Version classiqueVersion mobile

Savoirs locaux et agriculture durable au Yémen

Frédéric Pelat
Amin Al-Hakimi

Section I - Origins of agriculture and its techniques

The hydraulic set‑up of Bayt al‑Ashwal valley: a case study

Miquel Barceló et Josep Torró

Note de l’éditeur

Editor's note: the illustrations, at the end of the book on the printed version, have been integrated here in the body of the article.

Texte intégral

  • 1 Maite Márquez, a graduate student of the Universitat Autònoma of Barcelona, assisted helpfully in (...)

1The survey carried out during this campaign (17 November to 4 December)1 had the goal of ascertaining whether the modulation in size and technical sequence detected in the 1998 survey and analysis of hydraulic systems in the Zafâr area were equally discernible in adjoining valleys.

  • 2 In reference to this area, see Gibson, M., and T.J. Wilkinson, 1995, The Dhamar plain, Yemen: A pr (...)

2The hydraulic systems that had been studied earlier (see Figure 1) had turned out to have a noticeably smaller size than previous archaeological studies had supposed.2 It rapidly became obvious that we were in fact dealing with systems that could be observed in detail at the same resolution we had applied to al‑Andalus systems. The compactness, technical articulation and size (modal area about 10 ha) of these systems clearly indicated that the sayl water—the torrential run‑off following monsoon rains—was concentrated and made to run in a circular fashion, as with any other continuous‑flow water. Consequently, the procedures for study of sayl systems could be the same as those used for the study of continuous‑flow water. At the same time, the survey at Zafâr would allow for reasonable predictions about the ordering and functioning of the size of other nearby systems. It was also clear that larger samples of case studies would reveal the technical similarities of this stable agricultural construction, and also would highlight the regularity of population sizes of construction groups as expressed by surface area.

Figure 1. Irrigated spaces of Zafâr.

Figure 1. Irrigated spaces of Zafâr.
  • 3 al‑Hamdânî, Abu Muhammad al‑Hasan b. Ahmad (edited by Muhammad b. ‘Ali al‑Akwâ), 1986, Al‑Iklil vo (...)
  • 4 Khalid al‑Ansi, pers. comm. 1999.

3Al‑Hamdânî’s list of asdâd (sing. sadd) drawn up in the 10th century AD (mid‑4th century A.H.), in conjunction with the recent commentary of Muhammad b. cAli al‑Akwâ,3 established the orientation of the survey. The asdâd to which al‑Hamdânî referred were in his opinion already very old and certainly pre‑Islamic. Both the architecture and the samples of surface pottery from the survey area4 indicate a Himyarite‑period date. This result offered the opportunity to study the creation, some two millennia ago, of agricultural systems and to examine the technical reasons for their stability and durability. None of these issues had, to our knowledge, been analysed with clarity to date, it simply being assumed that Islam signified the start of a change that archaeologically had been neither identified nor measured.

The survey of Bayt al‑Ashwal

4The valley of Bayt al‑Ashwal lies two kilometres from Zafâr. The whole valley is organised as a hydraulic complex of converging series, each series originating in a sadd. The complex includes five asdâd, each inventoried by al‑Akwâ in his commentary on al‑Hamdânî.

5Our work consisted of detailed survey on the basis of aerial photography (1:10 000), and creation of sectional and elevational planimetric representations of the water containment walls, of some of the terraces, of the irrigation ditches (sawâqî, sing. sâqiya) of the maqlah‑s and of other elements of irrigation. Detailed planimetry was also undertaken of all the hydraulic systems, plot by plot, following the entire course of the water through the systems, ditch by ditch and watergate by watergate. This work is carried out on the basis of the aerial photography interpreted by on‑the‑ground tracing of the water flow (the latter is not directly visible in the photographs). In addition, the names of all the jirab (sing. jirbah, cultivable field) were taken and numbered, and will be philologically analysed in the future.

Figure 2. View of Bayt al‑Ashwal Valley.

Figure 2. View of Bayt al‑Ashwal Valley.

6The results obtained are the following:

Sadd al‑Gubayr (N 14° 11. 850; E 44° 25,300); water from this sadd irrigates 17 jirab with a total area of 7.3 ha
Sadd al‑Maliki (N 14° 11. 690; E 44° 25,405) 14 jirab 3.46 ha
Sadd al‑Nuqâq (N 14° 11. 965; E 44° 24,889) 6 jirab 2.8 ha
Sadd al‑Amra (N 14° 11. 646; E 44° 25,001) 9 jirab 1.71 ha
Sadd al‑cUmran (N 14° 11. 656; E 44° 25,341) 33 jirab 7.56 ha

7The map of this complex (Figure 3) clearly shows that the orientation of the valley’s irrigation is based upon two principal systems, each being larger than 7 ha: sadd Gubayr and sadd al‑ cUmran. The other systems (al‑Nuqâq, al‑Maliki, al‑Amra) are evidently secondary to the irrigation system as a whole, yet they are fundamental in ensuring the compactness of the overall system and in reducing to a minimum the erosive effects of uncontrolled torrential run‑off.

Figure 3. Map of Bayt al‑Ashwal Valley.

Figure 3. Map of Bayt al‑Ashwal Valley.

8The compactness of the complex is in fact fundamental to understanding the design and construction of the series. We observe in Figure 4 that the sadd al‑Gubayr system constitutes the central unit, both because of its total area and because of the exceptional extension of its terraces. We can also perceive the decisive function of its architecture in underpinning all the linked systems in the valley. But less obviously this centrality would not exist without the sadd al‑Nuqâq­ (nuqâq means ‘toad’) diverting, collecting and directing one of the strongest sayl flows in the valley. Al‑Nuqâq protects the al‑Gubayr terracing. Strictly speaking, it is al‑ Nuqâq that provides a stable existence for al‑Gubayr. To limit ourselves to the observation of al‑Nuqâq’s size (6 jirab, 2.8 ha) or to its apparently peripheral area would be a grave distortion of reality.

Figure 4. Sadd Gubayr.

Figure 4. Sadd Gubayr.

9In this way, the observation of the compactness and contingent stability of al‑Gubayr allows us to appreciate that the juxtaposition of these sub‑systems cannot have been produced at long chronological intervals. The stability of the al‑Gubayr subsystem requires a solution to the potential problems posed by the powerful al‑Nuqâq subsystem. This inherent linkage suggests that the hydraulic organisation of the valley was conceived as a whole, and that its realisation was successive but integrated. We should therefore abandon the attractions of evolutionary causality. For example, it is not empirically consistent to propose that demographic pressure gradually induced the extension of the integrated system across the valley. A more suitable explanation would consider the integrated ensemble congruent with the size of the initial population. This approach does not assume that all the subsystems within the valley have the same degree of necessary interrelation as that between al‑Gubayr and al‑Nuqâq. Indeed, the 14 jirab of al‑Maliki or the 9 jirab of al‑Amra are not necessary to the functioning of al‑Gubayr or al‑Nuqâq. Nor, similarly, is it at all certain that the 33 jirab of al‑ cUmran, as we shall see later in this discussion, are essential to the design, construction and maintenance of al‑Gubayr and al‑Nuqâq.

10This kind of interwoven articulation was expected from the studies of the hydraulic areas of al‑Andalus. The novelty lies in the fact that the arrangement of complementarity and compactness of spaces was undertaken not to controlconcentrate and distributecontinuous water flows, but instead was designed as temporary channels for monsoon run‑off. The monsoon itself introduces variations of range in hydraulic subsystems, variations that had not been detected in the modest continuous‑flow waters of al‑Andalus.

11In the zone bordering the sadd Gubayr and the first terraces of the Gubayr and Maliki (the eastern portions of Figure 3) lie 20 jirab dependent upon irrigation from sadd Harara. This dam, located at 2,760 m asl in the mountains overlooking the valley, is the true hydrographic nucleus of the basin. The Harara jirab are clearly additional to the valley’s hydraulic organisational ensemble. Part of the run‑off water that supplies the jirab is collected by the sadd Gubayr itself. The superimposed nature of the enormous sadd Harara water collection system was evident from the first moment. The sadd had an anomalous appearance, given its unnecessary connection with the valley’s hydraulic sequences, the logic of the latter being clearly discernible.

12Current peasant memory attributes malignity to the waters that cascade from the sadd Harara, situated at such a height. According to people now living in the area, this water would be too cold and would damage the sorghum harvest. They also claim that the reflection of the waters in the sadd, so high above, was so strong that it dazzled the farmers in the valley below, stopping them from working. These opinions—manifestly very old—reveal the perception of this sadd as something external; that perception is certainly correct. Perhaps we ought to think of sadd Harara as representing an interference in the hydraulic management of the valley, since it permitted control of the water supply to two of the sadd from afar. One of these, sadd Gubayr, was the most important in the entire valley sequence. Sadd Harara was also in a position to play a determining role in the hydraulic circuits of neighbouring valleys. Far more detailed fieldwork needs to be undertaken to assess the ways in which the existence of this enormous sadd regulated change in all these valley circuits.

13In principle, then, we believe that future fieldwork should be directed at determining how the sadd Harara mediated the functioning of all the hydraulic circuits in the valley below. Construction by the Himyarite state of Zafâr, or by their client groups, seems plausible. In that case, the hydraulic order of the valley bottom would have existed before Zafâr’s interference in local systems: if the local systems were not yet in existence, it would have been very difficult to establish intervention through control of the sporadic sayl channels from places outside the local order. This seems to us, all in all, to be an issue of greater complexity than the usual description of hydraulic areas.

14The creation of hydraulic organisation in the Bayt al‑Ashwal valley is possible from the systematic use of specific hydraulic knowledge. The empirical recognition of monsoon run‑off topography accompanied a knowledge of agricultural terrace construction. To simplify, we could say that the builders resolved three interdependent issues: construction of cultivation areas, concentration and distribution of water, and control of soil erosion. The technical expression of this issue is the maqlah, a depression, in the form of a deep channel transverse to the natural slope, in which the rainwater and channel filtration collected. Its function was to moderate a run‑off volume inevitably beyond the capacity of the channels, and to concentrate the infiltration of the irrigation water. The maqlah therefore contributed to the draining of the terraces and constituted a much needed element of protection for these areas. But they must also have had a basic function in terrace construction. The maqlah channel is dug into the ground, but on its upper section closest to the direction of the slope, there was usually an upper‑terrace retaining wall of considerable magnitude, in some cases reaching several metres’ width. Their function in the process of construction could be, therefore, that of halting or slowing and of accumulating sediment carried in the channels in order to create a “terrace‑able” surface just above the height at which they were situated. The direction of construction in the cultivating area is, then, from bottom to top, causing the deposit of sediments brought by the channels. For this reason, the maqlah‑s are sited at the key points of concentration for the earlier natural current, and may be used to identify phases of construction. Clearly they are a device the function of which may be confused with that of a deep irrigation channel. They seem to establish interstices in the irrigated area, yet do not constitute any type of break with the final circulation of the water (Figure 3). In order better to appreciate their importance to the hydraulic system, we represent only the circuit of maqlah connections (Figure 5), their role being manifestly clear in the construction, maintenance, and drainage of the terraces. The maqlah, therefore, is a constructive device combining a great deal of peasant knowledge that was essential both at the moment of its construction and when it began to carry out its function.

Figure 5. The maqlah

Figure 5. The maqlah

15There is the opportunity at Bayt al‑Ashwal to verify by excavation the specific use of this device. In jirab 13 and 14 in sadd al‑cUmran there are the ruins of channels three metres underground and dug into the rock that must have constituted an essential construction phase in the process of organising the terrace silting against the direction of the water current. One aim of the next campaign would be to verify by systematic excavation what are for the moment preliminary findings.


1 Maite Márquez, a graduate student of the Universitat Autònoma of Barcelona, assisted helpfully in the field work.

2 In reference to this area, see Gibson, M., and T.J. Wilkinson, 1995, The Dhamar plain, Yemen: A preliminary study of the archaeological landscape. Proceedings of the Seminar for Arabian Studies 25.

3 al‑Hamdânî, Abu Muhammad al‑Hasan b. Ahmad (edited by Muhammad b. ‘Ali al‑Akwâ), 1986, Al‑Iklil volume 8, Beirut.

4 Khalid al‑Ansi, pers. comm. 1999.

Table des illustrations

Titre Figure 1. Irrigated spaces of Zafâr.
Fichier image/jpeg, 85k
Titre Figure 2. View of Bayt al‑Ashwal Valley.
Fichier image/jpeg, 177k
Titre Figure 3. Map of Bayt al‑Ashwal Valley.
Fichier image/jpeg, 167k
Titre Figure 4. Sadd Gubayr.
Fichier image/jpeg, 87k
Titre Figure 5. The maqlah
Fichier image/jpeg, 193k


Universitat Autònoma de Barcelona, Departament de Ciencies de l’Antiguitat de l’Edat Mitjana

Universitat Autònoma de Barcelona, Departament de Ciencies de l’Antiguitat de l’Edat Mitjana

© Centre français de recherche de la péninsule Arabique, 2003

Conditions d’utilisation :


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search