1

Monitoring and measurement techniques on water fluxes and sediment export in dryland ecosystems Project by: all SESAMs Bedload sampling station A bedload trap was installed within the Ribera Salada Catchment (Pre-Pyrenees, Spain) by the SESAM partners from the University of Lleida and from the Forestry Centre of Catalonia (Prof. Ramon Batalla and colleagues) in July 2005 in the middle reaches of the catchment (with a contributing area of around 100 km2). The station is composed of three automatic bedload pit traps, a NEP-390-CBL Turbidity probe, an ISCO-3700 automatic sampler with 24 1-litre sample bottles and a water level actuator, and a water level sensor. The bedload trap was installed to enable a continuous measurement of bedload flux for single rainstorm events with a sampling period of at least five hours. The bedload trap consists of a concrete structure which contains a metal box with a capacity of 0.22 m3 supported on top of a water pillow (MSC Survival), which records the increase by weight of the entering bedload and transmits it by a means of a pressure sensor (PTX-1730) to a Campbell CR1000 data-logger. Sediment capacity of each trap is around 330 kg (submerged weight). Recording interval is 5 minutes. Pit traps have been preliminary calibrated and fully operate since mid-November 2005, together with the rest of the instrumentation. A water temperature sensor was installed during December 2005. Figure 1 displays photographs of the installation of the sediment transport station, with particular attention to the bedload pit trap. The specific design of the bedload trap was based on previous works done by Garcia et al. (2000). Besides permanent instrumentation, sediment transport devices will be deployed from the bridge above the station during floods to complement the automatic sediment transport records. Bedload will be measured by means of a 76 kg 152 mm-intake Helley-Smith Sampler and suspended sediment by means of a US DH-74 depth integrated sampler (for details see Vericat and Batalla 2005).

Figure 1: Installation of the bedload trap at the Ribera Salada Watershed

Figure 2 shows the bedload sampling station in progress, as inspected during a field excursion with a group of students in September 2006.

2

Figure 2: Functioning of the bedload trap (from left to right: traps after a flood event, entire sampling

station with equipment in metal box to the right side of the picture, filled traps with open lid) References: Garcia, C., Laronne, J. B., and Sala, M., 2000, Continuous monitoring of bedload flux in a mountain gravel-bed river: Geomorphology, v. 34, p. 23-31.

Vericat, D. and Batalla, R. J., (2005). Sediment transport in a highly regulated fluvial system during two consecutive floods (lower Ebro River, NE Iberian Peninsula): Earth Surface Processes and Landforms, v. 30, p. 385-402.

Water-stage sediment sampler Three water-stage sediment samplers were installed within the Bengue Basin in Brazil in 2005. The samplers enable the measurement of suspended sediments in the river water with sediment concentration being measured at different depths of water flow inside the river. The sampler is specifically designed for flow events in dryland, ephemeral rivers where runoff occurs only a few times each year. The samplers were constructed within the workshops of the GFZ (Geoforschungszentrum Potsdam). Figure 3 shows photographs of an installed sampler: the metal sampler holder in the dry, ephemeral riverbed, the box containing 15 sampling bottles, and the numbered sediment sampling bottles.

Figure 3: Automatic sediment sampler, left: dry riverbed, middle: sampling box, right: individual

bottles The water-stage sediment sampler needs to be emptied and reset after each flood event, which might become problematic if no local agency is available to do so consistently. Other problems included nesting of insects in the box (especially after extended periods without runoff) and the clocking of the tubes leading to the sampling bottles with organic litter. Water discharge, suspended sediment and turbidity data: A turbidity meter was installed in July 2005 at the outlet of the Isabena Catchment (at the gauging station Capella, no. 47 of the Ebro Water Authorities), as depicted in Figure 4, which enables the continuous measurement of turbidity with a temporal resolution of 15 minutes. The device is a NEP-390-CBL Turbidity probe 0 – 3000 NTU (-2.5 v to * 2.5 v dc). Data is collected by means of Campbell CR10X data-logger. At the same location, an ISCO-3700 Sampler with water level actuator

3

was installed that takes up to 24 1-litre water samples for a single flood event. Direct water samples are regularly taken to support the calibration of the turbidity and automatic sampler measurements.

Figure 4: Capella Station: Turbidimeter probe and ISCO-3700 sampler

Figure 5: Flood event at the Capella Station with substantial sediment transport on 14.09.2006

Temporary sediment storage in dryland riverbeds The temporary sediment storage along the Isabena River (Pre-Pyrenees, north-east Spain) was assessed with a simple technique after Hilton and Lisle (1993) for 78 cross-sections along a river stretch of about 33 km. The technique as shows in Figure 6 was applied to determine the role and the order of magnitude of in-channel storage in the annual sediment budget of the catchment. The total storage volume allows then the estimation of the residence time of fine sediments in the river system.

Figure 6: Spatial sampling procedure of temporary river storage (after Hilton and Lisle 1993) The following methodology was employed during the sampling of the sediment storage: 1. Sampling interval of cross-sections: 300-500 metres

4

2. Record station number, GPS location while standing at bankful level on the right side of the river (looking downstream, European Datum 1950 UTM Zone 30N)

3. Record approximate width, depth and form at bankful discharge (sketch river stretch) 4. Start survey of sediment storage at the right side of the river, at bankful level 5. Along 1-metre intervals (horizontal line) record perpendicular to the flow direction:

a) sampling interval in metres, b) sampling interval covered with water: yes/no, c) % fraction of fine sediments, d) 5 measurements of the height (in cm) of the sediment layer (from right to left), every 20

cm, using the graduated steel bar. The high-resolution height measurements of sediment thickness in the riverbed can then be transformed to sediment volume or mass (per unit metre width of the cross-section) to obtain estimates for the mass of sediment stored in a river stretch. The high spatial resolution data on riverbed storage volume can then be related to estimates of riverbed slope, shape, geological, geomorphological and cross-sectional characteristics along the longitudinal profile of the river to study geospatial units and pattern formation of the river system. Figure 7a shows as an example an ephemeral river stretch in which sediment from upslope badland areas was deposited after a large rainstorm event; Figure 7b shows two students at a typical river cross-section during the field campaign.

Figure 7: Temporary sediment storage and its measurement in the ephemeral riverbed of the Isabena

Catchment Publications: Hilton, Sue; Lisle, Thomas E. 1993. Measuring the fraction of pool volume filled with fine sediment . Research Note PSW-RN-414. Albany, California: Pacific Southwest Research Station, Forest Service, U.S. Department of Agriculture. 11 p. available online. Infiltration measurements with hood-infiltrometer Infiltration measurements were carried out on badland formations, agricultural fields and abandoned fields in the Isabena Catchment (Pre-Pyrenees, Spain) to obtain information on the saturated hydraulic conductivity of different soils. For the soil hydraulic measurements, a hood infiltrometer (UGT Müncheberg) was used consisting of an infiltration vessel, a U-tube-manometer, 2 hoods and cover bag. The advantage of the hood infiltrometer in comparison to other infiltration measurement devises is its minimal disturbance of the soil matrix and its small water requirement. The infiltration is done by placing a circular shaped hood filled with water directly on the soil surface. Thus, no contact layer on the soil surface is necessary. The pressure head in the water filled hood is regulated by a Mariotte water supply. The effective pressure head at the soil surface can be adjusted between zero and any negative pressure up to the bubble point of the soil. The pressure head and the bubble point can be measured directly via an U-pipe manometer and the stand pipe of the hood. For calculation of the hydraulic conductivity after Wooding’s equation (Wooding 1968, steady state infiltration) two hoods are used with a ratio of 2 : 1 in terms of the covered infiltration area at the soil surface. Wherever possible, the measurement was repeated at a depth of 30 cm below the soil surface. In addition, a soil sample was taken with a 100 cm³ ring cylinder at most sample sites to estimate bulk density and

5

particle size distribution. The saturated soil hydraulic conductivity can be calculated by using equations proposed by Gardener (1958) and Wooding (1968), also described in UGT (2004).

Figure 8: Hood-infiltrometer measurements in the Isabena Catchment

Measurements on the badland slopes proved to be difficult in the field, as the hood has to be set up in a horizontal position. On the badlands this could only be assured measuring on a ridge or by levelling the soil which disturbs natural soil surface conditions. Durations of individual measurements varied between 1 and 6 hours. The measured values for hydraulic conductivities were highly variable even for similar soil and vegetation cover conditions, which can be either explained by the high heterogeneity of dryland soils, or by the complicated installation and handling of the equipment. References: Appel, K. (2006) Characterisation of badlands and modelling of soil erosion in the Isábena watershed, NE Spain. Charakterisierung und Modellierung von Bodenerosion auf Badlands im Isábena-Einzugsgebiet, NO-Spanien. Unpublished Diplomarbeit, University of Potsdam, Germany

Gardner, W.R. (1958): Some steady-state solutions of unsaturated moisture flow equations with application to evaporation from a water table. In: Soil science 85, 228-232.

Wooding, R.A. (1968): Steady infiltration from a shallow circular pond. In: Water Resources Research 4, 1259 – 1273.

UGT (2004) Bedienungsanleitung für das Haubeninfiltrometer. Umweltgerätetechnik GmbH. Müncheberg.

River characteristics To enable the parameterisation of process-based river transport models, a spatial distributed data set on water flow and sediment-transport characteristics were collected for ten river stretches along the Isabena River (Pre-Pyrenees, Spain) ranging from very narrow, turbulent mountain to very wide, shallow lowland stretches. Data for the following parameters were collected:

Detailed cross-section and longitudinal slope profiles Flow velocity measurements for the determination of Manning’s n Riverbed gradation Vegetation type and cover inside and surrounding the river stretch Water temperature and oxygen content

1. River survey Cross-sections were surveyed with a high-precision total station with an interval of minimum 0.25 metres, and information on current water depth, bankful depth, and flood-prone areas were taken. The longitudinal height profile over a distance of 6 – 30 channel widths in length were measured for the derivation of slope estimates.

6

2. Measurement of flow velocity Flow velocity was measured with an OTT C2 small current meter followed the very detailed procedure of Harrelson et al. 1994. At least 25-30 readings were taken for each river cross-section with an interval of 0.25 to 100 cm. Each reading lasted at least 40 seconds and was taken at a depth of 0.6 times the total depth of current water level.

3. Riverbed gradation was estimated following two methods described by Harrelson et al. 1994 and

Kondolf et al. 2003: a) Photographic method: at one to three locations at each cross-section, a photograph was taken

from a height of approximately 1 – 1.5 metres with scale in the picture. Additionally, soil samples of finer sediments were collected and analysed in the laboratory for their particle size distribution.

7

b) Wolman-Pebble count (1954): For each cross-section, a 100-sample pebble count was carried

out using the random step-toe procedure. Samples on particle size were drawn randomly by wading through the river section close to the previously survey cross-section, and drawing the particles that are closest to the toe of the collector’s wader. The intermediate axes of the particle was measured (neither the longest nor shortest of the three mutually perpendicular sides of each particle picked up). If distinctly different homogenous facies exist, each facie was sampled individually. Particles were categorised by using Wentworth size classes in which the size doubles with each class (2, 4, 8, 16, 32, etc.) or smaller class intervals based on 1/2 phi values. The method is very well described in Harrelson et al. (1994). Data analysis on particle size distribution was carried out using the statistical methods by Bunte and Abt (2001).

4. Characterisation of river reach: Each river reach was classified according to the scheme of Montgomery and Buffington (1997) and Rosgen (1997). Montogmery and Buffington (1997) classified channel-reach morphology in mountain drainage basins into seven distinct reach types mainly as a function of sediment-transport and flow dyanmic characteristics: colluvial, bedrock, and five alluvial channel types (cascade, step pool, plane bed, pool riffle, and dune ripple). Their classification is based on a coupling of reach-level channel processes with the spatial arrangement of reach morphologies, their links to hillslope processes, and external forcing by confinement, riparian vegetation, and woody debris. Rosgen (1994) categorises rivers according to morphological stream characteristics such as entrenchment, gradient, width/depth ratio, sinuosity, bed material and landforms (Figure 9 and Figure 10).

8

Figure 9: Rosgen classification by pattern, profile and dimension (Rosgen 1994)

Figure 10: Rosgen classification by bed material and cross-section characteristics (Rosgen 1994)

In addition, following a rather descriptive approach, for each cross-section the dominant vegetation inside and along the river was determined as well as the % vegetation cover inside the main channel. Water temperature and oxygen content was measured with a multi-sensor device.

9

References Bunte, Kristin; Abt, Steven R. 2001. Sampling surface and subsurface particle-size distributions in wadable gravel-and cobble-bed streams for analyses in sediment transport, hydraulics, and streambed monitoring. Gen. Tech. Rep. RMRS-GTR-74. Fort Collins,CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 428 p. available online.

Harrelson, C. C., C. L. Rawlins, J. P. Potyondy (1994) Stream channel reference sites: An illustrated guide to field technique. United States Department of Agriculture. Forest Service General Technical Report RM-245. available online.

Kondolf, G. M., T. E. Lisle, G. M. Wolman (2003) Bed Sediment Measurement in: Tools in fluvial geomorphology edited by G. M. Kondolf and Herve Piegay, Wiley, Chichester

Montgomery, D. R. and J. M. Buffington (1997) Channel-reach morphology in mountain drainage basins. GSA Bulletin 109 (5), p. 596 – 611

Rosgen, D. L., A classification of natural rivers. CATENA 22 (1994): 169-199.