Wadi Zabid is one of the main agricultural Wadis in the Tihama Plain. Wide ranges of crops have been cultivated such as cereals, vegetables, fruits, and some cash crops like cotton, sesame and tobacco. It has provided the country with high valued food basket as well as exporting to the neighboring countries. The government realized the importance of Wadi Zabid and started in 1979 constructing diversion works and canals to maximize the agricultural output of the Wadi. During the past twenty years of operations, sediment deposition in front of the diversion works and along the canal system posed serious threats to the project network. Sediment deposition caused many problems such as blocking the off take pipes and gates, raising canal beds and reducing canals slope, increasing the field levels and reducing canal flow capacity. This study would assess the changes in canal sections and profiles at which the present canal profiles and sections are drawn and compared with the canal profiles and sections of the previous years to provide sound basis for problems assessment. Laboratory experiments and sieve analysis were conducted to analyze canal bed samples in order to draw recent grading curves and compared it with the previous grading curves of the canals. This study exposed the fact that canal sections and profiles are changed greatly due to the accumulation of sediments in the upper reaches of the canals whilst equilibrium sections and profiles existed in the middle parts with section erosion in the lower parts of the canal reaches. It also indicated that greater sediments sizes were deposited in the canal upper reaches, which meant that canal behavior is changed and does not work as it was designed for.

Key words: Sediments, Irrigation Canals, Profile, and Cross-section

1. Introduction

1.1 General Description of the Study Area

Yemen Republic is located in the southwestern corner of the Arabian Peninsula (figure 1.1). In the western part of Yemen is the Tihama coastal plain, bordered to the west by the Red Sea and to the east by the mountains. The Tihama plain is a semi-arid coastal plain, which is 25- 45 km wide. The total cultivated land is about 500,000 ha which is used to produce various cereal, vegetable and different crops. More than 30% of the Tihama plain consists of alluvial fans, crossed by seven major Wadis originating from the mountains in the east as shown in figure (1.1). These seven Wadis have steep bed slopes of 0.5 to 100 and their flows are characterized by series of multi-peak flash floods (spates). The water flows in a series of multi-peak flash floods.

Settlement in the Tihama plain has been dependent on flood-spreading techniques of Wadi flows for irrigation, which have been developed over many years. In order to make best use of the agricultural potential of the Tihama plain, the Ministry of Agricultural and Water Resources planned a single-wadi development strategy, starting from the three major wadis; Zabid, Rima and Mawr. Major irrigation development projects have been completed in the three wadis (Zabid (1979), Mawr (1984) and Rima (1988)).

1.1.1 Wadi Zabid

Wadi Zabid is the second major Wadi after Wadi Mawr in the Tihama plain (see table 1.1). Wadi Zabid runs east west direction from the foothills of Ibb Governorate with a total length of 250 km. It contains one of the oldest and most highly developed traditional spate systems in the Tihama plain and was the first area given funding for development.

Wadi Zabid's climate is generally described as tropical with a rainy summer and dry winter. The annual rainfall varies from 100 mm near the Red Sea coast to 600 mm in the foothills. The annual rainfall in the catchments area of the Wadi ranges from 400 to 600 mm/yr, which increases in some years up to 1000 mm [2].
There are two rainfall seasons and one dry season in Wadi Zabid:
1. The first wet season from March to May: rainfall occurs in the middle and upper catchments areas of the Wadi and generates floods.
2. The second wet season from May to September: rainfall occurs in the Tihama plain and in the Wadi catchments also [3].
3. The dry season from October to March: a strong southerly wind is predominate and there is seldom rainfall.

The annual average temperature in the area is 30.5 0C, the maximum is 43.6 0C and the minimum is 15.00C as measured in the FAO camp at Jerbah station in 1970. The annual sunshine duration was 2760 hours during the year 1970 (7.6 hrs on average). The relative humidity ranges from 15% to 98% with an annual average of 65% [3]

Wadi Zabid traditional irrigation system consists of 16 main supply canals, which divert flash floods and base flow from the Wadi to the fields by means of deflectors or dykes crossing the Wadi bed. The structures are made from poorly compacted earthen materials (soil, gravel, tree branches and boulders). Hence these structures are often damaged or completely destroyed by medium to large floods [4].

In Zabid area basin irrigation has been practiced for centuries and water was distributed by a field -to-field method. The water rights in the Tihama Wadis declare the priority of higher lying area over lower lying ones in diverting water from the Wadi for irrigation (Al Aala-Fal -Aala), [4]. In the Zabid area special water rights have been formulated approximately 500 years ago. These rights make maximum use of the base flow and periodic floods by means of a time and space relationship based on the experience of centuries. The irrigated areas supplied by the 16 canals were divided into three groups. The timing and frequency of water intake were defined for each group as shown in table (1.2). In 1996 the existing water rights law still recognized [3].

Wadi Zabid irrigation system has 5 diversion structures with 9 head regulators (see figure 1.3) serving 16 canals (see table 1.3), which allow the available spate flows to be allocated in accordance with the traditional water rights [4].

The average quantity of suspended sediment in an average year in Wadi Zabid is about 3,000,000 tons (1,600,000 ) and the bed load is about 430,000 tons (240,000 ) [1]. The sediment concentration especially in the sand size rages rises to more than 10% by weight during large floods [6].

1.1.2 Existing Problems in the Irrigation System

There are two kinds of problems, affecting the irrigation system namely operation and maintenance problems sedimentation problems. These problems would be described in the following sections.

1.1.2.1 Operation and Maintenance Problems
The main problem in the operation and maintenance is the inadequate control of the head works to divert the Wadi flows to the irrigation canals. This problem occurred by the following:
A. The erosion of the concrete in the head regulators and sluiceway: -
This erosion occurs as a result of filling of the headwork pool to the weir crest level by
sediment, which caused the flow velocity to increase until it exceeded the design
velocity. Also erosion was caused by big stones, which were carried by large floods and
passed through the sluiceway or over the weir crest causing severe damages to the crest,
the weir body and stilling basing.
B. The intake was incapable to divert sufficient irrigation water: -
i- The bed level has been increased in the head reach, especially behind the gates as shown in pictures (1)
ii- Decreased of the bed slope in the head reach behind the gates, which was caused by sediment deposition in this place.
iii- The level of the sluiceway and the intake was at the same level, which caused the sluiceway to flush sediments only to the level of the intake gates.
C. The erosion of the canal sides: -
i- Change the flow direction in the canals.
ii- The soil was too loose at the canal sides; there were no banks stabilization made from stones or stone mattresses.
iii- In some places along the canals the farmers plant crops on thecanal sides, which resulted in canal flows to erode soil banks.
D. Problems due to the lack of any light during the nights and due to the no operation of the sluiceways on Fridays and during holiday.

1.1.2.2 Sediment Problems in the Irrigation Canals

The sediment deposition was the main problem in the irrigation system, which was in fact a result of mismanagement of the control structures. The medium and large floods always carried heavy bed and suspended loads, which was deposited when the velocity decreased in the pool until the sediments accumulated up to the weir crest level and then the sediment is deposited in front and behind the intake gates of the head regulator. Then the sediment entered the irrigation canal and is deposited in the head reach and along the canals. This sediment deposition prevented the head regulator to control the Wadi flood flow in a correct way.

The sediment deposition caused the following problems
1. The sediment deposition upstream of the weirs increased the bed level of the pool up to the weir crest. See picture (2)
2. Canal bed level increased, especially in the head reaches;
3. Field levels increased; the level became higher than the water surface level in the main canal so the fields cannot be irrigated any more. See picture (3)
4. A reduction of the canal flow capacity lead to insufficient flows to meet the irrigation requirements;
5. Complete closure of the fields off takes. See picture (4)
6. High costs for the mechanical sediment removal, especially when the canal banks became very high due to the deposition of the removed sediments.
7. The coarse sediment and debris blocked off-takes pipes and the secondary canal bed rose especially in front of and behind the off-takes. See pictures (1 and 4).

2. Materials and Methods

2.1 Sediment transport studies in the Tihama plain

Several sediment transport studies were carried out in the Tihama wadis as part of the various visibilities / design studies of the irrigation systems in these wadis such as the following: -
1) TESCO -VIZTERV -VlTUKI, (1971): This firm investigated the suspended sediment and bed load transportation in Wadi Zabid main canals. They estimated the quantity of suspended sediments and bed load in an average year at 3 and 4 million tons/year.
2) TIPTON and KALMBACH, (1980): This firm used DH-59 sampler, with a series of single-stage samplers at the gauging site at Wadi Mawr. The sediment concentration measurements were made for discharges ranging from 2-90 . They defined that the suspended sediments, concentration varies from 150 to 50,000 ppm.
3) LAWRENCE, (1986 and 1987): This firm made a study to predict the sediments concentration in Wadi Zabid irrigation system concluded that the sediments concentration during high floods could reach 10% by weight.
4) NESPAK, (1989): This firm made a study to predict the sediments concentration in Wadi Siham main channel. The measurements were made at two stations with discharges of 2 and 12 . They concluded that the sediments concentrations at the two flows were 10,000 ppm and 23,000 ppm respectively.

2.2 Data collection and measurements
This study was partly carried out in the field at which measurements of the discharges, the particle size, the cross-section profiles and the longitudinal profiles for parts of the canals were investigated and partly in the labs. During the laboratory and field experiments several measuring equipment and apparatuses were used, such as current meter, bed load sediment sampler, suspended sediment sampler, leveling equipment, stopwatch, filter paper, etc. This part would provide a short outline of the types of measurements and the experiments conducted, their purpose and the equipment used.

2.2.1 Longitudinal Profile and Cross-sections Measurements

The longitudinal profile was measured to find the bed level elevation at several parts along the canal to estimate the sediment deposition in the canal bed. The profile was measured by leveling equipment (level, staff, etc). The measurements of the profile were taken at the centre of the canal bed and at an interval of 100 m along the canal. During the fieldwork two profiles were measured; one for the Bunny-Barry canal with 2700 m long (see figure 1.3) for the whole profile. The second profile was for the Mawi- Yusifi canal with a total length of 2100 m (see figure 1.4) for the whole profile. The cross-sections were measured to find the elevation and shape of the canal at specific points and at different dates.

2.2.2 Measurements of particle size, discharge, and velocity

2.2.3 Particle size measurements

A sufficient large portion of bed material (not less than 500 g) was taken from the sample and put in an oven at 110 °C for 24 hours. The dried sample was then put in a mechanical sieve apparatus and sieved for 15 minute. The weight retained and the percentages passing were determined and the grading curves were drawn. A comparison of the grading curves were compared with those of Lawrence (1983) [6], showed clearly that the bed material of the measured curves was coarser than those curves made by Lawrence (1983) [6], which was due to the following reasons:
a) The increase in bed slope.
b) The use of the total flow during small to medium floods without flushing of the sluice
gate, especially at the beginning of every flood during the irrigation period
c) Sediment accumulation in the pool behind the weir until it reaches the intake entrance.

A sufficient large portion (mass) from the sample, which was brought from the field as suspended load, was washed through the sieve no.200 (0.075µm) in order to obtain a mass of 500 g. This mass was placed in the oven at 110 °C for 24 hours to dry.

2.2.3.1 Discharge and velocity measurements

The discharge at a given section could be measured by several methods such as: current meter method, float method or dilution method. The choice of the method depended on the conditions present at the site. To facilitate comparison, the sites chosen for discharge measurements were the same as those selected by Lawrence during his 1982-1983 fieldwork. The accuracy of the discharge measurement depended on the number of verticals at which the depth and the velocity were measured. The position of the verticals should be in line with the variation in canal bed elevation and the horizontal variation in velocity. The width between any two verticals should not be more than 1/20 of the total width. The channel width was measured from a fixed reference point (usually the initial point on the bank). A graduated tape determined the distance between verticals and the depth was measured with the graduated metal rod of the current meter.

The velocity was measured at one or more points in each vertical. The velocity was then determined by counting the revolutions of the propeller of the meter during 50 seconds at every point. The average velocity was determined by the three-point method, which took the velocity observations in each vertical at 0.2, 0.6, and 0.8 intervals of the total water depth measured from the water surface. The average of the three values gave the mean or average velocity in the vertical.

3. Results And Discussion

3.1 Canals Longitudinal Profiles and Cross Sections

The elevation surveys of Bunny-Barry and Mawi-Yusifi canals were done during the fieldwork. From these surveys the bed levels were computed and compared with the designed bed levels (as shown in figures 1.3 and 1.4). A deposition of sediment occurred in all canals reaches, but it was most severe in the head reaches of canals close to the head regulator and at the upper portions of every reach before the drop structure. The deposition at the head reach caused the canal bed level to increase to the extent that water level became higher than that in the upstream of the head regulator.

3.1.1 The Bunny-Barry canal longitudinal profile shows:

There was much sediment deposition in the head reach of this canal and little erosion in the downstream reaches (as shown in figure 1.3). The present longitudinal 1996 profile for the downstream reaches appeared lower than the design and other profiles which were measured in the previous years. However, the head reach profiles for the years 1996 and 1983 survey were similar due to a high sediment deposition and the widening of the canal. The following observations can be made: -

i. Large sediment deposition in the head reach of the canal (from 0 to 600 mdownstream the
head regulator) especially the first 300 m. The deposition of sediments decreased along the
canal in a downstream direction until erosion occurred near the end of the reach.
ii. Less sediment deposition was observed in the 600 to 1250 m reach.
This sediment deposition also decreased in a downstream direction until erosion was
observed at the end reach;
iii. In the 1250 to 2000 m and 2150 to 2730 m reaches much erosion has been observed.
This erosion of the canal bed was due to an increase of the bed slope and flow
velocities.In the 2150 to 2730 m reach erosion increased with time (as shown in
figure 1.3) and due to the increase in bed slope which increased the flow velocity;
iv. In the 2000 to 2150 m reach a balanced condition for erosion and deposition was
observed, especially for the present survey.

The cross-sections profiles of Bunny-Barry canal showed both the rise in canal bed levels and the reduction in canal width in the downstream reaches. The reduction of canal width in the downstream canal cross-sections was partly due to the natural deposition of sediments in the canal side where the flow velocity is slower. The second reason for reduction of canal width was the disposal of sediment on the canal banks and there back filling into the canal sides. The cross section at the head reaches was enlarged forming a small settling basin immediately downstream of the head regulator. The bed level at section no.1 during the present survey was about 1.0 m above the design full supply level. The accumulation of sediment caused a reduction of both the canal intake capacity and the bed slope near the head regulator. Cross section no.1 (figure 1.5) shows the enlarge section since the 1987 survey, which could be the result of sediment removal from the canal bed to the canal bank. However, there was a reduction in the cross-section area and erosion in the canal bed at the downstream reaches (section. 2, 3 and figures 1.6 and 1.7). The erosion in the canal bed was due to the increase of bed slope and flow velocities, and the sediments clearance before the present measurements

3.1.2 The Mawi-Yusifi canal longitudinal profile shows

Substantial depositions of sediments occurred, especially at the head reach near the head regulator (0 to 429 m reach) and after drop structures 1 and 2. The deposition in these reaches was about 1.5 m thick at the upper part, which decreased, in a downstream direction forming a very steep bed slope (as shown in figure1.4). This figure also shows similar but thinner layer of sediment in the (429 to 800 m, 800 to 1580 m and 1580 to 2100 m) reaches.

From the longitudinal profile, the sediment deposition height reached about 1.5 m above the canal bed at some places (as shown in figure 1.4). This canal need sediments clearance to prevent the sediment transport of coarse materials through the downstream reaches and finally to the fields.

Comparing the existing canal bed slope for each reach (as determined from the profiles data) with the design or built bed slopes (see table 1.4), it became clear that the bed slop has increased with time due to the sedimentation on the canal bed and the erosion of this bed in some reaches due to the increased bed slopes and flow velocities. The results in table (1.4) show that the average in the bed slopes of Bunny-Barry was about 1.6 times the design bed slope except at the head reach where this increase was about 2.6 times the design bed slope.
The overall increase in bed slope of Mawi-Yusifi is about 9 times the built slopes in all the sections. At the head reaches and upper portion of some reaches the sediment deposit was around 1.5 m above canal built bed level.

Finally, for all canals the most significant impact of sedimentation occurred at the canal head reach, near the head regulator. This sedimentation may reduce the diverted water to the canals especially when the bed level closer to the intake becomes higher than the intake bed level. In this case the sediment had to be removed (mechanically) in order to increase the diverted wafer to the canals and to minimize transporting sediments downstream through the canal an finally to the fields.

3.2 Canal Bed Materials

Bed materials samples were collected from various locations along the Bunny-Barry canal and from the Mawi-Yusifi canals. The bed material of Bunny-Barry canal was found to consist of coarse sand and gravel at the head reach (see section 1 of figure 1.8) and coarse to fine sand at the downstream (see sections 2 and 3 of figures 1.9 and 1.10, respectively). The comparison of the grading curves of the Bunny-Barry canal profile observed by Lawrence and the present grading curve from Mawi-Yusifi canal (see figure 1.11), clearly demonstrated that the bed material of the measured curves was a little bit coarser than that mad by Lawrence (1983) (6). This due the following reasons:

a. The increase in canal bed slope.
b. The use of the total flow during small to medium floods, without sluiceway flushing especially at the beginning of every flood during the irrigation period.
c. Sediment accumulation in the pool until it reaches the intake entrance, which allows the coarse bed load to enter to canal head reach.

The close to the head reach (in the first 100 m) ranges from 30 to 17 mm in size and reduces in the downstream reaches to 0.36 mm. This bed material was a little bit coarser than that observed by Lawrence in 983.

The suspended sediments, which enter the fields, cause elevation of the soil surface, which eventually caused difficulties in irrigating the fields from the usual off-take. To overcome these difficulties, farmers usually lay down a temporary earth dykes to pond the water up to enter the fields. These dykes retained the sediment behind them and causing canal bed rising
The sediment quantities, which entered an irrigation canal, were estimated from the field measurement for two floods: one in Bunnay-Barry canal (section no.1 at 100 m downstream head regulator) and one for Mawi-Yusifi canal (section no. 1 at 100 m D/S head regulator). The rate of sedimentation was about 39 ton/ day for Bunnay-Barry canal and 126 ton / day for Mawi-Yusifi canal.

Moreover, from the economical point view, the field data indicates that the clearance of sediments from the canals costs 700 YR /ton (5.6 US $/ ton). So the sedimentation of one day will cost about 30,000 YR /day 240 US $/day) for Bunnay-Barry canal and 90,000 YR / day (720 US /dy) for Mawi-Yusifi canal. According to the rate sediment predicted during this study the total costs per season (for 45 days floods) are 1,350,000 YR /year (10,800 Us $ /year) and 4,050,000 YR / year (32,400 U5$/year) respectively.

4. Conclusions and Recommendations

4.1 Conclusions

Based on the results and discussion in the preceding sections, the following conclusions can be drawn:

1- The deposition of sediment along the irrigation canals in Wadi Zabid caused change in canal bed slope and cross section.
2- The operation of sluiceway was not sufficient to eliminate the bed load from the flow before entering the canal.
3- There was no use of sediment control structures.
4- Farmers should be prevented from constructing earth dykes across the canals to raise water levels.
5- High sediment concentration naturally existed in Wadi Zabid especially during medium to high floods (which reached 100,000 ppm in very high floods; according to Lawrence, 1986) (6) required thorough investigation in order to prevent sediment deposition in the systems.
6- Canal maintenance and removal of coarse sediments from the head reaches would prevent sediments from transporting further down stream the canal systems and to the fields. The coarse sediment transported to the fields caused blockage of the off-takes pipe and changed the fields soil uniformity, which will negatively affect the planting practices
7- The sediment deposition at the head reaches near the intakes resulted in a large reduction in the quantities of diverted water for irrigation especially during medium to large floods.
8- More water could be diverted to the canals if the openings intake gates were adjusted during medium to large floods and the sediments were always removed from the head reach.

4.2 Recommendations

In order to improve Wadi Zabid irrigation system (by reducing the sedimentation problem through the irrigation canals and minimizing the sediment clearance costs) the following recommendations are made:
1. The head reach of the irrigation canals needs more frequent sediment clearance than the down stream reaches in order to keep its bed level as constructed in order to increase the diverted water to the canals (especially during medium to large floods) and to reduce the coarse bed materials transportable to the downstream reaches
2. The canals intakes gates should be adjusted during medium to large floods and sluiceway should be operated according to the operating procedures. This will increase the diverted water to the canals (especially during medium to large floods) and to reduce the coarse bed martial that entered to the canals.
3. The heavy sediment load could be prevented from entering the irrigation canals by closing the head regulators intake gates and excluding all flows with heavy sediment concentration for 15 to 20 minutes at the beginning of floods (especially during medium to large floods)
4. In order to minimize the bed load sediment in the irrigation canal flow, sediment extractor (Vertex tube or Tunnel type) could be constructed at the canals head reach
5. Several settle basins could be constructed along the canals especially at the first two reaches from the head regulator where much sedimentation occurs. These settle basins would be easer to clean than clearing the entire canal reaches.
6. For maintenance purposes the irrigation canals could be divided into several sections with specific users being responsible for each section close to their lands. The maintenance and cleaning of the irrigation canals should be the responsibility of the various users, each in respect to his irrigated area. This would minimize the sedimentation problem and avoid the bed-load sediment transport to the fields, and would decrease the maintenance cost.

References

[1] TESCO, " Survey of the Agricultural Potential of Wadi Zabid in Yemen", Technical
Report No. 12, Budabest , Hungary ,1971.
[2] Scheitz, E. L., " Certain Aspects and Problems of Wadi Development ", Vizier
Consulting Company, Budabest, Hungary, 1987.
[3] NESPAK, " Wadi Siham Project Inception Report ", Mai Report, TDA, Yemen, 1989
[4] Tahir, T.M. " Traditional Water Rights Versus Water Availability- Case Study Wadi
Siham ", Water Resources In the Arab World Conference, Tripoli, Libya, 1996.
[5] Lawrence, P. et al, "Sediment Control in Wadi Irrigation Systems ", Hydraulic Research,
Wallingford, UK, 1986
[6] Lawrence, P. et al, " Wadi Zabid Diversion Structures-Field Performance Measurements
", Report No. 73, Hydraulic Research, Wallingford, UK, 1983

Appendix