top widths and three depths were measured at locations
where the gully cross-sections changed abruptly. Based on these
measurements an average gully volume of each segment could be estimated.
In addition, the shape of the gully was determined to understand the current
rate of gullying and to obtain whether parts of the gully stabilized naturally. The
cross-sectional shapes of the gully in the study site are almost trapezoidal or
V-shaped. They were usually larger in width than depth (large width to depth
ratio). The measurements were taken at the beginning of the rain (July,
considered as measurement of 2007) and end of the rainy season (October,
considered as measurement of 2008). Thus, the short-term erosion could be
estimated considering only the gully segment measurements taken during the
past rainy season. The eroded volume of each gully segment was calculated
using the cross sectional dimensions and the distance between cross
sections.
∑= ii ALV Equation 2
Where Li is the length of considered gully segment (m) and Ai is the
representative cross sectional area of the gully segment (m2).
Long-term gully erosion rates (t ha-1 yr-1) (RL) were calculated using the
estimated current volume (V) of the gully, the average bulk density (Bd) of
soils occurring in the contributing area, the time span of gully development in
years (T) and the watershed area in hectares (A).
16
Equation 3
The soil bulk density was estimated in different locations and depths
throughout the contributing area of the gully using a cylindrical core sampler
with a volume of 98 cm3. Six core samples were collected at the twenty-four
piezometer locations and 12 from within the gully at active erosion sites.
Samples were weighed then dried at 105oC for 24 hours and weighed again.
An average of all the samples was taken as the bulk density value for the gully
watershed.
Short-term erosion rates (t ha-1 yr-1) (Rs) were determined to estimate
the rate over the duration of this study.
Equation 4
Where V is the current volume of the gully and Vo is the initial volume at the
start of the study period.
Erosion per unit gully surface (t m-2), (RP), were estimated as:
Equation 5
Where V is the current volume of the gully and AP: Plan area of the gully (m2).
PP A
VBdR =
TABdVV
RS)( 0−
=
TAVBdRL =
17
CHAPTER THREE
3 RESULTS AND DISCUSSION
The Long term Evolution of Gully in Debre-Mewi
According to the respondents, in the earlier 1980’s, the valley bottom of
the study area was marshy, and grasses were grown all year long. The
hillslope area was covered with different tree species like Podocarpus
africana, Acacia spp. and other indigenous species. Two ephemeral and a
perennial spring were observed in the midslope gully segments. Currently
most tree species were cleared and eucalyptus plantations have expanded in
alarming rate since 1991. All respondents agreed on the incision location of
the current gully. Although one of the older and one of the younger
respondents disagreed in their statements, all other respondents confirm that
the locations of the two gully incisions were related to three springs in the
hillslope.
Since the beginning of the 1980s, one of the springs, which is located
approximately 272 m uphill from the current valley bottom gully (SP1 in Figure
4), showed running water throughout the year. One of the oldest respondents
mentioned that the water from this spring accumulated in a small pond and
served to hose down the maize seed and to prepare local beer for different
local ceremonies. According to the remaining respondents, it was also used as
a watering hole for cattle. Presently the pond is filled with “adebuha”,
sediments deposited from the surrounding agricultural lands and the slumping
of gully walls. According to the farmer, the time when the pond began to dry up
coincided with the incision of the gully in the valley bottom. It is hypothesized
that the drying up of the pond is likely caused through the lowered water table
in the gully.
18
In 1981 during the Derg regime of the new settlement period, passage
between the nearby town and the villages “Shanko Bahir” and “Yemariam
Wuha“to Debre-Mewi town was made via a footpath above the second spring.
The second spring is approximately 373m away from the valley bottom gully
and approximately 78m away from the first spring (section SP2, Figure 4).
According to the local informants, after the fall of the Derg regime in 1990 and
1991, the marshy area around the second spring changed into a branched
gully with a northern and a southern gully branch as soon as settlers returned
to their previous homeland ,“Shanko Bahir” and “Yemariam Wuha“ from the
Debre-Mewi town . Furthermore, the edges of the gully downstream of the
second spring were used as a border for the farmland after the land
distribution in 1996.
The third spring (SP3), located near the head cut areas of the southern
gully branch is still actively eroding into the hillside as indicated by newly
developed, ephemeral side branches (Figure 4) .These gully arms may have
developed due to seepage flow or interflow. Decades ago the spring water
was seeping from the start of the rainy season (June) until the following April.
Recently, this period of seepage has decreased from June to February.
Although the increasing livestock density and depletion of the vegetation cover
is easily to blame for the decrease in spring flow as proposed by (Øygarden,
2003), as we will see later it is much more likely that a decrease in
groundwater table due to drainage by a gullies is the cause of the decreased
water availably for the springs. Thus, the change management practice is only
indirectly the cause: less vegetation increased subsurface initiating the gulley
resulting lowered groundwater levels and ultimately decline in spring flow.
19
Short term Evolution of Gully in Debre-Mewi
The satellite image delineation of the gully system indicated the
planimetric area of the gully was 0.65 ha in 2005 while the GPS delineation of
the gully area showed the area had extended to 1.0 ha in 2007 and 1.43 ha in
2008 (Table 3 and Figure 5). The gully had increased by 43 and 60% within
three years and during the monitoring period of 2008, respectively. The total
length of the gully in the study site was 1319 m with a total volume of 24900
m3 in 2008 (Table 3). The northern gully branch is 427 m long has a volume of
1640 m3 in 2007 and increased by 41 % in 2008 to 2310m3. The southern
gully branch is 519 m long has a volume of 6670m3 in 2007 and increased by
41 % in 2008 to 8610m3. The intermediate gully, the most stabilized section of
the midslope gully has 186 m long and a volume of 3000 m3 and 3560 m3 in
2007 and 2008 respectively .It was increased by 19 %. The valley-bottom gully
segment was 159 m long and has a volume of 4816 m3 in 2007 in 2007 and
increased to 187 m long and 10,450 m3 in 2008, increased by 117%. The total
volume of the studied gully system increased approximately 46% during the
monitoring period. During the rainy season of 2008, a deep gully formed from
a former footpath that crossed the valley bottom branch of the gully. In this
gully branch, it was observed that the gully eroded at the beginning of the
rainy season due to sliding of the heavier and saturated topsoil on top of a
heavy silty clay layer, which functions as a semi-permeable layer and hinders
deeper percolation of infiltrated rainwater.
Most gully segments in the valley bottom and in the upper end of the
two gully branches in the midslope area were said to have been actively
eroding during the past 4 years as confirmed by one of the younger
informants.
20
Table 3: Long and short-term erosion rates in different gully locations of the Debre-Mewi watershed
Figure 5: The evolution of the studied gully system in Debre-Mewi (2005-2008).
NOTE: RL, Area specific long-term gully erosion rate since incision RS, Area specific short-term gully erosion rate (2007- 2008)
21
Saturated areas in the valley bottom, which were fenced by one of the
informants to grow grass for his cattle, as well as the communal grazing lands,
have shown accelerated gullying in recent years as shown in Figure 6. The
measurements with the piezometers, which are discussed in the later section
show that at these locations the water table is above the bottom of the gully.
Other changes that occurred in the watershed and could be possible causes
are a decrease of soil fertility and the expansion of croplands and greater
density of cattle
Figure 6: Actively forming gully at the most downstream end, the saturated valley bottom gully walls are actively eroding up to 5m deep and 26 m wide.
Gully Erosion Rate and Soil Loss
According to information obtained with the semi-structured interview the
approximate incision period of the gully segments ranged from one year for
the new gullies to 27 years for the old gullies. Thus, it was assumed that the
gully started to erode approximately 27 years ago. Local informants confirm
the alarming incision rate and the accelerated development of the gully since
2004.
22
The estimated long-term average soil loss rate (RL) due to gully erosion
of 17 t ha-1 yr-1 was estimated for the midslope gully segments since the
incision period of 1981 to 2007 and increased to 21 t ha-1 yr-1 in 2008 (Table
3). In comparison, the valley bottom gully’s long-term erosion rates were 13 t
ha-1 yr-1 from 1981 to 2007 and 27 t ha-1 yr-1 from 1981 to 2008. These values
are very high for the region compared to the result of other studies (Nyssen et
al., 2006). The percent increase in the erosion rate of the valley bottom gully
(109 %) compared to the midslope gully (24 %) within one rainy season is
likely caused by slumping of the gully walls and the surrounding soil at the
valley bottom. However, the erosion rates indicate that gully incision was
slower at the beginning of the incision period (1980s) compared to the current
rainy season. The volume of gully at the beginning of the monitoring period
(2007) was less than by 8800 m3 to the end the monitoring period (2008). The
area specific short-term gully erosion rates (RS) between 2007 and 2008 were
approximately 128 t ha-1 yr-1 for the midslope gully and contributes to 1.7 cm
soil loss for its 16.5 ha watershed and 402 t ha-1 yr-1 for the valley bottom gully
contributes to 3 cm soil loss of its 17.4 ha watershed. The recent increase in
measured gulley erosion is in accordance with local information.
Figure 7 shows the depth of soil (cm) eroded during the monitoring
time. The higher erosion rate estimated for the valley-bottom gully relative to
the upper part of the gully (midslope gully) is most likely caused by higher rate
of landslides and land slumps, which increased the sediment production, and
the presence of stabilized gully segments in the midslope gully section. The
higher erosion rate estimated for the southern gully branch than the northern
gully branch in the midslope gully is related with the presence of the springs.
The average soil loss per unit surface area of the gully was estimated to be
23
1.24 t m-2, based on measured gully volumes and the average soil bulk
density. The soil loss from the valley-bottom gully was estimated to be 2.9
t m-2 and 1.7 t m-2 for the midslope gully segments with an average of 2.2 t m-2
for the past 27 years. This indicates the severity of soil loss due to gully
erosion in the studied sub-catchment. The estimated erosion rates compare
well with gully erosion rates observed in other regions of the country, which
are in the range of 1 to 2 t m-2 (Daba et al., 2003).
Figure 7: The estimated soil loss from gully volume for the Valley bottom (VB) and midslope (southern (S) and northern (N) gully branches) gully in the two monitoring dates.
The average gully width and depth of the saturated valley bottom gully
segments was 22.1 m and 4.6 m, respectively, and 6.8 m and 2 m for the
midslope gully segments(including the north, south and intermediate gully
branches),. These measurements indicate that gullying occurred more
predominately in the valley bottom than in the hillslope areas.
24
Table 4 and Figure 8 show the average dimensions of the various
cross-sections of the various parts of the gully. The intermediate gully changed
least while the valley bottom increased in size the most. The top width of the
valley bottom doubled in just one year. The gully at the valley bottom was
actively progressing upstream and the average depth increase greatly as
shown in Table 4 and Figure 8. The other sections of the gully only slightly
increased in depth. Average top width for the north and south branch was in
the order of 25%.
Table 4: summary of the measured parameter average values and change in % for south, north intermediate and valley bottom gully branches.
25
Figure 8: The relative positions, widths (top and bottom) and depths of gully segments compared for two monitoring periods (2007 and 2008). (a) Northern gully branch( starting from intersection point to the north), (b) Southern gully branch (starting from intersection point to the south) and (c) Valley bottom gully(from the bottom end to splitting point) at a distance of 159 -187m away the valley bottom gully was not present in 2007.
26
Precipitation and Observed Discharge
The total precipitation from September 7, 2008 to September 26, 2008
was 97.5 mm. On September 13, it rained 23 mm with an average intensity of
17 mm hr-1 that caused the water to rise 70 cm at the weir flowing with a
velocity of 2.1 m sec-1. This rainfall event was similar to a late July storm that
caused gully wall slump. The hydropgraph of the September 13 storm is
depicted in Figure 9. The total runoff volume is 2,750 m3, which indicate within
the accuracy of the measurement that, nearly all the rain ran off. The
maximum suspended sediment concentration was 25.3 g l-1 occurring after the
peak flow.
Figure 9: A single storm event observed discharge for Debre Mewi gully system in Sep. 13 2008.
Gully Formation Mechanisms
In the next section, we will discuss how gullies are being formed and
widened. The mechanism for gully formation is different for the beginning of
the rainy season when the vertisols have large cracks that later in the rainy
27
season when the cracks are closed and the ground water table has increased.
We will first shortly discuss the initial gully erosion in the beginning of the rainy
season and the occurrence of pipes and then discuss extensively the effect of
water table on gully formation.
Gully Formation due to cracks and pipes
After the dry season, the vertisol around the gully has wide cracks is a 5
to 8 m rectangular pattern (Figure 10). In addition, there are a number of soil
pipes from 10 to 30 m long with diameters of 5 to 50 cm in mostly lower lying
areas usually near the gully (Figure 10). Especially in the saturated areas of
the valley bottom, several holes, vertical and horizontal pipes, were observed
at the interface of the organic A-horizon and the underlying red clay horizon
(heavy clay soil). After the first rains of the rainy season, surface runoff from a
high intensity rainfall fills the cracks and pipes filled with water, wetting up the
soil through these cracks. Since the soil was dry initially and was loose and
non-cohesive, the pipes and cracks could easily erode giving the water a high
sediment concentration. Deposition of sand at the end of the pipes clearly is
indicative of the erosion process. In addition subsurface soil pipes and tunnels
(Zhu, 1997), which are seen as initiators of gully erosion (Jones, 1987; Nyssen
et al., 2006).
Slumping of the gully walls is facilitated when the top organic A-horizon
becomes saturated and heavy and looses its stability due to the higher
potential of the groundwater exerting pressure onto the overlaying saturated
soil profile. The soil consequently looses its stability causing the slumping of
the gully walls and the surrounding soil (Zhu, 2003). The existence of grayish
colored clay at different depths of the gully wall and at the pipe outlets
28
indicates the level of the local long-term water table (Appendix 6). The relative
importance of this type of erosion becomes less as the rainy season
progresses because of closure of cracks.
Figure 10: The cracks, pipe and pipe out lets indicates the sub surface erosion and future incision of gullies.
Gully Formation due to High Water Table
During the progression of the rainy season, the water table builds up
and becomes the major cause for the gully widening and progressing uphill. In
the next section, we discuss the relation between water table height, gulley
depth for the actively widening gulley areas. For these active gully formation
areas the depth of the water table in piezometers is given in Figure 11.
Piezometers P23 and P24 are located in the valley bottom, P13 and P1 in the
29
south gully and P16 in the north gully. All the piezometers associated with the
active gully area are listed in Table 5 and their location is given in Figure 3.
Table 5: The location of representative piezometers in the studied gully system
The depths of the gully (Figure 12a) and the corresponding widths
(Figure 12b) bottom is depicted before (2007) and after (2008) the rainy
season for the valley bottom area as a function of the distance of point where
this gully joins the main branch. The average water table depth for the most
nearby piezometers( from bottom to top P24, P23, P22 and P26 and P17) are
shown as well and indicate that the valley bottom is saturated close to the
surface while further upstream the water table is below the gully bottom.
During the 2008 rainy season the gully advanced backward past the
187 m mark (Figure 12a) and increased up to 20 m in top width (Figure 12b).
In this region, the water table was near the surface and approximately 4 m
above the gully bottom (Figure 12a). Upstream of the 187 mark, the water
table is below the gully bottom (Figure 12a) and the gully is stable as can be
seen from Figure 12b since the width is not increasing.
The difference between the water table and gully bottom in the first 200
m of the gully varies from were 3.5 m to 4 m (Figure 12a). This means that
under static conduction the pore water pressure near the gully advance point
is 5 m causing the valley wall to fail repeatedly. In gully adjacent to piezometer
at 272 m the water table and the depth were both approximately 120 cm and
the gully was stable. More upstream at P17 the water table was below the
gully bottom. The gully walls did not fail at this location as well.
31
Figure 12: Gully dimensions before and after the 2008 rainy season for the valley bottom. a) Depths and average ground water table; b) change in top and bottom width and depth of the gully.
North Gully
The active gully erosion process in the North Gully is driven by the
ground water dynamics similarly to the valley bottom as shown in Figure 13.
0
200
400
600
50 159 187 244 272 322 373
Water de
pth to gu
lly bottom
and gully dep
th (cm
)
Gully length (m)
Gully depth Jul. 2008 Gully depth Oct. 2008
Average water table
Valley bottom
??Dry
??DryP24
P24 P23
P22 P26
P17
P17
a)
b)
32
The depth of the gully in 2008 is depicted in Figure 13a, and the change in
bottom and top width and depth during the 2008 rainy season in Figure 13b. In
addition, the height of maximum and average water table above the gully
bottom is shown in Figure 13 and is obtained by subtracting the water table
depth from the gully depth. Positive numbers mean that the water table is
above and negative numbers below the valley bottom.
Figure 13: Gully dimensions before and after the 2008 rainy season for the northern gully branch. a) Depths and average ground water table; b) change in top and bottom width and depth of the gully.
a)
b)
33
Although the relationship is not as dramatic as in Figure 12, the general
trend is the same. Where the water table is approximately 2 m below the
valley bottom (at 130 m) the gully does not change. When at 201 and 231m
above the junction with the south gully, the water table is the farthest
(approximately 75 cm) above the gully bottom and the water seeps into the
gully the gully dimensions increase most. For water tables near the gully
bottom, some small changes in gully dimension were observed.
South Gully
The gully widening in the south gully is influenced by a sapprolite and
rock close to surface. The rock outcrop (called dyke) and sapprolite together
with the gully depth is shown in Figure 14a. Figure 14b shows that the most
active gully formation occurred at 263 m from the junction with the south
branch just uphill from the dyke. Ground water table flow is blocked by the
dyke and is therefore above the gully bottom. Only downhill from the dyke at
115 m from the junction is water table is below the ditch bottom.
Unlike the north and valley bottom gully where the water table above
the valley bottom caused collapse of the walls, there is no widening in the gully
in the section 300 to 400 m from the junction with south gully despite that the
water table is between 1 and 2 m above the gully bottom. The only severe
widening of the gully occurred just upslope of the dyke (187 m from the
junction) where the water table was approximately 3 m above the gully bottom
(Figure 14). Here the water cannot flow further downhill due to the dyke and
seeps into the gully and causing the bank to slip. Further upslope the soil is
underlain by the sapprolite and water is seeping through the sapprolite to the
34
gully making the bank stable.
Figure 14: Gully dimensions before and after the 2008 rainy season for the southern gully branch. a) Depths and average ground water table; b) change in top and bottom width and depth of the gully.
Erosion of Stabilized Gullies
In stable areas of the gully where the water table is generally below the
channel bottom, some widening of the top of the gully is occurring. The typical
U shape of the active eroding gully with the 90 degrees wells becomes V
b)
a)
35
shaped due to grazing animals in and around the gully and the inflow of direct
hillslope runoff into the gullies. The direct runoff originates in many cases from
the cultural ditches used by farmers in the cropland for erosion control and by
concentrated flow of breached stone terraces similarly to that reported by
(Esser et al., 2002). In some other cases rills developed in the middle slope
gully section formed with an intervals of 0.2 to 1 m and caused gully bank
erosion. Thus, although concentrated runoff from the surrounding
farmlands both widen the gully and forms new gully branches as proposed by
(Øygarden, 2003), high ground water is the main cause of rapidly expanding
gully systems in the flatter areas of the landscape.
36
CHAPTER FOUR
4 CONCLUSIONS AND RECOMMENDATIONS
The present and past gullying process in the Debre Mewi watershed
was related to environmental changes that induce surface and subsurface
runoff from the hillside to the valley bottoms. Clearance of indigenous
vegetation in the 1990’s decreased annual evaporation causing increased
subsurface flow of water, thereby upsetting the delicate balance of
subsurface and surface flows in the watershed.
Gully erosion processes are driven by the local hydrology and
differences in soil properties. The actively eroding sections especially in the
valley bottom water table were in general closer to the ground surface on the
gully shoulder than in stabilized sections. Therefore, this research
suggests that the water table depths above the gully bottom are
responsible for the establishment of newly formed gullies in the vertisols
located flatter areas of the landscape. Once the gully is formed the water
table is being drawdown by the gully itself and the gully becomes more
stable and changes from a U to a V shape due to erosion of direct runoff
entering the gully. Additionally, erosion is caused by grazing animals or
food. Besides gully slumping due to high water tables, gully walls slide in the
beginning of the rainy season by water entering the wide cracks that are
formed during the dry season in the vertisols. Finally, many soil
pipes in the watershed play a role in the formation of gullies.
These findings suggest that erosion due to gully formation can be
decreased through conservation and protection measures that alter the
runoff response and the water balance so that the water table is lowered
below the gully bottom. This can be accomplished by reestablishing
37
deep-rooted permanent vegetation on the hillsides and, grasses in the
saturated valley bottoms. Artificial drainage will lower the water table
but is only of limited applicability under Ethiopian conditions. Finally,
improved land use practices will decrease erosion after the gully is
formed but likely not prevent the initial gully formation after the
landscape is cleared.
38
CHAPTER FIVE
5 REFERENCES
Billi, P., Dramis, F., 2002. Geomorphological investigation on gully erosion in
the Rift Valley and the northern highlands of Ethiopia.
Bocco, G., 1991. Gully erosiuon :processes and models. Progress in Physical
Geography 15,392-406.
Capra, A., Mazzara, L.M., Scicolone, B., 2005. Application of EGEM Model to
Appendix 2: The relation ship of daily discharge including base flow and rainfall recordings for twenty consecutive days (Sep. 7, 2008-Sep 27, 2008).
45
Appendix 3: The studied saturated valley bottom gully water table was near or at the surface resulted the slumping of the gully walls and the surrounding soil (SG=Gully Section and P=Piezometer).
46
Appendix 4: At the southern gully branch piezometer P19 down slope of the dyke, show a water table below the ditch bottom (SG 17 and P19) and the gully section was semi stabilized through out the monitoring period. While Piezometers P13 upslope of the dyke show a water table above the ditch bottom near or at the soil surface resulted the slumping of cropland.
47
Appendix 5: Head cut migration and undercut deepening was contributed for the development of northern gully branch at a point where the water table was below the ditch bottom (GS 8-10 and P21) and the active gully section Piezometer P16 water table was above the ditch bottom.
48
Appendix 6: The existence of grayish colored clay at different depths of the gully wall and at the pipe outlets indicates the level of the local long-term water table.
49
Appendix 7: The springs are the splitting points of the studied gully system at different locations.