- Email: [email protected]
Monitoring and analysis of suspended sediment transport dynamics in the Carapelle torrent (Southern Italy)
Catena 80 (2010) 1–8
Contents lists available at ScienceDirect
Catena j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c a t e n a
Monitoring and analysis of suspended sediment transport dynamics in the Carapelle torrent (Southern Italy) F. Gentile, T. Bisantino ⁎, R. Corbino, F. Milillo, G. Romano, G. Trisorio Liuzzi PROGESA Department, University of Bari, Via Amendola 165/A Bari, Italy
a r t i c l e
i n f o
Article history: Received 3 July 2008 Received in revised form 21 May 2009 Accepted 7 August 2009 Keywords: Continuous river monitoring Sediment transport Turbidimeter Hysteretic loops
a b s t r a c t In Northern Puglia watershed's high rates of suspended sediment loads occur along the hydrographic network during intense rainfall events. In order to monitor this phenomenon an automated station, equipped with a turbidity probe, has been set up in the Carapelle torrent. A laboratory testing stage of the turbidity probe was preliminary designed to evaluate the dual functionality of the instrument (turbidity and suspended solids) in relation with the sediment concentration and the grain size distribution. Successively a field calibration of the instrument was carried out to determine the relationship between optical and gravimetric data and to check the housing device. Afterwards, the high temporal resolution data collected over a 2-year period (2007–2008) were used to analyze the sediment transport dynamics at event scale. © 2009 Elsevier B.V. All rights reserved.
1. Introduction The sediment transport that affects rivers in the semi-arid environment of Northern Puglia is mainly characterised by suspended materials. The Hydrological Service of Puglia Region measured the suspended load of these water courses by means of hand sampling during the period 1933–1989. On the basis of the available data a good correlation between discharges and sediment load was observed in those torrents (Marchi et al., 1986). This agrees with the observations made by other authors (Walling et al., 1992) who observed that a large amount of suspended sediments is transported during flood events. The relationship between discharge and sediment concentration presents a hysteretic loop, the shape of which allows to identify the sediment source areas (Williams, 1989). The rate of transport changes from the rising limb to the falling limb of the hydrographs and this effect is induced by the unsteadiness of the flow (De Sutter et al., 2001). An adequate assessment of the sediment load during flood events can be made through continuous measuring. In recent years the use of the optical technology for continuous river monitoring largely spread as it is considered a good “surrogate” to measure suspended sediment (Gippel, 1995; Lewis, 1996). Several optical instruments are based on the scattering theory. In simple terms, turbidity is the optical property that results when light is scattered by suspended particles (silt, clay, algae, organic matter, etc.). The intensity of the scattered light increases with the suspended solids (Grayson et al., 1996). The scattered light can be measured by one or more photodetectors posi-
⁎ Corresponding author. Tel.: +39 080 5442957. E-mail address: [email protected] (T. Bisantino). 0341-8162/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.catena.2009.08.004
tioned at α= 90° (nephelometric scattering), α<90° (forward scattering) or at α>90° (backscattering) to the direction of the incident beam. The light source can vary in wavelength and type. These instruments are sensitive to the grain size distribution of the material. This feature has little relevance in small watersheds because they are generally characterised by homogenous pluviometric regimes and erodibility but can be significant in large watersheds (Lenzi and Marchi, 2000). Variations in grain size distribution, associated with floods, can lead to nonlinear turbidity/suspended sediment concentration relationships. In practice, the degree of curvature is usually very small and linear models perform quite well in estimating loads (Lewis, 2003). Further interferences associated with optical measuring consist of the colour of the particles, the watery medium, the ratio between the size and shape of particles, the density of the particles and the content of dissolved particles which absorb light in some bands of the visible spectrum, altering the characteristics of the transmissed light (Gippel, 1989; Foster et al., 1992; Sadar, 1998). These interferences are minimized by the ratio detection system that uses a combined system of photodetectors and a specific algorithm that calculates the intensity of the scattered light (EPA, 1999; Sadar, 1999). Based on such considerations an experimental station was set up to measure the suspended sediment transport in the Carapelle stream (basin area: 506 km2). The expected maximum sediment concentrations were high (SSC > 15 g/l) and for this reason the Hach–Lange Solitax sc optical probe was chosen to measure the suspended sediment concentration (Bisantino et al., 2006). The probe was preliminary laboratory tested to verify its suitability to monitor suspended sediment transport in medium-size watersheds. The instrument operates in two different modes: the turbidity-nephelometric mode with detection at 90° and the suspended solids-ratio
2
F. Gentile et al. / Catena 80 (2010) 1–8
Fig. 1. The Carapelle watershed mouthed at the sediment transport experimental station.
detection system with detection at 140° and 90°. The tests were executed in order to evaluate the relationships between optical and gravimetric data varying the sediment concentration, and to assess the influence of the coarser materials (sand) on measurements (Gentile et al., 2007). Afterwards, the instrument was field tested through a calibration stage. The check of the housing device was also carried out in order to investigate any bias in concentration due to the sheltering tube. The most relevant flood events were then considered and the suspended sediment concentration, monitored at half-hourly scale, was plotted versus discharge to analyze the sediment transport dynamics.
2. Materials and methods 2.1. Experimental station measuring suspended sediment concentration
The station, already supplied with an ultrasound stage meter and a stage recorder, was in addition equipped with the Hach–Lange Solitax probe. The instrument has a LED (light-emitting diode) light source that emits a beam of infrared light into the stream at an angle of 45° to the probe face. A pair of photodetectors detects light scattered at 90° and 140° to the incident beam. In the turbidity-nephelometric mode the 90° photodetector solely works while in the suspended solids-ratio detection system the backscatter photodetector (140°) is also active. In the suspended mode the instrument's microprocessor uses an algorithm to ratio the signals from each detector and to convert them into suspended sediment concentrations (Sadar, 1998). This method enables a broad range of concentrations to be examined (0.001–4000 Nephelometric Turbidity Units – NTU – for turbidity and 0–150 g/l for suspended solids) and compensates for the
The experimental station measuring suspended sediment concentration is located in the Carapelle torrent (at Ordona-Castelluccio Dei Sauri bridge), one of the main water courses in Northern Apulia (Fig. 1; Table 1). The torrent originates in the Apennine mountains and crosses the Tavoliere flood plain before flowing into the Adriatic sea. The watershed is characterised by clayey-sandy Plio–Pleistocene sediments in the alluvial fan and by flyschoid formations in the mountainous areas, which are subject to erosion. The plain and the low hilly areas are mainly used for olive growing, whereas the higher slopes are occupied by woods and pasture. The climate is typically Mediterranean, with rainfalls ranging from 450 to 800 mm/year and average temperatures from 10 to16 °C. Table 1 Main characteristics of the Carapelle watershed. Watershed area Average altitude Main channel length Main channel slope Mean watershed slope
A Hm L ic iw
506.2 km2 466 m a.s.l. 52.16 km 16.3% 8.2%
Fig. 2. Sediment transport measuring station and probe housing device.
F. Gentile et al. / Catena 80 (2010) 1–8
3
Fig. 3. Data acquisition system and telemetry equipment.
effects of the stray light, organic matter and watery medium colour. Furthermore, the instrument is fitted with a programmable time interval screen wiper that prevents fouling. Time-consuming multipoint calibrations are replaced by a single correction factor. Through a pulley, a float and a counterweight group, the probe is housed in a tube to protect it from the impact of any flowing coarse material and from any potential measuring errors caused by incident radiant energy straying into the infrared field (Fig. 2). The probe is controlled by a data acquisition set and a telemetry system is also provided for remote measurements (Fig. 3). 2.2. Laboratory testing and field calibration of the optical probe The laboratory experiments were conducted to evaluate, with reference to river sediments, the functionality of the Solitax sc probe. The potential applications of the instrument range from clear water to strongly turbid liquid. It can be used in turbidity-nephelometric mode to predict SSC in low turbid water and in suspended sediment-ratio detection system mode when measuring in high concentrated sediments. The two-detector optical system allows to measure high
Fig. 4. Laboratory equipment.
SSCs and compensates for colour in the sample, light fluctuation, and stray light. In order to determine the relationships between the optical data and the gravimetric one, some tests were executed for 31 suspensions having fixed granulometric mixtures and 36 suspensions with varying ratios between sandy and fine fractions. The material was collected from the riverbed of the Carapelle stream. Granulometric analysis of the sediments, carried out by means of gravimetric method, revealed that the material was predominantly sandy (87.5%) and contained a small percentage of silt (8.8%) and clay (3.7%). The sample was split into three batches. The first one (OW—Original Wet) was kept wet and in its original granulometry, the second one (OD—Original Dried) was kept in its original granulometry too but it was oven-dried at 110 °C, the third one (SD—Separated Dried) was split into its fine and sand fractions (diameter d = 0.063 mm—UNI, BS Standards). From the OW batch 11 suspensions were prepared by picking up different amounts of sediment whose weights were determined at the end of the session through gravimetric analysis. The use of wet material ensured that no alterations to the grain structure were induced in the suspensions preparation. From the OD batch 20 suspensions with known SSC were constituted. Such preparation somehow altered the grain structure, but it allowed to better homogenize the sample and to obtain exactly the desired amount of sediment (in weight).
Fig. 5. Flow depth series and suspended sediment concentrations of the samples collected in the Carapelle stream.
4
F. Gentile et al. / Catena 80 (2010) 1–8
Fig. 6. Counterclockwise and clockwise loops.
The suspensions obtained from OW and OD batches have the same grain size distribution and sediment concentrations ranging from 0.1 to 150 g/l. Such a broad range was established taking into account the maximum values of sediment concentrations measured in the water courses of northern Puglia (90 g/l), and in order to test the manufacturer's specifications (0–150 g/l). From the SD batch 36 suspensions were prepared combining six weights and six sand percentages. The suspensions were prepared independently, rather than by successive dilutions of the same sediment batch, for two reasons. The former is that during stirring the sample did not remain in its initial conditions because of the separation of the aggregates. The latter is that progressive dilutions would require increasing volumes and then different boundary conditions. Equal volumes could be obtained by subsampling, but this often causes a biased sample to be extracted (Gray et al., 2000). The suspensions were stirred (Fig. 4) for ten minutes and the measures at five minute intervals were considered as representative data in order to get well mixed suspensions but not too altered by over-stirring. The first test was carried out on the OD and OW suspensions and the readings of the probe for turbidity and SSC were used to determine the relation between the probe outputs, in both the operating modes, and the gravimetric SSCs. The second test was executed on the SD suspensions and the readings of the probe for turbidity and SSC were used to evaluate the influence of the sand fraction on measures. The aims of the field calibration stage were to evaluate the efficacy of the housing system, to identify a calibration curve of the instrument for the specific torrent and to assess the type of the relationship between the optical and the gravimetric SSCs. This stage was executed monitoring nine flood events and collecting 65 hand samples (Fig. 5). The SSC of all samples was measured with a gravimetric method and compared with the data observed by the optical probe.
Fig. 7. Relationships between turbidity as measured by the instrument and gravimetric SSC.
Furthermore 17 pairs of samples were taken outside and inside the tube in order to check the housing device. Their SSCs were then compared in the laboratory to evaluate the occurrence of any alterations caused by the protective tube. Afterwards the calibration curves were compared with the results obtained during the laboratory stage. 2.3. Flood events During the monitoring seasons 2007–2008 nine significant events, having continuous data and absence of anomalies in the time series, were selected. The data, registered at a half-hourly scale, have been used to analyze the relationships between the SSC and the discharge (Q) during flood events. The loop-shaped curves, existing between SSC and Q, depend on the sediment availability that derives from the sediment supply and/ or from possible sediment deposits (Williams, 1989). In the clockwise loop a rapid increase of the suspended sediment at the beginning of the flood event occurs due to the presence of available sediment in the water course. If the sediment concentration decreases before the falling limb of the hydrograph, the sediment source areas can be considered limited. In the counterclockwise loop, instead, the sediment source areas are mainly located on hillslopes and the material moves slower than the peak discharge (Lenzi and Marchi, 2000). Fig. 6 reports two different flood events having counterclockwise and clockwise loops between suspended sediment concentration and discharge. 3. Results 3.1. Laboratory and field testing Figs. 7 and 8 show the results of the first laboratory tests. Using the instrument in turbidity-nephelometric mode the measurement increases up to a top SSC-value ranging between 40–50 g/l. As the concentrations further increase, the turbidity decreases. This trend
Fig. 8. Relationships between SSC as measured by the instrument and gravimetric SSC.
F. Gentile et al. / Catena 80 (2010) 1–8
5
Fig. 9. Relationships between the turbidity measured by the instrument and gravimetric SSC (OW and OD batches). Fig. 11. Comparison between the concentrations observed inside and outside the protective tube.
reveals the probe blindness occurring in concentrated suspensions already observed by other authors (Sadar, 1998; EPA, 1999). Using the instrument in suspended solids-ratio detection system blindness does not occur and the measurement has a linear increment with the sediment concentration. The results point out the good performance of the two-detector optical system especially for monitoring sediment transport in streams having high sediment concentrations as the Carapelle torrent. Figs. 9 and 10 report the results of the second tests. For mixtures at the same concentrations, the instrument's response depends on the percentage of sand. As expected according to the literature (Foster et al., 1992), the signal was greater in the fine material suspensions, except for the turbidity mode in the curve limbs affected by probe blindness. In particular, in the turbidity-nephelometric mode, by varying the percentage of sand in the mixture, the probe blindness occurs as from SSC = 7–8 g/l for the finest mixtures to higher values in the predominantly sandy ones (Fig. 9). In the suspended solids-ratio detection system blindness does not occur and the relationship between optical and gravimetric SSC is linear for each mixture (Fig. 10). These trends were already observed by other authors at lower SSCs (Sadar, 1998). The results of this stage show that the suspended solids-ratio detection system gives more reliable SSC values and that a linear conversion factor can be used for mixtures having the same granulometric composition. During the field calibration stage 65 samples were collected during the flood events. They mainly consisted of fine materials: the first 28 samples, joined together, contained 9.4% of sand, 32.9% of silt and 57.7% of clay. A comparison between the sediment concentrations inside and outside the protective tube was made, by taking 17 pairs of simultaneous samples. The external concentrations proved to be equal to the internal ones (R2 = 1), and the samples taken from both positions have been considered as part of the whole dataset used during the calibration stage (Fig. 11).
Nine flood events, having peak discharges between 0.94 and 44.62 m3/s, were examined. The relationship between sediment concentration and discharge revealed the existence of clockwise, counterclockwise and mixed-shaped hysteretic loops (Figs. 13–15). The counterclockwise hysteresis (Fig. 13) is predominantly related to events of moderate intensity that occur during the entire flood season. During these events small amounts of sediments move downstream and SSC increases in the falling limb of the hydrograph due to the distance between the sediment sources and the measuring station as observed by other authors (Brasington and Richards, 2000). It can be observed comparing the February 27 and the subsequent March 7 and 31, 2007 floods that the first flood of the year has higher rates of SSCs at equivalent flows than the following events. The reason can be referred to the dry weather preceding the first event and to the sediment depletion that influences the last ones (De Sutter et al., 2001; Bača, 2008). Considerable rates of suspended sediment were also observed in the moderate flood that occurred at the beginning of the flood season 2008 (January 24). A different behaviour was observed for the April 9, 2007 flood that starts on the falling limb of
Fig. 10. Relationships between the SSC measured by the instrument and gravimetric SSC (OW and OD batches).
Fig. 12. Linear relationship existing between the optical SSC and the gravimetric data.
The sediment concentrations of all the samples proved to be correlated with the data measured by the instrument, as shown in Fig. 12. In the range of the monitored discharges the results confirmed the existence of a very good linear relationship (R2 = 0.99) between the gravimetric and the optical data. This outcome allows to neglect the possible grain size variations, occurring during flood events, that can determine a curvilinear relationship between optical and gravimetric SSC.
3.2. Analysis of flood events
6
F. Gentile et al. / Catena 80 (2010) 1–8
Fig. 13. Counterclockwise loops.
F. Gentile et al. / Catena 80 (2010) 1–8
7
Fig. 14. Clockwise loop.
Fig. 15. Mixed loops.
the previous intense event (April 5, see Fig. 15), and for this reason suspended sediment concentration markedly increases. The most intense events (Figs. 14 and 15) are characterised by a clockwise or mixed loop. The clockwise loop occurs during a simple flood while the mixed loop is typical of graded floods. In the clockwise loop a rapid increase of the SSC in the raising limb of the hydrograph occurs. The SSC reaches a peak value, then decreases and the advance of the SSC maximum value towards the peak discharge can be observed. This condition occurs when the sediment availability in the watershed is limited, especially at the end of the rainy season when the depletion of sediments occurs as a consequence of the previous floods (Campbell, 1985). To estimate the dependency of the loop shape on the event intensity, the time delay T occurring between SSC and discharge peak
Fig. 16. Time delay between discharge and SSC peaks.
values was plotted versus peak discharge Qp. A decreasing trend (coefficient of determination R2 = 0.7) was observed (Fig. 16). The boundary between negative and positive values of the time T is representative of the condition at which the flow exerts a strong influence on the sediment load and determines the depletion of the watershed. 4. Conclusions The study examines the results of a sediment transport monitoring programme using high temporal resolution data. An experimental station was set up in the Carapelle torrent (Northern Apulia), equipped with an optical immersion probe which measures turbidity and suspended sediment concentration. The instrument was preliminary laboratory tested. The results showed that the two-detector optical system produces the best results and is more suitable for monitoring high suspended sediment transport in rivers because it provides a linear relationship between the optical and the gravimetric data. The field testing stage confirmed the results achieved in the laboratory, evidencing the linear relationship between the optical and the gravimetric data during flood events. The instrument housing device did not interfere with the measuring process. Nine flood events, collected over a 2-year period (2007–2008), were then analysed to investigate the relationships between SSC concentration and discharge (Q) in flood flows. The SSC–Q curves
8
F. Gentile et al. / Catena 80 (2010) 1–8
revealed that in the Carapelle stream clockwise, counterclockwise and mixed loops are possible and can be related to the event intensity and to the sediment supply of the watershed. In particular the counterclockwise loops occur during moderate events while clockwise and mixed loops are typical of intense floods that determine the sediment depletion of the watershed from fixed discharges. References Bača, P., 2008. Hysteresis effect in suspended sediment concentration in the Rybárik basin. Slovakia Hydrol. Sc. 53 (1), 224–235. Bisantino, T., Corbino, R., Gentile, F., Grittani, A., Milillo, F., Romano, G., Trisorio Liuzzi, G., & Zanframundo, P., 2006. Monitoraggio del trasporto solido nei bacini della Puglia settentrionale tra il Candelaro e l'Ofanto. Quaderni di Idronomia Montana. Ed. Bios. 26: 193–204. Brasington, J., Richards, K., 2000. Turbidity and suspended sediment dynamics in small catchments in the Nepal Middle Hills. Hydrol. Process. 14, 2559–2574. Campbell, G.S., 1985. Soil physics with basic transport models for soil–plant systems. Developments in Soil Science, vol. 14. Elsevier, New York. De Sutter, R., Verhoeven, R., Krein, A., 2001. Simulation of sediment transport during flood events: laboratory work and field experiment. Hydrol. Sci. 46 (4), 599–610. Foster, I.D.L, Millington, R., Grew, R.G., 1992. The impact of particle size controls on stream turbidity measurement; some implications for suspended sediment yield estimation. Erosion and Sediment Transport Monitoring Programmes in River Basins, vol. 210, pp. 51–62. Gentile, F., Bisantino, T., Milillo, F., Romano, G., Trisorio Liuzzi, G., 2007. Caratteristiche funzionali di una sonda a raggi infrarossi per la misura del trasporto solido in sospensione nei corsi d'acqua. Quaderni di Idronomia Montana. Ed. Bios. 27:421–432. Gippel, C. J., 1989. The use of turbidity instruments to measure stream water suspended sediment concentration. Monograph Series No. 4, Dept of Geography and
Oceanography, University College, University of New South Wales, Australian Defence Force Academy, Canberra, 204 pp. Gippel, C.J., 1995. Potential of turbidity monitoring for measuring the transport of suspended solids in streams. Hydrol. Process. 9, 83–97. Gray, J.R., Glysson, G.D., Torcios, L.M., Schwartz, G., 2000. Comparability of suspendedsediment concentration and total suspended solids data. Water Resources Investigation Rep. 00-4191. InUSGS, Washington, DC. Grayson, R.B., Finlayson, B.L., Gippel, C.J., Hart, B.T., 1996. The potential of field turbidity measurements for the computation of total phosphorus and suspended solids loads. J. Environ. Manag. 47, 257–267. Lenzi, M.A., Marchi, L., 2000. Suspended sediment load during floods in a small stream of the Dolomites northeastern Italy. Catena 39, 267–282. Lewis, J., 1996. Turbidity-controlled suspended sediment sampling for runoff-event load estimation. Water Resour. Res. 32 (7), 2299–2310. Lewis, J., 2003. Turbidity-controlled sampling for suspended sediment load estimation. In: Bogen, J., Tharan, Fergus, Des, Walling (Eds.), Erosion and Sediment Transport Measurement in Rivers: Technological and Methodological Advances. Proc. Oslo Workshop, 19–20 June 2002, vol. 283. IAHS Publ, pp. 13–20. Marchi, L., Trisorio-Liuzzi, G., Zanframundo, P., 1986. Analisi dei deflussi torbidi nei piccoli bacini del Sub- Appennino dauno. Quaderni di Idronomia Montana. Ed. Bios. 6:95–121. Walling, D.E., Webb, B.W., Woodward, J.C., 1992. Some sampling considerations in the design of effective strategies for monitoring sediment-associated transport. IAHS Spec. Publ. 210, 279–288. Williams, G.P., 1989. Sediment concentration versus water discharge during single hydrologic events in rivers. J. Hydrol. 111, 89–106. Sadar, M.J., 1998. Turbidity science. In: Technical Information Series. Hach Company. 26 pp. http://www.hach.com. Sadar, M.J., 1999. Turbidimeter Instrument Comparison: Low-level Sample Measurements. Technical Information Series. Hach Company. 56 pp. http://www.hach.com/. U. S. Environmental Protection Agency, 1999. Basic turbidimeter design and concepts. Guidance Manual Turbidity Provisions, chap. 11.
Contents lists available at ScienceDirect
Catena j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c a t e n a
Monitoring and analysis of suspended sediment transport dynamics in the Carapelle torrent (Southern Italy) F. Gentile, T. Bisantino ⁎, R. Corbino, F. Milillo, G. Romano, G. Trisorio Liuzzi PROGESA Department, University of Bari, Via Amendola 165/A Bari, Italy
a r t i c l e
i n f o
Article history: Received 3 July 2008 Received in revised form 21 May 2009 Accepted 7 August 2009 Keywords: Continuous river monitoring Sediment transport Turbidimeter Hysteretic loops
a b s t r a c t In Northern Puglia watershed's high rates of suspended sediment loads occur along the hydrographic network during intense rainfall events. In order to monitor this phenomenon an automated station, equipped with a turbidity probe, has been set up in the Carapelle torrent. A laboratory testing stage of the turbidity probe was preliminary designed to evaluate the dual functionality of the instrument (turbidity and suspended solids) in relation with the sediment concentration and the grain size distribution. Successively a field calibration of the instrument was carried out to determine the relationship between optical and gravimetric data and to check the housing device. Afterwards, the high temporal resolution data collected over a 2-year period (2007–2008) were used to analyze the sediment transport dynamics at event scale. © 2009 Elsevier B.V. All rights reserved.
1. Introduction The sediment transport that affects rivers in the semi-arid environment of Northern Puglia is mainly characterised by suspended materials. The Hydrological Service of Puglia Region measured the suspended load of these water courses by means of hand sampling during the period 1933–1989. On the basis of the available data a good correlation between discharges and sediment load was observed in those torrents (Marchi et al., 1986). This agrees with the observations made by other authors (Walling et al., 1992) who observed that a large amount of suspended sediments is transported during flood events. The relationship between discharge and sediment concentration presents a hysteretic loop, the shape of which allows to identify the sediment source areas (Williams, 1989). The rate of transport changes from the rising limb to the falling limb of the hydrographs and this effect is induced by the unsteadiness of the flow (De Sutter et al., 2001). An adequate assessment of the sediment load during flood events can be made through continuous measuring. In recent years the use of the optical technology for continuous river monitoring largely spread as it is considered a good “surrogate” to measure suspended sediment (Gippel, 1995; Lewis, 1996). Several optical instruments are based on the scattering theory. In simple terms, turbidity is the optical property that results when light is scattered by suspended particles (silt, clay, algae, organic matter, etc.). The intensity of the scattered light increases with the suspended solids (Grayson et al., 1996). The scattered light can be measured by one or more photodetectors posi-
⁎ Corresponding author. Tel.: +39 080 5442957. E-mail address: [email protected] (T. Bisantino). 0341-8162/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.catena.2009.08.004
tioned at α= 90° (nephelometric scattering), α<90° (forward scattering) or at α>90° (backscattering) to the direction of the incident beam. The light source can vary in wavelength and type. These instruments are sensitive to the grain size distribution of the material. This feature has little relevance in small watersheds because they are generally characterised by homogenous pluviometric regimes and erodibility but can be significant in large watersheds (Lenzi and Marchi, 2000). Variations in grain size distribution, associated with floods, can lead to nonlinear turbidity/suspended sediment concentration relationships. In practice, the degree of curvature is usually very small and linear models perform quite well in estimating loads (Lewis, 2003). Further interferences associated with optical measuring consist of the colour of the particles, the watery medium, the ratio between the size and shape of particles, the density of the particles and the content of dissolved particles which absorb light in some bands of the visible spectrum, altering the characteristics of the transmissed light (Gippel, 1989; Foster et al., 1992; Sadar, 1998). These interferences are minimized by the ratio detection system that uses a combined system of photodetectors and a specific algorithm that calculates the intensity of the scattered light (EPA, 1999; Sadar, 1999). Based on such considerations an experimental station was set up to measure the suspended sediment transport in the Carapelle stream (basin area: 506 km2). The expected maximum sediment concentrations were high (SSC > 15 g/l) and for this reason the Hach–Lange Solitax sc optical probe was chosen to measure the suspended sediment concentration (Bisantino et al., 2006). The probe was preliminary laboratory tested to verify its suitability to monitor suspended sediment transport in medium-size watersheds. The instrument operates in two different modes: the turbidity-nephelometric mode with detection at 90° and the suspended solids-ratio
2
F. Gentile et al. / Catena 80 (2010) 1–8
Fig. 1. The Carapelle watershed mouthed at the sediment transport experimental station.
detection system with detection at 140° and 90°. The tests were executed in order to evaluate the relationships between optical and gravimetric data varying the sediment concentration, and to assess the influence of the coarser materials (sand) on measurements (Gentile et al., 2007). Afterwards, the instrument was field tested through a calibration stage. The check of the housing device was also carried out in order to investigate any bias in concentration due to the sheltering tube. The most relevant flood events were then considered and the suspended sediment concentration, monitored at half-hourly scale, was plotted versus discharge to analyze the sediment transport dynamics.
2. Materials and methods 2.1. Experimental station measuring suspended sediment concentration
The station, already supplied with an ultrasound stage meter and a stage recorder, was in addition equipped with the Hach–Lange Solitax probe. The instrument has a LED (light-emitting diode) light source that emits a beam of infrared light into the stream at an angle of 45° to the probe face. A pair of photodetectors detects light scattered at 90° and 140° to the incident beam. In the turbidity-nephelometric mode the 90° photodetector solely works while in the suspended solids-ratio detection system the backscatter photodetector (140°) is also active. In the suspended mode the instrument's microprocessor uses an algorithm to ratio the signals from each detector and to convert them into suspended sediment concentrations (Sadar, 1998). This method enables a broad range of concentrations to be examined (0.001–4000 Nephelometric Turbidity Units – NTU – for turbidity and 0–150 g/l for suspended solids) and compensates for the
The experimental station measuring suspended sediment concentration is located in the Carapelle torrent (at Ordona-Castelluccio Dei Sauri bridge), one of the main water courses in Northern Apulia (Fig. 1; Table 1). The torrent originates in the Apennine mountains and crosses the Tavoliere flood plain before flowing into the Adriatic sea. The watershed is characterised by clayey-sandy Plio–Pleistocene sediments in the alluvial fan and by flyschoid formations in the mountainous areas, which are subject to erosion. The plain and the low hilly areas are mainly used for olive growing, whereas the higher slopes are occupied by woods and pasture. The climate is typically Mediterranean, with rainfalls ranging from 450 to 800 mm/year and average temperatures from 10 to16 °C. Table 1 Main characteristics of the Carapelle watershed. Watershed area Average altitude Main channel length Main channel slope Mean watershed slope
A Hm L ic iw
506.2 km2 466 m a.s.l. 52.16 km 16.3% 8.2%
Fig. 2. Sediment transport measuring station and probe housing device.
F. Gentile et al. / Catena 80 (2010) 1–8
3
Fig. 3. Data acquisition system and telemetry equipment.
effects of the stray light, organic matter and watery medium colour. Furthermore, the instrument is fitted with a programmable time interval screen wiper that prevents fouling. Time-consuming multipoint calibrations are replaced by a single correction factor. Through a pulley, a float and a counterweight group, the probe is housed in a tube to protect it from the impact of any flowing coarse material and from any potential measuring errors caused by incident radiant energy straying into the infrared field (Fig. 2). The probe is controlled by a data acquisition set and a telemetry system is also provided for remote measurements (Fig. 3). 2.2. Laboratory testing and field calibration of the optical probe The laboratory experiments were conducted to evaluate, with reference to river sediments, the functionality of the Solitax sc probe. The potential applications of the instrument range from clear water to strongly turbid liquid. It can be used in turbidity-nephelometric mode to predict SSC in low turbid water and in suspended sediment-ratio detection system mode when measuring in high concentrated sediments. The two-detector optical system allows to measure high
Fig. 4. Laboratory equipment.
SSCs and compensates for colour in the sample, light fluctuation, and stray light. In order to determine the relationships between the optical data and the gravimetric one, some tests were executed for 31 suspensions having fixed granulometric mixtures and 36 suspensions with varying ratios between sandy and fine fractions. The material was collected from the riverbed of the Carapelle stream. Granulometric analysis of the sediments, carried out by means of gravimetric method, revealed that the material was predominantly sandy (87.5%) and contained a small percentage of silt (8.8%) and clay (3.7%). The sample was split into three batches. The first one (OW—Original Wet) was kept wet and in its original granulometry, the second one (OD—Original Dried) was kept in its original granulometry too but it was oven-dried at 110 °C, the third one (SD—Separated Dried) was split into its fine and sand fractions (diameter d = 0.063 mm—UNI, BS Standards). From the OW batch 11 suspensions were prepared by picking up different amounts of sediment whose weights were determined at the end of the session through gravimetric analysis. The use of wet material ensured that no alterations to the grain structure were induced in the suspensions preparation. From the OD batch 20 suspensions with known SSC were constituted. Such preparation somehow altered the grain structure, but it allowed to better homogenize the sample and to obtain exactly the desired amount of sediment (in weight).
Fig. 5. Flow depth series and suspended sediment concentrations of the samples collected in the Carapelle stream.
4
F. Gentile et al. / Catena 80 (2010) 1–8
Fig. 6. Counterclockwise and clockwise loops.
The suspensions obtained from OW and OD batches have the same grain size distribution and sediment concentrations ranging from 0.1 to 150 g/l. Such a broad range was established taking into account the maximum values of sediment concentrations measured in the water courses of northern Puglia (90 g/l), and in order to test the manufacturer's specifications (0–150 g/l). From the SD batch 36 suspensions were prepared combining six weights and six sand percentages. The suspensions were prepared independently, rather than by successive dilutions of the same sediment batch, for two reasons. The former is that during stirring the sample did not remain in its initial conditions because of the separation of the aggregates. The latter is that progressive dilutions would require increasing volumes and then different boundary conditions. Equal volumes could be obtained by subsampling, but this often causes a biased sample to be extracted (Gray et al., 2000). The suspensions were stirred (Fig. 4) for ten minutes and the measures at five minute intervals were considered as representative data in order to get well mixed suspensions but not too altered by over-stirring. The first test was carried out on the OD and OW suspensions and the readings of the probe for turbidity and SSC were used to determine the relation between the probe outputs, in both the operating modes, and the gravimetric SSCs. The second test was executed on the SD suspensions and the readings of the probe for turbidity and SSC were used to evaluate the influence of the sand fraction on measures. The aims of the field calibration stage were to evaluate the efficacy of the housing system, to identify a calibration curve of the instrument for the specific torrent and to assess the type of the relationship between the optical and the gravimetric SSCs. This stage was executed monitoring nine flood events and collecting 65 hand samples (Fig. 5). The SSC of all samples was measured with a gravimetric method and compared with the data observed by the optical probe.
Fig. 7. Relationships between turbidity as measured by the instrument and gravimetric SSC.
Furthermore 17 pairs of samples were taken outside and inside the tube in order to check the housing device. Their SSCs were then compared in the laboratory to evaluate the occurrence of any alterations caused by the protective tube. Afterwards the calibration curves were compared with the results obtained during the laboratory stage. 2.3. Flood events During the monitoring seasons 2007–2008 nine significant events, having continuous data and absence of anomalies in the time series, were selected. The data, registered at a half-hourly scale, have been used to analyze the relationships between the SSC and the discharge (Q) during flood events. The loop-shaped curves, existing between SSC and Q, depend on the sediment availability that derives from the sediment supply and/ or from possible sediment deposits (Williams, 1989). In the clockwise loop a rapid increase of the suspended sediment at the beginning of the flood event occurs due to the presence of available sediment in the water course. If the sediment concentration decreases before the falling limb of the hydrograph, the sediment source areas can be considered limited. In the counterclockwise loop, instead, the sediment source areas are mainly located on hillslopes and the material moves slower than the peak discharge (Lenzi and Marchi, 2000). Fig. 6 reports two different flood events having counterclockwise and clockwise loops between suspended sediment concentration and discharge. 3. Results 3.1. Laboratory and field testing Figs. 7 and 8 show the results of the first laboratory tests. Using the instrument in turbidity-nephelometric mode the measurement increases up to a top SSC-value ranging between 40–50 g/l. As the concentrations further increase, the turbidity decreases. This trend
Fig. 8. Relationships between SSC as measured by the instrument and gravimetric SSC.
F. Gentile et al. / Catena 80 (2010) 1–8
5
Fig. 9. Relationships between the turbidity measured by the instrument and gravimetric SSC (OW and OD batches). Fig. 11. Comparison between the concentrations observed inside and outside the protective tube.
reveals the probe blindness occurring in concentrated suspensions already observed by other authors (Sadar, 1998; EPA, 1999). Using the instrument in suspended solids-ratio detection system blindness does not occur and the measurement has a linear increment with the sediment concentration. The results point out the good performance of the two-detector optical system especially for monitoring sediment transport in streams having high sediment concentrations as the Carapelle torrent. Figs. 9 and 10 report the results of the second tests. For mixtures at the same concentrations, the instrument's response depends on the percentage of sand. As expected according to the literature (Foster et al., 1992), the signal was greater in the fine material suspensions, except for the turbidity mode in the curve limbs affected by probe blindness. In particular, in the turbidity-nephelometric mode, by varying the percentage of sand in the mixture, the probe blindness occurs as from SSC = 7–8 g/l for the finest mixtures to higher values in the predominantly sandy ones (Fig. 9). In the suspended solids-ratio detection system blindness does not occur and the relationship between optical and gravimetric SSC is linear for each mixture (Fig. 10). These trends were already observed by other authors at lower SSCs (Sadar, 1998). The results of this stage show that the suspended solids-ratio detection system gives more reliable SSC values and that a linear conversion factor can be used for mixtures having the same granulometric composition. During the field calibration stage 65 samples were collected during the flood events. They mainly consisted of fine materials: the first 28 samples, joined together, contained 9.4% of sand, 32.9% of silt and 57.7% of clay. A comparison between the sediment concentrations inside and outside the protective tube was made, by taking 17 pairs of simultaneous samples. The external concentrations proved to be equal to the internal ones (R2 = 1), and the samples taken from both positions have been considered as part of the whole dataset used during the calibration stage (Fig. 11).
Nine flood events, having peak discharges between 0.94 and 44.62 m3/s, were examined. The relationship between sediment concentration and discharge revealed the existence of clockwise, counterclockwise and mixed-shaped hysteretic loops (Figs. 13–15). The counterclockwise hysteresis (Fig. 13) is predominantly related to events of moderate intensity that occur during the entire flood season. During these events small amounts of sediments move downstream and SSC increases in the falling limb of the hydrograph due to the distance between the sediment sources and the measuring station as observed by other authors (Brasington and Richards, 2000). It can be observed comparing the February 27 and the subsequent March 7 and 31, 2007 floods that the first flood of the year has higher rates of SSCs at equivalent flows than the following events. The reason can be referred to the dry weather preceding the first event and to the sediment depletion that influences the last ones (De Sutter et al., 2001; Bača, 2008). Considerable rates of suspended sediment were also observed in the moderate flood that occurred at the beginning of the flood season 2008 (January 24). A different behaviour was observed for the April 9, 2007 flood that starts on the falling limb of
Fig. 10. Relationships between the SSC measured by the instrument and gravimetric SSC (OW and OD batches).
Fig. 12. Linear relationship existing between the optical SSC and the gravimetric data.
The sediment concentrations of all the samples proved to be correlated with the data measured by the instrument, as shown in Fig. 12. In the range of the monitored discharges the results confirmed the existence of a very good linear relationship (R2 = 0.99) between the gravimetric and the optical data. This outcome allows to neglect the possible grain size variations, occurring during flood events, that can determine a curvilinear relationship between optical and gravimetric SSC.
3.2. Analysis of flood events
6
F. Gentile et al. / Catena 80 (2010) 1–8
Fig. 13. Counterclockwise loops.
F. Gentile et al. / Catena 80 (2010) 1–8
7
Fig. 14. Clockwise loop.
Fig. 15. Mixed loops.
the previous intense event (April 5, see Fig. 15), and for this reason suspended sediment concentration markedly increases. The most intense events (Figs. 14 and 15) are characterised by a clockwise or mixed loop. The clockwise loop occurs during a simple flood while the mixed loop is typical of graded floods. In the clockwise loop a rapid increase of the SSC in the raising limb of the hydrograph occurs. The SSC reaches a peak value, then decreases and the advance of the SSC maximum value towards the peak discharge can be observed. This condition occurs when the sediment availability in the watershed is limited, especially at the end of the rainy season when the depletion of sediments occurs as a consequence of the previous floods (Campbell, 1985). To estimate the dependency of the loop shape on the event intensity, the time delay T occurring between SSC and discharge peak
Fig. 16. Time delay between discharge and SSC peaks.
values was plotted versus peak discharge Qp. A decreasing trend (coefficient of determination R2 = 0.7) was observed (Fig. 16). The boundary between negative and positive values of the time T is representative of the condition at which the flow exerts a strong influence on the sediment load and determines the depletion of the watershed. 4. Conclusions The study examines the results of a sediment transport monitoring programme using high temporal resolution data. An experimental station was set up in the Carapelle torrent (Northern Apulia), equipped with an optical immersion probe which measures turbidity and suspended sediment concentration. The instrument was preliminary laboratory tested. The results showed that the two-detector optical system produces the best results and is more suitable for monitoring high suspended sediment transport in rivers because it provides a linear relationship between the optical and the gravimetric data. The field testing stage confirmed the results achieved in the laboratory, evidencing the linear relationship between the optical and the gravimetric data during flood events. The instrument housing device did not interfere with the measuring process. Nine flood events, collected over a 2-year period (2007–2008), were then analysed to investigate the relationships between SSC concentration and discharge (Q) in flood flows. The SSC–Q curves
8
F. Gentile et al. / Catena 80 (2010) 1–8
revealed that in the Carapelle stream clockwise, counterclockwise and mixed loops are possible and can be related to the event intensity and to the sediment supply of the watershed. In particular the counterclockwise loops occur during moderate events while clockwise and mixed loops are typical of intense floods that determine the sediment depletion of the watershed from fixed discharges. References Bača, P., 2008. Hysteresis effect in suspended sediment concentration in the Rybárik basin. Slovakia Hydrol. Sc. 53 (1), 224–235. Bisantino, T., Corbino, R., Gentile, F., Grittani, A., Milillo, F., Romano, G., Trisorio Liuzzi, G., & Zanframundo, P., 2006. Monitoraggio del trasporto solido nei bacini della Puglia settentrionale tra il Candelaro e l'Ofanto. Quaderni di Idronomia Montana. Ed. Bios. 26: 193–204. Brasington, J., Richards, K., 2000. Turbidity and suspended sediment dynamics in small catchments in the Nepal Middle Hills. Hydrol. Process. 14, 2559–2574. Campbell, G.S., 1985. Soil physics with basic transport models for soil–plant systems. Developments in Soil Science, vol. 14. Elsevier, New York. De Sutter, R., Verhoeven, R., Krein, A., 2001. Simulation of sediment transport during flood events: laboratory work and field experiment. Hydrol. Sci. 46 (4), 599–610. Foster, I.D.L, Millington, R., Grew, R.G., 1992. The impact of particle size controls on stream turbidity measurement; some implications for suspended sediment yield estimation. Erosion and Sediment Transport Monitoring Programmes in River Basins, vol. 210, pp. 51–62. Gentile, F., Bisantino, T., Milillo, F., Romano, G., Trisorio Liuzzi, G., 2007. Caratteristiche funzionali di una sonda a raggi infrarossi per la misura del trasporto solido in sospensione nei corsi d'acqua. Quaderni di Idronomia Montana. Ed. Bios. 27:421–432. Gippel, C. J., 1989. The use of turbidity instruments to measure stream water suspended sediment concentration. Monograph Series No. 4, Dept of Geography and
Oceanography, University College, University of New South Wales, Australian Defence Force Academy, Canberra, 204 pp. Gippel, C.J., 1995. Potential of turbidity monitoring for measuring the transport of suspended solids in streams. Hydrol. Process. 9, 83–97. Gray, J.R., Glysson, G.D., Torcios, L.M., Schwartz, G., 2000. Comparability of suspendedsediment concentration and total suspended solids data. Water Resources Investigation Rep. 00-4191. InUSGS, Washington, DC. Grayson, R.B., Finlayson, B.L., Gippel, C.J., Hart, B.T., 1996. The potential of field turbidity measurements for the computation of total phosphorus and suspended solids loads. J. Environ. Manag. 47, 257–267. Lenzi, M.A., Marchi, L., 2000. Suspended sediment load during floods in a small stream of the Dolomites northeastern Italy. Catena 39, 267–282. Lewis, J., 1996. Turbidity-controlled suspended sediment sampling for runoff-event load estimation. Water Resour. Res. 32 (7), 2299–2310. Lewis, J., 2003. Turbidity-controlled sampling for suspended sediment load estimation. In: Bogen, J., Tharan, Fergus, Des, Walling (Eds.), Erosion and Sediment Transport Measurement in Rivers: Technological and Methodological Advances. Proc. Oslo Workshop, 19–20 June 2002, vol. 283. IAHS Publ, pp. 13–20. Marchi, L., Trisorio-Liuzzi, G., Zanframundo, P., 1986. Analisi dei deflussi torbidi nei piccoli bacini del Sub- Appennino dauno. Quaderni di Idronomia Montana. Ed. Bios. 6:95–121. Walling, D.E., Webb, B.W., Woodward, J.C., 1992. Some sampling considerations in the design of effective strategies for monitoring sediment-associated transport. IAHS Spec. Publ. 210, 279–288. Williams, G.P., 1989. Sediment concentration versus water discharge during single hydrologic events in rivers. J. Hydrol. 111, 89–106. Sadar, M.J., 1998. Turbidity science. In: Technical Information Series. Hach Company. 26 pp. http://www.hach.com. Sadar, M.J., 1999. Turbidimeter Instrument Comparison: Low-level Sample Measurements. Technical Information Series. Hach Company. 56 pp. http://www.hach.com/. U. S. Environmental Protection Agency, 1999. Basic turbidimeter design and concepts. Guidance Manual Turbidity Provisions, chap. 11.