DETERMINATION OF INFLUENCE OF VEGETATION ON THE FRICTION FACTORS OF THE BIEBRZA RIVER

Dorota Swiatek¹, Ronny Verhoeven², Jarosław Chormański¹, Tomasz Okruszko¹, Stefan Ignar¹, Robert Banasiak^2
1 Department of Hydraulic Engineering and Environmental Reclamation, Warsaw Agricultural University, Poland
² Hydraulics laboratory, TW15, Ghent University - Belgium

ABSTRACT

The Biebrza River, situated in the north-eastern part of Poland, flows through the last extensive, fairly undisturbed river-marginal wetlands in Europe. Its valley intensively drains the surrounding plateau and the outwash plain into the river and creates a very complex ground and surface water system. This study focuses on the upstream part of the river between Lipsk and Sztabin with a length of about 35 km and deals with the problem of determining the Manning friction factor as a function of vegetation and discharge or water levels. During intensive field measurement campaigns information on the river topography and on the hydraulic characteristics of river flow has been gathered. Special attention is paid to to the impact of dense vegetation on the discharge characteristics. The hydraulic data analysys focuses on the determination of the friction coefficients using four methods for steady state conditions.

Key words: flow resistance, wetlands, GIS techniques.

INTRODUCTION

The Biebrza Wetlands are situated in the northeastern part of Poland in an ice-marginal valley, some 195 000 hectares in area; the wetlands occupy some 116 000 hectares. It forms the last extensive, fairly undisturbed river-marginal peatland in Europe, containing endangered plant and animal species in a large variety of fully developed ecosystems. It is internationally recognised as a reference area for restoration of deteriorated peatlands. Despite of extensive reclamation of its tributaries the main course of the Biebrza River remains almost untact. Since establishing the Biebrza National Park (BNP) in 1993 even dregging and water plant removal is forbidden in the whole river channel.

Biebrza Valley, according to the geomorphologic description, is an extensive depression formed during the last glacial period and is filled with several thick deposits of fluvioglacial sands. The Biebrza Valley was divided [8] into basins using a relation of the higher order morphologic features such as ancient and young morainic plates, glacier outwash plains or river flood terraces. There are three basins, respectively identified as the Upper Basin with the Augustow outwash, reaching from the springs of Biebrza to the mouth of the Netta River; the Middle Basin covering the area from Netta to the mouth of Rudzki Channel and the Lower Basin situated in the southern part of the valley up to the alluvial cone of the Narew River (Fig. 1).

The area of this study, the Upper Basin, is relatively long (40 km) and narrow (2–3 km), with two basin-lake widening: in its middle part and in the transition zone to the Middle Basin. The valley intensively drains the surrounding plateau and the outwash plain into the river, forming mires of so called soligenic type of hydrological feeding. As a result of this most of the valley is covered by predominates, the deep layers of sphagnum peat, only in the viscinity of the river reed peat and alluvial soils mirror the area flooded by the surface waters.

Fig. 1. Map of the three basin of the Biebrza river: upper, middle and lower

The land reclamation systems, which drain over 50% of the surface of the valley, are not intensive; moreover their impact strictly depends on hydraulic conductivity of the drainage channels and the Biebrza riverbed, which serves as a water recipient. During the summer water plant development increases the water stages and controls the drainage (Photo 1) intensity. There is a natural expectation from the farmers to have the drainage ditches free from water plants and mud.

The ban on cleaning the ditches and river channel imposed by the BNP is justified as the water plants serve as a fish shelter and as an obstacle maintaining the high water level in the river and thus the high soil moisture of the peat in the valley. On the other hand the extensive farming activity i.e. haymaking on the natural meadows is needed in controlling the scrub encroachment, which eliminates the open habitat of the sedge moss fen. So, optimum moisture of the peat in this area means not only the optimum water conditions for the peat development, but also for extensive, grasslands farming. In this case a question of the impact of the water plants development on the river stages is crutial for formulation of the management options.

Photo 1. Intensive plant development in the upper course of the Biebrza River

BRIEF OVERVIEW OF THE RESEARCH PROJECT

The research project on the Biebrza Wetlands was performed in the frame of a bilateral international scientific and technological cooperation agreement concluded between the Polish and the Flemish government and entitled ‘Environmental river catchment protection by natural or artificial wetlands’. The main purpose of the project was developing the tools helpful in water management for this and similar sites. At first step three Flemish and a Polish research teams gathered existing knowledge on main water management problems and extended it by extensive field measurement campaign. Once the information on the topography of the Upper Basin was available and the first calculations of the water surface profile were made, it became clear that there was a huge influence of the value of the friction factor and that this aspect needed special attention for the future development of an accurate mathematical simulation model. So it become clear that gathering topological data, building of accurate DEM and calculation of friction coefficients is necessary prerequisite for any further modelling efforts.

THE FIELD MEASUREMENT CAMPAIGNS AND ELABORATION OF THE DATA BASE

During the field measurement campaigns topographical and the hydraulic data have been collected. Five intensive field-measuring campaigns have been performed in the area. Unfortunately all campaigns, respectively in September1999, May 2000, May 2001, June 2002 and April 2003 have been performed under low flow conditions. Basically, the measurements were focussed on two main objectives: the collection of topographical information on a large number of cross-sections and the determination of surface water flow data under different conditions [7].

TOPOGRAPHICAL DATA COLLECTION

As the numerical model both needs to simulate, the flow in the river and in the complex system of floodplains, detailed topographical information is needed. For this, a Digital Elevation Model (DEM) of the study area was developed, using the ArcInfo 7.2 TOPOGRID procedure [6]. This method, originally developed to interpolate the DEM for hydrologic modelling purposes, was also several times successfully applied to construct a model of floodplain topography [1,2]. For the elaboration of the DEM, the first necessary data were collected by field measurements and map analysis. Consequently, the Digital Elevation Model has been created as a raster model by spatial interpolation. The main data source for the model generation was a set of contour lines elaborated in digital form on the basis of topographic maps with a scale of 1: 25 000. The DEM of the research area was created as a raster network with a cell resolution of 25 meters. Because the topographic maps do not accurately show the flat areas at the valley bottom, the construction of a proper elevation model needs extension of the available data with topographical information from field measurements. As the flow mainly happens in the minor riverbed, the efforts were focussed on the area close to the river channel. The information gathered during the field campaigns is then combined with data available from topographical maps.

Fig. 2 shows a way of the topography measurements on example of about 20 km long part of the Biebrza valley between Jastrzebna and Sztabin. Data on six cross-sections have been collected. In-between each two cross-sections the water level and riverbank elevation was measured every few hundred meters in selected spots. For this a GPS system (Topcon Legacy E receiver) working in Differential RTK mode was used, which offering accuracies of 0.5–1cm in the horizontal plane and 2-3cm in the vertical direction [3].

Fig. 3 shows the digital elevation model (DEM) of the same part of the valley, which combines measured data with map data on the floodplain. This DEM was used for determination of the additional elevation cross-sections, The accuracy of the latter is less than the data obtained from GPS, but still largely sufficient as far as the simulation of flood plain flow and storage is concerned. Fig. 4 gives an example of the cross-section at Ostrowie realized by combination of both above-mentioned techniques GPS measurement and DEM analysing.

Fig. 2. Cross-sections and points of elevation measurements in the upstream part of the Biebrza River between Jastrzebna (Ostrowie) and Sztabin

Fig. 3. The DEM of the upstream part of the Biebrza River valley between Jastrzebna and Sztabin

Fig. 4. An example of the cross-section geometry at Ostrowie location as a combination of GPS and GIS data collection

SURFACE WATER FLOW DATA COLLECTING

During the measurment campaigns local surface water flow was recorded in several, discrete cross-sections along the Biebrza River and its main tributaries. Fig. 5 gives a schematic overview of the measurement locations and of all discharges that have been recorded during the field campaigns. The local water surface levels and the longitudinal channel shape measurements were performed with assessment of a general geodetic levelling, making use of, gauging nails, adjusted to the water surface level at the start of each discharge measurement and by this making the hydraulic gauging team independent from the geographical one. The discharge measurement equipment contains, depending on local physical and flow conditions, a range of velocity meters and level gauges: several OTT propeller meters, a one dimensional electromagnetic current meter OTT Nautilus 2000 and a two-dimensional electromagnetic current meter Valeport 802.

Fig. 5. Discharges as measured during the five field campaigns

The measurements were applying the method of integration of the velocity field over the cross-section. As in many locations severe difficulties with vegetation were encountered, first the gauging section had been cleared from plants then two discharge recordings were made, each of them in different verticals. Averaging the result of both recordings should lead to more reliable discharge values. In some cases (sections) the quality of measurements was badly influenced by the very slow velocities and/or dense vegetation. Although the measurement campaign periods were focusing on different flow regimes, up till now the research teams found themselves confronted with low to medium flow. The track length of the river segments between consecutive cross-sections has been measured in ArcView GIS using the digital topographical maps with a scale of 1:10 000.

DETERMINATION OF THE FRICTION COEFFICIENTS

Once all discharge data have been processed, the friction coefficient of each part of the rivernetwork was determined in a different ways. In this study four approaches to calculate the value of the Manning n have been used and evaluated. As previously mentioned the evaluation is restricted to the 35 km long upper course of the Biebrza River. The three cross-sections where used in order to crefully compare the results of this exercise. There are sections of Lipsk, Ostrowie and Sztabin presented on Figs. 2 and 6.

Fig. 6. Cross-section at Lipsk, Ostrowie and Sztabin

UNIFORM FLOW METHOD – TWO APPROACHES

As the field campaigns were performed in periods of stable water conditions, one can suppose the flow regime do be in a steady state with uniform flow and apply the Manning formula in each observation point to calculate the friction coefficient.

where:
S_f – friction slope [-],
n – Manning coefficient [m^-1/3s],
P – wetted perimeter cross section [m],
A – wetted area cross section [m²],
Q – discharge [m³·s^-1].

The friction slope value of the upstream reach was used for the calculation of n in the different cross-sections, except for the most upstream section, where S_f from the downstream reach was applied. The n-values in the nods between to sections were linearly interpolated. Table 1 gives an overview of the Manning n coefficients as calculated for the different locations and discharges. The n values vary substantial from one section to the other and also from one year to the other. The reason for this should be found in the fact that the cross-sections are quite different, that the vegetation is very unequally present over the years, and that the friction slope S_f is averaged over a very big distance (Lipsk – Ostrowie: 17.89 km; Ostrowie – Sztabin: 17.11 m). As a result, the quality of the simulation of the steady state water surface profile on the basis of these Manning coefficients is rather doubtful.

Table 1. Friction coefficient n calculated from Manning’s formula (‘x’ means that the field data proved to be unreliable)

Cross section name	1999	2000	2001	2002	2003	1999	2000	2001	2002	2003
	Q[m3·s-1]					n[m-1/3s]
Lipsk	0.497		1.01	0.488	1.79	0.137		0.030	0.276	0.040
Ostrowie	0.779	0.860	1.09	0.538	2.36	0.276	0.106	0.069	0.651	0.127
Sztabin	1.42	1.15	0.890	x	3.04	0.234	0.090	0.369	x	0.079

In order to overcome the problem of averaging of n over too long distances, during the 2003 field campaign, the local friction slopes have been additionally measured at each cross-section. It was done by levelling the water stages 300m up- and downstream of each section. Table 2 compares the results obtained by this approach with the previous one.

Table 2. Friction coefficient n calculated from Manning’s formula using different friction slope definitions

2003	Q[m3·s-1]	Full reach (FR)	Local slope (LS)	Difference
		n	n	%
Lipsk	1.79	0.040	0.052	23
Ostrowie	2.36	0.127	0.172	26
Sztabin	3.04	0.079	0.114	31

COWAN’S METHOD

The second method used to determine the n values was developed by Cowan [4]. He defines n as:

n = (n₀ + n₁ + n₂ + n₃ + n₄)· n₅

where:
n₀ – influence of bed material: 0.020–0.028,
n₁ – degree of irregularity of the longitudinal profile: 0.000–0.020,
n₂ – degree of variation of the cross-section: 0.000–0.015,
n₃ – relative effect of obstructions: 0.000–0.060,
n₄ – influence of vegetation: 0.005–0.100,
n₅ – degree of meandering: 1.000–1.300.

It is obvious that this approach leaves a big degree of freedom to the user and that for the application of this method a lot of practical experience is needed. In case of the upper Biebrza, we can state that as the bed consists mainly of coarse sand, n₀ can be put to 0.024; as the longitudinal profile is very irregular, we take 0.015 for n₁; the cross-section varies quite a lot, so n₂=0.012; as there are not too many obstructions n₃ will be rather small (=0.01); as there is a lot of vegetation n₄ will have a high value and as the reach between Lipsk and Ostrowie has smaller cross-sections and water depths the n₄ values will be higher for this part (=0.100)than for the stretch between Ostrowie and Sztabin(=0.075); and finally as the Biebrza river is meandering a lot, we put n₅ to 1.2.

Applying these values in the Cowan formula leads to a Manning n value of 0.193 for the reach Lipsk – Ostrowie and of 0.163 for the part between Ostrowie and Sztabin.

WATER SURFACE PROFILE THEORY

The third method uses the theory of hydraulic water surface profiles. Starting from the data on the geometry, the water levels and the discharge, the Bresse equation [5] allows calculating the friction coefficient n in a river reach between two gauging locations.

where:
h – water depth [m];,
x – distance along the river [m],
S₀ – bottom slope [–],
n – Manning’s factor [m^-1/3s],
g – gravitational acceleration[m·s^-2],
R – hydraulic radius [m]; Q – discharge [m³·s^-1],
A – flow area [m²],
B – width at the water surface [m].

Friction coefficient n obtained by this approach is presented in Table 3.

Table 3. Friction coefficient n calculated from water surface profiles

2003	Q[m3·s-1]	n
Lipsk	1.79	0.06
Ostrowie	2.36	0.10
Sztabin	3.04	0.11

RESULTS

The four methods calculation of friction factor leads to the different Manning’s n values. Values determined by Cowan method significantly exceed coefficients calculated by others methods (Fig. 7).

Fig. 7. Calculated values of Manning friction factor for selected methods

In Fig. 8 the results of the water surface line calculation with n obtained from both approaches based on the Manning formula are compared with the measured water profile of 2003. The first approach where n is determined using the overall friction slope clearly offers better results than the first-one. From this one can conclude that the method making use of a linear interpolation of the friction factors, which have been calculated from the overall situation between 2 gauging cross-sections, leads to a more reliable prediction of the water surface profile than the method, which determines n as a local value at each gauging cross-section. This can be explained by the fact that averaging n over a whole reach has the advantage of taking account of all friction effects, which influence the slope of the surface profile.

Fig. 8. Comparison measured surface profiles with those obtained with a friction factor value calculated from Manning’s formula

Fig. 9 shows the results with n obtained from Cowan’s method and from calibration based on Bresses equation. It is clear that from comparison of all four methods the latter offers the best results. This can be explained by the fact that determining the friction factor by calibration using Bresses equation, based on field observations offers the opportunity to account for all local influences, however it “fits” to the certain set of the measured points.

Fig. 10 comparing the quality of the results from all four methods puts this conclusion to evidence.

Fig. 9. Comparison of measured surface profiles with those obtained with a friction factor value calculated from Cowan’s method and from application of Bresses formula

Fig. 10. Root mean square error for the four methods

Fig. 11. Water surface profiles as measured during the five field campaigns

Fig. 11 gives an overview of the water surface profiles as measured during the five field campaigns. One easily sees that there is small difference in the water stages over the five years, although Fig. 5 shows quite different discharges. In 2003 the discharges measured in late April, with no vegetation in the river were about the triple from those measured in June 2002, with full vegetation development. From this it becomes clear that vegetation development has a huge impact on the friction factor and that reliable hydrodynamic modelling of these kinds of rivers requires a sound knowledge of the friction factor variation as a function of the time of the year. As plants start to grow in late April till early May and come to full growth in the month of June, field campaigns must focus on this period in order to earn reliable data on the impact of vegetation growth on the friction factor value.

CONCLUSIONS

In this paper the basic requirements to develop a database to support a numerical model of a wetland river and its flood plains are discussed. It is shown how an adequate combination of GPS and GIS techniques helps to develop a topographical database with reasonable effort. The combination of both techniques offers the opportunity to assure great accuracy in these parts of the river basin, which are of major importance for the simulation of flow, and to guarantee sufficient information on those parts of the basin where storage capacity is important.

From different field measurement campaigns it became clear that vegetation growth is seriously influencing the variation of the friction characteristics. Different approaches for the determination of the friction factor value have been evaluated. From this it became clear that calibration of the friction factor on the basis of Bresses equation offers very good and reliable results. The knowledge of the variation of the friction factor as a function of the period of the year is an aspect of major importance for the elaboration of a numerical hydrodynamic water resources management model.

REFERENCES

1. Cera T.B., Tremwel T.K., Burleson R.W., 1996. Use Arc/INFO, EPA-SWMM, and UNIX text processing tools to determine flood extend. AWRA Symposioum on GIS and Water Resources. Sept 22-26, 1996FT. Lauderdale, USA. http://www.awra.org/proceedings/gis32/cera/index.html.

2. Chormanski J., 2003. Methodology for flood extent determination in the Lower Biebrza Basin. PhD thesis. Warsaw Agricultural University.

3. Chormański J., Kowalewski K., Mazippus M., 2003. Application of GPS techniques for water stage measurements and river slope calculation in wetland area of Upper Biebrza, Measurements techniques and data assessment in wetlands hydrology. Eds. S. Ignar, P. Nowakowski, T. Okruszko, Warsaw Agricultural Press, Warsaw 53-60.

4. Cowan, W. L., 1956. Estimating hydraulic roughness coefficients, Agricultural Engineering 37(7), 473-475.

5. French R.H., 1985. Open channel hydraulics. Mc Graw-Hill, New York.

6. Hutchinson M.F., 1996. A locally adaptive approach to the interpolation of digital elevation model. Third International Conference/Workshop on Integrating GIS and Environmental Modeling, Santa Fe, NM, January 21-26, 1996. Santa Barbara, CA, National Center for Geographic Information and Analysis. http://www.ncgia.ucsb.edu/conf/SANTA_FE_CD-ROM/sf_papers/hutchinson_ michael_dem/local.html.

7. Huygens M., Okruszko T., Batelaan O., Verhoeven R., 2000. Field monitoring in the upper catchments of the Biebrza wetlands. Conf. On ‘Monitoring and modelling catchment water quantity and quality’, Ghent 27-29/09/2000, 263-266.

8. Zurek S., 1984. Relief, Geologic structure and Hydrography of the Biebrza ice-marginal Valley. Polish ecological studies 10(3,4), 39-51.

 

Accepted for print: 25.10.2007

---

Dorota Swiatek
Department of Hydraulic Engineering and Environmental Reclamation,
Warsaw Agricultural University, Poland
Nowoursynowska 159, 02-787 Warsaw, Poland
Phone: (+48 22) 593 53 13
email: dorotams@levis.sggw.waw.pl

Ronny Verhoeven
Hydraulics laboratory, TW15, Ghent University - Belgium


Jarosław Chormański
Department of Hydraulic Engineering and Environmental Reclamation,
Warsaw Agricultural University, Poland
Nowoursynowska 159, 02-787 Warsaw, Poland

Tomasz Okruszko
Department of Hydraulic Engineering and Environmental Reclamation,
Warsaw Agricultural University, Poland
Nowoursynowska 159, 02-787 Warsaw, Poland

Stefan Ignar
Department of Hydraulic Engineering and Environmental Reclamation,
Warsaw Agricultural University, Poland
Nowoursynowska 159, 02-787 Warsaw, Poland

Robert Banasiak
Hydraulics laboratory, TW15, Ghent University - Belgium

---

Responses to this article, comments are invited and should be submitted within three months of the publication of the article. If accepted for publication, they will be published in the chapter headed 'Discussions' and hyperlinked to the article.

---

Main  -  Issues  -  How to Submit  -  From the Publisher  -  Search  -  Subscription