Climate and hydrology
The regional climate in the central Haouz is semiarid with a high spatial and temporal heterogeneity. In the plain, the semiarid climate is characterized by low and heterogenous rainfall, with episodic drought periods (Fniguire et al., 2017). The mean annual precipitation in Marrakech weather station in the period between 1962 and 2015 is 184 mm/year (Hajhouji et al., 2018).
Temperature in Ourika basin varies from -7.2 °C to 48.2 °C with an annual average of 27.8 °C depending on the altitude (Figure 2-3). The hottest months are July and August and the coldest months are December and January. The annual precipitation in the Aghbalou station is 527 mm/year (Figure 2-3) but can exceed 700 mm/year in the headwater catchment (Saidi et al., 2010). An important part of precipitation in headwater catchments of the central High-Atlas fall in form of snow. The watershed is drained by Ourika wadi, one of the important Atlasic wadis in term of flow. The Aghbalou gaging station located at 974 m monitors Ourika wadi. The mean flow of the wadi is 4.9 m3/s. The analysis of Ourika wadi flow reveals two annual peaks following rainfall trends (Figure 2-4). The first peak of November is due to the raining in the area, and the first peak, which is the highest, is in April with conjugal effect of rain and snowmelt (Bouimouass et al., 2020).
Figure 2- 3 : Interannual means of precipitation and temperature in Aghbalou and Oukeimden stations located in Ourika Watershed.
Mean annual rainfall in Tahanaout station is 356 mm/year and increases with altitude (Hajhouji et al. 2018) (Figure 2-5). The Rheraya wadi, monitored by the Tahanaout station at 1030 m, drains Rheraya watershed. The mean annual streamflow at Tahanaout station is 1.15 m3/s. The peak flow occurs in April (Figure 2-6) due to rain-on-snow events due to the rise of temperature (Zkhiri et al., 2017). Numerous studies have been carried out to assess snow resources in the High-Atlas within the frame of the SUDMED project (Chehbouni et al., 2008) and the LMI TREMA (Jarlan et al., 2015). These studies showed that snow contributes with an important fraction, ranging from 30 to 50%, to the annual flow of the Atlasic wadis (Boudhar et al., 2009; Marchane et al., 2015; Baba et al., 2018).
Figure 2- 5: Interannual means of precipitations in Tahanaout gage station (Rheraya watershed).
Figure 2- 6: Interannual means of Rheraya wadi streamflow at Tahanaout gage station.
The Haouz basin is a sedimentary basin with a tectonic origin. The basin is filled during the Neogene and Quaternary by detrital deposits issued from the erosion of the High-Atlas. The complex geology of the Haouz and the High-Atlas is summarized following the geological timescale (Figure 2-7):
x The Precambrian: It forms the axial zone of the High-Atlas of Marrakech (HAM). The Precambrian, highly metamorphosed, is formed of crystalline rocks such as granodiorite, amphibolite, shale, quartzite and granite (Michard, 1975).
x The Paleozoic: The Primary formations is widespread in the HAM. Pre-Hercynian formations (between the Cambrian and the Carboniferous) are formed of sandstone, clay, shale and limestone. The Visean is deformed and metamorphosed by the Hercynian orogenesis (Proust, 1961).
x The Mesozoic: The Permo-Triassic of the HAM is formed of a thick series, exceeding 1000 m, of clay and evaporates covered often by 200 m of doleritic basalts (Biron, 1982). The Jurassic outcrop in the lower part of the mountain is formed of conglomerate (15 to 30 m), siltstone (10 to 20 m) and sandstone with conglomeratic channels (80 to 100 m). The Cretaceous in the HAM is formed by the Albian, Cenomanian, Cenomano-Turonian and the Senonian. It is formed essentially of limestone, dolomitic limestone and clay.
x The Cenozoic: The Eocene is present in the bordure of the High-Atlas, and outcrops in some places especially in the Ourika and Zat watersheds. It is formed of limestone and dolomite in alternation with marls and sandstone. During the Mio-Pliocene, a very intense tectonic compression occurred giving birth to the actual High-Atlas chain (Proust, 1961; Biron, 1982). Furthermore, the newly elevated chain is being dismantled and the sediments have accumulated in the adjacent depressions (such as Haouz basin). The Haouz depression began to be filled with detrital deposits transported by the Atlasic wadis (Sinan, 2006).
Figure 2- 7: Geology of the High-Atlas of Marrakech (Ourika basin, Ouanaimi , 2011)
The Haouz plain (6200 km2) hosts the large phreatic aquifer of Haouz (Figure 2-2) and some confined aquifers encompassed within the limestones of the Eocene, Cenomano-Turonian and Jurassic deposits underlying the alluvial aquifer (Moukhchane, 1993; Sinan, 2000). In the central Haouz, the confined Cenomano-Turonian and Eocene aquifers cover an area between 160 and 170 km2 in the southern border of the plain, adjacent to the High-Atlas (Moukhchane, 1983). These aquifers are recharged by precipitations and streamflow in the mountain-front where they outcrop to the surface along the border of the High-Atlas between Amezmiz and Ait Ourir (Sinan et al., 2000). The alluvial Haouz aquifer is considered the most important source of groundwater in the whole Tensift basin. This unconfined aquifer is limited from the bottom by the marly Miocene and, in some areas, by clayey Triassic and Paleozoic shales constituting the substratum of the aquifer. It is formed by Quaternary alluvium with high vertical and lateral heterogeneity. Vertically, the aquifer can be schematized as a series of layers of lenticular structures of clay and marls in alternation with others of coarser elements. Laterally, many structures of deposits can be found in the Haouz aquifer such as paleo-channels, paleo-reliefs, and scree cones (Sinan, 2000).
The aquifer is highly heterogeneous (Sinan and Razack, 2006). Measured transmissivity varies between 5 10-5 m2/s and 9 10-2 m2/s with a mean of 6.7 10-3 m 2/s. Groundwater flows from the south to the north (Figure 2-2). The aquifer is believed to be primarily recharged by flood water infiltration within wadi channels and irrigation, irrigation returns, inflow from the underlying confined aquifer (Abourida, 2007), and direct infiltration from rainfall (Sinan, 2000).
There is no natural perennial surface water in the Haouz, hence, the majority of water demands for irrigation and drinking water supplies in the central Haouz relies on groundwater of the alluvial aquifer. More than 24000 pumping wells (inventory carried out between 2003 and 2006) consistently pump water from the Haouz aquifer (Le Page et al. 2012).
In order to support groundwater resources, an artificial canal (Rocade canal) was built in the 70s to transport water 120 km away from Moulay Youssef dam in the Lakhdar wadi to irrigate the agricultural perimeters and to provide the City of Marrakech (2 million inhabitants) with 90% of its domestic and drinking water demands. The Lala Takerkoust dam was FRQVWUXFWHG LQ LQ WKH 1¶ILV ZDGL
In the mountain front, local inhabitants have taken advantage of floods and snowmelt-driven runoff for irrigation purposes by developing a large network of gravity-fed surface irrigation channels (locally named Seguias) that divert the runoff (Figure 2-8). To avoid conflicts over water, a regulated irrigation system was established, inspired from traditional water rights. This irrigation system is widespread in Morocco, but particularly within the south and north piedmont areas of the High-Atlas Mountains. Each major channel, managed by users’ associations, irrigates a given area ranging from hundreds to thousands of hectares. The secondary channels have their own delegates who distribute water between farmers. The duration of irrigation deliveries to a single farmer initially depended on the land area. Upstream channels have priority to access water and are fed greater volumes than downstream channels. The channels located upstream also have access to the streamflow during low discharge periods in summer and autumn or during extended dry periods. Where, the downstream channels are supplied only during floods and high spring flow. Consequently, crops in the downstream areas rely on groundwater irrigation during the dry periods.
Figure 2- 8: Water resources in the central Haouz
Worldwide, a shifting from snow to rain in the total precipitation of snow-dependent basins is occurring due to climate change (Berghuis et al., 2014). Snow cover reduction and increase in temperature will reduce streamflow volumes and will be shifted from irrigation months in late spring and early summer (Malek et al., 2020). The agricultural activities in the Haouz plain relies significantly on snowmelt water, however, spring streamflow decrease during the last two decades has forced the farmers to shift from seasonal agricultural practices (vegetables) to annual crops (cereals) relaying mainly on rain water and trees relying on groundwater pumping. At the other hand, groundwater has been taking all the pressure due to the population growth, which caused a severe withdrawal (Bouimouass et al., 2020).
Groundwater recharge sources in the mountain-front of the High-Atlas: case study of the Ourika watershed
As mentioned in chapter 1, the main sources of recharge in semi(arid) mountain-fronts are focused infiltration from streambeds and subsurface inflow from the mountain block. Despite the importance of mountain front recharge for semi(arid) groundwater resources, there remain significant challenges in distinguishing those different sources and their specific contribution (Bresciani et al., 2018). This is particularly the case in irrigation areas, where the groundwater recharge may also include inflow from irrigation leakage (herein referred to as irrigation recharge IR) that is often a redistribution of both local- (pumping from alluvial aquifers) and mountain- (floods and streamflow generated from rainfall or snowmelt) derived water.
Previous studies have used groundwater fluctuation and seasonal piezometric mounds to analyze recharge processes in mountain-front recharge areas (Shanafield and Cook, 2014; Bresciani et al., 2018). However, although hydraulic head data record the groundwater recharge events, the discrimination of different sources of recharge generally requires the use of environmental tracers. Commonly used tracers include stable isotopes and major ions (Liu and Yamanaka 2012; Zhu et al., 2018). The main advantage of using stable isotopes to analyze the mountain front recharge is that the signature of local rainfall on the plain are likely to significantly vary from the signature of rainfall in the mountains due to rainout effects (Clark and Fritz, 1997; Scanlon et al., 2002; Lambs, 2004; Kalbus et al., 2006). Similarly, for major ions, since changes in recharge rates and flow pathways can result in significant differences in the composition of the major ions, the latter could provide insights into mixing between different groundwater recharge sources (Liu et al., 2020).
The objective of this chapter is the use of water table fluctuation data and stable isotopes to determine the sources contributing to groundwater recharge in the mountain-front of the High-Atlas. Seasonal variations of piezometric levels and isotopic composition of meteoric water, surface water and groundwater in the Ourika watershed and its mountain-front area, as well as historical piezometric data and streamflow diversion for irrigation will be analyzed and discussed.
Field measurements, sampling and analytical methods
A total of 91 samples were collected from rainfall (Rn), snow (Sn), streamflow (St), irrigation channels (Ir), private wells (W) and springs (Sp) (locations shown in Figure 3-1). Rain, snow and streamflow were sampled in the mountain, at different elevations. A flood event was sampled in the piedmont in March 2018; Three samples were collected during the recession from the main stream (Ourika) and its tributaries (Igerifrouan and Elmaleh). In addition, 11 rain samples collected in 2014 (unpublished data) in the plain, the mountain-front and the High-Atlas were used to characterize the spatial variation in stable isotopes of precipitation.
Streamflow was sampled during 07 sampling campaigns between September 2017 and March 2018 on a monthly basis (except for October 2017). Irrigation water was collected from a concrete irrigation channel (Ir1) and an earthern irrigation channel (Ir2). Groundwater was collected during four field campaigns (1-5 September 2017, 10-15 November 2017, 25-30 December 2017 and 25-30 March 2018) covering dry and wet seasons. The samples were taken from two springs in the high mountain (n=11), one spring and three wells from the hillslope, and 27 private irrigation wells (n=56) from the alluvial aquifer in the mountain front area and further downstream. The wells are generally shallow, open hand-dug wells used mostly for irrigation with immersed permanent pumps. The results of all samples are presented in Table 1 (Appendix).
Figure 3- 1: Maps showing (a) the location of Tensift basin in Morocco, (b) the localization of the study area in the Tensift basin, and (c) the localization and types of the sampling points in the Ourika basin.
The groundwater samples were collected during or immediately after pumping. The physical parameters (temperature, pH, electrical conductivity) were measured in the field. Water was sampled for cations (filtered at 0.45 Pm and acidified with HNO3), anions and stable isotopes, and was analyzed at the Laboratory of Hydrogeology of the University of Avignon. The alkalinity was measured using a HACH digital titrator, and stable isotopes were analyzed using a Picarro Analyser L 2130-I. For the VWDEOH LVRWRSHV WKH HUURU ZDV IRU 18O and ð+ Groundwater hydraulic heads were measured twice in September 2017 and March 2018 in a large network of 55 wells. The static level measurements were performed in stable wells or at least 2 days after pumping. The objective was to characterize and draw groundwater elevation and fluctuation maps.
The volumes diverted for irrigation were provided by the ORMVAH (Office Régional de Mise en Valeur Agricole du Haouz) and the multi-year groundwater level records by the ABHT (Agence de Bassin Hydraulique du Tensift).
Irrigation diversion impacting streamflow
During low streamflow, upstream channels divert almost the totality of the streamflow leaving a dry streambed. It is only during high streamflow and floods that there are surface water volumes in excess of mountain front irrigation diversions, and therefore streamflow resources for the downstream irrigation areas. The channels named Tassoultant and Taoualt in the mountain-front divert a monthly mean volume of 1326986 m3 (14.4% of the diverted volume) and 557286 m3 (6.1% of the diverted volume) respectively, whilst the Cherrifia located further downstream only diverts a monthly mean volume of 45228 m3 (0.5% of the diverted volume) (Figure 3-2). On average the whole irrigation network annually diverts more than 65% of the streamflow volumes (Figure 3-3).
In addition to the spatial variation of diverted streamflow, there is also a seasonal variation. The diverted volumes exhibit two peaks; one in November coinciding with rainfall floods, and one in April (the highest peak) coinciding with snowmelt floods. However, the peak in April has decreased during the last decade attesting of a possible reduction in snowmelt contribution to streamflow. Figure 3-2: Monthly mean (2001–2016) volumes of water diverted by the major irrigation channels in the study area (locations shown in Figure 3.1). The channels Tasoultant (a), Taoualt (b) are located in the mountain front and divert water to the irrigation area on the left side of Ourika stream. The Tihilit (c) is also located in the same area as the channels Tasoultant and Taoualt but divert water to the right side of the Ourika stream. The Cherrifia (d) is located downstream.
Figure 3-3: Time series (2000–2016) showing monthly Ourika runoff volumes measured at Aghbalou gauge station and the diverted volumes to the irrigation channels.
Groundwater flow and fluctuation
The piezometric maps of September 2017 (dry season) and March 2018 (wet season) indicate groundwater flowing from the south to the north, that is from the High-Atlas mountain front towards the Haouz plain (Figure 3-4a). From September 2017 to March 2018, the water table rose by 3.5 m upstream to 0.5 m downstream (Figure 3-4b). This attests for a variable seasonal groundwater recharge over the study site. The region with the highest water table fluctuations between the wet and dry season is located in the mountain front area, and extends west from the Ourika stream. As presented previously in the text, the upstream western irrigation area receives relatively high volumes of streamflow during the wet seasons and it is prioritized during the dry seasons. Therefore, within the mountain front region the irrigation leakage has generated more groundwater recharge compared with recharge via the streambed.
Table of contents :
Chapter 1: Introduction
Chapter 2: Study area
Chapter 3: Groundwater recharge sources in the mountain-front of the High-Atlas : case in the Ourika watershed
I. General background
II. Scientific background
II.1. Mountain front recharge (MFR)
II.2. Mountain-block recharge
II.3. Stream losses within ephemeral and intermittent streams in semiarid areas
III. Research questions and objectives
IV. Thesis structure
II. Presentation of the central Haouz area
II.1. Localization and geomorphology
II.2. Climate and hydrology
II.3. Regional geology
IV. Water resources
II. Field measurements, sampling and analytical methods
III.1. Irrigation diversion impacting streamflow
III.2. Groundwater flow and fluctuation
III.3. Stable isotopes
IV.1. Hydraulic heads and stable isotopes as combined tools
IV.2. Impact of irrigation on the groundwater recharge and availability
IV.3. Groundwater recharge sources in the mountain front
II. Material and methods
II.1. Sampling and laboratory analyses
II.2. Data analysis
II.3. Water suitability for domestic use and irrigation
III.1. Hydrochemical properties from the mountain to the piedmont
III.2. Origins of ions in mountains water and groundwater
III.3. Origins of nitrate
III.4. Quality of drinking and irrigation water
IV.1. Hydrochemical processes from the mountain to the piedmont
IV.2. Impact of the traditional irrigation on the groundwater quality
IV.3. Piedmont traditional irrigation and sustainability
II. Material and methods
II.1. Experimental site
II.2. Streamflow detection
II.3. Measurements of changes in temperature and sediment water content
II.4. Calculating vertical infiltration fluxes with 1-Dimensional model of heat transfer
III.1. Sediments characteristics
III.2. Detection of streamflow events
III.3. Water content changes
III.4. Streambed-sediments temperature changes
III.5. Delineating potential recharging events
III.6. Numerical modeling of heat for recharging events
IV.1. Near surface temperature as a proxy to infer streamflow presence and duration
IV.2. Effects of the sediment moisture on the infiltration processes
Chapter 4: Effects of traditional irrigation practices on groundwater chemistry and quality in a semiarid piedmont
Chapter 5: Seasonality in intermittent streamflow losses beneath a semiarid Mediterranean wadi
IV.3. Seasonal variation of the potential groundwater recharge
IV.4. Lateral recharge beneath the channel
IV.5. The pattern of the intermittent streamflow losses
II. Objectifs et structure de la thèse
III. Résumé de chaque chapitre
III.1. La zone d’étude
III.2. Les sources de recharge des eaux souterraines dans le piémont du Haut-Atlas
III.3. Evolution hydrochimique et qualité de l’eau
III.4. Infiltration dans les lits des oueds