Scholarly article on topic 'Controls on runoff generation along a steep climatic gradient in the Eastern Mediterranean'

Controls on runoff generation along a steep climatic gradient in the Eastern Mediterranean Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Keywords
{Mediterranean / "Catchment hydrology" / "Semi-arid areas" / "Climatic gradient" / "Physical catchment properties" / "Runoff generation processes"}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Fabian Ries, Sebastian Schmidt, Martin Sauter, Jens Lange

Abstract Study region Lower Jordan River. Study focus The main aim of this study was to identify differences in catchment runoff reactions across a variety of scales and a strong climatic gradient and to correlate them to physical catchment properties. For this purpose we observed rainfall and runoff responses on a hillslope (1000m2) and in several nested catchments (3.2–129km2) over a period of five years. Catchment characteristics and surface cover types were derived from high-resolution aerial images. To gain process understanding a single high magnitude event was analysed in detail using information from soil moisture plots. New hydrological insights for the region Our results show that runoff in the semi-arid headwater area is strongly related to long lasting rainfall events of high amounts and is predominantly generated by saturation excess overland flow (SOF). Observations from the arid runoff plot indicated a strongly contrasting behaviour with dominating Hortonian overland flow (HOF). At catchment scale we found an accentuated runoff response when we compared arid with semi-arid conditions, which can be attributed to different geological substrate, more abundant rock surfaces, shallower soil and sparser vegetation cover. Identified strong correlations between event rainfall and runoff volumes may provide promising options for the assessment and management of surface runoff as a water resource.

Academic research paper on topic "Controls on runoff generation along a steep climatic gradient in the Eastern Mediterranean"

ELSEVIER

Contents lists available at ScienceDirect

Journal of Hydrology: Regional Studies

journal homepage: www.elsevier.com/locate/ejrh

Controls on runoff generation along a steep climatic gradient in the Eastern Mediterranean

Fabian Riesa b *, Sebastian Schmidtb, Martin Sauterb, Jens Langea

a Chair of Hydrology, University of Freiburg, Fahnenbergplatz, 79098 Freiburg, Germany b Applied Geology, Geoscience Centre, University of Göttingen, Goldschmidtstrasse 3, 37077 Göttingen, Germany

ARTICLE INFO ABSTRACT

Study region: Lower Jordan River. Study focus: The main aim of this study was to identify differences in catchment runoff reactions across a variety of scales and a strong climatic gradient and to correlate them to physical catchment properties. For this purpose we observed rainfall and runoff responses on a hillslope (1000m2) and in several nested catchments (3.2-129km2) over a period of five years. Catchment characteristics and surface cover types were derived from highresolution aerial images. To gain process understanding a single high magnitude event was analysed in detail using information from soil moisture plots.

New hydrological insights for the region: Our results show that runoff in the semi-arid headwater area is strongly related to long lasting rainfall events of high amounts and is predominantly generated by saturation excess overland flow (SOF). Observations from the arid runoff plot indicated a strongly contrasting behaviour with dominating Hortonian overland flow (HOF). At catchment scale we found an accentuated runoff response when we compared arid with semi-arid conditions, which can be attributed to different geological substrate, more abundant rock surfaces, shallower soil and sparser vegetation cover. Identified strong correlations between event rainfall and runoff volumes may provide promising options for the assessment and management of surface runoff as a water resource.

© 2016 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

CrossMark

Article history:

Received 31 March 2016

Received in revised form 28 October 2016

Accepted 1 November 2016

Keywords:

Mediterranean

Catchment hydrology

Semi-arid areas

Climatic gradient

Physical catchment properties

Runoff generation processes

1. Introduction

In the Mediterranean region, runoff generation and its relationship to climate, rainfall, soil cover and landuse has considerable influence on phenomena such as soil erosion, land degradation, desertification and flooding that impose considerable problems on society (García-Ruiz, 2010; Hill et al., 2008; Martinez-Mena et al., 1998; Llasat et al., 2010). For that reason, it has gained increased attention in recent years (e.g. Cantón et al., 2011; Cammeraat et al., 2010; Ziadat and Taimeh, 2013; Marchamalo et al., 2016; Ochoa et al., 2016). Furthermore, in the light of water scarcity, surface runoff may be regarded as an additional water resource (Thornes and Wainwright, 2004; Kalogeropoulos and Chalkias, 2013), which is however not regularly used. Additionally, flow in ephemeral channels usually leads to transmission losses (e.g. Lange, 2005) i.e. focussed aquifer recharge, with implications for water resources management and protection. Semi-arid catchments often comprise a climatic transition from wetter headwaters to dryer downstream sections. In combination with different scales of observation (plot, hillslope and catchment scale) this poses challenges for hydrological process studies.

* Corresponding author at: Chair of Hydrology, University of Freiburg, Fahnenbergplatz, Freiburg 79098, Germany. E-mail address: fabian.ries@hydrology.uni-freiburg.de (F. Ries).

http://dx.doi.org/10.1016/j.ejrh.2016.11.001

2214-5818/© 2016 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

On the plot scale, overland flow is first triggered on isolated rock outcrops or soil surfaces with limited infiltration capacities due to e.g. crusts, from where it contributes to saturation of adjacent soil pockets (Cantón et al., 2002; Lange et al., 2003; Sohrt et al., 2014). Once soil patches are saturated or local infiltration rates exceeded, surface runoff can continue downslope. A major impact on hydrological response is often assigned to the presence or absence of vegetation cover (Cantón et al., 2011), which has led to the concept of vegetation-driven spatial heterogeneity (Puigdefábregas, 2005). Vegetation is usually reported to enhance infiltration (e.g. Cerda, 1997; Quinton et al., 1997; Cantón et al., 2011), while in some rare cases, even the opposite may be true due to the reduction in infiltration as a result of the presence of organic litter and its hydrophobic effect (e.g. Nicolau et al., 1996). Further, the spatial distribution of vegetation patches, which is often associated with livestock grazing (Kropfl et al., 2013), influences runoff generation (Calvo-Cases et al., 2003; Puigdefábregas, 2005). With certain patterns overland flow resistance may be reduced and thus flow velocity increased. With increasing aridity, vegetation and other biotic factors become less and abiotic factors more important for infiltration and runoff behaviour (Lavee et al., 1998; Ruiz-Sinoga et al., 2011). At the same time runoff generation gradually changes from saturation excess (SOF) to Hortonian overland flow (HOF) as shown by Cerda (1997,1998).

On the hillslope and sub-catchment scale, physiographic factors such as the connectivity of runoff generating areas, as well as rainfall characteristics are relevant for runoff generation. Runoff coefficients were observed to drop with increasing slope length mainly due to run-on effects and increased infiltration losses (Puigdefábregas et al., 1998). Hillslope scale runoff generation and amounts were found to be mainly controlled either by rainfall intensity, rainfall depth and antecedent precipitation (Li et al., 2011), or by surface properties (Yair and Kossovski, 2002; Arnau-Rosalén et al., 2008). Puigdefábregas et al. (1998) described two types of runoff generation, HOF at the beginning of short intensive storms and SOF by saturation of shallow soil patches or locations with decreasing permeability with depth during long-duration rainfall events with high amounts.

On the catchment scale, stream runoff is composed of contributions from sub-catchments and single hillslopes, sometimes with considerable variations in climatic conditions and surface properties, such as soil type, thickness and bedrock permeability (Ries et al., 2015). Yair and Raz-Yassif (2004) found that ephemeral streams in arid regions receive most runoff from headwater areas. Generally, runoff coefficients decrease with increasing catchment size (e.g. Goodrich et al., 1997; Yair and Raz-Yassif, 2004). This is partly due to the limited size of storm cells, runoff concentration time, differences in litholog-ical composition, slope length and morphology as well as transmission losses in tributaries and the main channel alluvium (e.g. Kirkby et al., 2002; Yair and Kossovsky, 2002; Lange, 2005; Zanon et al., 2010). Also human caused transformation of the landscape such as urbanization (Grodek et al., 2011; Perrin and Tournoud, 2009; Lange et al., 2001) or land cultivation (Koulouri and Giourga, 2007; Kosmas et al., 1997; García-Ruiz, 2010) might play an important role on the catchment runoff response.

At all spatial scales in arid and semi-arid areas, runoff generation is characterized by high variability in time and space. Despite considerable advances in the understanding of runoff generation on the plot and hillslope scale, there is still a lack of knowledge on how small scale runoff patterns are reflected in catchment scale runoff response.

Our aim was to identify differences of dominant runoff generation processes within and between several carbonate rock catchments along a strong climatic gradient in the Eastern Mediterranean. For this purpose, we installed a dense monitoring network in hitherto basically ungauged basins (three major ephemeral streams including four headwater basins) to measure rainfall, wadi runoff, hillslope overland flow and soil moisture in high temporal resolution. Runoff events covered a period of five years and were correlated with amounts, intensities as well as spatial and temporal distribution of rainfall. The runoff response was further related to soil, rock lithology, land use and surface cover. This analysis provides new insights into runoff generation across different spatial scales and sub-climate types.

2. Study area

The study catchments are located in the central and eastern parts of the West Bank, northwest of the Dead Sea and east of Jerusalem (Fig. 1). Deeply incised ephemeral streams (Wadis) drain the area and discharge into the endorheic Lower Jordan River/Dead Sea basin. Catchment elevations range from 1016m a.s.l. to 240 m b.s.l. Gauged catchment sizes for the main watersheds (Wadi Auja, Wadi Nueima and Wadi Qilt) vary between 54.8 and 129 km2 and between 3.2 and 14.5 km2 for individually gauged headwaters of Wadi Auja (H1-H4).

Precipitation shows a pronounced seasonality with a rainfall period from October to April. During this period, most rainfall occurs within few intense storm events originating from large-scale low-pressure systems (mainly Cyprus-Lows) over the Eastern Mediterranean Sea (Goldreich, 2003). These storms usually last for several days, cover large areas and provide high overall rainfall amounts of mainly low intensity. Nevertheless, they can be accompanied by localized high-intensity rainfall for short periods of time (Dayan and Morin, 2006). Convective rainfall events with high precipitation intensity but relatively low depths occur in spring and autumn and are often associated with air masses originating from the south (Active Red Sea Troughs). Seasonal rainfall amounts are highest in the elevated mountain range (semi-arid conditions) and decrease rapidly towards the Jordan Valley (arid conditions) eastward due to the topographic gradient (rain-shadow desert). Mean annual long-term precipitation is 532 mm in Jerusalem (810 m a.s.l.) and only 156 mm in Jericho (290 m b.s.l.) (Morin et al., 2009), with a high interannual variability. Potential evapotranspiration largely exceeds seasonal precipitation with annual values of 1350 mm in the mountains and up to 1650 mm in the Jordan Valley (Israel Meteorological Service - http://www.ims.gov.il).

Fig. 1. Study area with installed observation network and geological characteristics. Isohyets display long-term average annual rainfall (20 years) according to data from ANTEA (1998). Coordinates in the detailed map are in Palestinian Grid format.

Fractured layers of well bedded to massive limestone and dolomite from the Upper Cretaceous with minor intercalations of marl and chalk (Begin, 1975) dominate the geology of the mountain area and constitute the main regional aquifers. In the southeast, the limestone and dolomite rocks are covered with Senonian chalks of low hydraulic conductivity (Rofe and Raffety, 1963). Quaternary sediments (alluvium and Lisan marls) form the planes of the Jordan Valley.

Soils developed from residual clay minerals of weathered carbonate rocks and from aeolian input of dust from the Sahara (Yaalon, 1997). The most abundant soil types are Terra Rossa and Brown Rendzina, characterized by high clay contents. In general, Rendzina soils are thin and still developing, while Terra Rossa soils were formed under more humid climatic conditions in the past (Shapiro, 2006). Difference in dissolubility and weathering of the underlying bedrock result in a heterogeneous and patchy terrain where massive bedrock exposures constantly alternate with lose rock fragments and soil pockets of different dimensions, shapes and depths. Soil depth measurements on two hillslope catenas in the headwater areas of Wadi Auja (Sohrt et al., 2014) showed a high variability of soil thickness and mean depths of 26-39 cm at the two investigated hillslopes with few locations with depths of more than 1 m. In chalk areas on sloped terrain, soils tend to be thin as soil accumulation is strongly limited due to strong erosion (Singer, 2007). In valley planes alluvial soils developed with depths up to several meters. Towards the more arid zones in the east and especially in the southeast on chalk bedrock, soils gradually merge into Brown Rendzina and silty, brown, calcareous lithosols (Dan and Koyumdjisky, 1963). The presence of soil crusts were reported for the more arid part of the study area (Kutiel et al., 1998). Land cultivation, deforestation, terracing and other human activities have intensively transformed soils (Yaalon, 1997). Annual plants and Mediterranean shrubs are the dominant vegetation types growing more and more sparsely towards the arid zones in the east (Fig. 2). In the upper mountain area, olives grow on terraced slopes and valley planes are used for rainfed or irrigated agriculture. Densely populated areas of Jerusalem and Ramallah are located at the south-western watershed divide of Wadi Qilt and Wadi Nueima, while villages characterized by scattered built-up areas, prevail in the northern part (Wadi Auja). Populated areas in the arid region of the catchments are negligible.

3. Methods

3.1. Hydrometeorological observations, soil moisture measurements and soil texture characterization

A rain gauge network with 15 tipping buckets (HOBO RG3-M, Onset Computer Cooperation) connected to data loggers (HOBO Pendant Event Data Logger UA-003-64, Onset Computer Corporation) with a resolution of 0.2 mm was installed from 2007 on. At two high altitude stations additional heated tipping bucket precipitation sensors were installed (Thies clima, data logger: DLx-MET, resolution: 0.1 mm) in order to accurately measure snow events. All gauges were calibrated prior to employment according to the manufacturers specifications, maintained and cleaned twice a year. Shorter gaps in the time

Fig. 2. Landscape transformation along the strong precipitation gradient from west to east. The images display: a) Olive terraces on rocky terrain close to Kafr Malek in the headwaters area (Auja headwater 3); b) The valley plane of Ein Samia in the centre of the Wadi Auja catchment; c) The desert "badland"-type landscape on Senonian Chalks in the lower Wadi Qilt area. Locations of the images are displayed in Fig. 1.

series of the single stations were filled by linear correlation with neighbouring stations. To improve the spatial coverage of our network we additionally included rainfall data from the Israel Meteorological Service database (http://www.data.gov.il/ims).

To calculate catchment rainfall, point precipitation data was first transferred to the same topographic elevation using an elevation-rainfall gradient calculated on a daily time step. Then, point values were interpolated by inverse distance weighing (IDW). Finally, interpolated rainfall was extrapolated to the original topographic altitude by the same gradient and a digital elevation model. Within our study period, three major snowstorms occurred, which necessitated a concentration on stations where precipitation occurred as liquid or on stations, which were equipped with a working heating system. For the heaviest snowfall event in December 2013 we could determine the snow line at an altitude of 650 m a.s.l. from MODIS surface reflectance scenes (LP DAAC, 2013) at December, 15 2013, the day after the main snow fall event. Snowfall was excluded from event precipitation, as runoff data showed that subsequent snowmelt occurred over several days and did not contribute to surface runoff generation.

In order to be able to separate the rainfall time series consistently into individual events, we defined that two rainfall events are independent from each other if the period between two rainfall spells exceeds 24 h. Soil moisture was observed at several depths at four different locations (SM-1-SM-4) with a thickness of the soil layer ranging between 45 and 100 cm. More details on soil moisture probe installation can be found in Ries et al. (2015).

We collected undisturbed soil samples at 29 locations covering mainly Wadi Auja and the entire climatic gradient, which were analysed in the lab for soil texture. The soil samples were grouped into a high (>600 m a.s.l.), medium (300-600 m a.s.l.) and low (<300 m a.s.l.) elevation band.

3.2. Measurement of ephemeral stream discharge

We gauged runoff in the three major Wadis and four subcatchments by means of concrete flumes, weirs or at road culverts in order to guarantee a stable rating curve (Fig. 1). Water level and temperature were measured with automatic pressure transducers (Micro-Diver D1601, Eijkelkamp; HOBO U20-001-01, Onset Computer Corporation; Dipper-3, SEBA) at a temporal resolution of five minutes. Water level data was subsequently corrected for barometric pressure fluctuations (Baro-Diver D1601, Eijkelkamp) and then converted into discharge either by hydraulic equations (two flumes and two weirs) or by the Manning-Strickler formula at three road culverts. Where spring flow initiated during major rainfall events and discharged into the Wadi, surface runoff was separated from karst spring discharge using either the measured temperature difference between spring water and surface runoff (Wadi Qilt, H1) or by abstracting measured spring discharge from total discharge in the ephemeral stream (Wadi Auja, Schmidt et al., 2014).

Fig. 3. Runoff plot (1000 m2) close to the southeast catchment divide ofWadi Auja (a); Detail of the runoff collector tank with the attached tipping bucket (b); Soil plot close to the runoff plot with surface crust with a thickness of approximately 2 cm (c).

A runoff plot with an approximate size of 1000 m2 (Fig. 3) was installed close to the southeastern catchment boundary ofWadi Auja (Fig. 1) at the transition zone of limestone and chalk outcrops. Surface runoff from the plot was diverged by a low v-shaped wall into a metal tank, which served as a trap for course sediments. Runoff was measured using a self-made stainless-steel tipping-bucket with a volume of 5.61 per tip attached to the tank outlet. Tips were continuously recorded with an event data logger (HOBO Pendant Event Data Logger UA-003-64, Onset Computer Corporation). Data from the runoff plot was only included into the analysis of the January 2013 event, due to larger gaps within the remaining observation period.

3.3. Analysis of runoff event characteristics

We compared signatures of all runoff events for the entire 5-year observation period (01.11.2010-30.09.2015). Small events with a runoff below 1 mm were disregarded because they were considered to originate from local rainfall events. For the remaining events, runoff volumes were correlated with rainfall amounts, antecedent precipitation, rainfall intensities and thresholds for runoff initiation. A runoff event on January 2013 was selected for more detailed analysis. It consisted of several high intensity rainfall spells over a six-day period initiating the largest runoff event within the observation period at Wadi Qilt and Wadi Nueima.

3.4. Delineation of catchment characteristics

Physiographic catchment properties were delineated from a digital elevation model with a cell size of 25 m. Geological characteristics were allocated using a 1:200,000 geological map (Sneh et al., 1998) and main landuse types were digitized from aerial images. To determine the spatial distribution of the highly variable patchy surface cover we applied supervised classification on orthophotos with a pixel size of 0.5 x 0.5 m following a procedure from Ghosh et al. (2013).

First, surface classes at 1500 randomly sampled points were manually identified as bare soil, rock outcrop (including built-up areas) and vegetation to be used as a training sample for the succeeding steps of the supervised classification. RGB pixel values for the single classes were then separated using support vector machines, and bootstrapping was applied to finally carry out the actual classification procedure. Identified rock outcrop areas, which primarily included also urban areas were separated using landuse information. Fig. 4 shows examples of aerial images and the correspondent classified surface cover types from two locations within Wadi Auja.

4. Results

4.1. Catchment characteristics

The largest proportion of the study area can be classified as bare soil, rock outcrop or vegetation, only 2% had to be excluded due to shading by e.g. steep cliffs. Bright soil colour in chalk areas reduced the reliability of the classification scheme for Wadi Qilt and Wadi Nueima and classification was therefore only applied for the upper catchment areas with limestone and dolomite. Details on the proportions of surface cover type and other catchment characteristics for the single

Fig. 4. Aerial image and surface cover derived from supervised classification for two locations within Wadi Auja study area with a cell size of 0.5 x 0.5 m a) Village of Kafr Malik at the watershed divide of Auja headwater I; b) Valley plane of Ein Samia at the confluence of Auja headwater I and headwater II.

catchments are given in Table 1. While the share of the combined rock outcrop/build-up areas remained similar between the single Wadi Auja headwater areas, vegetation cover decreased significantly from the headwaters down to the more arid parts, thus reflecting the pronounced climatic gradient. Vegetation cover was densest where climatic conditions favoured the plantation of olive orchards on terraced slopes within the headwater area, on the small valley planes cultivated with annual crops and on north-facing slopes within the more arid downstream parts. Rock outcrops were more or less equally distributed over the entire area and their scattered appearance was connected to geology and topographic characteristics such as slope inclination. Only Wadi Qilt showed a considerable degree of urbanization with built-up areas, which were concentrated close to the uppermost drainage divide at the outskirts of northeast Jerusalem and Ramallah/El-Bireh.

Soil textural analysis showed a reduction of clay (43-31%) and silt (41-34%) content from the upper to the lower parts of Wadi Auja and an increase of sand (16-35%). In general, soil textures showed relatively low variation with soil depth, however, a slight increase in clay content with depth was recognizable.

4.2. Hydrometeorological conditions

At Jerusalem station, mean seasonal rainfall of 495 mm for the observation period was slightly below the long-term average of 535 mm (1981-2010; IMS, 2015) with strong annual variations between 324 mm and 630 mm (Fig. 5). In our data more than 60 % of the overall rainfall occurred in the months January to March. Seasonal rainfall decreased slightly from north to south and much more pronounced from west to east, dropping down to 85 mm in the Jordan valley.

A total of 58 rainfall events with rainfall sums greater than 5 mm were identified for Jerusalem and between 55 and 64 for the individual catchments. Further event rainfall characteristics are given in Table 1. Rainfall exceeded an intensity of

Table 1

Catchment characteristics for Wadi Qilt, Wadi Nueima, Wadi Auja and associated sub-catchments. Numbers in the table sections for physiographic characteristics represent mean (minimum/maximum) values.

Qilt Nueima Auja Headwater 1 Headwater 2 Headwater 3 Headwater 4

Physiography1

Area (km2) 129.2 85.9 54.8 14.5 8.4 4.2 3.2

Elevation (m) 496 (-240/911) 492 (-190/1010) 594 (-11/1010) 780 (434/1010) 603 (419/730) 892 (670/1010) 803 (650/951

Slope angle (◦) 12.4 (0/62.1) 12.1 (0/62.5) 14.0 (0/49.1) 11.8 (0/46.8) 12.0 (0/34.3) 9.8 (0/32) 11.6 (0/36.7)

Slope length (m) 210(0/1515) 211 (0/1285) 252 (0/1615) 255 (0/1086) 225 (0/831) 280 (0/924) 277 (0/1025)

Drainage density (km/km)2 1.84 1.85 1.82 1.52 1.56 1.42 1.37

Hydrology3

Mean seasonal rainfall (mm) 336 314 379 472 367 491 493

Number of rainfall events >5 mm 57 56 57 64 55 64 64

Maximum daily rainfall (mm) 58 56 76 89 80 90 95

Maximum event rainfall (mm) 142 133 167 202 170 200 209

Share of 10 largest events on total (%) 49 51 54 52 55 51 50

Measurement structure parshall flume road culvert round-crested weir V-notch weir road culvert H-flume road culvert

Number of observed runoff events 15 20 5 7 8 6 5

Mean specific discharge (l/s/km2)4 46.2 26.6 18.9 19.8 13.3 22.5 25.1

Max peak discharge (m3/s)5 42.3 28.5 9.7 2.1 0.72 0.8 0.96

Max specific discharge (l/s/km2)5 327 332 177 145 86 190 300

Max event runoff coefficient (%)5 10.5 12.6 4.2 4.5 4.1 5.4 6.7

Mean annual runoff (mm) 8.1 8.5 2.8 4.7 2.9 3.9 4.4

Mean annual runoff volume (103 m3) 1047 730 153 68 24 16 14

Total runoff coefficient (%) 2.4 2.7 0.7 1.0 0.8 0.8 0.9

Event January 2013

Rainfall sums (mm) 129 120 157 196 174 213 225

Max point rainfall intensity (mm/h) 21,7 19,8 26,0 23,7 22,0 24,3 26,0

Event runoff (mm) 13,5 15,1 4,6 7,3 4,4 6,6 9,5

Event runoff coefficient (%) 10.5 12.6 2.9 3.7 2.5 3.1 4.2

Geology (%)6

Limestone and dolomite 56 62 100 100 100 100 100

Chalk 44 35 0 0 0 0 0

Alluvial sediments 0 3 0 0 0 0 0

Land use (%)7

Stone quarry 1 0 1 1 3 0 0

Olives terraces 5 11 10 30 2 17 54

Arable land 4 8 10 8 13 3 2

Build-up area 15 3 4 11 5 14 15

Bare desert soil 44 44 20 0 0 0 0

Shrubland 31 34 55 52 77 65 29

Surface cover (%)8

Soil 33 50 57 36 65 30 34

Rock outcrops 22 21 16 11 11 8 5

Urban areas 23 5 4 10 6 14 14

Vegetation 21 24 22 41 14 47 48

Undefined 1 0 1 1 1 1 2

1 Physiographic characteristics are derived from a 25 x 25 m digital elevation model.

2 Drainage densities are based on a 25 x 25 m flow accumulation grid and a channel initiation threshold of 0.1 km2 contributing area.

3 Rainfall and runoff data are related to the period November 2010-September 2015.

4 Mean specific discharge is calculated forthe period November 2010-September 2014 where observations cover all catchments.

5 Maximum values of peak discharge, specific discharge and event runoff coefficients correspond to two events: event E1 in March 2012 (Auja and Headwaters) and E2 in January 2013 (Qilt and Nueima) where observations cover all catchments.

6 Shares based on geological maps at scale 1:200,000 (Sneh et al., 1998).

7 Landuse was manually digitalized from aerial images and describe larger connected units of common landuse characteristics.

8 Surface cover types for Wadi Qilt and Wadi Nueima were only classified for areas with limestone and dolomite (excluding chalk). In difference the assessment of land use, small-scale differences are included and proportions cannot be simply compared to land use cover.

60mm/h only occasionally. Differences between the catchments in terms of maximum rainfall depth as well as intensities were attributed to the small spatial extension of single rain bursts during large-scale frontal rainfall events, which cover small headwaters with higher probability than the main catchments that also include the arid zones of the Jordan Valley.

4.3. Runoff events

Runoff observations at the Wadis Qilt, Nueima, Auja H1 and H3 were available for the entire study period, while observations for Wadi Auja ended in September 2014 and for the stations Auja H2 and H4 in March 2014, respectively. Spring discharge contributions reaching maximum values between 0.3 (Ein Samia) and 2 m3/s (Wadi Qilt springs) during single events, were separated from surface runoff. General runoff characteristics are summarized in Table 1 and an overview of the complete runoff time series of all catchments is given in Fig. 5. Runoff events were only observed between October and April. Approximately 84% of overall runoff occurred within the months of January to March receiving 62% of seasonal rainfall, while only 2% of runoff was observed within the months April and October. Apart from channel sections downstream of

Fig. 5. Overview on rainfall and runoff events during the monitoring period 2010-2015. Note the high interannual rainfall variability and the restriction of runoff reactions to Wadi Qilt and Wadi Nueima during the convective rainfall events LE1 and LE2.

partly intermittent karst springs in Wadi Qilt and Wadi Auja, ephemeral streams were dry for most of the year. Stream flow was restricted to a few hours for small, often convective and localized rainfall events, and up to 8 days for long lasting and large-scale events. Highest frequency of runoff events was observed in the main streams of Wadi Qilt and Wadi Nueima and considerably less in Wadi Auja and the Auja subcatchments (Table 1). Over the entire observation period seven major events were identified (E1-E7 in Fig. 5), causing runoff in all observed streams. The corresponding event rainfall at Jerusalem station was between 96 and 221 mm. Three events in late spring and late summer could clearly be identified as being generated by localized, convective rainfall events (LE1-LE3 in Fig. 5).

Fig. 6. Comparison of ephemeral stream discharge during the event in January 2013 in the main catchments (a) and the headwaters of Wadi Auja (c); surface near soil moisture patterns at four locations (b), vertical soil moisture observations at plot SM-1 (d) and runoff patterns at the hillslope runoff plot and observed rainfall close to the plot location (e). Hourly and cumulative rainfall represents catchment and plot rainfall values, respectively. The locations of runoff gauges and soil moisture plots are shown in Fig. 1.

4.4. The high magnitude runoff event of January 2013

Based on the rainfall and runoff distribution, the runoff event of January 2013 could be separated into six sections (I-VI) (Fig. 6a-e). Prior to the event, no rainfall occurred during a period of about 2 weeks, soil moisture remained below field capacity and streambeds were completely dry. Total catchment rainfall of the event was between 120 and 225 mm during a period of 1 week. For most of the time rainfall intensities were moderate but reached single peaks of up to 26 mm/h (Table 1).

During section I, 26 mm catchment rainfall was calculated for the main catchments and 38 mm for the headwater area, but no runoff was generated and near-surface soil moisture displayed only a slight increase (Fig. 6b). Instead, SM-1 indicated

fast infiltration and percolation to the bottom of the soil profile (80 cm) and a gradual saturation from the bottom to the top. Runoff at Wadi Qilt and Nueima was initiated in event section II following a single peak of high rainfall intensity contributing around 20 mm. No runoff occurred in Wadi Auja and its headwaters. Soil moisture reactions were similar to the previous period but remained on higher moisture levels, especially at greater depths of SM-1.

Event section III-V received the largest amount of the rainfall with cumulative sums of 93 mm in the main catchments and 126 mm in the headwaters. Wadi Nueima reacted first to the new rainfall spell, followed by Wadi Qilt and, with greater lag time, by Wadi Auja and its headwaters. While peak flow rates decreased for Wadi Qilt and Nueima from section III-V, an increase was observed for Wadi Auja and its headwaters. Soil moisture data suggested soil saturation at 10 cm depth at plots SM-3 and SM-4 and between 40 and 80 cm at SM-1. The hydrograph of all streams clearly reflects soil moisture patterns at plot SM-1 at 40 cm depth. During the final event section VI, a last minor rainfall peak generated a small runoff peak in Wadi Qilt and Nueima but no response in the remaining catchments.

In general, Wadi Qilt, Nueima and Auja H2 showed an earlier reaction of stream runoff to rainfall intensity peaks than Wadi Auja and H1 and H3. From all headwaters, H4 was the last in reacting but showed the most pronounced rise in the hydrograph. From the main catchments, Wadi Auja had lowest runoff and also showed a slower reaction than its headwaters. Event runoff coefficients reached values between 2.5 and 12.6%.

A divergent pattern was found for the plot scale (1000 m2) in the arid downstream region of the study area (Fig. 6e).The hydrograph in Fig. 6e shows six isolated short runoff peaks, each lasting for only 5-30 min and occurring two hours before the runoff peaks of Wadi Qilt and Nueima. A first small runoff peak from the plot was already measured after 3.5 mm of rainfall. Yet, total runoff summed up to only 2.1 mm yielding an event runoff coefficient of 2.7% with peak runoff rates of 5.5 mm/5 min.

4.5. Catchment runoff responses

Overall runoff quantity was low and reflected the accentuated interannual rainfall variability. Seasonal runoff coefficients ranged from practically no flow in all catchments in 2010/2011 up to 6.4% in Wadi Nueima in 2012/2013 and up to 4 % in H1-H4 in 2011/2012. Runoff coefficients for Wadi Qilt and Wadi Nueima over the entire observation period were 2.4 and 2.7% respectively, but could reach up to 12.6% for single events. In contrast, runoff coefficients for Wadi Auja and the subcatchments were notably lower and reached only half the values of Wadi Qilt and Nueima (Table 1). Substantial differences between these two groups were further evident in maximum specific discharge, which were twice in Wadi Qilt and Nueima compared to the rest except for H-4.

Most events were characterized by a considerable time lag between the start of the rainfall event and the initiation of runoff in ephemeral streams. Only very few events that made up less than 5% of the total runoff volume showed a distinct, rapid runoff response, which was characterized by fast rising hydrograph limbs (from zero up to peak flow in less than 30 min), and a short duration of a few hours only. None of the headwaters did respond during these short, convective rainfall events. Another example of such localized high intensity rainfall is event LE3 in May 2014 (Fig. 5). Although none of our 15 rainfall stations recorded any rainfall, runoff was measured in all three main catchments.

Fig. 7 correlates event runoff with different rainfall parameters. Events below 1 mm made up less than 10% of the total runoff during the 5-year observation period and were excluded from the analysis. An apparent threshold level of 50 mm event rainfall was found for all catchments. However, this threshold (intersection of the regression line with the x-axis) is not the true event rainfall threshold for the initiation of stream flow, as the corresponding event rainfall is a cumulative value that also includes precipitation after the onset of runoff.

Significant correlations were found between event precipitation and runoff (r2 = 0.62-0.96; p< 0.01) and between mean rainfall intensity and runoff (r2 = 0.55-0.77; p<0.05). Correlations between runoff and 14day antecedent precipitation (API14) were less significant (r2 = 0.22-0.45; p>0.2). No correlation could be established for APIs of 5, 7 or 30 days. The slopes of the regression lines indicated a pronounced similarity between Wadi Qilt and Nueima, while Wadi Auja and its headwaters formed a second group.

5. Discussion

5.1. Interannual rainfall and runoff variability

An analysis of the rainfall time series revealed a high degree of interannual rainfall variability, which is typical for most of the Mediterranean (Dunkeloh and Jacobeit, 2003). However, precipitation sums were poorly reflected in seasonal runoff generation. During the year 2012/2013, rainfall was in the midrange of the five observed seasons, but overall runoff amounts and coefficients were highest. This suggests that the magnitude of runoff generation follows the characteristics of single rainfall events rather than seasonal rainfall amounts. In our area the largest fraction of runoff was generated during long-duration events with high rainfall amounts during the mid-winter season, which gives a first indication that runoff generation in our study area is predominantly controlled by saturation excess overland flow (SOF) rather than by Hortonian overland flow (HOF). The three main catchments showed differences in event frequency, average and maximum runoff coefficients as well as in their sensitivity to short convective rainfall events. These differences are a result of variable climatic attributes (e.g. aridity), physical catchment properties (e.g. soil water storage capacity) and dominating runoff generation processes. Our

Fig. 7. Correlation between event runoff heights and a) event rainfall, b) 14-day antecedent rainfall and c) rainfall intensity. The darker background area indicates a rainfall threshold of 1 mm. Runoff events in overall Wadi Auja are plotted in the same graph of its headwaters but separately marked in green. All linear correlations refer to runoff events >1 mm.

overall runoff coefficients of around 2% are common in semi-arid environments and similar values were already observed in the area (e.g. Wittenberg et al., 2004; Gunkel et al., 2015). The low runoff coefficients reflect the generally high infiltration rates of Mediterranean soils and the dominance of evapotranspiration during most of the year. A seasonal difference of runoff coefficients could not be shown by this study as demonstrated by e.g. Cerda (1996,1997) for the semi-arid Spain due to the limited length of the observation period and superimposing effects of runoff influencing factors on the catchment scale.

5.2. Soil moisture and runoff generation processes during a high magnitude event

In Wadi Auja, our soil moisture data suggested that the subsurface wetting front propagation was limited by a low bedrock permeability, which might be surprising given the dominance of karstified limestone and dolomite in the headwater areas (Rofe and Raffety, 1963; Begin, 1975; Hughes et al., 2008; Mimi and Assi, 2009). Furthermore, we found no evidence for reduced infiltration rates at the surface, since saturated conditions developed first at the bottom of the soil profile and then moved upwards but never appeared at the surface alone. This suggests SOF as a dominant runoff generation process, which was confirmed by artificial sprinkling experiments, where high infiltration capacities and preferential flow paths within the soil could even compensate for additional surface runoff from adjacent rock outcrops (Sohrt et al., 2014). Also in the main Wadi stream flow was only initiated after a larger amount of cumulative rainfall.

In the course of the event discharge peaks Wadi Nueima and Qilt decreased from section III-V, while Wadi Auja showed the opposite (Fig. 7a). This suggests that HOF was an important runoff contributor and that already during the first peak cumulative rainfall was high enough to saturate the main SOF contributing areas in Wadi Nueima/Qilt, while in Wadi Auja saturation was more gradual. The distribution of soil depths and bedrock percolation rates at locations with SOF dominated

runoff response might therefore play a dominant role for runoff magnitude and the shape of the hydrograph. SOF-dominated runoff response was reported by various Mediterranean studies (e.g. Latron et al., 2009; Schnabel and Gutiérrez, 2013), but so far only very few studies synchronized catchment scale runoff behaviour with measured soil moisture data.

At our runoff plot, located in the lowermost part of Wadi Auja-, the sharp isolated runoff peaks suggested that HOF is the dominating runoff generation process. The accumulated event rainfall before runoff initiation (a few mm only) was not enough to saturate the unconsolidated topsoil. Instead we assume that the crusty surface (Fig. 3c) acted as a sealing top layer impeding infiltration. Comparably, very small rainfall thresholds for runoff initiation on the plot and hillslope scale were found in other studies in similar environments (e.g. Greenbaum, 1986; Lavee et al., 1998) and also in relation with surface crusts (Chamizo et al., 2012). However, observations from the runoff plot cannot be translated into catchment scale runoff patterns without detailed knowledge on the spatial distribution of soil properties within the single catchments. Information that is largely missing for the study area.

Based on soil moisture data, observed rainfall thresholds for runoff initiation and on hydrograph shapes, we can conclude that both SOF and HOF occur in our study area. Hydrographs at the catchment outlets in the Jordan Valley consist of superimposed signals from different areas along the climatic gradient, where dominating runoff generation processes gradually change from SOF to HOF. Thus, our investigations corroborate the observations made by Lavee et al. (1998) and Imeson et al. (1998) on the transition of runoff generation processes along climatic gradients in semi-arid areas. However, most of runoff events are generated by SOF mechanisms.

Regardless of the observed differences in catchment reactions, a common characteristic shared by all three basins is the considerable time lag between the start of a rainfall event and the initial detection of runoff at the gauging station. This is likely related to the fact that single runoff generating areas require rainfall events to last a certain amount of time to become connected across spatial scales, a well-known issue in (semi-)arid areas (Puigdefábregas et al., 1998; Lange et al., 2003; Marchamalo et al., 2016).

5.3. Runoff in relation to rainfall and catchment characteristics

Antecedent precipitation indices calculated for the 2-week period preceding the single events showed positive although relatively weak correlations with event runoff. This finding may be attributed to the typical Mediterranean rainfall regime with prominent dry spells between two rainfall events (e.g. Martin-Vide and Gomez, 1999; Saaroni et al., 2014) and generally high evapotranspiration rates in semi-arid areas. When we used shorter (5, 7 days) or longer (30 days), we could not find any correlation. This indicates that in our area the probability of large rainfall amounts less than two weeks prior to an event is decreasing, while at the same time rainfall that fell a month before an event has little influence on runoff generation, because most water had already drained or evaporated. This ambiguous correlation of antecedent precipitation may also be one reason for the contradictory findings of hydrological process studies in semi-arid and arid areas that either found a strong relationship between antecedent catchment state and runoff (Greenbaum et al., 2006; Mclntyr et al., 2007; Camacho Suarez et al., 2015), or a weak correlation (Lane et al., 1971; Zhang et al., 2011). Conditions may be different during exceptional high magnitude events (e.g. Marchi et al., 2010), but these are not included in our data.

The correlation of runoff with rainfall amounts and intensities is less ambiguous. Here, the slopes of the regression lines of Wadi Qilt and Wadi Nueima are very similar and constitute a clear and persistent contrast to Wadi Auja and its headwater catchments. Besides common topographic characteristics (slope angle, mean slope length and drainage density, Table 1), Wadi Qilt and Nueima share some important spatial catchment characteristics, which are distinct from the remaining catchments. Differences in aridity along a strong climatic gradient are a common feature of all catchments. However, Wadi Qilt and Wadi Nueima extend deeper into the arid climate zone than Wadi Auja, which leads to lower annual catchment rainfall and a patchier rainfall distribution. Both catchments are lithologically divided and show a composed runoff pattern of limestone/dolomite and chalk areas. Despite lower rainfall, Wadi Qilt and Nueima have higher overall runoff coefficients and over three times higher mean specific annual discharge than Wadi Auja (Table 1). The relation of aridity to runoff response is generally known (e.g. Costa and Soares, 2012; Schick, 1988), but there are additional factors that explain these marked differences.

Only Wadi Qilt and Nueima include large outcrops of chalks in their lower catchments, while Wadi Auja is entirely devoid of chalk (Fig. 1 and Table 1).This Senonian chalk is known as the major impermeable regional aquitard (Toll et al., 2009) and its outcrops are characterized by weak soil development, low water storage capacity and hence accentuated runoff response. Moreover, our surface cover classification scheme suggested that rock outcrop areas cover over 20% of Wadi Qilt and Nueima but only 16% of Wadi Auja (Table 1). In addition, the sparsely vegetated areas were more pronounced in the downstream stretches of Wadi Qilt and Nueima than in Wadi Auja. Both factors, a high portion of rock outcrops and a sparse vegetation cover are commonly connected to higher rates of runoff generation (Martinez-Mena et al., 1998; Yair and Kossovsky, 2002; Arnau-Rosalén et al., 2008; Li et al., 2011).

Only Wadi Qilt includes noteworthy urban areas in its headwaters. Recent measurements by Grodek et al., 2011) documented ambiguous hydrological effects of urbanization in the region. They found that the urban runoff response was more accentuated during small to medium events, while similar runoff responses were measured in natural and urban catchments during high magnitude events. Our results corroborate these findings, since Wadi Nueima had a more accentuated runoff response than Wadi Quilt despite a much smaller urban area (Table 1). In Wadi Auja H-3 and H-4 are characterized by a clearly different coverage of terraced olive plantations but show no difference in runoff response. This suggests that olive

Events < 1 mm

■ Wadi Qilt

20 - ♦ Wadi Nueima

* Wadi Auja

▼ Headwaters

Wadi Qilt + Nueima x *

event precipitation (mm)

local hortonian events

Fig. 8. Effect of physiographic catchment properties and rainfall characteristics on the relationship between rainfall and runoff heights forthe two catchment groups Wadi Qilt/Wadi Nueima and Wadi Auja with its headwaters.

terraces per se do not reduce runoff generation, while cultivation and soil management practices might hydrologically be important, as found by Kosmas et al. (1997), Gómez et al. (2009) and García-Ruiz (2010).

To summarize the main identified influencing factors discussed above, Fig. 8 combines the events measured in Wadi Qilt and Wadi Nueima as one group and Wadi Auja and its headwaters in another group.

Following the ideas of Graham and McDonnell (2010), we hypothesize that the distinct slopes of the rainfall-runoff regression lines are caused by the spatial characteristics of the geology, bedrock permeability, by the fraction of rock outcrops, soil thickness and corresponding soil water storage. Temporal variations of rainfall intensity and rainfall distribution lead to a spread of the events along the lines, but not to differences in their slopes with the exception of small, isolated convective storm events that do not account for the majority of runoff events (<1 mm) in our study area. As the basic concept was developed in a humid setting (Graham and McDonnell, 2010), more information also from other catchments is required to prove its general applicability in semi-arid regions. This is also true for the characterization of the soil zone to allow for a more comprehensive conclusion on the relationship of soil characteristics and runoff processes. Our data show a change in composition along the elevation gradient, however, soil texture is just one soil property that effects runoff reactions.

In this study we present a unique 5-year data set of soil moisture and rainfall-runoff observations on multiple spatial scales in high temporal resolution for the data scarce (semi-) arid Lower Jordan River region. Our data enabled us to identify dominating runoff processes and responses in a complex environment with contrasting geology, heterogeneous surface characteristics and a strong climatic gradient. The precise identification of catchment properties and surface cover types, in particular carbonate rock outcrops, allowed us to relate the dissimilar runoff responses to individual catchment characteristics.

In Wadi Auja and its headwaters, temporal runoff patterns indicate that runoff was predominantly generated by saturation excess overland flow during events with high rainfall depths. Soil moisture data corroborated these findings and suggest a gradual saturation from the bottom to the top, indicating reduced bedrock permeability despite the dominant limestone lithology in the study area. Overall, the runoff response in Wadi Auja was lower than in Wadi Qilt and Nueima, that both showed a more accentuated runoff response also with contributions of Hortonian overland flow. We related this difference to a different type of lithological composition, more abundant bare rock surfaces, greater aridity, and to an associated shallower soil and sparser vegetation cover. These particular patterns were evident both in our plot and catchment scale data.

We did not find hydrological urbanization effects in our area, since runoff reactions in Wadi Qilt were very similar to those in Wadi Nueima, despite the five times higher fraction ofbuilt-up areas in the headwaters ofWadi Qilt. Nevertheless, outskirts of cities like Ramallah in Wadi Qilt as well as the few small villages in the other catchments shows a very open and scattered arrangement with only small amounts of densely constructed land, that could explain the not observed effect of urbanization on runoff generation. Moreover, a comparison of two headwater catchments within Wadi Auja with a significantly different proportion of olive plantations did not show a hydrological effect of this specific land use type either.

6. Conclusion

While differences in patterns of individual rainfall events resulted in a strong variability of catchment responses in the headwaters, the relationship between depth and intensity of rainfall events and catchment runoff response becomes more and more linear once the catchment size exceeds a certain threshold. Our correlations may therefore be used for a simple estimation of catchment scale runoff amounts, important for the successful assessment and management of available water resources. Surface runoff represents a potentially valuable water source, but as of today remains widely unused due to the lack of reliable information regarding quantity and temporal variability.

Acknowledgements

This work was conducted within the framework of the multi-lateral research project "SMART - Sustainable Management of Available Water Resources with Innovative Technologies" funded by the German Federal Ministry of Education and Research (BMBF), references No. 02WM0802 and 02WM1081. The article processing charge was funded by the German Research Foundation (DFG) and the Albert Ludwigs University Freiburg in the funding programme Open Access Publishing. We would like to thank Awad Rashid, Anwar Zuhluf and Amer for their substantial support and assistance during installation and maintenance of our observation network in the West Bank and Clemens Messerschmid for his suggestions regarding design of field installations and valuable comments on the manuscript. We also thank Noam Greenbaum/University of Haifa for soil analysis. Helpful discussions with Michael Stölzle and Tobias Schütz on the early manuscript are likewise acknowledged. Finally, we thank A. Cerdá and two anonymous reviewers for their valuable comments that helped to improve the manuscript.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ejrh.2016.11.001.

References

ANTEA, 1998. Well Development Study of the Eastern Aquifer Basin, Northern Districts of Palestine, vol. 1, Interim Report, Conceptual Model, unpublished ANTEA Report No. A11903.

Arnau-Rosalén, E., Calvo Cases, A., Boix-Fayos, C., Lavee, H., Sarah, P., 2008. Analysis of soil surface component patterns affecting runoff generation. An

example of methods applied to Mediterranean hillslopes (Alicante, Spain). Geomorphology 101 (4), 595-606. Begin, Z.B., 1975. The Geology of the Jericho Sheet. Geological Survey of Israel, Jerusalem, Bulletin No. 67.

Calvo-Cases, A., Boix-Fayos, C., Imeson, A.C., 2003. Runoff generation, sediment movement and soil water behaviour on calcareous (limestone) slopes of

some Mediterranean environments in southeast Spain. Geomorphology 50 (1), 269-291. Camacho Suarez, V.V., Saraiva Okello, A.M.L., Wenninger, J.W., Uhlenbrook, S., 2015. Understanding runoff processes in a semi-arid environment through

isotope and hydrochemical hydrograph separations. Hydrol. Earth Syst. Sci. 19 (10), 4183-4199. Cammeraat, L.H., Cerda, A., Imeson, A.C., 2010. Ecohydrological adaptation of soils following land abandonment in a semi-arid environment. Ecohydrology 3 (4), 421-430.

Cantón, Y., Domingo, F., Solé-Benet, A., Puigdefábregas, J., 2002. Influence of soil-surface types on the overall runoff of the Tabernas badlands (south-east

Spain): field data and model approaches. Hydrol. Processes 16 (13), 2621-2643. Cantón, Y., Solé-Benet, A., De Vente, J., Boix-Fayos, C., Calvo-Cases, A., Asensio, C., Puigdefábregas, J., 2011. A review of runoff generation and soil erosion

across scales in semiarid south-eastern Spain. J. Arid Environ. 75 (12), 1254-1261. Cerda, A., 1996. Seasonal variability of infiltration rates under contrasting slope conditions in southeast Spain. Geoderma 69 (3-4), 217-232. Cerda, A., 1997. The effect of patchy distribution of Stipa tenacissima L. on runoff and erosion. J. Arid Environ. 36 (1), 37-51.

Cerda, A., 1998. Effect of climate on surface flow along a climatological gradient in Israel. A field rainfall simulation approach. J. Arid Environ. 38,145-159,

http://dx.doi.org/10.1006/jare.1997.0342. Chamizo, S., Cantón, Y., Rodríguez-Caballero, E., Domingo, F., Escudero, A., 2012. Runoff at contrasting scales in a semiarid ecosystem: a complex balance

between biological soil crust features and rainfall characteristics. J. Hydrol. 452,130-138. Costa, A.C., Soares, A., 2012. Local spatiotemporal dynamics of a simple aridity index in a region susceptible to desertification. J. Arid Environ. 87, 8-18. Dünkeloh, A., Jacobeit, J., 2003. Circulation dynamics of Mediterranean precipitation variability 1948-98. Int. J. Climatol. 23 (15), 1843-1866. Dan, J., Koyumdjisky, H., 1963. The soils of Israel and their distribution. J. Soil Sci. 14 (3) 12-20.

Dayan, U., Morin, E., 2006. Flash flood-producing rainstorms over the Dead Sea: a review. Geol. Soc. Am. Spec. Pap. 401, 53-62.

Gómez, J.A., Sobrinho, T.A., Giráldez, J.V., Fereres, E., 2009. Soil management effects on runoff, erosion and soil properties in an olive grove of Southern

Spain. Soil Tillage Res. 102 (1), 5-13. García-Ruiz, J.M., 2010. The effects of land uses on soil erosion in Spain: a review. Catena 81(1), 1-11.

Ghosh, A., Fassnacht, F., Joshi, P.K., Koch, B., 2013. A framework for mapping tree species combining Hyperspectral and LiDAR data: role of selected

classifiers and sensor across three spatial scales. Int. J. Appl. Earth Observ. Geoinform. 39, 49-63. Goldreich, Y., 2003. The Climate of Israel: Observation Research and Application. Klewer Academic Press, London, 270 pp.

Goodrich, D.C., Lane, L.J., Shillito, R.M., Miller, S.N., Syed, K.H., Woolhiser, D.A., 1997. Linearity of basin response as a function of scale in a semiarid

watershed. Water Resour. Res. 33 (12), 2951-2965. Graham, C.B., McDonnell, J.J., 2010. Hillslope threshold response to rainfall:(2) Development and use of a macroscale model. J. Hydrol. 393 (1), 77-93. Greenbaum, N., Ben-Zvi, A., Haviv, I., Enzel, Y., 2006. The hydrology and paleohydrology of the Dead Sea tributaries. Geol. Soc. Am. Spec. Pap. 401, 63-93. Greenbaum, N., 1986. Point Runoff in an Extremely Arid Region. Infiltration-Runoff Tests on Small Plots in the Southern Arava and Their Hydrological and

Pedological Implications. Department of Physical Geography, Hebrew University, Jerusalem (in Hebrew), M. Sc. Thesis. Grodek, T., Lange, J., Lekach, J., Husary, S., 2011. Urban hydrology in mountainous middle eastern cities. Hydrol. Earth Syst. Sci. 15 (3), 953-966. Gunkel, A., Shadeed, S., Hartmann, A., Wagener, T., Lange, J., 2015. Model signatures and aridity indices enhance the accuracy of water balance estimations

in a data-scarce Eastern Mediterranean catchment. J. Hydrol.: Reg. Stud. 4,487-501. Hill, J., Stellmes, M., Udelhoven, T., Röder, A., Sommer, S., 2008. Mediterranean desertification and land degradation: mapping related land use change

syndromes based on satellite observations. Global Planet. Change 64 (3), 146-157. Hughes, A.G., Mansour, M.M., Robins, N.S., 2008. Evaluation of distributed recharge in an upland semi-arid karst system: the West Bank Mountain Aquifer, Middle East. Hydrol. J. 16 (5), 845-854.

IMS (Israel Meteorological Service), 2015. Online Database of the Israel Meteorological Service, Bet Dagan, available at: http://data.gov.i1/ims/3 (last Accessed: 1 October 2015).

Imeson, A.C., Lavee, H., Calvo, A., Cerda, A., 1998. The erosional response of calcareous soils along a climatological gradient in Southeast Spain. Geomorphology 24 (1), 3-16.

Kalogeropoulos, K., Chalkias, C., 2013. Modelling the impacts of climate change on surface runoff in small Mediterranean catchments: empirical evidence from Greece. Water Environ. J. 27 (4), 505-513.

Kirkby, M.J., Bracken, L., Reany, S., 2002. The influence of land use, soils and topography on the delivery of hillslope runoff to channels in SE Spain. Earth Surf. Processes Landforms 27,1459-1473.

Kosmas, C., Danalatos, N.G., Cammeraat, L.H., Chabart, M., Diamantopoulos, J., Farand, R., Gutierrez, L., Jacob, A., Marques, H., Martinez-Fernandez, J., Mizara, A., Moustakas, N., Nicolau, J.M., Oliveros, C., Pinna, G., Puddu, R., Puigdefabregas, J., Roxo, M., Simao, A., Stamou, G., Tomasi, N., Usai, D., Vacca, A., 1997. The effect of land use on runoff and soil erosion rates under Mediterranean conditions. Catena 29,45-59.

Koulouri, M., Giourga, C., 2007. Land abandonment and slope gradient as key factors of soil erosion in Mediterranean terraced lands. Catena 69 (3), 274-281.

Kröpfl, A.I., Cecchi, G.A., Villasuso, N.M., Distel, R.A., 2013. Degradation and recovery processes in semi-arid patchy rangelands of northern Patagonia, Argentina. Land Degrad. Dev. 24 (4), 393-399.

Kutiel, P., Lavee, H., Ackermann, O., 1998. Spatial distribution of soil surface coverage on north and south facing hillslopes along a Mediterranean to extreme arid climatic gradient. Geomorphology 23 (2), 245-256.

LP DAAC (Land Processes Distributed Active Archive Center), 2013. MOD09CMG MODIS and the Terra Surface Reflectance Daily L3 Global 0.05Deg CMG V006. Version 6. NASA EOSDIS Land Processes DAAC, USGS Earth Resources Observation and Science (EROS) Center, Sioux Falls, South Dakota (https://lpdaac.usgs.gov), accessed December 16, 2013, at http://dx.doi.org/10.5067/MODIS/MOD09CMG.006.

Lane, L.J., Diskin, M.H., Renard, K.G., 1971. Input-output relationships for an ephemeral stream channel system. J. Hydrol. 13, 22-40.

Lange, J., Leibundgut, C., Husary, S., Nativ, R., Hassan, M.A., Schick, A.P., 2001. A field-based hydrological model to study the impacts of urbanization on regional water resources. IAHS Publ. 268, 255-261.

Lange, J., Greenbaum, N., Husary, S., Ghanem, M., Leibundgut, C., Schick, A.P., 2003. Runoff generation from successive simulated rainfalls on a rocky, semi-arid, Mediterranean hillslope. Hydrol. Processes 17(2), 279-296.

Lange, J., 2005. Dynamics of transmission losses in a large arid stream channel. J. Hydrol. 306 (1), 112-126.

Latron, J., Llorens, P., Gallart, F., 2009. The hydrology of Mediterranean mountain areas. Geogr. Compass 3 (6), 2045-2064.

Lavee, H., Imeson, A.C., Sarah, P., 1998. The impact of climate change on geomorphology and desertification along a Mediterranean-arid transect. Land Degrad. Dev. 9, 407-422.

Li, X.Y., Contreras, S., Solé-Benet, A., Cantón, Y., Domingo, F., Lázaro, R., Linf, H., Van Wesemael, B., Puigdefábregas, J., 2011. Controls of infiltration-runoff processes in Mediterranean karst rangelands in SE Spain. Catena 86 (2), 98-109.

Llasat, M.C., Llasat-Botija, M., Prat, M.A., Porcú, F., Price, C., Mugnai, A., Lagouvardos, K., Kotroni, V., Katsanos, D., Michaelides, S., Yair, Y., Savvidou, K., Nicolaides, K., 2010. High-impact floods and flash floods in Mediterranean countries: the FLASH preliminary database. Adv. Geosci. 23 (23), 47-55.

Marchamalo, M., Hooke, J.M., Sandercock, P.J., 2016. Flow and sediment connectivity in semi-arid landscapes in SE Spain: patterns and controls. Land Degrad. Dev. 27 (4), 1032-1044, http://dx.doi.org/10.1002/ldr.2352.

Marchi, L., Borga, M., Preciso, E., Gaume, E., 2010. Characterisation of selected extreme flash floods in Europe and implications for flood risk management. J. Hydrol. 394(1), 118-133.

Martin-Vide, J., Gomez, L., 1999. Regionalization of peninsular Spain based on the length of dry spells. Int. J. Climatol. 19 (5), 537-555.

Martinez-Mena, M., Albaladejo, J., Castillo, V.M., 1998. Factors influencing surface runo generation in a Mediterranean semi-arid environment: Chicamo watershed, SE Spain. Hydrol. Process 12, 741-754.

McIntyr, N., Al-Qurashi, A., Wheater, H., 2007. Regression analysis of rainfall-runoff data from an arid catchment in Oman. Hydrol. Sci.J. 52 (6), 1103-1118.

Mimi, Z.A., Assi, A., 2009. Intrinsic vulnerability, hazard and risk mapping for karst aquifers: a case study. J. Hydrol. 364 (3), 298-310.

Morin, E., Jacoby, Y., Navon, S., Bet-Halachmi, E., 2009. Towards flash-flood prediction in the dry Dead Sea region utilizing radar rainfall information. Adv. Water Resour. 32,1066-1076.

Nicolau, J.M., Solé-Benet, A., Puigdefábregas, J., Gutiérrez, L., 1996. Effects of soil and vegetation on runoff along a catena in semi-arid Spain. Geomorphology 14 (4), 297-309.

Ochoa, P.A., Fries, A., Mejía, D., Burneo, J.I., Ruíz-Sinoga, J.D., Cerda, A., 2016. Effects of climate, land cover and topography on soil erosion risk in a semiarid basin of the Andes. Catena 140,31-42.

Perrin, J.L., Tournoud, M.G., 2009. Hydrological processes controlling flow generation in a small Mediterranean catchment under karstic influence. Hydrol. Sci.J. 54 (6), 1125-1140.

Puigdefábregas, J., del Barrio, G., Boer, M.M., Gutierrez, L., Sole, A., 1998. Differential responses of hillslope and channel elements to rainfall events in a semi-arid area. Geomorphology 23 (2), 337-351.

Puigdefábregas, J., 2005. The role of vegetation patterns in structuring runoff and sediment fluxes in drylands. Earth Surf. Processes Landforms 30 (2), 133-147.

Quinton, J.N., Edwards, G.M., Morgan, R.P.C., 1997. The influence of vegetation species and plant properties on runoff and soil erosion: results from a rainfall simulation study in south east Spain. Soil Use Manage. 13,143-148.

Ries, F., Lange, J., Schmidt, S., Puhlmann, H., Sauter, M., 2015. Recharge estimation and soil moisture dynamics in a Mediterranean, semi-arid karst region. Hydrol. Earth Syst. Sci. 19,1439-1456, http://dx.doi.org/10.5194/hess-19-1439-2015.

Rofe and Raffety Consulting Engineers, 1963. Jerusalem and District Water Supply: Geological and Hydrological Report, Report to the Central Water Authority of the Hashemite Kingdom of Jordan, Rofe and Raffety Consulting Engineers, London.

Ruiz-Sinoga, J.D., Galeote, M.G., Murillo, J.M., Marín, R.G., 2011. Vegetation strategies for soil water consumption along a pluviometric gradient in southern Spain. Catena 84 (1), 12-20.

Saaroni, H., Ziv, B., Lempert, J., Gazit, Y., Morin, E., 2014. Prolonged dry spells in the Levant region: climatologic-synoptic analysis. Int. J. Climatol. 35, 2223-2236.

Schick, A.P., 1988. Hydrologic aspects of floods in extreme arid environments. In: Baker, V.R., Kochel, R.C., Patton, P.C. (Eds.), Flood Geomorphology. John Wiley, New York, pp. 189-203.

Schmidt, S., Geyer, T., Guttman, J., Marei, A., Ries, F., Sauter, M., 2014. Characterisation and modelling of conduit restricted karst aquifers-example of the Auja spring, Jordan Valley. J. Hydrol. 511, 750-763.

Schnabel, S., Gutiérrez, A.G., 2013. The role of interannual rainfall variability on runoff generation in a small dry sub-humid watershed with disperse tree cover. Cuadernos de investigación geográfica 39, 259-285.

Shapiro, M.B., 2006. Soils of Israel. Eurasian Soil Sci. 39,1170-1175.

Singer, A., 2007. The Soils of Israel. Springer, Berlin and Heidelberg, 306 pp.

Sneh, A., Bartov, Y., Weissbrod, T., Rosensaft, M., 1998. Geological Map of Israel, 1:200,000,4 Sheets. Geological Survey of Israel, Jerusalem.

Sohrt, J., Ries, F., Sauter, M., Lange, J., 2014. Significance of preferential flow at the rock soil interface in a semi-arid karst environment. Catena 123,1-10.

Thornes, J.B., Wainwright, J., 2004. Environmental Issues in the Mediterranean: Processes and Perspectives from the Past and Present. Routledge, London, 479 pp.

Toll, M., Messerschmid, C., Wolfer, J., Hötzl, H., Ali, W., Sauter, M., 2009. Aquifers in the western Jordan valley. In: Hötzl, H., Möller, P., Rosenthal, E. (Eds.), The Water of the Jordan Valley. Springer, Berlin Heidelberg, pp. 265-286.

Wittenberg, L., Greenbaum, N., Paz, S., Kutiel, H., 2004. Frequency of large floods in the eastern Mediterranean region: a comparison between two 12-Year

periods - Nahal Oren, Mt. Carmel, Israel. Contemp. Israeli Geogr. (Special issue of Horizonts in Geography) 60-61,129-138. Yaalon, D.H., 1997. Soils in the Mediterranean region: what makes them different? Catena 28,157-169.

Yair, A., Kossovsky, A., 2002. Climate and surface properties: hydrological response of small arid and semi-arid watersheds. Geomorphology 42 (1), 43-57. Yair, A., Raz-Yassif, N., 2004. Hydrological processes in a small arid catchment: scale effects of rainfall and slope length. Geomorphology 61 (1), 155-169. Zanon, F., Borga, M., Zoccatelli, D., Marchi, L., Gaume, E., Bonnifait, L., Delrieu, G., 2010. Hydrological analysis of a flash flood across a climatic and geologic

gradient: the September 18, 2007 event in Western Slovenia. J. Hydrol. 394 (1), 182-197. Zhang, Y., Wei, H., Nearing, M.A., 2011. Effects of antecedent soil moisture on runoff modeling in small semiarid watersheds of southeastern Arizona.

Hydrol. Earth Syst. Sci. 15 (10), 3171-3179. Ziadat, F.M., Taimeh, A.Y., 2013. Effect of rainfall intensity, slope, land use and antecedent soil moisture on soil erosion in an arid environment. Land Degrad. Dev. 24 (6), 582-590.