Hydrological assessment of some old dams in Saudi Arabia under the current climate, environmental conditions, and the use of the advanced technology

Abstract

Dams were built in Saudi Arabia (SA) starting from the late fifties. The rainfall record at that time is rather short since the early rainfall records in the study area were in the mid-sixties. Therefore, due to the recent changes in climatic conditions worldwide and having longer rainfall records nowadays, it is very important to assess the existing dams under the current conditions. The assessment is based on some criteria: the length of the rainfall records, the rainfall storm hyetographs, storm duration, the type of the rainfall frequency distribution, topographic data, and the current land use/land cover conditions. Fourteen dams in the Riyadh region are chosen, as a pilot study, to perform the proposed dam evaluation procedure. Later, the study can be extended to any region in SA. The results show that increasing the record length leads to a convergence of the rainfall to an asymptotic value. The minimum record length to produce stable statistical rainfall is about 20 years. Nine out of fourteen dams have a storage capacity less than the 5-year return period. The statistical analysis showed that the measured rainfall in some past years corresponds to 50 and 100 years return periods. Gumbel Type I distribution, which was used in the analysis of the old dams, does not seem to be the best distribution under the recent climate data. The dams should also be assessed under the effects of climate change and urban expansion.

Introduction

Dams are built in Saudi Arabia (SA) starting from the late fifties. The rainfall recodes at that time is rather short since the early rainfall records were in the mid-sixties. Therefore, due to the recent changes in climate conditions and longer rainfall records it is of great importance to evaluate whether the old dams can still withstand the current climatic situation, and the recent environmental conditions in terms of land use changes. Moreover, recently we equipped with recent software technology for dam design, the scientific questions might be asked: (1) do these new techniques provide different design values than the old techniques? and (2) Can these dams accommodate the changes in the current climate and land use/land cover changes?

In the USA, 245 of 256 dam failure events happened due to high inflow discharges as published by the National Performance of Dams program (2007). These high inflow discharges were due to unexpected rainfall events. Adamo et al. (2017) stated that dam assessment is now shifted to risk management. Old school in dam assessment considers technical standards and legislative frameworks as stipulated by each authority. However, the evaluation procedure covering hydrological risks due to climate change and re-evaluation of the design of old dams seems more beneficial.

Areu-Rangel et al. (2017) presented a study for the hydrologic safety of dams. The safety of a dam is linked to the selection of the return period of the design storm during the design stage. The proposed design was found to impose hydrological risks with a 15-year return period. A modification of the design was necessary to ensure the dam’s safety. Tsakiris et al. (2010) in their study of embankment dam failure, pointed out that one of the reasons for such events is overtopping due to under-estimation of flood volume.

Gebregiorgis and Hossain (2012) pointed out that new flow data for old dams might be used to assess the safety of these dams. Atkins (2013) in their report studied the effect of climate change on dams and reservoirs. They pointed out that old dams in the United Kingdom are affected by increased precipitations in recent decades and should be assessed as well considering the future forecasted climate.

Fluixa-Sanmartin et al. (2019) reported a detailed study on the increased failure risk due to climate change. They applied the study to the Santa Teresa dam built in 1960. It was concluded that an effective assessment of the failure risk should include other factors besides climate change. Ehsani et al. (2017) presented a study about the effect of climate change on the operation of dams and reservoirs. They concluded that current operating characteristics should be changed to cope with foreseen changes in rainfall characteristics to be able to achieve demands. Some of the old dams in Saudi Arabia has been breached due to several reasons such as land use/land cover changes and extreme weather conditions see e.g. Azeez et al. (2020). To the best of the authors’ knowledge, there is no suffient literature studies on the topic. In the current study, some dams were chosen in the Riyadh region as a pilot study to perform such a dam evaluation study. Later, the study can be extended to any location in SA. This paper seems to be the first attempt to investigate this issue especially in arid region in general and in SA in particular.

The main objective of this research is to perform a hydrological assessment of some existing dams in SA as a pilot study, to evaluate whether these dams can cope with the recent rainfall data obtained in recent years. Also, a comparison between the use of new technologies such as digital elevation models, automatic delineation of the basins, stream networks, and the use of topo maps in dam design is considered. The items for such an evaluation are presented in Table 1. The table shows the criteria used in the evaluation procedure given in the first column, while the second column provides the information used for the design of the old dams, and in the third column the items considered in the recent evaluation. Details regarding these items are presented in the methodology part.

Table 1 Criteria used for dams’ evaluation

Study area

Geographical location and the dams considered in the analysis

The study area is located within the Riyadh region, Kingdom of Saudi Arabia (KSA). The study area consists of Ad-Diriyah, Huraymila, and Thadiq governorates which cover an area of about 2300 Km² which represents about 0.1% of the total area of Saudi Arabia. The area is bounded between 24° 44′ and 25° 19′ N and 45° 50′ and 46° 50′ E. Fourteen dams were selected to perform the current study because of the data available as shown in Table 2. Except for two, all of them were constructed before 1985. These dams were constructed on the major basins in the area. Figure 1 shows the location and photographs of the selected dams and basins.

Table 2 General characteristics of the studied dams

Fig. 1

Locations and some images of the dams, spillways, and basins in the study area are projected on Google Earth Image as a base map

Climate conditions

In general, the climate of the study area is hot and arid. On average, the temperature range is between 20° and 45° Celsius while the humidity range is between 18 and 45%. The average rainfall depth is between 32 and 310 mm with an average annual precipitation of 107 mm (Tarawneh and Chowdhury 2018). Figure 2 shows the annual rainfall depth (mm) in Saudi Arabia. Nine rainfall stations are located around and within the study area. Table 3 shows the data of the rainfall stations. During the evaluation of dams under study, only 4 stations (R102, R103, R105, and R106) were used, while in the present analysis, the nine stations reported in Table 3 are used and investigated to provide an overview of the rainfall pattern in the region.

Fig. 2

The distribution of annual rainfall (mm) over Saudi Arabia based on a 50-year surface rainfall climatological record (Hijmans et al. 2005)

Table 3 Characteristics of the daily rainfall of the stations in the area

Geological settings

The KSA is made up of two dissimilar geologic regions. The oldest is the Arabian Shield, on the western side of the Arabian Peninsula. The youngest is the Arabian Platform or Shelf, located on the middle and eastern side of the peninsula Fig. 3. The Arabian Shield comprises igneous and metamorphic rocks of Precambrian age and volcanic rocks (harrats) of Tertiary and younger age. In the valleys, alluvial basin fillings exist. On the Arabian Shield, only secondary aquifers can be found. The igneous and metamorphic bedrocks have in general a low yield. Moderate yields can be found in the harrats and the alluvial basin fillings. However, because of the limited extent of these aquifers, they are only of local importance.

Fig. 3

Geological map of SA (Al Saud and Rausch 2012)

The geology of the Arabian Shelf is characterized by a thick sedimentary succession that is lying on the basement rocks. It constitutes almost two-thirds of the Arabian Peninsula and is divided into three structural units that have inner layers with a similar tendency which include many of the canyons, basins, and several watersheds. The sediments range from the Cambrian age to the recent. They consist of carbonates, sulphates, shales, marls, and sandstones. Deposition occurred during several transgressive–regressive cycles. In the lower part of the succession clastic sediments like sandstones dominate while in the upper part limestones, dolomites and sulphates prevail. The sequence of the geological formations in the Arabian shelf is from west to east, from the oldest to the newest starting with the Sak formation and ending with the Kharj formation.

Methodology and data collection

The methodology and the criteria that are adopted to evaluate the dams consist of the following steps.

Rainfall frequency analysis

The frequency analysis of the maximum daily precipitation is performed using the common probability distribution namely: Gumbel, generalized extreme value (GEV), 2- and 3-parameter log-normal, Person Type III and Log-Person Type III distributions (see Table 4). The root-mean-square error (RMSE) is used as a measure of the best fit (Kite 1977, Niyazi et al. 2014, Elfeki et al. 2017, Noor and Elfeki 2018). The root-mean-square error (RMSE) is given by,

$${\text{RMSE}} = \sqrt {\left( {\frac{1}{n}\sum\limits_{i = 1}^{n} {\left[ {\hat{P}_{i} - P_{i} } \right]^{2} } } \right)}$$

(1)

where $P_{i}$ is the observed precipitation depth, $\hat{P}_{i}$ is the predicted precipitation depth, and n is the number of observations. (Table 4).

Table 4 The common pdfs, f(x), used in the analysis

The effect of the rainfall record length

To study the effect of record length, dams’ information is collected specifically: the type of dam and the year of construction, the rainfall stations in the vicinity of each dam which are supposed to be used for the dam design in the past from the construction of the dam. Table 5 shows the dams’ names in the study region, the type of dam, the construction year, the rainfall stations close to each dam, and their recent record length. The table also shows the number of years that the dam could be analyzed for the design in the past based on the same stations. Frequency analysis is performed based on the aforementioned distributions (Table 4) to find the best distribution that fits the data. However, for studying the effect of the record length, Gumbel Type I is used since it is the common distribution used in the past for the design of these dams (e.g. Binnie & Partrers Consulting Engineers 1978; Zuhair 1983; Saudconsult consulting services 1397). The frequency analysis is repeated for each station based on the number of recorded years according to the dam construction year (as shown in Table 5).

Table 5 Dams in the study region and the rainfall stations close to each dam used for the dam design

Rainfall temporal distribution and the design of storm duration

For design purposes, it is required to distribute the design rainfall depth over the design storm duration. Traditionally, engineers use the SCS-type II (USDA-SCS 1985) in SA since there are no specific design hyetograph curves for SA. Recently, Elfeki et al. (2013) and Ewea et al. (2016) followed Huff (1967) methodology to classify storms based on four classes namely, the first, second, third, and fourth quartiles. However, only two quartiles are extracted from the data namely the first quartile (storms less than 6 h) and second quartile (storms of durations 6–12 h) of the study region. These two quartiles are the only ones obtained due to the parsimonious of the data that allowed to extract storm up to 12 h durations. Figure 4 shows a comparison between these patterns: the 2nd quartile (for Riyadh region), and the SCS-type II which are significantly different. The old dam design is based on the SCS-type II, while the current dams’ evaluation is based on the 2nd quartile storm pattern.

Fig. 4

Dimensionless storm patterns: vertical axis is the dimensionless cumulative rainfall depth (p/P), and the horizontal axis is the dimensionless elapsed time from the start of the storm (t/D)

Basin delineation using WMS

Watershed modeling system (WMS), version 7.1, is originally developed by the Environmental Modeling Research Laboratory of Brigham Young University (2004) and now is distributed by AQUA VEO (2021). All aspects of basin hydrology and hydraulics can be modeled by the software’s comprehensive graphical modeling environment. It performs many procedures such as basin delineation, GIS overlay computations, calculations of the geometric parameter, terrain cross-section, floodplain delineation, and more. In the current study, WMS 7.1 is used to extract hydrologic, topographic, and topologic parameters from digital spatial data. Then, it transfers the generated data layers in ASCII file format to the basin and precipitation modules of the HEC-HMS 4.6.1 interface.

Hydrological analysis using HEC-HMS based NRCS-method

The Natural Resources Conservation Services Curve Number (NRCS-CN) method introduced by the USDA-SCS (1985) is commonly used in Saudi Arabia in flood studies and dam design (e.g. Binnie & Partrers Consulting Engineers 1978; Zuhair 1983; Saudconsult consulting services 1397). Therefore, it has been used herein for the evaluation of the dams in the current study. The method is used to estimate the direct runoff from rainfall data given by:

$$R=\left\{\begin{array}{c}\frac{{\left(P-0.2S\right)}^{2}}{P - 0.8S}, P>0.2S\\ 0, P\le 0.2S\end{array}\right.$$

(2)

where P is the total design rainfall depth, R is the expected direct runoff depth from the rainstorm, and S is the potential retention of the basin estimated by,

$$S=\frac{25400}{\mathrm{CN}+254}$$

(3)

where CN is the curve number.

The traditional approach to estimating CN is based on the NRCS-CN table (USDA-NRCS 2004). In the current study, the same approach is used however, estimation of CN relied on soil maps for soil types and land cover (LC) and Satellite images and base maps for land use (LU).

The NRCS-CN method uses a specific dimensionless unit hydrograph (NRCS-UH) as given by a Gamma pdf function and can be represented by Hann Equation (Eq. 4) and presented in Fig. 5.

$$\frac{Q\left(t\right)}{{Q}_{\mathrm{p}}}={\left[\left(\frac{t}{{t}_{\mathrm{p}}}\right){e}^{1-\left(\frac{t}{{t}_{\mathrm{p}}}\right)}\right]}^{k}$$

(4)

where Q(t) is the discharge at time t, Q_p is the peak discharge, t is the time since the start of the hydrograph, t_p is the time to peak, and k is an exponent which is equal to 3.77 for NRCS-UH.

Fig. 5

Dimensionless UH of the NRCS method. The red curve is the hydrograph, and the blue curve is the mass curve

Since there are limited studies on the unit hydrograph (UH) derived for the Saudi arid environment (Albishi et al. 2016, 2017; Elfeki et al. 2020), researchers and engineers use the NRCS-HU method as a synthetic UH for the design of hydraulic structures and flood mitigation in the arid region of SA. Moreover, the cited studies are in the southwestern part of SA which has different hydrological conditions from the middle part of SA where the current study is concerned. Therefore, the NRCS-HU method was used in the current study for dam evaluation.

US Army Corps of Engineers developed the Hydrologic Modeling System (HEC-HMS), version 4.6.1 (HEC 2000). Rainfall-runoff processes of dendritic basin systems are simulated by HEC-HMS. The program consists of three main modules, basin model, meteorological model, and control specifications component. The first two modules handle the input data while the third controls the running period. The model enables parameter estimation using optimization theory.

Dam reservoir analysis

Dam reservoir analysis is performed by constructing the relationships between the basins’ area and volume against the water depth in the reservoir using WMS software. To unify these relationships for comparison reasons, dimensionless curves are obtained. Mohammadzadeh-Habili et al. (2009) developed dimensionless equations for modeling reservoir-depth relationships. The dimensionless volume–depth relationship and the dimensionless surface area–depth relationships are given by the following formulas (Mohammadzadeh-Habili et al. 2009),

$$\frac{V}{{V}_{\mathrm{m}}}= {\left({e}^{\mathrm{ln}(2)\frac{h}{{h}_{\mathrm{m}}}}-1\right)}^\frac{1}{N}$$

(5)

$$\frac{A}{{A}_{\mathrm{m}}}= \frac{1}{2} {e}^{\mathrm{ln}(2)\frac{h}{{h}_{\mathrm{m}}}} {\left({e}^{\mathrm{ln}(2)\frac{h}{{h}_{\mathrm{m}}}}-1\right)}^{\frac{1-N}{N}}$$

(6)

where V_m is the reservoir maximum volume, A_m is the reservoir maximum surface area, h_m is the maximum depth of the basin, and N is known as the reservoir coefficient and is given by:

$$N=2\mathrm{ln}\left(2\right)\frac{{V}_{\mathrm{m}}}{{A}_{\mathrm{m}}{h}_{\mathrm{m}}}$$

(7)

Reservoirs can be classified by many methods. Broland and Miller (1958) introduced a classification criterion as shown in Table 6. The reservoir shape M can be evaluated by the following equation as introduced by Mohammadzadeh-Habili et al. (2009):

Table 6 Reservoir classification

$$M={\left(\frac{N}{1.075}\right)}^{-1.103}$$

(8)

This approach is followed to evaluate the dam reservoirs as explained in the following sections.

Inundation maps in dam reservoirs

It is very important to delineate the flooded area in the dam reservoir to protect the dam lake region from urbanization. This delineation is considered a nonstructural measure of the flood protection scheme. To achieve this task, the hydrological analysis using HEC-HMS and the dam reservoir analysis described above shall be utilized to estimate the inundation area. The inundation area in the dam reservoir is evaluated based on storms of different return periods of 5, 10, 50, and 100 years. The approach followed in this study is similar to the work by Elfeki et al. (2017).

Results and discussions

Rainfall frequency analysis

Rainfall analysis has been performed in its common approach to test various probability distributions. Nine stations located within the study area were selected to perform the analysis. The root-mean-square error (RMSE) has been used as a criterion for the best probability distribution (Kite 1977; Haan 1977; Niyazi et al. 2014). To be consistent in our analysis we also used Gumbel Type I distribution in the analysis since old dam design reports relied on it (Binnie & Partrers Consulting Engineers 1978; Zuhair 1983; Saudconsult consulting services 1397). Figure 6 shows the results of the analysis where for each station there are two figures the first one shows all the probability distributions fitted to the station data while the second figure illustrates the best probability distribution with the minimum RMSE for each station. Table 7 shows the minimum RMSE for each station. It is noticed that no station followed Gumbel Type I distribution. Three stations followed the Log-Pearson distribution, two stations followed General Extreme Value (GEV) distribution and two stations followed two Parameter Log-Normal distribution. These results show that Gumbel Type I distribution should not be taken for granted in rainfall analysis in arid regions, but data should be examined by various distributions to obtain the best representative distribution for accurate predictions.

Fig. 6

Fitting probability distribution functions (given in Table 4) to the maximum daily rainfall data of each station (a), and the best distribution, (b) based on the criterion of minimum RMSE (given in Table 7)

Table 7 RMSE for the fitting of the probability distributions

Table 8 shows the expected rainfall depth in mm for return periods 2, 5, 10, 25, 50, 100, and 200 years. It also shows the maximum recorded rainfall depth for each station with the year of record and the recorded length. The table shows for most of the stations (seven out of nine) that the expected value of rainfall for a return period of 50 years exceeds the maximum recorded values. Figure 7 illustrates rainfall distribution for the 10, 25, 50, and 100 years return periods. The figure shows that the highest rainfall is located in the northern part of the study area while the lowest rainfall is located in the south and west parts of the region.

Table 8 Summary of the rainfall analysis

Fig. 7

Rainfall distribution over the study area for 10, 25, 50, 100 years return periods

Figure 8 depicts the relationship between the expected rainfall depth for the 2, 10, 25, 50, 100, and 200 years return periods with the length of records for four stations. It can be concluded from the figure that the values become almost constant for the record length exceeds 20 years. Therefore, to produce a reliable prediction, a minimum of 20 years of records is needed.

Fig. 8

Relation between the rainfall depth at different return periods (2, 5, 10, 25, 50, 100, and 200 years) and the record length at the rainfall station

The results of the rainfall analysis show that the effects of climate change are evident in the expected values of rainfall depths for different return periods when compared to the maximum recorded values. It is also reflected in the various distributions that best fit the recorded data.

Basin delineation and stream network extraction

Drainage networks were extracted from the Digital Elevation Models (DEM) provided from ASTER DEM which is distributed for free. ARC GIS 10.0 software was used to map the basins and the drainage network (Masoud et al. 2014). Fourteen basins were delineated using DEM with a spatial resolution of 30 m. Figure 9 shows a sample delineation of four basins as an example of the basin delineation in this study. The figure shows the delineation overlaid on the topographic maps of scale 1:50,000. The figure shows that the stream network derived from the automatic delineations with WMS matches well with the stream network observed from the topo maps. This gives confidence in the DEM model used in the analysis. Figure 10 illustrates the delineation of all basins on the soil maps and the road network maps for further analysis of the estimation of land use and land cover estimation.

Fig. 9

Delineation of some dams and their projections on topo maps for comparisons between stream networks obtained from the DEM and the observed in the topo maps

Fig. 10

Projection of the dam basins on the soil map (A), and the road network map and Satellite image (B), for land use/land cover estimation

Morphological and hydrological analyses

The results of the morphological analysis of the basins are presented in Table 9. The table shows the basin areas that vary between 13.7 Km² for Ghubyra’ dam to 858.23 Km² for Al 'Ulb dam with cv = 1.34 which is a considerable variation in the basin area, it helps in the analysis of a wide range of dam basins distributed over SA. The basin slope varies between 0.019 for Al Uyayna dam to 0.127 for Al 'Ammariyah dam with cv = 0.37 showing a relatively moderate range of slope variability. The maximum stream length varies between 8153.38 m for Al Qirinah dam to 71,145.56 m for Al 'Ulb dam with cv = 0.67. High variability can still be observed in the maximum stream length which has an impact on the flood arrival time to the outlet. The maximum stream slope varies between 0.004 for Al 'Ulb dam and 0.017 for Al Hariqah dam with cv = 0.5. This variability also affects the arrival time of flood from the uphill to the basin outlet. The basin length varies between 7595.1 m for Al Qirinah dam, and 45,655.1 m for Al 'Ulb dam with cv = 0.52 with the same order of magnitude with the cv = 0.5 of the maximum stream slope. The basin average elevation varies between 750.9 m for Al Ghubyra', and 895.3 m for Al 'Ammariyah dam with cv = 0.06. This is considerably very low variability in the average elevations between these dams which indicates that the basins are on the same plateau.

Table 9 Morphological characteristics of dam basins and average CN values

The CN values are estimated based on the delineation of land use and land cover as presented in Fig. 11a and b, respectively, using soil maps and satellite images as explained in the methodology. The curve number is assigned to each zone based on the NRCS-CN table (USDA-NRCS 2004) and a weighted average value is obtained. Table 9 last column shows the average CN values of the study basins (also see Fig. 11c). The values range from CN = 81 for Al Uyayna dam to CN = 88 for many dams as shown in the table. The majority of the basins have CN = 88. The cv = 0.02 shows a very low variability of CN between the basins. The same can be said about the basin average elevation and basin length where cv = 0.06 and 0.52, respectively.

Fig. 11

Land use, (a), land cover (soil type), (b), and CN estimation, (c) maps based on NRCS-CN tables

HEC-HMS software (HEC 2000) was then used to compute the hydrographs of the study basins. Figure 12 shows a sample of the obtained hydrographs for some dams at 5, 10, 25, 50, and 100 years return periods. It is obvious the increase of the hydrograph shapes by increasing the return period. Table 10 summarizes the parameters of these hydrographs regarding the peak discharge and the runoff volume. The table shows the design dam capacity, the peak discharge, and the runoff volumes at different return periods and the design evaluation through the comparison of the cam capacity and the runoff volume at different return periods. Moreover, the basic statistics of the peak discharge and the runoff volume are estimated at different return periods. For 5 year return period, the minimum Q_p is 4.3 m³/s for Al Uyayna dam, and the maximum Q_p is 69.4 m³/s for the Sulbukh dam. The minimum runoff volume is 118,000 m³ for Ghubyra' dam, and the maximum runoff volume is 3,801,000 m³ for Al 'Ulb dam. For 10-year return period, the minimum Q_p is 7.6 m³/s for Al Uyayna dam, and the maximum Q_p is 119.2 m³/s for Al 'Ulb dam. The minimum runoff volume is 180,000 m³ for Ghubyra' dam, and the maximum runoff volume is 6,592,000 m³ for Al 'Ulb dam. For 25-year return period, the minimum Q_p is 12.6 m³/s for Al Uyayna dam, and the maximum Q_p is 193 m³/s for Al 'Ulb dam. The minimum runoff volume is 265,000 m³ for Ghubyra' dam, and the maximum runoff volume is 10,512,000 m³ for Al 'Ulb dam. For 50 year return period, the minimum Q_p is 16.5 m³/s for Al Uyayna dam, and the maximum Q_p is 294.4 m³/s for Al 'Ulb dam. The minimum runoff volume is 332,000 m³ for Ghubyra' dam, and the maximum runoff volume is 13,455,000 m³ for Al 'Ulb dam. For 100-year return period, the minimum Q_p is 20.6 m³/s for Al Uyayna dam, and the maximum Q_p is 306.4 m³/s for Al 'Ulb dam. The minimum runoff volume is 398,000 m³ for Ghubyra' dam, and the maximum runoff volume is 16,434,000 m³ for Al 'Ulb dam. The cv for all cases is relatively high indicating high variability between peak flow and volume of these dams. The last column in the table shows dam evaluation based on the comparison of the dam capacity and the volume at different design return periods. Nine dams out of 14 dams have a storage capacity of less than the 5 years return period (Malham, Sulbukh, Al Uyayna, Al Haysiyah, Al 'Ulb, Al Hariqa, Ghubyra', and Safar). Two dams are less than 10 years return period (Thadiq and Huraymila'), a dam less than 25 years return period (Al 'Ammariyah), a dam less than 50 years return period (Al Mutrafiyah), and a dam above 100 years return period (Al Qirinah). These results reveal that most of the dams are in danger of failure due to their small storage capacity. Seven of the nine dams were built before 1975 where at that time there were about 10 years of rainfall records. Yet, two dams were built in 2010 where enough records existed. Therefore, the dams that are designed for less than 100 years return period should be checked for dam stability under 100 years return period conditions. Moreover, they should also be checked for the effect of climate change and urban sprawl.

Fig. 12

Flood hydrographs for some dams as example: a Safar Dam, b Al Mutrafiyah Dam, c Al 'Ammariyah, and d Ghubyra' Dam reservoir

Table 10 Summary results of the hydrological analysis

Dam reservoir analysis

The relationships between the basins’ area and volume with respect to the reservoir depth are presented in Fig. 13a and b for all basins. Reservoirs can be classified by many methods. Broland and Miller (1958) introduced classification criteria as shown in Table 6. The reservoir shape factor M (Eq. 8) is used for such classification.

Fig. 13

Dimensionless relationships for the basins: Volume–Depth (a), and Area–Depth (b)

Table 11 presents the results of the reservoir analysis. It shows that eight reservoirs out of the 14 reservoirs are of type III which represents about 60% of the reservoirs. Figures 13a and b are constructed such that the reservoir coefficient N is adjusted to give the best fit to the reservoir data. Table 11 shows the values of N based on Eq. 7 and the adjusted values to give the best fit. The results of this analysis show that the N estimated from Eq. 7 is close to the N estimated from the fitting of the volume-depth equation rather than the area-depth equation, therefore, it is recommended to use the N from Eq. 7 to model the reservoir volume for further studies such as reservoir life and sedimentation in the reservoir. This analysis is needed to model the reservoirs in order to be able to prepare the inundation maps.

Table 11 Reservoir classification and the calculated and adjusted values of reservoir coefficient, N

Figure 14 shows the difference between the inundation area of Al-Ulb and Al-Matrafya dams, as example, in Al-Ulb the area is concentrated in the mainstream of the watershed bounded by high land on the flood plains while in Al-Matrafya dam appeared in the mainstream and spread over a large area in the flood plains which indicated a flatter area. So, this indicates that during urban development in the region, the authorities should protect these regions otherwise they would be in danger from the risk of flooding.

Fig. 14

Sample maps of inundation area in the dam reservoir due to storms of different return periods from top to bottom 5, 10, 50, and 100 years: the left column is Al-Ulb dam and the right column is AL-Ammariyah dam

Summary and conclusions

This study deals with the evaluation of the safety of some existing dams from a hydrologic point of view under the current conditions of climate, land use and land cover changes, the use of the advanced technology in GIS, DEM, and satellite images. The study assesses the dams based on some criteria: (1) the type of rainfall distribution used in the frequency analysis, (2) the design storm duration, (3) the effect of the length of the rainfall records on the reliability of the prediction of the rainfall depth for different return periods, (4) the type of the temporal rainfall distribution over the storm duration, (5) topo maps versus DEM, and (6) land use/land cover changes and SCS-CN estimation methods. It revealed several conclusions that can be summarized as follows:

• The Gumbel Type I distribution, that was used in the analysis of the old dams, does not seem to be the best distribution under the recent rainfall data. Therefore, in recent dam studies, engineers have to test different distributions and decide on the best based on some statistical criterion as explained in the current study.

• The results show that increasing the record length leads to a convergence of the values of the expected rainfall at a given return period. It reveals a minimum of 20 years record is needed to produce convergence of the predicted rainfall for any return period to a stationary value.

• The minimum record length to produce stable statistical rainfall values is about 20 years at a given return period.

• The rainfall statistical analysis in the study area shows that the rainfall that happened in some years in the past falls within the critical rainfall values of the 50 and 100 years return periods. However, it does not reach the 200 years return period.

• It is recommended to use the power N for reservoir capacity modeling from Eq. 7 since it is close to the N estimated from the fitting of the volume-depth equation rather than the area-depth equation, Therefore, the reservoir type can be better evaluated. 

• The hydrological analysis shows that nine dams out of 14 dams have a storage capacity of less than the 5 years return period. Two dams are less than 10 years return period, one dam is less than 25 years return period, one dam is less than 50 years return period and one dam is above 100 years return period. Therefore, the dams that are less than 100 years return period should be checked for dam stability. Moreover, they should also be checked for the effect of climate change and urban expansion. This study can further extended to other dams in the Kingdom.