A new method for the determination of Holocene palaeohydrology

Journal of Hydrology 420–421 (2012) 1–16

Contents lists available at SciVerse ScienceDirect

Journal of Hydrology journal homepage: www.elsevier.com/locate/jhydrol

A new method for the determination of Holocene palaeohydrology A.J. Wade a,⇑, S.J. Smith b, E.C.L. Black c, D.J. Brayshaw c, P.A.C. Holmes d, M. El-Bastawesy e, C.M.C. Rambeau f, S.J. Mithen f a

Department of Geography and Environmental Science, University of Reading, Reading RG6 6DW, UK Department of Anthropology and Geography, Oxford-Brookes University, Oxford OX3 0BP, UK c Department of Meteorology, University of Reading, Reading RG6 6BB, UK d Water Resource Associates Ltd., PO Box 838, Wallingford OX10 9XA, UK e Department of Geography, Umm Al Quara University, Mecca, Saudi Arabia f Department of Archaeology, University of Reading, Reading RG6 6AB, UK b

a r t i c l e

i n f o

Article history: Received 11 April 2011 Received in revised form 15 September 2011 Accepted 29 October 2011 Available online 15 November 2011 This manuscript was handled by Laurent Charlet, Editor-in-Chief, with the assistance of Martin Beniston, Associate Editor Keywords: Holocene Climate change Environment Hydrology Semi-arid Jordan

s u m m a r y This paper describes a new method for the assessment of palaeohydrology through the Holocene. A palaeoclimate model was linked with a hydrological model, using a weather generator to correct bias in the rainfall estimates, to simulate the changes in the flood frequency and the groundwater response through the late Pleistocene and Holocene for the Wadi Faynan in southern Jordan, a site considered internationally important due to its rich archaeological heritage spanning the Pleistocene and Holocene. This is the first study to describe the hydrological functioning of the Wadi Faynan, a meso-scale (241 km2) semi-arid catchment, setting this description within the framework of contemporary archaeological investigations. Historic meteorological records were collated and supplemented with new hydrological and water quality data. The modelled outcomes indicate that environmental changes, such as deforestation, had a major impact on the local water cycle and this amplified the effect of the prevailing climate on the flow regime. The results also show that increased rainfall alone does not necessarily imply better conditions for farming and highlight the importance of groundwater. The discussion focuses on the utility of the method and the importance of the local hydrology to the sustained settlement of the Wadi Faynan through pre-history and history. Ó 2011 Elsevier B.V. All rights reserved.

1. Introduction Water is an essential requirement for life and the relationship between climate and water availability has been fundamental to human activities in the past. The reconciliation of palaeoenvironmental and archaeological studies has focused on matching dates of major climate changes and key cultural developments but this view is over-simplistic (Maher et al., 2011). To gain a full understanding of societal advances, such as the onset of farming, careful consideration of all environmental factors is required. Consideration of climate alone may lead to an over-simplification that when the climate is warm and moist society flourishes and when cold and dry, society declines. The regional variations in the global climate are significant, as demonstrated by the present-day temperature and rainfall distribution across the Middle East, yet few studies of human development consider this, with many studies reliant on global temperatures reconstructed from ice-cores, or temperatures and rainfall extrapolated across a region such as ⇑ Corresponding author. Tel.: +44 (0)118 378 7315; fax: +44 (0)118 378 6666. E-mail address: [email protected] (A.J. Wade). 0022-1694/$ - see front matter Ó 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jhydrol.2011.10.033

the Levant from cave speleothems or lake levels (e.g. Bar-Matthews et al., 2003; Byrd, 2005). The oxygen isotope variability of such cave speleothems as a function only of near surface air temperature and/or rainfall fluctuations has been challenged (e.g. Frumkin et al., 1999). The Wadi Faynan in southern Jordan has a rich archaeological heritage, covering much of the Pleistocene and Holocene periods, and this has been comprehensively described in several recent volumes (Barker et al., 2007a,b; Finlayson and Mithen, 2007; Hauptmann, 2007; Mithen and Black, 2011). The Neolithic archaeology in Wadi Faynan is of key importance for studies investigating the origins of sedentism and agriculture (Finlayson and Mithen, 2007), whilst archaeological remains from later periods provide some of the most important evidence for early metallurgical activity in SW Asia (Hauptmann, 2007). To gain a complete understanding of water availability and security at key times in the past in relation to long periods of occupation in southern Jordan, it is necessary to appreciate the interplay between climate, geology (and in turn the soils) and vegetation, since climate controls precipitation and evaporation, geology defines the groundwater resource and soil character, and vegetation modifies flow pathways

2

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

and can provide an additional water store (Bull and Kirby, 2002). The influence of climate, geology and vegetation on hydrology can be established at a regional scale, such as for the Jordan Valley, but to appreciate the subtlety of the interplay between these three factors and the resultant influence on past cultural developments, it is anticipated that a study at a more local scale would be instructive. As such, this work focuses on the hydrology of the Wadi Faynan in relation to altered climate and vegetation cover at key times during the Holocene and the development of a modelling approach to couple simulations of past climate change to the hydrological response. The effect of soil on the catchment hydrology is included implicitly in this study through the consideration of the geology and vegetation. This hydrological study is based on an analysis of existing and recently collected hydrological and water quality data from the wadi and environs (Mithen and Black, 2011; Wade et al., 2011; Smith et al., 2011). The data are used to construct a conceptual model of the main water stores and pathways within the Faynan catchment and this model is then realised as a calibrated version of the Pitman hydrological model, a computer program devised for the hydrological simulation of semi-arid environments (Pitman, 1973). The numerical model is subsequently used to explore the hydrological functioning of the wadi under different rainfall and vegetation regimes, and the results are interpreted in terms of water availability and security and crop cultivation at key times during the Late Pleistocene and Holocene based on a sensitivity analysis and palaeoclimate model hindcasts of rainfall. In particular, the flood regime and the security of the groundwater source, which provides the perennial base flow, are considered. This modelling framework is a novel method for investigating catchment palaeohydrology. It has been proposed that the extent of woodland vegetation was much greater in the Neolithic than that evident today (Barker et al., 2007a; Mithen et al., 2007a; Hunt et al., 2004). Many studies, summarised by Beven (2002), have concluded that vegetation is the dominant control of runoff generation in semi-arid environments. The effect of a more extensive vegetation cover on the catchment hydrology is investigated. The infiltration capacity of soils in semi-arid regions can be reduced as the impact of raindrops will loosen the soil surface and rearrange the soil particles to form a crust. This crust promotes infiltration-excess overland flow, though the surface runoff generated may re-infiltrate as it flows to an area of higher permeability. In semi-arid areas runoff coefficients show a decrease as either the hill-slope length or catchment area increases. This occurs because the rainfall is often highly localised, and though runoff may be generated locally in response to the rainfall, as the water flows across the soil surface it re-infiltrates as it moves to an area of higher permeability (Yair and Klein, 1973; Yair and Lavee, 1976; Yair, 1987). A vegetation canopy or groundcover helps to protect the soil surface from forming a crust during high-intensity rainfall events, and roots form conduits within the soil thereby aiding infiltration. In addition, rain falling on a vegetation canopy may evaporate or may drip to the soil below either directly, as throughfall, or by trickling down the stems or trunks, as stem flow. Thus vegetation can reduce surface runoff by channelling water to the soil and by storing water which subsequently evaporates. To shape this investigation of climate-geologyvegetation interactions and the societal consequences, the research questions posed are: how did vegetation interact with rainfall change to control the hydrological functioning of the Wadi Faynan; and was this interaction significant in terms of water provision for cultural development? Present and ancient water management technologies in Wadi Faynan are not considered in detail, the focus being on the hydrology rather than the hydraulic constructs used to manipulate the water resource. Barker et al. (2007a) review archaeological

investigations of the water management in Faynan and environs, Crook (2009) discusses the irrigation of the ancient field-system, Darmame et al. (2011) review the contemporary situation and Yair (1983) considered the broader issue of water harvesting in deserts.

2. Study area and data resource The Wadi Faynan drains the eastern scarp slope of Wadi Araba, south of the Dead Sea (latitude: 30°380 N, longitude: 35°290 E; Fig. 1). The Wadi Faynan disgorges to Wadi Araba (at 300 m below sea level) after passage through the Jebel Hamrat al Fidan, an aplite-granite mass located at the mouth of the Wadi Fidan; the Wadi Fidan is the name given to the extension of Wadi Faynan between Al Qurayqira and Jebel Hamrat al Fidan. Wadi Araba runs north–northeast/south–southwest between the African and Arabian tectonic plates. Measured from the northwest side of Jebel Hamrat al Fidan, the catchment area of Faynan is 241 km2 and the main channel is approximately 25 km long flowing from east to west from the plateau down the scarp slope and past the town of Al Qurayqira. The predominant storm track for rainfall over Jordan is from the Mediterranean (Alpert et al., 2004; Enzel et al., 2008). The rain gauge network in southern Jordan is too sparse to provide a detailed spatial characterisation of the rainfall patterns but highlights a general increase in mean annual rainfall from approximately 400–700 mm year1 in the north-west of Jordan which reduces toward the desert centre to the east, with a mean annual average rainfall of approximately 50 mm year1. Rainfall patterns in the region of Faynan are dominated by the orographic effect of the rift escarpment, and the area of highest annual rainfall follows a north–south line between Kerak, Tafilah and Petra in the Wadi Musa with mean annual rainfall between 200 and 400 mm per year (Fig. 2). The climate of Wadi Faynan is currently classified as semiarid since mean annual potential evaporation, measured as 1978 mm year1 at Tafilah on the plateau, exceeds the mean annual precipitation measured as 50 and 312 mm year1 on the rift valley floor and plateau respectively and this demonstrates a large contrast in mean annual rainfall between the upper and lower reaches of the Wadi Faynan (Al-Qawabah et al., 2003; Tarawneh and Kadioglu, 2003; Fig. 2). Rainfall generally occurs between October and May, as elsewhere in Jordan. During winter, the precipitation can fall as snow on the plateau. The temperature extremes measured in the Dana Reserve, which are assumed representative of those in the Wadi Faynan, are typically a mean of 9 and 27 °C during January and August respectively (AlQawabah et al., 2003). East of the Jordanian Highlands, first steppe and then desert result from the decrease in relative humidity and precipitation and the increase in temperature (Hunt et al., 2004). Evidence of Lower and Middle Palaeolithic occupation occurs in the form of scatters of stone tools (including bifacial hand-axes and Levallois points) which have been found on many gravel terraces within the Faynan catchment (Rollefson, 1981; Mithen et al., 2007a). Evidence for later, Upper Palaeolithic and Epipalaeolithic occupation is much less common and no convincing evidence of Epipalaeolithic occupation has been found in the Faynan area (Mithen et al., 2007a), although as Epipalaeolithic occupation is common in the areas around Wadi Faynan, this absence of evidence may not be an accurate reflection of the settlement history of the area. Following this possible hiatus in settlement, the beginnings of the Holocene see clear evidence for occupation of the wadi by early (aceramic) Neolithic groups at sites such as WF16 (Finlayson and Mithen, 2007; Mithen et al., 2011) and Ghuwayr 1 (Simmons and Najjar, 2000) located near the confluence of Wadis Dana, Ghuwayr and Shaqyr as well as in the upper reaches of the wadi system at Wadi Badda (Fujii, 2007). There is

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

Fig. 1. A schematic map of the Wadi Faynan, its major tributaries and settlements.

Fig. 2. Rainfall patterns (isohyets) in the region of the Wadi Faynan (marked by the grey circle). Source: Department of Civil Aviation, Jerusalem, 1937–1938.

3

4

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

also evidence that the region was occupied during later, Pottery Neolithic periods, with a major settlement at the site of Tell Wadi Faynan (Al-Najjar et al., 1990). Following the occupation of the region by these early farming groups, human activity in the wadi system begins to focus more intensively upon exploitation of the region’s copper resources, with concurrent developments in water management systems. Evidence for Chalcolithic occupation occurs at Tell Wadi Faynan (Al-Najjar et al., 1990) as well as in the form of lithic scatters (Mithen et al., 2007a) and evidence for Chalcolithic-period metallurgical activity in the area (Barker et al., 2007b). During the Early Bronze Age, there appears to have been a step change in the occupation of the Wadi Faynan, as shown by the large (c. 12 ha) site at WF100 as well as by the presence of a plethora of smaller

occupation, industrial, agricultural and ritual sites and lithic scatters (Barker et al., 2007b; Mithen et al., 2007a). During this period there is the first evidence for water management, with the development of simple systems for managing floodwater for agricultural and domestic purposes (Barker et al., 2007b). Following a decline in human activity during the Middle and Late Bronze Ages (mirroring the general situation in the southern Levant), the wadi appears to have been intensively occupied once more during the Iron Age, when the large site of Khirbet an-Nahas was occupied, the field system (WF4) was developed and exploitation of copper resources greatly increased (Levy et al., 2004; Mattingly et al., 2007a). During the subsequent Nabatean and Roman/Byzantine periods the area was even more intensively exploited, with the growth of the large site of Khirbet Faynan, the industrial-scale exploitation of copper

Fig. 3. The geology of the Wadi Faynan area. Source Geological Map of Jordan 1:250,000, prepared by F. Bender, Bundesanstalt für Geowissenschaften und Rohstoffe, Hannover 1968 [Sheet: Aqaba-Ma’an and Amman]. Reproduced with permission. Not to scale. Ó Bundesanstalt für Geowissenschaften und Rohstoffe.

5

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

resources, and the further development and expansion of the WF4 field system (Mattingly et al., 2007b). Following the Classical period peak in human activity and occupation, evidence suggests that during the Islamic and Ottoman periods industrial activity rapidly declined and the region was mainly occupied by pastoral groups, a tradition which has continued up to the present day (Newson et al., 2007). Even from this brief review of the archaeology of the region, it is clear that the Wadi Faynan has been the scene of major periods of human activity for many millennia. Of particular interest, in the present context, is how human groups, drawn by the region’s mineral resources, attempted to manage the slender water resources of this semi-arid location. Today, the Bedouin located in the Wadi Faynan use the perennial water to irrigate their fields near the villages of Al Qurayqira and Rashida by conveying the water in plastic pipes under gravity from the Wadi Ghuwayr (Fig. 1). There are no major urban centres in the catchment, with the population being sparse in comparison with that of northwest Jordan where the rainfall and topography are more favourable for irrigation. The Wadi Faynan fringes the Dana National Park, which is maintained by the Royal Society for the Conservation of Nature (RSCN), and for which a detailed characterisation of the quantity and quality of the water resource is planned (Al-Qawabah et al., 2003). In this study, geological and hydrological information were derived from digital and paper maps. Hydrological measurements were collated from previous academic, government agency and engineering studies. These data were produced from national monitoring programs overseen by the Ministry of Irrigation, now subsumed by the Ministry of Water and Irrigation and the Jordan Valley Authority, and from engineering projects for planned irrigation schemes by Hazra Overseas Engineering Company (1978) and the development of the Arab Potash Company evaporation pans by Binnie and Partners (Overseas) Limited (1979). The historical data were supplemented by contemporary hydrological and water chemistry data gathered during field visits in 2006, 2007 and 2008. There was recourse to satellite imagery to confirm the presence of specific geological structures and to derive the catchment boundaries of the study area. The following mapping sources were used: topographic mapping at 1:50,000, Soviet military series sheets; source: Royal Jordanian Geographic Centre; landsat imagery 28 m resolution; SPOT satellite imagery provided by the National Imagery and Mapping Agency (NIMA), containing 10 m resolution Digital Orthorectified Imagery (DOI-10M) derived from data obtained from the SPOT Image Corporation under an unrestricted licence; the Geological Map of Jordan 1:250,000, prepared by Bundesanstalt für Geowissenschaften und Rohstoffe, Hannover 1968 (Sheet: Aqaba-Ma’an and Amman); Google Earth™ (http:// earth.google.com). The geology, soils, climate, vegetation and hydrology are described in detail in Wade et al. (2011). The two main tributaries of the Wadi Faynan, the Wadi Ghuwayr and the Wadi Dana are formed on two fault lines (Fig. 3). The geology of the Wadi Faynan is constituted by fluvial deposits and aeolian sands from the Quaternary period; limestones from the Tertiary (Eocene/Palaeocene epochs) and Cretaceous periods; sandstones from the Cambrian period; and porphyrite and aplite-granite from the Precambrian. There is also an outcrop of basalt from the Quaternary on the northeast rim of the catchment which forms the Jebel al Afa’ita (Fig. 3). The Wadi Ghuywar and Dana have contrasting geology, with both sustaining a perennial flow fed by a series of springs in the middle and upper reaches, though only flow in the Wadi Ghuywar reaches the lower sections due to more abstraction in, and around, the Dana village. The Wadi Ghuwayr and the Wadi Dana are steep. The range in altitude on the scarp slope varies from approximately 300 m asl at the Ghuwayr–Dana confluence on the alluvial plain to 1300 m asl on the plateau.

Fig. 3 (continued)

Springs in the Wadi Ghuwayr and Wadi Dana occur at the contact between the Cenomanian–Turonian (Cretaceous) limestones and the Cenomanian–Cretaceous sandstone, and at the fault-contact between the Cambrian sandstones and the underlying Precambrian volcanic rocks. It is also possible that springs occur at the

6

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

Fig. 3 (continued)

contact between the Eocene–Palaeocene and Cenomanian–Turonian limestones. The contact between the Cenomanian–Turonian limestones and the Cenomanian Cretaceous sandstone runs along

the eastern scarp slope, and springs on this interface are also found at Beidha and Petra, 30 and 35 km to the south respectively. Cherts (flints) are found in the plateau limestone and these are washed

7

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

downstream and copper and manganese veins are found in the igneous rocks. To a first approximation, the extent of the Cretaceous geology probably demarcates the groundwater divide (Figs. 3 and 4). The rocks of the scarp dip downwards towards the Wadi Araba whilst, on the plateau side of the mountain crest, the drainage is either northeast or east–southeast towards the desert (Agrar und Hydrotechnik GmbH, 1977). Further investigation is required to determine whether there is a hydrological connection to the continental aquifer and if so, what the water movements are. At this juncture, it is assumed that there is no flow of water from the continental aquifer to the Wadi Faynan, because of the dip of geological strata away from escarpment (Fig. 3). Rather, the springs in the Wadi Faynan are fed by relatively small aquifers whose extent equates approximately to the area of the surface water catchment. The volcanic ridge and associated faulting of the Jebel Hamrat al Fidan may have also influenced ancient settlement in the Wadi Faynan. The ridge of Hamrat al Fidan forms a barrier to the course of the Wadi Faynan channel, causing a constriction in the wadi along its pathway to the rift floor. This constriction has allowed a low-gradient alluvial plain to form between the volcanic ridge and the base of the escarpment. The alluvial plain may have improved the past fertility of the valley. Water is likely to be retained in the sand and gravel deposits for longer than would have occurred if the wadi emerged from the escarpment to the rift floor unhindered. The presence of the granitic barrier appears to be an isolated case along the rift valley south of the Dead Sea, although a few minor barriers are also present west of Petra. Satellite images of the Wadi Faynan show a number of Acacia (Acacia spp.) trees in, and close to, the constriction, perhaps indicating a longer retention of moisture following a flood. It is difficult to determine the key aquifer recharge mechanism in the wadi without recourse to a detailed hydrogeological study. The headwaters of both the Wadi Ghuwayr and the Wadi Dana are formed mainly from limestone with a partial soil mantle on the plateau. These headwaters are likely to act as the main recharge area. During flood events, aquifer recharge will also take place through transmission loss whereby flood water percolates from the channel substrate, which is predominantly constituted by unconsolidated sands and gravels, to the aquifers below. In the lower reaches, especially the alluvial plain in the Wadi Fidan, this water will be transferred deeper into the sedimentary detritus. These transmission losses are highly variable in space and time and likely vary significantly between rainfall events (Selaolo, 1998). Soil profiles dug in the alluvial terraces of the Wadi Fidan, 9.5 km east of the Hamrat al Fidan (35°270 3700 E; 30°370 1600 N; Aridisol; Typic Calciorthid; Yermic-Haplic Calcisol), as part of the project Jordan Soils and Land Management (2006) and 1.5 km NE of Bir Khidad (35°330 1200 E; 30°270 1000 N; Inceptisol; Calcixerollic Xerochrepts; Calcaric Cambisol) on the plateau suggest that the channel alluvium and the plateau soils are both well drained. The alluvial and plateau soils are classified as having slow and medium surface-runoff respectively, though the behaviour of all the catchment soils is likely to be very different under storm conditions in comparison with drainage following a storm and the catchment has many steep, bare rock slopes (United States Department of Agriculture – USDA, 1999). The 1:50,000 topographic mapping shows vegetation growth along the main channel in the Wadi Ghuwayr and a widespread area with sparse vegetation cover in the upper catchment of both the Wadi Ghuwayr and the Wadi Dana. Hammam Adethni, a tributary of the Wadi Ghuwayr, was found to be densely vegetated close to the wadi channel in which the flow was sustained by spring water (Lancaster et al., 2007; Mithen et al., 2007b). Hammam Adethni lies on the more southerly of the two major faults in the Wadi Faynan. A major spring was found in Hammam Adethni during the 2007 field season though its location was difficult to identify

exactly owing to a lack of available detailed mapping. An approximate location was determined on return from the field, using Google Earth™, as 30°360 5700 N, 35°320 4900 E. The channel of Hammam Adethni lies at the contact between the sandstone and porphyrite which is likely to form a spring line. The spring water temperature was 29.7 °C in Hammam Adethni which is difficult to interpret as either water derived from a deep source, perhaps transmitted upward along the geological fault, or alternatively as water derived from a shallower geological contact which has then ponded in a surface pool and been subsequently warmed by solar radiation. The present-day woodland of Hammam Adethni and the archaeobotanical remains at WF16 are described by Mithen et al. (2007b) and the flora of the Dana Biosphere reserve by Al-Qawabah et al. (2003). The vegetation of Hammam Adethni was divided into four distinct zones by Mithen et al. (2007b): woodland, maquis, porphyry steppe and sandstone steppe. The densely vegetated valley bottom contains little diversity, and in the woodland the vegetation mainly consisted of Salix acmophylla, Populus euphratica, Nerium oleander, Tamarix amplexicaulis and Ficus carica. The area of maquis vegetation was dominated by Phoenix dactylifera, Smilax aspera, Phragmites australis and Arundo donax. The vegetation growing on the porphyry contains fewer species than on the sandstone steppe south of Hammam Adethni, and the vegetation on the steppes was sparser than in the valley bottom next to the running water. Mithen et al. (2007b) found evidence of a sufficient overlap between the species located at Hammam Adethni and those present in archaeobotoanical assemblages from the early Holocene site of WF16 to imply that the riverine woodland provides an analogue for the predominant type of woodland in the Early Neolithic at the Wadi Faynan. Mithen et al. also go further, speculating that this type of woodland was more extensive in the Wadi Ghuwayr, especially in the lower areas and on the land immediately adjacent to WF16. Today, in the area of the Wadi Faynan between the Khirbet Faynan, the ancient field-system and Hamrat al Fidan, the landscape is denuded and there is an absence of soil organic matter (Tipping, 2007). The true extent of the ancient woodland in the wadi is unknown. Archaeological evidence suggests that trees were present in the lower reaches, though it is debated whether these were removed predominantly during the Roman–Byzantine period (Barker et al., 2007a), or in the early twentieth century by the Ottomans to build the Hijaz Railway (Palmer et al., 2007). The former argument is supported by a reduction in tree pollen in the sediments (Hunt et al., 2004). No flow data were readily available for the Wadi Faynan. The base flows in the Wadi Ghuwayr and the Wadi Nichel were measured during the 2006, 2007 and 2008 field visits (Table 1, Fig. 5) Table 1 Baseflow measurements in the Wadi Ghuwayr made during the 2006, 2007 and 2008 field seasons. Location

Northing

Easting

Sample date

Base flow (m3 s1)

Ghuwayr_2008_4

30°360 3400

35°310 5100

0.025

Ghuwayr_2008_3

30°360 4500

35°310 4500

07 May 2008 07 May 2008 07 May 2008 07 May 2008 29 April 2007 24 April 2006 24 April 2006 09 May 2008

0

00

0

Ghuwayr_2008_2

30°36 34

Ghuwayr_2008_1

30°370 0800

35°300 4700

Ghuwayr_2007_WGDS3

30°360 4500

35°310 4500

Ghuwayr_2006_WGDS2

30°370 1400

35°300 3400

0

00

Ghuwayr_2006_WGDS1

30°37 10

2008_Fidan

30°390 1100

35°31 51

00

0

35°30 43

00

35°230 4800

0.016 0.032 0.020 0.093 0.018 0.024 0.012

8

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

by the velocity–area method. In total eight sites were surveyed. Changes in the stream bed cross-section occurred between years owing to the disturbance of the bed during flood events. The peak flows were estimated using the Darcy–Weisbach technique as described by White (1995). The observed base flow measured close to WF16 showed a range of 0.02–0.09 m3 s1. Water chemistry data were collected during the 2007 and 2008 field visits and the sampling method and analysis is described fully in Wade et al. (2011). The data are limited in time and space but form a baseline survey and provide preliminary information on the flow pathways and water sources. The d18O values increase from 6.1‰ to 5.3‰ along the channel in the Wadi Nichel at the sandstone/carbonate contact as water flows from its spring source and evaporates as it progresses downstream to the confluence with Hammam Adethni. Immediately downstream of the confluence between Hammam Adethni and the Wadi Ghuwayr, the d18O values decrease to 7.5‰, owing to mixing with the waters from Hammam Adethni which have a lower proportion of 18O due to the relatively short distance between the spring source and the confluence. The d18O ratio of the stream water then increases to approximately 5.8‰ as the water evaporates as it flows along the Wadi Ghuwayr before disappearing from the surface as it infiltrates into the alluvial plain downstream of WF16. The spring at Hamrat al Fidan also has a d18O ratio of 5.8‰ which provides evidence that the spring water is derived from water that infiltrates into the alluvial plain near WF16, and then moves at depth without further evaporation, to resurface as it follows the upthrusted granitic barrier. A deep source of the Fidan spring would be expected to cause a lower measured d18O ratio. The patterns in the d18O data from the Wadi Nichel to the Jebel Hamrat al Fidan spring are repeated in the d2H data. The anions show a general increase in concentration between surface water samples taken in the Wadi Ghuwayr and the spring water at Jebel Hamrat al Fidan. The increase in anion concentrations is most apparent in the chloride measurements which increase from approximately 70–220 mg l1 between the Wadi Ghuwayr and the Fidan spring, nitrate increases from 1.7 to 8.1 and sulphate from 75 to 226 mg l1 (reported as mg L1 of ion). An increase is also evident in the cations: sodium increases from 51 to 103, potassium from 2.5 to 4.0, magnesium from 35 to 61 and calcium from 64 to 121 mg l1. A possible explanation for this increase is evaporation. However, if this is the mechanism for these increases then it contradicts the interpretation of the stable isotope data, which suggests little or no evaporation along the sub-surface pathway from Wadi Ghuwayr to the Fidan spring. Another possible explanation, consistent with the observed water quality, is that evaporation is minimal at depth but the anion and cation concentrations are modified by the Wadi Fidan alluvium or possibly altered by leachates from wastes at Al Qurayqira entering the groundwater during storm events.

A conceptual model of the key water pathways and stores in the Wadi Faynan is shown in Fig. 4. This model is built by integrating all the knowledge ascertained from the review of the existing and newly collected data. The following description and assumptions were used to build the conceptual model:  the major aquifers are defined by the catchment boundary and the major aquifers are the limestone and the sandstone;  groundwater recharge of the limestone and sandstone aquifers occurs through the limestone and colluvium in the upper reaches of the Wadi Faynan around Dana and Shawbak;  this recharge is supplemented by transmission losses from the main wadi channels to the underlying aquifers and the shallow channel alluvium;  springs occur at the contact between the limestone and sandstone and between the sandstone and Precambrian volcanic rocks;  the Precambrian volcanic rocks act as an impermeable layer, keeping the water near the surface as it flows past WF16 before the sand and gravels in the channel deepen in the Wadi Fidan alluvial plain and the water flows beneath the surface, possibly along the contact with the underlying aplite-granite;  the key pathways are lateral perennial flows through the limestone and sandstone with surface overland flow generated during rainfall events;  snow does fall during winter in the headwaters of the catchment but it is assumed that this will infiltrate into the welldrained soils upon melting. 3. Methods An overview of the modelling framework is shown in Fig. 6. The framework chains together a global climate model, a regional climate model, a weather generator and a hydrological model. 3.1. Palaeoclimate modelling The global climate model (GCM) and regional climate model (RCM) used in this study were HadSM3 and HadRM3 respectively. HadRM3 is a regional-scale climate model with a spatial resolution of 0.44° (50 km) for both latitude and longitude developed by the UK Hadley Centre. The RCM is formally compared with observations from 1961 to 1990 in Black (2009, 2011) and Brayshaw et al. (2011a) and summarised here. For the control period, 1961–1990, the modelled temperatures in the eastern Mediterranean are reasonably well represented. The modelled control-period precipitation data show that whilst the spatial pattern of precipitation over the 30 year control period is generally modelled well, the resolution of the HadRM3 model is too coarse to capture the subtle variations in rainfall across the rift valley and northern Jordan,

Fig. 4. A schematic diagram of the geological cross-section made 5 km to the north of the Wadi Faynan. The geology forms the basis for the conceptual model of the key water stores and pathways, shown by arrows, in the Wadi Faynan. Based on the Geological Map of Jordan 1:250,000, prepared by F. Bender, Bundesanstalt für Geowissenschaften und Rohstoffe, Hannover 1968 [Sheet: Aqaba-Ma’an and Amman].

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

9

Fig. 5. Sample site locations in the Wadi Faynan from the 2006, 2007 and 2008 field seasons.

Fig. 6. A schematic of the modelling framework used to investigate catchment palaeohydrological functioning through the Holocene. The boxes outlined with single, solid lines are the models, the dashed box represents an estimate based on data and the box with the double-line outline represents the final output from the framework.

likely because the RCM cannot resolve the topographic variations in this region. Moreover, a comparison with rain gauge measurements shows that the intensity of rainfall is under-estimated, leading to the biases in the annual totals (Black et al., 2011). Thus, these results highlight the need for a weather generator, both to interpolate the RCM simulations to a point and to correct the model bias (see Black, 2011 for further detail on this). Climate projections were extracted from HadRM3 simulations of the 1961–1990 control and key representative scenarios of 20-year periods through the Holocene and Late Pleistocene at 2, 4, 6, 8, 10 and 12 k years BP. These scenarios, generated using a consistent modelling framework driven by a global climate model (HadSM3) are discussed in full by Brayshaw et al. (2011a,b) and the modelled rainfall is compared with pre-Industrial (c. 1800) observations. As the modelled outcome for rainfall for each scenario corresponds to a simulated

20-year period, the results were interpreted cautiously, namely as plausible scenarios for Holocene and Late Pleistocene climate rather than definitive reconstructions. Brayshaw et al. (2011a,b) identified a technical error in the GCM simulations prior to 8 ky BP but, based on the discussion in those papers, this error does not significantly affect the results or conclusions of this article. A weather generator was used for both the sensitivity studies and the future scenarios. The weather generator is described and evaluated in Black (2011), where it was shown that, while it was capable of reproducing the main features of the observed rainfall seasonal cycle, there were some biases – with more rain at the margins of the rainy seasons than was observed. The weather generator derives daily rainfall stochastically for a 100-year period, according to underlying patterns of daily rainfall inferred from the model and observations. In the weather generator used here, the patterns of daily rainfall were described statistically through the mean rain per rainy day (rainfall intensity) and the rainfall occurrence (rainfall probabilities). For the scenario studies, the rainfall probabilities were derived from the observed time series and then adjusted as follows. For the past scenario integrations, the probabilities of rain given rain and of rain given no rain were determined from the RCM, for the control period, 1961–1990. The same probabilities were determined for the scenario periods, and these hindcast probabilities were bias-corrected using a factor derived by comparing the control model integration with the observed probabilities. The distribution of rainfall intensities (rain per rainy day) was based on the observed time series and the model projections of changes in intensity (which were in practice generally small). Because the output time series were considerably longer than the training series, an extra parameterisation for extreme rainfall events, based on a theoretical distribution was incorporated into the statistical model of intensity.

10

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

3.2. Hydrological modelling The conceptual model was realised as a numerical model through an implementation of the Pitman model (Pitman, 1973). The Pitman model was chosen because it was developed in South Africa for semi-arid hydrological conditions and forms a trade-off between model complexity, data requirements and useful model output appropriate for aims of this study. The Pitman model is designed to be applicable at the catchment scale and was applied at the daily time-step in this study to investigate flood characteristics. The purpose of this model application is not to quantify flood flows exactly. That cannot be achieved in this case because of a lack of observed time-series flow data with which to rigorously assess model performance. Rather, the purpose is to define a hydrological model as a best estimate of how the hydrology functions and then run scenarios to explore how changes in rainfall amounts and vegetation characteristics affect flood characteristics and the base flow. The latter provides a smaller but more secure water source in the Wadi Faynan. The Pitman model was applied to the Faynan catchment, defined from a point on the channel network adjacent to the ancient field-system and immediately downstream of the confluence of the Wadi Ghuwayr and the Wadi Dana (Fig. 5). At this point the upstream contributing area is 115 km2. This was done since the lowest point at which the measurements of both the base flow and peak floods were made was the outflow of the Wadi Ghuwayr and the Wadi Dana, thus allowing estimates of the base flow and peak flow to be made at the confluence. It was extremely difficult to survey the channel downstream of the confluence where the alluvial plain flares to a width of approximately 1 km. Modelling the catchment flows to a point on the channel network adjacent to the ancient field-system also removes the complication of simulating the transmission losses and water residence (or transit) times in the alluvial plain of the Wadi Fidan and eliminates the need for a definitive understanding of the source of the Fidan spring within this model-based assessment. This application of the Pitman model required daily estimates of rainfall and the potential evaporation. The use of the rainfall data from Tafilah was appropriate for the model application. Tafilah is located on the plateau 18 km north of Dana and receives similar rainfall to the upper reaches of the Wadi Faynan. Moreover, a substantial daily record of rainfall was available from 1 October 1937 to 30 April 1974 at Tafilah, allowing the model to be run for a relatively long period which is important when making an assessment of flood magnitude and frequency. An estimate of the annual variation in potential evaporation was derived from monthly measurements of wind speed, sunshine hours, relative humidity and air temperature available for a 2 year period using the Penman equation. The data were available for the meteorological station at Ma’an. This 2 year estimated time-series was repeated to form a time-series of 36 years, the same length as the observed rainfall time-series. This repetition of the 2 year timeseries, whilst a clear over-simplification of the changes expected during the 36 year period, was done to allow progression with the model application. It is practically impossible to separate the effects of transmission loss from those of infiltration and deep percolation to the underlying aquifer when estimating groundwater recharge, so no attempt was made to do so within this model-based assessment. Rather, the combined effects of transmission loss, infiltration to the soil and subsequent deep percolation were considered together as a single groundwater-recharge mechanism. To apply the model, estimates of the groundwater volume and infiltration rate were required. From the available geological mapping, the depth of the Cenomanian–Turonian limestones was estimated as 10 m and the horizontal extent as 60  100 m2. Similarly,

the depth of the sandstone was estimated as 40 m and the extent as 60  100 m2. Assuming a porosity of 0.25, this gave an ‘effective’ groundwater depth of 6.5 m for the aquifers in the catchment calculated using the equations of the Pitman model. The ‘effective’ depth is that in which the groundwater interacts with the surface channel network. Observations from the soil pits dug in, and close to, the catchment suggest a soil depth of approximately 1 m in the alluvial terraces and on the soil mantle covering the headwaters between Dana village and Shawbak. These observations contrast with those made during journeys along the Wadi Ghuwayr and Wadi Nichel, where the landscape often consists of steep, sometimes vertical, rock faces or shallow soils on steep slopes covered with stones and gravel, and typically little vegetation cover. As such, the soil depth in the model application was assumed to be 50 cm. With an assumed soil porosity of 0.45, the maximum soil moisture capacity was 225 mm. Infiltration is notoriously difficult to measure in the field. Double-ring infiltrometers were used during the 2006 and 2008 field visits, but it was felt that the results were misleading. The technique depends on filling the infiltrometer to a depth of approximately 20 cm and then recording the drop over a 30 s time interval. The infiltrometer was topped up after a 5 cm drop in water level to prevent differences in hydraulic head affecting the measurements. In reality, such a head of water would never exist during a rainfall event except, perhaps, in areas of depression storage. A recent study by Crook (2009) measured infiltration rates in the WF4 field-system, obtaining values of 29 and 90 mm h1. A review of the literature was also undertaken to help determine infiltration rates for use in the Pitman model application to the Wadi Faynan. In a study of a Typic Calciorthid soil (Aridisol) which is similar to the soils found in the lower Wadi Faynan, an experiment was done in which soil was packed into columns and then watered with simulated rainfall for 120 min (Mandal et al., 2008). It was observed that infiltration decreased as the cumulative rainfall increased owing to seal formation. After 20 mm of rainfall the infiltration rate was 37.5 mm h1; after 40 mm of rainfall the rate dropped to 12 mm h1; after 70 mm of cumulative rainfall, the infiltration rate was below 5 mm h1. Other studies using single- and double-ring infiltrometers suggest a range for soils of 10–14 mm h1 for fine, crusted soil to 8–24 mm h1 on an interfluve slope (Abu-Awwad and Shatanawi, 1997; Al-Qinna and Abu-Awwad, 1998, 2001). Abu-Awwad and Shatanawi (1997) studied three field sites with sparse vegetation cover and observed a range of values from 9 to 35 mm h1. The model parameter representing the maximum infiltration rate, ZMAX, was set to 8 mm day1. This value and the minimum infiltration rate ZMIN, set to 0 mm day1, were determined through model calibration so as to obtain a runoff coefficient between 30% and 40%, thereby within the range of estimated runoff coefficients for studies in arid zones in Spain (Puigdefàbregas and Sánchez, 1996), southwestern USA (Dunne, 1978), a catchment in the semi-arid zone of Australia (Williams and Bonell, 1988), and within the range of runoff coefficients given in a general review by Wheater (2002). The total simulated runoff over the 36 hydrological years (1 October 1937–30 September 1973) modelled was 3640 mm, producing a long-term mean of 101 mm year1. This compared with a rainfall input over the same period of 10,019 mm, producing a long-term mean of 278 mm year1. In terms of the field and laboratory infiltration measurements, the calibrated value of 8 mm day1 was lower than the values obtained from the literature review. However, it is difficult to interpret the laboratory and point estimates of infiltration in terms of a value representative of the entire catchment. Moreover, the concerns about the methodology for measuring infiltration are also valid for the studies of infiltration rate found in the literature. The

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

Pitman model operates on a daily step, yet rainfall events may last only a few hours. If the modelled and measured infiltration rates are thought broadly representative of a rainfall event then the two rates are similar. Whilst the absolute value of the infiltration rate representative of the catchment can be debated, the value of 8 mm day1 is still useful because it provides a benchmark estimate for present-day conditions and to progress the study it can be assumed that sparser or denser vegetation will lower or increase the rate respectively. The model was also calibrated to give a base flow of 0.03 m3 s1 immediately upstream of the ancient field-system, which was within the range of the recent observations (Table 1). The threshold for the onset of simulated percolation from the bottom of the soil profile was set to 40 mm by calibration to achieve a situation whereby, over the simulated 36 year period, there was only a minimal change in the stored groundwater. It is unknown whether the groundwater store is truly stable or whether there is a net loss or gain of water with time. However, for the purposes of this study a stable groundwater store provides a basis with which to compare changes in the groundwater during scenario exploration. Anecdotal evidence suggests that a good year in terms of water availability occurs when there are five rains over the headwaters of the catchment. In discussions regarding hydrological conditions in the Wadi Faynan and the lands around Shawbak and Dana, Lancaster and Lancaster (1999) note: People say five rains for the plateau and slopes, or five flowings for the low-lying land above the Wadi Araba, are needed for good crops and natural grazing; these occur on average one year in five. People expect a good year, two or three ordinary years, and one or two poor years. (Lancaster & Lancaster citing Bedouin informants 1999 p. 118). This anecdotal evidence was used to establish a definition of ‘good hydrological conditions’ in the Wadi Faynan for the present farming and grazing regimes. Specifically a flow threshold, determined from the simulation results, was defined as 12 m3 s1. This value was chosen to give a probability of 0.2 that five or more flows occurred in 1 year over a 5 year period. 3.3. Hydrological modelling sensitivity analysis A sensitivity analysis was done to explore the catchment response to changes in the annual rainfall amounts and vegetation

11

cover using simple scenarios whereby the total rainfall was scaled by multiplying by the daily values in the time series by 20%, 40%, 60%, 80%, 120%, 140%, 160% and 200%; the model parameter ZMAX, which describes the maximum infiltration, was increased through the series 0, 5, 8, 35 and 100 mm day1; and combinations of altered amounts of rainfall coupled with different vegetation cover the rainfall and maximum infiltration rates were applied. These scenarios are simplified and arbitrary, however, given the lack of a flow time-series with which to calibrate the model, it was deemed more appropriate to keep the scenarios simple. The purpose was to gain an overview of the sensitivity of the hydrological response to a range of rainfall amounts and changes in vegetation density rather than precise flood estimates. Air temperature and evaporation were kept constant to explore the sensitivity to rainfall only. 3.4. Palaeoclimate and land cover modelling Following the sensitivity analysis for the rainfall, vegetation then rainfall-vegetation combinations, the modelled paleaorainfall from HadRM3, biased corrected and extended to a 100-year simulation period using the weather generator, was used to investigate the changes in rainfall input through the Late Pleistocene and Holocene and compared with the reconstruction of palaeorainfall from the Soreq cave speleothem and pre-Industrial (c. 1800) rainfall. To simulate the change in vegetation the infiltration rate was varied and the vegetation was assumed to change as follows. At the start of the Holocene 12 ky BP vegetation was sparse resultant from the cooler global air temperatures and the infiltration was assumed to be low (8 mm day1). By 10 and 8 ky BP the vegetation established itself as global temperatures warmed and the infiltration was set to 100 mm day1. By 6 ky BP vegetation began to be cleared in the Faynan due to onset of farming (80 mm day1) and by 4 and 2 ky BP major clearance of vegetation for pit props and smelting reduced the infiltration to 35 and 8 mm day1 respectively. 4. Results 4.1. Palaeoclimate model results The palaeorainfall rainfall sequences reconstructed from the Soreq speleothem suggests that the early Holocene and Late Pleistocene from 8 to 14 ky BP was wetter than present day (1961–1990) and pre Industrial conditions (c. 1800) and this is

Fig. 7. The flow duration curve based on the flows simulated in the Wadi Faynan over the period 1937–1973. The highest simulated flow of 98 m3 s1 occurred in response to a mean daily rainfall event of 90 mm.

12

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

Table 2 Peak floods in the Wadi Ghuwayr and the Wadi Dana estimated using open-channel techniques. Location

Northing 0

Ghuwayr_A_2006 Ghuwayr_A_2008 Ghuwayr_B_2008 Ghuwayr_C_2008 Ghuwayr_D_2008 Dana_2006

Easting 00

0

30°37 31 30°370 3000 30°370 2600 30°360 3900 30°360 4900 30°370 4100

00

35°29 46 35°290 3900 35°300 1300 35°310 4300 35°310 5200 35°290 1100

Sample date

Annual (m3 s1)

Bank full (m3 s1)

Overspill (m3 s1)

26 05 05 05 05 26

4 12 2 0.4 2 10

89 31 22 4.5 15 88

103 339 209 – – 97

April 2006 May 2008 May 2008 May 2008 May 2008 April 2006

Table 3 Surface and groundwater flow statistics for the Wadi Faynan generated by running the Pitman hydrological model with modified rainfall inputs and/or a modified infiltration rate parameter. ⁄ The groundwater effective depth and flow are the modelled values on the last day of the 37 year simulation. Rainfall (% present day)

Maximum infiltration rate (mm)

Flow statistics (m3 s1)

Groundwater ⁄

Effective depth (m)

Peaks over threshold ⁄

Flow (m3 s1)

Number of years with five or more flows >12 m3 s1

Max

Q10

Q50

Q95

8 8 8 8 8

5 53 98 144 189

0.046 0.046 0.047 0.050 0.053

0.044 0.045 0.047 0.049 0.050

0.042 0.043 0.046 0.047 0.047

6.1 6.2 6.5 6.9 7.0

0.04 0.04 0.05 0.05 0.05

0 2 7 18 28

0 5 35 100

109 101 84 40

0.047 0.047 0.057 0.061

0.044 0.045 0.053 0.054

0.042 0.044 0.048 0.048

6.0 6.2 7.5 7.9

0.04 0.04 0.06 0.06

26 13 2 1

Combination scenarios 200 5 200 8 200 100

214 211 152

0.057 0.054 0.115

0.047 0.050 0.083

0.047 0.048 0.051

6.5 7.2 12.0

0.05 0.05 0.12

31 29 2

Rainfall scenarios 20 60 100 140 180 Infiltration scenarios 100 100 100 100

reproduced by the palaeoclimate modelling, though not with the same magnitude for four grid squares of the HadRM3 model from 30.75° latitude, 35.75° longitude (bottom left) to 31.15° latitude 36.25° longitude (top right). The latitude and longitude quoted relate to the rotated HadRM3 grid. After approximately 8 ky BP there is evidence of a drier climate though the palaeoclimate model suggests a wetter later Holocene. Comparison of the palaeorainfall is difficult because the climate model provides a snap shot of discrete 20 year periods whereas the Soreq sequence provides a more continuous record. Moreover there will be errors in the output of the Regional Climate Model due to structure and parameter uncertainties as well as natural (internal) variability and the dating of the Soreq speleothem is subject to uncertainties.

4.2. Hydrological simulation results The output from the hydrological model shows that the calibrated flows are within the base flow and peak flow constraints identified by field observations (Fig. 7, Table 2). The base flow in the catchment was observed to be in the range 0.02–0.09 m3 s1 and the model replicates this. The simulated annual flood ranges from 2 to 98 m3 s1 and floods that generally occur every year or two years were simulated to have magnitudes in the range 2– 17 m3 s1, which is within the range estimated by survey. The extreme floods with return periods of 12–37 years also lie within the estimated flood flow range when the channel is flowing full. The flow-duration curve also replicates the ephemeral characteristic of the catchment which is dominated for 95% of the time by base flow (Fig. 7). For only 5% of the time-period considered do the flows increase in response to precipitation events, and this hydrological behaviour is typical of semi-arid catchments (Bull and

Kirby, 2002). However, during floods the increase in flow is rapid with the simulated magnitudes reaching a maximum of 98 m3 s1 in response to the largest observed rainfall event at Tafilah of 90 mm day1.

4.3. Sensitivity analysis – rainfall All the flow-duration curves (not shown) demonstrated a prolonged period dominated by base flow, even those where the rainfall is 180% or 200% that of the observed time-series. This occurs because the magnitude only was changed in the rainfall scenarios. Neither the frequency nor duration of rainfall occurrence was altered (Table 3). Under the scenario of extreme low rainfall (56 mm year1), specifically 20% of the rainfall amount measured at Tafilah from 1 October 1937 to 30 September 1973, the groundwater depletes and there are few floods. Most of the rainfall that falls over the catchment infiltrates but this is insufficient to recharge the groundwater. Even at 60% of observed rainfall (169 mm year1), the flood occurrence is limited and the groundwater depletes, suggesting that this water resource is sensitive to changes in climate. In contrast, with 180% or 200% of the observed Tafilah rainfall (500 or 556 mm year1, respectively) which is more typical of present-day northwest Jordan, the hydrology is more responsive. The floods become larger and more frequent, and in at least 28 years of the 36, flood events exceed the 12 m3 s1 threshold (Table 3). Specifically, as the rainfall increases from 20% to 200% of the observed, then the number of years where floods are greater than 12 m3 s1 increases from 0 to 29. Between 100% and 180% of the observed mean annual rainfall, the relationship between mean annual rainfall and flood occurrence is linear. The number of floods then increases only marginally as rainfall is increased above a threshold of 500 mm year1 (180% of the

13

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

exploration of the relative changes in flood magnitude and frequency to determine the hydrological sensitivity of the Wadi Faynan. The modelled palaeorainfall suggests an early Holocene at 12 ky BP wetter than today and given a limited vegetation as it establishes itself, the result is high magnitude flood events (Table 4). The maximum flood is lower than the control period simulation but the number of large (>12 m3 s1) flood events is greater than today due to the more frequent, high intensity rainfall at 12 ky BP. At 10 and 8 ky BP the mean annual palaeorainfall amount is higher than that of the control period but more vegetation leads to smaller flood flows and more groundwater recharge. The effective groundwater depth at the end of the 100 year simulation period was 10 and 9 m for 10 and 8 ky BP respectively thereby showing an increase in the groundwater levels. The median flow is also higher for 10 and 8 ky BP than for the control period suggesting a higher base flow. At 6 to 4 ky BP, it is assumed that trees are felled to clear land for farming and vegetation cover is further reduced due to the increasing aridity. This relatively rapid reduction in vegetation leads to reduced infiltration and increased flood flows and lowered aquifer recharge, though flow derived from the aquifer is able to sustain the base flow and thereby the perennial flow is maintained. By 2 ky BP the landscape is denuded as the Romans further removed trees for pit props and to support smelting activity to extract copper ore. This removal of vegetation is assumed to further lower the infiltration rate and thereby increase floods and lower aquifer recharge. Over the whole of the Holocene there is only a very limited simulated change in the lowest flows; the minimum and Q95 flows remain approximately the same.

currently observed mean annual rainfall). This occurs because at this magnitude of rainfall, nearly every rainstorm produces channel runoff of 12 m3 s1 or greater. With increased rainfall the groundwater store also grows as deep percolation, which recharges the aquifers, is enhanced (Table 3). 4.4. Sensitivity analysis – infiltration With no infiltration, the groundwater depletes from 6.5 to 6 m owing to a lack of recharge, and surface runoff is the dominant hydrological pathway. Of the 36 years simulated, 26 produce events with a flow of 12 m3 s1 or more. As the infiltration rate is progressively increased from 0 mm day1 representative of limited sparse vegetation to 35 and 100 mm day1 indicative of greater vegetation abundance, then the groundwater recharge increases, and flood frequency and magnitude is reduced (Table 3). Specifically, with infiltration rates of 35 and 100 mm day1, the simulated effective groundwater depth increased to 7.5 and 7.9 m respectively; the base flow increased from 0.046 m3 s1 at calibration to 0.057 and 0.061 m3 s1; the maximum flood during the 36 year period decreased from 98 to 84 and 40 m3 s1, and the number of years with events greater than 12 m3 s1 decreased from 7 to 2 and 1. As the infiltration rate increases, for a given annual rainfall amount, the flow-duration curve (not shown) becomes less steep as the floods become less extreme. 4.5. Sensitivity analysis – combination scenarios With low rainfall and low infiltration, the system is responsive with floods generated albeit of low magnitude (Table 3). As the vegetation cover increases with a commensurate increase in infiltration rate, then groundwater recharge also increases, and flood magnitude and frequency decrease. If rainfall is high (200%) and the catchment vegetated (infiltration 100 mm day1), then the simulated maximum flood increases from 98 to 152 m3 s1 but the base flow and groundwater effective depth also increase to 0.12 m3 s1 and 12 m respectively, and the flood frequency drops to 2 of 36 years having a flow of 12 m3 s1 or greater.

5. Discussion The development of a chain of process-based models to simulate the palaeoclimate, bias-correct the rainfall estimates and then model the hydrology allowed a rational exploration of research questions regarding the altered hydrological functioning of an archaeological site at key times during the Late Pleistocene and Holocene. Inevitably there are many uncertainties within the approach but the advantage is that the modelling framework formalises assumptions and hypotheses and integrates existing and new data to create a conceptual understanding of the system. Moreover, the uncertainties are listed and here the main ones are: how well the climate model and weather generator represent the regional differences in rainfall, and the assumed infiltration rates for the hydrological modelling. This is instructive, as it helps focus future research effort. Such models chains are becoming commonplace for exploring the hydrological response to future climate change (e.g. Samuels et al., 2009), here it is demonstrated that such an approach can be implemented to look at past hydrological functioning for snapshots through the Holocene. This examination of the hydrology and hydrogeology of the Wadi Faynan reveals why the location was potentially attractive

4.6. Palaeoclimate-hydrological modelling The control period is defined as from 1937 to 1973 when the hydrological model was calibrated. Based on the modelled rainfall time-series from the GCM/RCM/weather generator, the simulated maximum flood flow was 66 m3 s1 for the control period (Table 4). This maximum flood differs in magnitude to that generated during hydrological model calibration (98 m3 s1) because the calibration used observed, rather than simulated, rainfall (Table 2, Fig. 6). This result implies that the simulated flood magnitude using the modelled rainfall is likely to be lower than in reality, but the results of the palaeoclimate-hydrological model are still useful as an

Table 4 The modelled rainfall and flow characteristics of the Wadi Faynan for 100 year periods at 2000 year intervals through the Holocene. Period

Infiltration rate

Rainfall

Max

(ky BP)

(mm day1)

Percentage of present day

(m3 s1)

12000 10000 8000 6000 4000 2000 Present day

8 100 100 80 35 8 8

119 125 111 107 119 127 100

60 6.3 5.0 19 31 66 66

Q50

Q95

Groundwater effective depth (m)

Number of floods >12 m3 s1

0.05 0.08 0.07 0.07 0.07 0.05 0.05

0.05 0.05 0.05 0.05 0.05 0.05 0.04

7.0 10.2 9.2 8.8 8.9 7.2 6.2

251 0 0 2 22 270 204

14

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

for early settlement. The presence of springs, due to the geological faults and contacts between different lithologies, makes the locality attractive to sedentary farmers and transient herders because of the reliability of the flow in a semi-arid landscape. Faynan has other attractions based on the geology. These include a ready source of cherts (flints) in the limestone of the plateau which are washed down to the lower reaches, and veins of copper and manganese within the igneous rocks, the former exploited on an industrial scale during Classical periods. Thus, following the initial attraction of perennial water, the subsequent finds of cherts and metals would have provided further motivation for settlers to maintain their presence within the Wadi Faynan. The aplite-granite barrier of the Jebel Hamrat al Fidan is unique along the rift valley between Aqaba and the Dead Sea, and this feature is likely to aid moisture retention in the alluvial plain of the Wadi Fidan. The catchment also lies north of other settlements on the scarp slope of the Wadi Araba where spring sources are located, for example Beidha and Petra. Possibly together these would have provided a network of watering points for travellers or herders. The gorges of the Wadi Ghuwayr and the Wadi Dana also provide a route up the scarp slope to the plateau above, where the land is more productive than in the valley being a Calcaric Cambisol with a depth of approximately 1 m and having a higher rainfall. The analysis of the collated data enabled a conceptual hydrological model to be created, although this is not definitive. In particular, further water quality measurements are required to confirm the degree of evaporation in the alluvial plain of the Wadi Fidan and to provenance the Hamrat al Fidan spring water. Thus, as this time, it is not possible to confirm that water flows through the Wadi Fidan alluvial plain at depth before emerging at the spring adjacent to Jebel Hamrat al Fidan, being forced to rise to the surface or near surface by the thrusted aplite-granite. However, it is strongly suspected that this is the case. The hydrological simulations suggest that, with the same mean annual rainfall (278 mm year1) as observed from 1937 to 1973 at Tafilah, but with a more dense vegetation cover giving an increased infiltration rate, then the hydrology of the Wadi Faynan was much more conducive to sedentary farming and settlement. With a modelled maximum infiltration rate of 35 mm day1 indicative of more dense vegetation cover than observed at present, water accumulates in the groundwater store producing a higher perennial base flow, and the flood magnitude and frequency are reduced. The simulated removal of the vegetation in the Wadi Faynan caused a severe, negative impact on the water resource. Specifically, the removal had two key effects due to reduced infiltration. The first was a depletion of the groundwater resource through reduced recharge, and the second was greater flood magnitude and frequency. Increased flood frequency occurred in addition to increased flood magnitude because surface runoff that would have previously re-infiltrated into the soil remained on the surface and flowed to the channel network. Though not tested in the applications of the model, it is also possible that the removal of the vegetation canopy will reduce throughfall and stem flow, thus increasing flood magnitude and frequency. A vegetated Wadi Faynan would have been much easier to farm because the vegetation would have helped to provide a higher base flow and promoted the retention of soil moisture. The feedback between vegetation, hydrology and soil erosion is important. The removal of the vegetation cover is shown to cause increased flooding, probably leading to soil erosion. This, in turn, would have made conditions more difficult to re-establish the natural cover and crops, reinforcing the negative spiral of larger flood events and reduced soil cover. This study is novel in its consideration of the interactions between climate, geology and vegetation in the Wadi Faynan and the results from the model-based assessment are suggestive only,

simulating the key alterations in the water stores and pathways in the catchment. It is unclear whether the reduction in rainfall caused the decline in vegetation, or if the vegetation was cleared by the Romans for pit props or smelting, or by the Ottomans for railway sleepers. These further details regarding the extent and timing of the changes in vegetation cover are required to provide a more detailed assessment of the hydrological functioning of the Wadi Faynan at key times during the Holocene. 6. Conclusions This study is also amongst the first to both describe and model the hydrology of the Wadi Faynan. Prior to this, relatively little work had focused on the hydrological functioning of this archaeologically important, semi-arid catchment. A major component of this work was the collation of a substantial database with which to develop an understanding of the catchment hydrology and hydrogeology. Together, new and historic data were analysed to understand the catchment hydrological and geographic characteristics. This analysis has highlighted the importance of the climate, geology and soils in the determination of the catchment hydrology, the vegetation cover and the establishment of sedentary communities in what is now a harsh landscape with low rainfall and sparse vegetation. Along the scarp slope of the rift valley between Aqaba and the Dead Sea, the tilting of the volcanic rocks and the associated faults of the Wadi Faynan give rise to springs, such as in Hammam Adethni. Other springs occur at the contact between limestone and sandstone in the mid-reaches of the Wadi Ghuwayr. Thus, of the rain falling on the plateau, some will be used to water crops, some evaporated, some will run off to the channel, but most importantly some will recharge the aquifers below and ultimately reappear as perennial spring flow further down the Wadi Faynan. Unbeknown to the earliest settlers, these were the causes of the availability of water and continued security of supply. The construction of the conceptual hydrological model and the application of the Pitman model demonstrated the inter-relationship between rainfall, vegetation and water availability in the Wadi Faynan. The results of the modelled scenarios show that increased rainfall can, in addition to providing more moisture for crop growth, result in larger, more frequent floods which may have an adverse impact on farming in a landscape denuded of vegetation. However, the presence of vegetation when coupled with higher rainfall is beneficial as more water is retained in the catchment soils and groundwater, and the spring flows are greater. A slightly wetter, but more densely vegetated Wadi Faynan would have been far more favourable to sedentary farming than the barren landscape present today. The results indicate that, in terms of water availability, changes in vegetation have the potential to have major effects on catchment hydrological functioning and should always be considered when assessing the impact of climate change in semi arid regions. At present vegetation is seen as responding to climate change, or used as a passive indicator of climate change, whereas the conclusion here is that vegetation plays an active role in defining the climate change impact. The groundwater was shown to provide a consistent supply of water despite changes in climate through the Holocene and this source has been exploited by the local inhabitants of the Wadi Faynan. The work presented provides a novel method for studying the interactions between climate and land cover at key times during the Late Pleistocene and Holocene. The modelling framework which chains global and regional palaeoclimate models, a weather-generator and a hydrological model provides a tool-kit with which to formalise the conceptual understanding of hydrological functioning at key times throughout the Holocene. Whilst there are uncertainties, mainly due to characterising variations in regional rainfall and catchment infiltration rates under differing

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

vegetation covers, the method formalises the assumptions allowing the key uncertainties to be identified. Perhaps most importantly, the toolkit allows a move away from the simple arguments that ‘when the climate is warm and moist society flourishes and when cold and dry, society declines’ to lines of inquiry that consider the subtle variations of regional rainfall, that ground cover may dampen or amplify the hydrological response to the prevailing climate and groundwater can buffer a site from changes in rainfall input. Acknowledgements The authors would like to thank the following for their contributions to this work: Anne Dudley and Dave Thornley for water sample analysis; Ian Thomas for the spatial analysis of the water quality data; Joshua Guest for the initial preparation of illustrations and hydrological model runs; Sameeh Nuimat and Khalil Jamjoum for provision of climate data for meteorological stations in Jordan; Ron Manley for provision of the HYSIMM model with which to estimate potential evaporation; Frank Farquharson and Helen Houghton-Carr from Centre of Ecology and Hydrology, Wallingford for access to hydrological records in the overseas archive; Nicola Flynn for compilation of infiltration rate estimates; Bill Finlayson and Nadja Qaisi from the Council for British Research in the Levant for assistance with the organisation of the three field seasons; and the inhabitants of the Wadi Faynan for their generosity of time and willingness to discuss the hydrology of their locality. References Abu-Awwad, A.M., Shatanawi, M.R., 1997. Water harvesting and infiltration in arid areas affected by surface crust: examples from Jordan. J. Arid Environ. 37, 443– 452. Agrar und Hydrotechnik GmbH, 1977. National Water Master Plan of Jordan. Hashemite Kingdom of Jordan. Natural Resources Authority, Amman. Al-Najjar, M., Abu-Dayya, A., Suleiman, E., Weisgerber, G., Hauptmann, A., 1990. Tell Wadi Faynan: the first Pottery Neolithic Tell in the South of Jordan. Annu. Depart. Antiquit. Jordan 34, 27–56. Alpert, P.I., Osetinsky, B.Z., Shafir, H., 2004. Semi-objective classification for daily synoptic systems: application to the eastern Mediterranean climate change. Int. J. Climatol. 24, 1001–1011. Al-Qawabah, M.S., Johnson, C., Ramez, H., Al-Fattah, M.Y.A., 2003. Assessment and evaluation methodologies report: Dana Biosphere Reserve, Jordan. Sustainable management of marginal drylands (Sumamad). In: Second International Workshop [on Combating Desertification: Sustainable Management of Marginal Drylands (SUMAMAD). UNESCO-MAB, Shiraz, Islamic Republic of Iran. Al-Qinna, M.I., Abu-Awwad, A.M., 1998. Infiltration rate measurements in arid soils with surface crust. Irrig. Sci. 18, 83–89. Al-Qinna, M.I., Abu-Awwad, A.M., 2001. Wetting patterns under trickle source in arid soils with surface crust. J. Agric. Eng. Res. 80, 301–305. Barker, G., Gilbertson, D., Mattingly, D., (Eds.), 2007a. Archaeology and Desertification: The Wadi Faynan Landscape Survey, Southern Jordan, Wadi Faynan Series 2, Levant Supplementary Series 6. Council for British Archaeology in the Levant/Oxbow Books, Oxford, 510pp. ISBN 1-84217-286-7. Barker, G., Adams, R., Creighton, O., et al., 2007b. Chalcolithic (c.5000–3600 cal.BC) and Bronze Age (c. 3600–1200 cal. BC) settlement in Wadi Faynan: metallurgy and social complexity. In: Archaeology and Desertification: The Wadi Faynan Landscape Survey, Southern Jordan, Wadi Faynan Series 2, Levant Supplementary Series 6. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 227–270. ISBN 1-84217-286-7. Bar-Matthews, M., Ayalon, A., Gilmour, M., Matthews, A., Hawkesworth, C.J., 2003. Sea–land oxygen isotopic relationships from planktonic foraminifera and speleothems in the Eastern Mediterranean region and their implications for paleorainfall during interglacial intervals. Geochim. Cosmochim. Acta 67, 3181– 3199. Beven, K., 2002. Runoff generation in semi-arid areas. In: Bull, L.J., Kirby, M.J. (Eds.), Dryland Rivers – Hydrology and Geomorphology of Semi-Arid Channels. Wiley, Chichester, pp. 57–105. Binnie and Partners (Overseas) Limited, 1979. Mujib and Southern Ghors Irrigation Project – Feasibility Report. Ministry of Planning, Jordan Valley Authority, Amman. Black, E.C.L., 2009. The impact of climate change on daily precipitation statistics in Jordan and Israel. Atmos. Sci. Lett. 10, 192–200. Black, E.C.L., 2011. Connecting climate and hydrological models for impacts studies. In: Mithen, S.J., Black, E.C.L. (Eds.), Water, Life and Civilisation: Climate, Environment and Society in the Jordan Valley. International Hydrology Series. Cambridge University Press, pp. 63–69, ISBN 9780521769570.

15

Black, E.C.L., Brayshaw, D.J., Slingo, J., Hoskins, B. 2011. Future climate of the Middle East. In: Mithen, S.J., Black, E.C.L. (Eds.), Water, Life and Civilisation: Climate, Environment and Society in the Jordan Valley. International Hydrology Series. Cambridge University Press, pp. 51–62, ISBN 9780521769570. Brayshaw, D.J., Rambeau, C.M.R., Smith, S.J., 2011a. Changes in the Mediterranean climate during the Holocene: insights from global and regional climate modeling. Holocene 21, 15–31. Brayshaw, D.J., Hoskins, B., Black, E., 2011b. Some physical drivers of changes in the winter storm tracks over the North Atlantic and Mediterranean during the Holocene. Philos. Trans. Roy. Soc. A 368, 5185–5223. Bull, L.J., Kirby, M.J., 2002. Dryland river characteristics and concepts. In: Bull, L.J., Kirby, M.J. (Eds.), Dryland Rivers: Hydrology and Geomorphology of Semi-Arid Channels. Wiley, Chichester, pp. 3–15. Byrd, B., 2005. Reassessing the emergence of village life in the near East. J. Archaeol. Res. 13 (3), 231–290. Crook, D., 2009. Hydrology of the combination irrigation system in the Wadi Faynan, Jordan. J. Archaeol. Sci. 36, 2427–2436. Darmame, K., Nortcliff, S., Potter, R.B., 2011. The role of water and land management policies in contemporary socio-economic development in Wadi Faynan. In: Mithen, S.J., Black, E.C.L. (Eds.), Water, Life and Civilisation. Cambridge University Press and UNESCO, Cambridge, p. 544, ISBN 9780521769570. Dunne, T., 1978. Field studies in hillslope flow processes. In: Kirby, M.J., Chorley, R.J. (Eds.), Hillslope Hydrology. Wiley, Chichester, pp. 227–293. Enzel, Y., Arnit, R., Dayan, U., et al., 2008. The climatic and physiographic controls of the eastern Mediterranean over the late Pleistocene climates in the southern Levant and its neighboring deserts. Global Planet. Change 60, 165–192. Finlayson, B.L., Mithen, S.J., 2007. Excavations at WF16. In: Finlayson, B.L., Mithen, S. (Eds.), The Early Prehistory of Wadi Faynan, Southern Jordan: Archaeological Survey of Wadis Faynan, Ghuwayr and al-Bustan and Evaluation of the PrePottery Neolithic A Site of WF16, Wadi Faynan Series 1, Levant Supplementary Series 4. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 145–202, ISBN 1-84217-212-3. Frumkin, A., Ford, D.C., Schwarcz, H.P., 1999. Continental oxygen isotopic record of the last 170,000 years in Jerusalem. Quaternary Res. 51, 317–327. Fujii, S., 2007. Wadi Badda: a PPNB settlement below Fjaje escarpment near Shawbak. Neolithics 1/07, 19–24. Hauptmann, A., 2007. The Archaeometallurgy of Copper: Evidence from Faynan. Springer, New York. Hazra Overseas Engineering Company, 1978. Jordan Valley Irrigation Project, Stage II Feasibility Study, vol. II.. Jordan Valley Authority, Hashemite Kingdom of Jordan. Hunt, C.O., Elrishi, H.A., Gilbertson, D.D., et al., 2004. Early-Holocene environments in the Wadi Faynan, Jordan. The Holocene 14, 921–930. Jordan Soils and Land Management, 2006. (retrieved July 2009). Lancaster, W., Lancaster, F., 1999. People, Land and Water in the Arab Middle East: Environments and Landscapes in the Bilad Ash-Sham. Harwood Academic Publishers, Amsterdam. Lancaster, N., Emberson, H., Mithen, S.J., 2007. The modern vegetation of Hammam Adethni and its palaeo-economic implications. In: Finlayson, B.L., Mithen, S. (Eds.), The Early Prehistory of Wadi Faynan, Southern Jordan: Archaeological Survey of Wadis Faynan, Ghuwayr and al-Bustan and Evaluation of the PrePottery Neolithic A Site of WF16, Wadi Faynan Series 1, Levant Supplementary Series 4. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 437–446, ISBN 1-84217-212-3. Levy, T.E., Adams, R.B., Najjar, M., et al., 2004. Reassessing the chronology of Biblical Edom: new excavations and C-14 dates from Khirbat en-Nahas (Jordan). Antiquity 78, 865–879. Maher, L.A., Banning, E.B., Chazan, M., 2011. Oasis or mirage? Assessing the role of abrupt climate change in the prehistory of the Southern Levant. Camb. Archaeol. J. 21 (1), 1–29. Mandal, U.K., Bhardwaj, A.K., Warrington, D.N., et al., 2008. Changes in soil hydraulic conductivity, runoff, and soil loss due to irrigation with different types of saline–sodic water. Geoderma 144, 509–516. Mattingly, D., Newson, P., Grattan, J., et al., 2007a. The making of early states: the Iron Age and Nabatean periods. In: Archaeology and Desertification: The Wadi Faynan Landscape Survey, Southern Jordan, Wadi Faynan Series 2, Levant Supplementary Series 6. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 271–303, ISBN 1-84217-286-7. Mattingly, D., Newson, P., Creighton, O., et al., 2007b. A landscape of imperial power: Roman and Byzantine Phaino. In: Archaeology and Desertification: The Wadi Faynan Landscape Survey, Southern Jordan, Wadi Faynan Series 2, Levant Supplementary Series 6. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 305–348. Mithen, S.J., Black E.C.L. (Eds.), 2011. Water, Life and Civilisation – Climate, Environment and Society in the Jordan Valley. International Hydrology Series. Cambridge University Press. ISBN 9780521769570. Mithen, S., Finlayson, B.L., Pirie, A., Smith, S., Whiting, C., 2007a. Archaeological survey of Wadis Faynan, Ghuwayr, Dana and al-Bustan. In: Finlayson, B.L., Mithen, S. (Eds.), The Early Prehistory of Wadi Faynan, Southern Jordan: Archaeological Survey of Wadis Faynan, Ghuwayr and al-Bustan and Evaluation of the Pre-Pottery Neolithic A Site of WF16, Wadi Faynan Series 1, Levant Supplementary Series 4. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 47–114, ISBN 1-84217-212-3. Mithen, S.J., Austen, P., Kennedy, A., et al., 2007b. Early Neolithic woodland composition and exploitation in the Southern Levant: a comparison between

16

A.J. Wade et al. / Journal of Hydrology 420–421 (2012) 1–16

archaeobotanical remains from WF16 and present-day woodland at Hammam Adethni. Environ. Archaeol. 12, 49–70. Mithen, S.J., Finlayson, B.L., Smith, S.J., Jenkins, E., Maricevic, D., 2011. An 11,600 year-old communal structure from the Neolithic of Southern Jordan. Antiquity 85, 1–15. Newson, P., Mattingly, D., Daly, P., et al., 2007. The Islamic and ottoman periods. In: Archaeology and Desertification: The Wadi Faynan Landscape Survey, Southern Jordan, Wadi Faynan Series 2, Levant Supplementary Series 6. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 349–368, ISBN 1-84217286-7. Palmer, C., Gilbertson, D., el-Rishi, H., et al., 2007. The Wadi Faynan today: landscape, environment, people. In: Indhu, K. (Ed.), Archaeology and Desertification: The Wadi Faynan Landscape Survey, Southern Jordan, Wadi Faynan Series 2, Levant Supplementary Series 6. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 25–57, ISBN 1-84217286-7. Pitman, W.V., 1973. A Mathematical Model for Generating River Flows from Meteorological Data in South Africa. Report 2/73: Hydrological Research Unit, University of the Witwatersrand. Johannesburg, South Africa. Puigdefàbregas, J., Sánchez, G., 1996. Geomorphological implications of vegetation patchiness on semi-arid slopes. In: Anderson, M.G., Brooks, S.M. (Eds.), Advances in Hillslope Processes. Wiley, Chichester, pp. 1027–1060. Rollefson, G., 1981. The Late Acheulean site at Fjaje, Wadi el-Bustan. Paleorient 7, 5– 21. Samuels, R., Rimmer, A., Alpert, P., 2009. Effect of extreme rainfall events on the water resources of the Jordan River. J. Hydrol. 375, 513–523. Selaolo, E.T., 1998. Tracer studies and groundwater recharge assessment in the eastern fringe of the Kalahari. Unpublished PhD thesis, Vrije Universiteit. Simmons, A.J., Najjar, M., 2000. Preliminary report of the 1999–2000 excavation season at the Pre-Pottery Neolithic settlement of Ghwair 1, Southern Jordan. Neolithics 1 (00), 6–8. Smith, S.J., Wade, A.J., Black, E.C.L., et al., 2011. From global climate change to local impact in Wadi Faynan, southern Jordan: ten millennia of human settlement in its hydrological context. In: Mithen, S.J., Black, E.C.L. (Eds.), Water, Life and Civilisation – Climate, Environment and Society in the Jordan Valley.

International Hydrology Series. Cambridge University Press, ISBN 9780521769570. Tarawneh, Q., Kadioglu, M., 2003. An analysis of precipitation climatology in Jordan. Theor. Appl. Climatol. 74, 123–136. Tipping, R., 2007. Long-term landscape evolution of the Wadis Dana Faynan and Ghuwayr. In: Finlayson, B.L., Mithen, S. (Eds.), The Early Prehistory of Wadi Faynan, Southern Jordan: Archaeological Survey of Wadis Faynan, Ghuwayr and al-Bustan and Evaluation of the Pre-Pottery Neolithic A Site of WF16, Wadi Faynan Series 1, Levant Supplementary Series 4. Council for British Archaeology in the Levant/Oxbow Books, Oxford, pp. 14–46, ISBN 1-84217-212-3. United States Department of Agriculture (USDA), 1999. Soil Taxonomy: A Basic System of Soil Classification for Making and Interpreting Soil Surveys. United States Government, Washington, DC. Wade, A.J., Holmes, P.A., el-Bastawesy, M., et al., 2011. The hydrology of the Wadi Faynan. In: Mithen, S.J., Black, E.C.L. (Eds.), Water, Life and Civilisation – Climate, Environment and Society in the Jordan Valley. International Hydrology Series. Cambridge University Press, ISBN 9780521769570. Wheater, H.S., 2002. Hydrological processes in arid and semi-arid areas. In: Wheater, H.S., Al-Weshah, R.A. (Eds.), Hydrology of Wadi Systems. UNESCO, Paris, pp. 5–22. White, K., 1995. Field techniques for estimating downstream changes in discharge of gravel-bedded ephemeral streams: a case study of southern Tunisia. J. Arid Environ. 30, 283–294. Williams, J., Bonell, M., 1988. The influence of scale of measurement on the spatial and temporal variability of the Philip infiltration parameters – an experimental study in an Australian savannah woodland. J. Hydrol. 104, 33–51. Yair, A., 1983. Hillslope hydrology water harvesting and areal distribution of some ancient agricultural systems in the northern Negev Desert. J. Arid Environ. 6, 283–301. Yair, A., 1987. Environmental effects of loess penetration into the northern Negev Desert. J. Arid Environ. 13, 9–24. Yair, A., Klein, M., 1973. The influence of surface properties on flow and erosion processes on debris covered slopes in an arid area. Catena 1, 1–18. Yair, A., Lavee, H., 1976. Runoff generative processes and runoff yield from arid talus mantled slopes. Earth Surf. Process. 1, 235–247.