Science of the Total Environment 511 (2015) 756–766

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Impact of papyrus wetland encroachment on spatial and temporal variabilities of stream flow and sediment export from wet tropical catchments N. Ryken a,b,⁎, M. Vanmaercke a,c, J. Wanyama d, M. Isabirye e, S. Vanonckelen a, J. Deckers a, J. Poesen a a

Department of Earth and Environmental Sciences, KU Leuven, Belgium Department of Soil Management, Ghent University, Belgium c Research Foundation Flanders (FWO), Brussels, Belgium d Department of Agricultural and Biosystems Engineering, Makerere University, Uganda e Busitema University, Namasagali, Uganda b

H I G H L I G H T S • • • • •

Eight representative catchments were selected in the Lake Victoria basin. Runoff discharge and suspended sediment yield (SY) were monitored during one year. The presence and status of papyrus wetlands greatly influenced runoff discharge and SY. Intact wetlands buffer discharge, trap sediments and decrease connectivity. Wetlands can play a crucial role in catchment management strategies.

a r t i c l e

i n f o

Article history: Received 1 October 2014 Received in revised form 9 December 2014 Accepted 17 December 2014 Available online 21 January 2015 Editor: D. Barcelo Keywords: Runoff Suspended sediment yield Riparian vegetation Gully Riverbank erosion Uganda

a b s t r a c t During the past decades, land use change in the Lake Victoria basin has significantly increased the sediment fluxes to the lake. These sediments as well as their associated nutrients and pollutants affect the food and water security of millions of people in one of Africa's most densely populated regions. Adequate catchment management strategies, based on a thorough understanding of the factors controlling runoff and sediment discharge are therefore crucial. Nonetheless, studies on the magnitude and dynamics of runoff and sediment discharge are very scarce for the Lake Victoria basin and the African Rift region. We therefore conducted runoff discharge and sediment export measurements in the Upper Rwizi, a catchment in Southwest Uganda, which is representative for the Lake Victoria basin. Land use in this catchment is characterized by grazing area on the high plateaus, banana cropping on the slopes and Cyperus papyrus L. wetlands in the valley bottoms. Due to an increasing population pressure, these papyrus wetlands are currently encroached and transformed into pasture and cropland. Seven subcatchments (358 km2–2120 km2), with different degrees of wetland encroachment, were monitored during the hydrological year June 2009–May 2010. Our results indicate that, due to their strong buffering capacity, papyrus wetlands have a first-order control on runoff and sediment discharge. Subcatchments with intact wetlands have a slower rainfall–runoff response, smaller peak runoff discharges, lower rainfall–runoff ratios and significantly smaller suspended sediment concentrations. This is also reflected in the measured annual area-specific suspended sediment yields (SYs): subcatchments with encroached papyrus swamps have SY values that are about three times larger compared to catchments with intact papyrus vegetation (respectively 106–137 ton km−2 y−1 versus 34–37 ton km−2y−1). We therefore argue that protecting and (where possible) rehabilitating these papyrus wetlands should be a corner stone of catchment management strategies in the Lake Victoria basin. © 2014 Elsevier B.V. All rights reserved.

1. Introduction ⁎ Corresponding author at: Department of Soil Management, Ghent University, Belgium.

http://dx.doi.org/10.1016/j.scitotenv.2014.12.048 0048-9697/© 2014 Elsevier B.V. All rights reserved.

The large and rapidly growing population and the high level of poverty in the Lake Victoria basin have led to a high pressure on

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

the environment (e.g. Odada et al., 2004; Muyodi et al., 2010), resulting in intense soil erosion and land degradation (e.g. Ntiba et al., 2001; Knapen et al., 2006). Although quantitative data on sediment fluxes are scarce, several studies clearly indicate that deforestation and intensified agriculture in this region have led to a strongly increased sediment input into Lake Victoria over the last decades (e.g. Ntiba et al., 2001; Verschuren et al., 2002; Odada et al., 2004; Swallow et al., 2009). Various additional challenges are associated with this increased sediment input. For example, sediment-bound nutrients led to important eutrophication problems in Lake Victoria, such as the very extensive bloom of water hyacinths (Witte et al., 1992; Kitaka et al., 2002; Swallow et al., 2003). This, in combination with a decrease in water transparency due the increased turbidity, the onset of deep water anoxia and the introduction of the Nile Perch, caused a collapse of the natural ecosystem of Lake Victoria (e.g. Witte et al., 1992; Hecky, 1993; Lehman, 1996; Ogutu-Ohwayo et al., 1997; Verschuren et al., 2002). Also sediment-fixed pollutants (e.g. mercury) form an important ecological threat to the lake, while the increases in sediment loads have important impacts on water treatment costs (e.g. Mwamburi, 2003; Campbell et al., 2003; Muyodi et al., 2010). Apart from, but closely linked to these sediment-related impacts, the occurrence of floods forms an important problem in the Lake Victoria basin (e.g. Swallow et al., 2009). The frequency, magnitude and impacts of these floods are expected to increase as a result of the predicted climate changes, growing population pressure, the resulting urbanization and other land use changes (e.g. Douglas et al., 2008; Swallow et al., 2009; Lwasa, 2010; Khan et al., 2011). Overall, these problems pose important threats to food and water security, the livelihood of the rapidly increasing population and the socio-economic stability of the region (Ntiba et al., 2001; Verschuren et al., 2002; Odada et al., 2004; Morrison and Harper, 2009; Swallow et al., 2009; Muyodi et al., 2010). Reliable data on runoff and sediment transport are essential to address these problems of flooding, high sediment concentration and eutrophication in the Lake Victoria basin (e.g. Swallow et al., 2009). However, measured data are very scarce for Eastern and sub-Saharan Africa in general and the Lake Victoria basin in particular (Walling and Webb, 1996; Odada et al., 2004; Verschuren et al., 2002; Vanmaercke et al., 2014a). As an alternative, several studies have attempted to model soil erosion and sediment yield in the region of Lake Victoria (e.g. Lufafa et al., 2003; Cohen et al., 2005; Swallow et al., 2009). However, results obtained from such studies are commonly associated with very large uncertainties, since the applied models were not calibrated for the environmental conditions under consideration (e.g. Maetens et al., 2012; De Vente et al., 2013; Vanmaercke et al., 2014a). For example, Cohen et al. (2005) evaluated the effectiveness of the USLE model in Kenya with field observations and concluded that this model only predicts a small fraction (~30%) of the observed erosion patterns. Field observations therefore remain essential for the calibration and validation of erosion and catchment sediment yield models (e.g. De Vente et al., 2013). One aspect that remains particularly unclear is the impact of wetlands on runoff and sediment discharge. Wetlands of Cyperus papyrus L. are important regulators of nutrient cycles (e.g. Brix, 1994; Kelderman et al., 2007), play a significant role in the conservation of local biodiversity (e.g. Kansiime et al., 2007) and serve as habitat for fish larvae (Mnaya and Wolanski, 2002). Apart from these ecological functions, it has also been argued that these wetlands function as sediment trap and buffer peak discharges (e.g. Cooper et al., 1987; Swallow et al., 2003). Hence, they may be of vital importance in addressing the sediment- and flood-related problems discussed above. Unfortunately, many of the papyrus wetlands in the Lake Victoria basin are under large pressure. At numerous locations, papyrus wetlands are drained, burned and cleared for human settlement and agricultural activities (crop production, cattle grazing), exploited for fuel or building material or as landfill (e.g. Bolwig, 2002; Swallow et al.,

757

2003, 2009; Rwakakamba, 2009; Muyodi et al., 2010; Nakiyemba et al., 2013). Whereas the effects of land use on sediment input in Lake Victoria is recognized (but understudied), the role of the papyrus wetlands and its encroachment is rarely indicated as a potential threat (e.g. Cooper et al., 1987; Boar and Harper, 2002; Mwanuzi et al., 2003; Van Dam, et al., 2007; Kelderman et al., 2007; Morrison and Harper, 2009). This may be partly explained by the fact that no quantitative data exist on the potential runoff buffering or sediment-trapping capacity of these wetlands. As a result, the potential effects that the degradation of these wetlands may have on river flow and sediment discharges are poorly understood. Hence, there is an urgent need for quantitative runoff and sediment discharge data for catchments in the Lake Victoria basin to allow for the development of effective and feasible catchment management strategies and (in particular) to better understand the possible impact of papyrus wetland encroachment on runoff and sediment discharge. Therefore, the objectives of this study were (i) to obtain runoff and sediment discharge measurements for 8 selected subcatchments in the Lake Victoria basin, having different degrees of wetland encroachment; (ii) to explore the spatial and temporal variabilities in runoff and sediment discharge of these subcatchments; and (iii) to evaluate the potential effect of papyrus wetland encroachment on runoff and sediment discharge. 2. Study area This study was conducted in the upper Rwizi catchment. This catchment of 2282 km2 is located in the South-West of Uganda and is part of the Lake Victoria basin (Fig. 1). Seven selected subcatchments, having different degrees of wetland encroachment, of the Upper Rwizi were monitored during one hydrological year (1 June 2009–31 May 2010). Six of these subcatchments are nested (Fig. 1 and Table 1). The Upper Rwizi catchment is characterized by a rugged topography in the north and the south of the catchment, while the central part consists of a plain. Altitudes in the catchment range between 1248 m and 2159 m a.s.l. (Table 1). The geology of the upper Rwizi catchment is dominated by phyllite rocks. Only in the northern part some granite and sandstone occur (Harrop, 1960). Dominant lithologies per subcatchment are shown in Table 1. The mayor soil types of the upper Rwizi catchment are Ferralsols in the central plain, Regosols in the mountainous northern and southern part of the catchment and Histosols in the papyrus wetlands (Chenery, 1960). The average annual precipitation in the upper Rwizi is ca. 1150 mm (FAO, 2005). Rainfall is spatially relatively homogenous and ranges from 1216 mm to 1683 mm between the different subcatchments (Table 1), but is characterized by strong seasonal contrasts: average monthly rainfall shows a bimodal distribution with rainfall maxima from March to May and from September to November. Two dry periods separate these rainy seasons. The mean annual air temperature is 20 to 21 °C and varies relatively little throughout the year (FAO, 2005). Land use in the Upper Rwizi shows a typical topographic sequence, with grazing areas on the high plateaus, banana cropping at the foot slopes and C. papyrus L. wetlands in the river valley bottoms (alluvial plains). Many of these papyrus wetlands are currently encroached and transformed into cropland or grazing area (e.g. Bolwig, 2002; Rwakakamba, 2009; Vanonckelen, 2009; Nakiyemba et al., 2013). Table 1 shows the relative area which is occupied by the different land use types for each subcatchment. These numbers are based on an edited and field-verified version of the Africover map (2002), by Vanonckelen (2009). Due to the low resolution of this map and the rapid rate at which wetlands are converted in some regions, exact numbers on the extent of wetland encroachment in each subcatchment were unavailable. Therefore, the state of the wetlands in each catchment was described using three robust classes (Table 1; Fig. 2): severely encroached (less than 20% of the wetlands are remaining), moderately encroached (20–80% of the wetlands are remaining) or minimally

758

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

Fig. 1. Location, hydrography and topography of the upper Rwizi catchment within the Lake Victoria basin and the seven subcatchments with their gauging stations. See Table 1 for the characteristics of the subcatchments.

encroached (more than 80% of the wetlands remain). These classes were assigned, based on field surveys conducted before and during the monitoring campaign (June 2009–May 2010). The surveys mainly focused

on the 2 to 5 km long river reaches upstream of the measuring stations, as these reaches were expected to be most relevant with respect to sediment trapping and runoff discharge buffering.

Table 1 Characteristics of the studied Upper-Rwizi sub-catchments. Monitoring station

Ruharo

Rugurama

Nyeihanga

Ndaija

Buhanama

Rwamunena

Koga

Abbreviation Monitored tributaries Location (S°), (E°)

RUH RUG, NYE, NDA, BUH, RWA, KOG 0.61525 30.68681 1348–2169 13.18 2120.8

RUG NYE, NDA, BUH, RWA,

NYE NDA, BUH, RWA,

NDA BUH, RWA,

BUH RWA,

RWA

KOG

0.63134 30.44625 1392–2015 14.94 1092.3

0.68251 30.38942 1404–2015 16.71 859.1

0.7233 30.35022 1426–2015 16.74 683.2

0.74976 30.31859 1435–2015 16.00 529.9

0.67486 30.2783 1447–2015 15.13 358.5

0.57882 30,45062 1409–2169 14.65 378.7

0.7 1.2 93 5.1

1.2 1.9 91.1 5.8

2.4 4.2 81.1 12.3

3.8 7.1 70.5 18.6

4.3 7.9 70.6 17.2

6.1 11.6 69.3 13.0

0 0 84.3 15.7

0.7 40.4 47.2 11.5 0.2 Severe encroachment

0.5 40.1 49.9 9.4 0.1 Severe encroachment

0.6 41.7 49.3 8.2 0.2 Moderate encroachment

0.3 38.3 52.6 8.8 0 Moderate encroachment

0.4 31.5 62.5 5.6 0 Moderate encroachment

0.5 27.1 67.8 4.6 0 Minimal encroachment

1.8 22.4 70.3 5.2 0.3 Minimal encroachment

Altitude min–max (m.a.s.l) Average slope (°) A (km2) Lithology SST (%) GRA (%) PHY (%) ALL (%) Land use Forest (%) Grass (%) Crop (%) Swamp (%) Buildup (%) Status of papyrus wetland

A: Total catchment area; SST: sandstone; GRA: granite; PHY: phyllite; ALL: alluvium; altitude, slope and catchment area were based on ASTER GDEM 2009 data, land use was based on Vanonckelen (2009) and lithology was based on Harrop (1960).

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

759

Fig. 2. a) A severely encroached papyrus wetland, leaving the river bank prone to erosion, near Ruharo measuring station. b) An intact wetland at Rwamunena covering the entire valley bottom (August 2010).

obtained with R2 values ranging between 0.81 and 0.97. Only for the Buhanama station, a lower R2 (0.57) was obtained. Nonetheless, also at the latter station, the obtained rating curve was expected to give robust estimates of the corresponding runoff discharge.

3. Materials and methods 3.1. Runoff discharge Runoff discharge measurements were conducted at the outlets of the seven selected subcatchments during one hydrological year (1 June 2009–31 May 2010). Pressure divers, installed at the measuring stations, recorded the total (water plus atmospheric) pressure every 30 min, while three barometric pressure divers were installed near Ruharo, Ndaija and in between Rwamunena and Koga (Table 1). The total recorded pressure measurements were converted to flow depths by corrected for the atmospheric pressure, using the nearest installed barometric pressure diver, and by taking into account the water temperature. This calibration was done automatically with the Diveroffice© software. The obtained flow depths were than compared with manual flow depth measurements and if necessary corrected (e.g. for the height difference between the installed pressure diver and the lowest point in the channel cross-section). For different manually observed flow depths (d, [m]) at each measuring station, also the corresponding instantaneous runoff discharge (Q, [m3 s−1]) was measured, using the area-velocity method (Bartram and Ballance, 1996; Vanmaercke et al., 2010). Based on these d and Q observations (between 31 and 151 for each station), depth–discharge rating curves were established allowing us to convert the recorded flow depth series into continuous runoff discharge series. These rating curves were of the type: b

Q ¼ad

ð1Þ

where Q is the instantaneous runoff discharge, d is the measured flow depth (m) and a and b are empirical constants that were fitted using a non-linear least square regression procedure and based on all available d and Q measurements (between 31 and 151 observations, depending on the station; Table 2). Overall, very good regression results were Table 2 Values of the coefficients in the depth (d)–runoff discharge (Q; Eq. (1)) and the suspended sediment concentration (SSC)–runoff discharge (Q, Eq. (2)) rating curves for each monitoring station. n is number observations; ME is the model efficiency (Nash and Sutcliffe, 1970). d-Q rating curves

Ruharo Rugurama Nyeihanga Ndaija Buhanama Rwamunena Koga

SSC-Q rating curves

At each measuring station, between 25 and 72 suspended sediment concentrations (SSC, [g m−3]) samples were taken, yielding a total of 379 samples (Table 2). These samples were taken as much as possible on a flow-proportional basis. The water samples were taken in a depthintegrated manner (Bartram and Ballance, 1996) and filtered using Whatman filter paper grade 6 (3 μm pores). The retained sediment was oven-dried for 24 h at 105 °C and weighed. Next, the suspended sediment concentration was determined as the ratio of the suspended sediment mass and the sample volume (ca. 500 ml). Sediment rating curves are a generally accepted and frequently applied method to estimate the SSC at a given moment, based on the instantaneous runoff discharge (e.g. Asselman, 2000; Moliere et al., 2004; Gray and Simões, 2008; Vanmaercke et al., 2010). These sediment rating curves are typically of the type: SSC ¼ f  Q

g

b

n

ME

f

g

n

ME

0.621 8.3762 0.8329 18.4432 0.3303 4.0065 2.5929

3.7744 1.2485 2.689 2.2799 4.0948 1.5786 1.1108

51 151 78 67 62 78 31

0.972 0.856 0.817 0.94 0.572 0.960 0.920

180.99 307.76 335.93 374.51 344.47 197.20 199.84

0.28 0.07 −0.07 0.03 −0.04 0.18 0.55

25 72 67 59 52 61 43

0.076 0.003 0.004 0.001 0.002 0.025 0.067

ð2Þ

where f and g are empirical constants. We therefore tried to fit Eq. (2) to the available Q and corresponding SSC values at each station, using a non-linear least-square regression procedure (Asselman, 2000). However, overall we observed no significant relationships between Q and SSC for the studied subcatchments. R2 values of the fitted rating curves ranged between 0.002 and 0.08 (Table 3). Also subdividing the data according to season, as was done in previous studies (e.g. Walling, 1977; Moliere et al., 2004; Vanmaercke et al., 2010), did not yield more reliable rating curves. Since SSC showed no clear correlation with Q (Table 3) and since the number of SSC observations was fairly low, we calculated the total annual sediment load at each station, using the weighted concentration method (Phillips et al., 1999): Xns Q s;y ¼ K Xi¼1 ns

a

a and b are the coefficients of Eq. (1) and f and g the coefficients of Eq. (2).

3.2. Suspended sediment export

Ci Q i

i¼1

Qi

Qr

ð3Þ

where Q s,y is the annual sediment load, K is a conversion factor to take the period of record into account, ns is the number of samples, Q r is the mean discharge for the period of record, Q i is the measured instantaneous runoff discharge and Ci is the corresponding suspended sediment concentration (Phillips et al., 1999). Previous studies indicate that this method overall yields robust estimates of the total sediment export,

760

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

Table 3 Measured suspended sediment concentrations (SSC) and runoff discharge characteristics (Q) at each monitoring station for the hydrological year June 2009–May 2010 and per season. Catchment

Ruharo Rugurama Nyeihanga Ndaija Buhanamaa Rwamunena Koga a

Min SSC (g/m3)

Median SSC (g/m3)

Mean SSC (g m3)

Max SSC (g/m3)

Discharge weighted mean SSC (g/m3)

Min Q (m3/s)

Median Q (m3/s)

Mean Q (m3/s)

Max Q (m3/s)

35.62 3.48 9.40 4.04 7.25 14.00 6.70

296.37 287.59 273.67 362.42 290.68 161.51 113.87

300.20 333.93 304.40 405.94 321.26 216.05 280.04

774.00 1585.09 934.75 1533.68 1010.00 786.00 2861.67

328.13 361.70 292.56 380.50 294.34 231.49 361.31

2.25 0.72 1.27 0.84 1.46 0.49 0.18

6.07 8.31 4.29 3.77 4.98 1.72 0.68

7.06 9.09 5.21 4.52 6.32 2.40 0.77

19.93 23.70 21.22 26.58 24.55 8.35 1.89

Mean seasonal SSC (g m3) 1 Jun–31 Aug

1 Sep–31 Dec

1 Jan–15 Feb

16 Feb–31 May

291 382 433 536 404 255 477

– 205 216 244 249 105 154

– 756 165 224 100 117 57

389 148 258 169 250 203 133

Due to uncertainties on the flow depth–runoff discharge relationship, results for this station should be interpreted with caution.

even if sampling frequencies are less than once per week (Phillips et al., 1999; Moatar et al., 2006).

catchment with minimally encroached wetlands) showed a similar trend (Fig. 4).

3.3. Rainfall data

4.2. Suspended sediment concentration and sediment export

Rainfall was recorded on a daily basis at 8 rainfall gauges in the study area (Fig. 1). These data were used to estimate the total rainfall depth in each subcatchment by means of the Thiessen polygon method (WMO, 1994). This approach is justified since the Upper Rwizi catchment is dominated by a large central area with limited topographic variation. However, due to the limited number of rain gauges available, catchment-specific rainfall estimates are subjected to considerable uncertainties. Based on these estimates of total annual rainfall, annual rainfall–runoff ratios (RR, [%]) were calculated as the ratio of the total runoff discharge and the total annual rainfall in each subcatchment.

Observed SSC values ranged between 3.5 g m−3 and 2862 g m−3. The average SSC of each catchment varied between 216 and 406 g m−3, while the median SSC values ranged between 114 and 362 g m−3 and the discharge-weighted averaged SSC varied between 231 and 380 g m−3 (Table 3). Overall, the two catchments with minimally encroached wetlands have significant (p b 0.05) lower SSC values than the catchments with severely or moderately encroached wetlands (Table 2, Fig. 4). Nonetheless, these differences are fairly limited. Sediment concentrations also showed a temporal variability, with SSC values that were clearly higher during the long dry season (June–August) (Table 2). Annual area-specific suspended sediment yields (SYs), calculated based on the discharge-weighted sediment concentration method (Eq. (3)) range between 34 and 137 ton km−2 y− 1 (Table 4). Here also, the smallest SY-values correspond to the two catchments with minimally encroached wetlands (i.e. Rwamunena and Koga). The highest SY-value was observed for the Rugurama catchment, which has severely encroached wetlands (Table 4).

4. Results 4.1. Runoff discharge Annual runoff depths for the seven subcatchments varied between 102 and 437 mm y−1, while rainfall–runoff ratios varied between 8% and 30% (Table 4). Rwamunena and Koga, i.e. the catchments with intact wetlands near their outlets, clearly show smaller runoff depths (respectively 147 and 102 mm y−1) and rainfall–runoff ratios (respectively 9% and 8%) compared to the other catchments (Table 4). Overall, temporal variability in runoff depth shows only little correspondence with rainfall patterns throughout the hydrological year. This is demonstrated in Fig. 3 for RUG (a catchment with severely encroached wetlands) and for RWA (minimally encroached). Although there are distinct periods of low and high rainfall, monthly runoff depths at RUG mainly tend to increase throughout the hydrological year and reflect only very poorly differences in monthly rainfall (see also Fig. 4). Other catchments with severely or moderately encroached wetlands showed a very similar pattern (Fig. 4). At RWA, temporal changes in runoff depth are even smaller. Also the increase in runoff depths throughout the year is much smaller here. KOG (the other

5. Discussion 5.1. Spatial variation in annual runoff and sediment yield Evidently, the measurements obtained in this study are subject to uncertainties. Especially for the Buhanama station, the runoff data (and by consequence the sediment export data) should be interpreted with some caution because the depth–discharge rating curve of the Buhanama station was subject to considerable scatter (Table 2) and only few runoff discharge measurements could be conducted during large flow depths at this station. Also the SY-values (Table 4) are subject to uncertainties. Firstly, our measurements only consider the suspended sediment load and do not include bedload which typically also constitute 10 to 20% of the total sediment load (Turowski et al., 2010). Secondly, the SSC sampling frequency is fairly low (see Table 2). Previous studies have

Table 4 Rainfall, runoff and sediment yield data for each subcatchment and for the hydrological year June 2009–May 2010.

Ruharo Rugurama Nyeihanga Ndaija Buhanamaa Rwamunena Koga a

Rainfall depth (mm/y)

Total runoff discharge (m3/y)

Mean runoff discharge (m3/s)

Runoff depth (mm/y)

Rainfall–runoff ratio (%)

Total susp. sediment export (ton/y)

Sediment yield (t/km2/y)

1535 1437 1451 1452 1476 1683 1330

6.83E + 08 4.15E + 08 3.71E + 08 2.10E + 08 2.31E + 08 5.28E + 07 3.86E + 07

21.66 13.16 11.76 6.66 7.32 1.67 1.22

322 380 431 307 437 147 102

21 26 30 21 30 9 8

2.24E + 05 1.50E + 05 1.08E + 05 7.99E + 04 6.81E + 04 1.22E + 04 1.39E + 04

106 137 126 117 128 34 37

Due to uncertainties on the flow depth–runoff discharge relationship, results for this station should be interpreted with caution.

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

761

Fig. 3. The monthly rainfall and corresponding monthly runoff depth for the severely encroached Rugurama catchment (left) and for the minimal encroached Rwamunena catchment (right).

shown that SY estimates based on low sampling frequencies are subject to important uncertainties and more likely to underestimate the actual SY, because of a larger probability of excluding low-frequency but highmagnitude events (e.g. Phillips et al., 1999; Moatar et al., 2006). For the same reason, also the short measuring period (i.e. one hydrological year), makes it likely that the SY values reported in Table 4 are not entirely representative for the long-term average SY (Vanmaercke et al., 2012). Rainfall measurements indicate that the considered hydrological year was somewhat wetter than the average (Table 4). However, it is uncertain if this implies that also the obtained SY values are above average. These limitations and sources of error not only affect our study, but are typical for most runoff and SY measuring campaigns. Since our methods were very similar to those of other studies across Africa, we expect that also the degree of uncertainty on our results is comparable (for a review on this matter, we refer to Vanmaercke et al. (2014a)). Nonetheless, our results allow for a direct comparison between the seven monitored subcatchments and indicate that both annual runoff depth and SY are clearly smaller for the two catchments with minimally encroached wetlands, compared to the catchments with moderately or severely encroached wetlands (Table 4). While observations from seven catchments are insufficient to statistically pinpoint the factors explaining the observed variation, we found very clear indications that

the smaller runoff and SY of the Rwamunena and Koga catchment are to a large extent attributable to their nearly intact wetlands. For example, in literature SY is often negatively correlated to catchment area but positively correlated to catchment topography and erosion-prone land cover or lithological conditions (e.g. Syvitski and Milliman, 2007; De Vente et al., 2013; Vanmaercke et al., 2014a). Scatter plots for the seven subcatchments of the Upper Rwizi clearly indicate that the lower SY values of the two catchments with minimally encroached wetlands are not attributable to catchment area, average catchment slope, land use or lithology (Fig. 6). Scatter plots based on other lithological units or land use types resulted in the same conclusion and are therefore not shown. Also climatic conditions or the degree of seismic activity may influence SY (e.g. Syvitski and Milliman, 2007; de Vente et al., 2013; Vanmaercke et al., 2014b). However, since these factors varied very little between the studied subcatchments, they are unlikely to explain why the catchments with nearly intact wetlands have clearly lower SY-values. The large agreement between the scatter plots for annual runoff depth and SY (Fig. 6) and the fact that differences in runoff depths (Fig. 4) are generally much larger than the observed differences in SSC (Fig. 5) further indicate that the lower SY values of the catchment with minimally encroached wetlands are mainly attributable to their smaller runoff depths (Fig. 4).

Fig. 4. The monthly cumulative precipitation depth of the seven catchments versus the corresponding annual cumulative runoff depth, with indication of the level of encroachment of the wetlands in the catchments.

762

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

Fig. 5. Box plots of the suspended sediment concentration per catchment, bars are shaded according to the encroachment status of the wetlands in the river reaches upstream of the gauging station.

Temporal trends in monthly rainfall and runoff showed important delays in the rainfall–runoff response for all catchments (Figs. 3 and 4). This indicates that runoff dynamics in the studied catchments are to a large extent controlled by catchment lithology through the infiltration and exfiltration of rain water. However, we have no reasons to assume that the lower runoff depths of the two catchments with minimally encroached wetlands are merely attributable to differences in catchment lithology. Lithological differences between all studied catchments are relatively small: all catchments are dominated by phylites. Furthermore, the lithological composition of catchments with minimally encroached wetlands is very similar to those of the other catchments (Table 1). Also other factors potentially controlling runoff (e.g. land use, topography) provide no clear reason why observed runoff depths are clearly lower for the two catchments with minimally encroached wetlands. Hence, it is very likely that these smaller runoff depths are indeed attributable to the presence of nearly intact wetlands. Evapotranspiration most likely plays a dominant role in explaining this difference. Fig. 4 and Table 4 indicate that catchments with minimally encroached wetlands have runoff depths that are at least 150 mm smaller than the smallest runoff depth for the catchment with severely or moderately encroached wetlands. Saunders et al. (2007) estimated that evapotranspiration rates of papyrus wetlands in the Lake Victoria basin can reach up to 4.75 kg of water per square meter and per day, a value that is about 25% higher than the evaporation rate of open water surfaces in the area. Based on recent aerial photos in Google Earth, we estimated that the intact wetland area near the outlet was at least 13.6 km2 for Rwamunena and about 26.6 km2 for Koga. Taking into account the total area of these catchments (Table 1) and assuming the evapotranspiration rate of Saunders et al. (2007), these wetlands could evaporate at least 66 mm y−1 at Rwamunena and 122 mm y−1 at Koga. Although evapotranspiration may explain a very large proportion of the runoff difference between catchments with moderately to severely impacted and minimally encroached wetlands, an important part of the difference remains unaccounted for. One possibility is that also other factors such as specific catchment conditions or local rainfall patterns contribute to the lower runoff depths of the Rwamunena and Koga catchment. Also runoff depths for catchments with severely or moderately encroached wetlands differ up to 120 mm y−1. Secondly, the fact

that wetlands significantly increase the retention time of the river discharge might also cause a significant part of the water to percolate into and through deeper lithologic units. However, due to the limited amount of data available, we were unable to identify these other potentially controlling factors or investigate the role of percolation. Despite the fact that most of the differences in SY are attributable to differences in runoff depth, the lower SYs of catchments with intact wetlands are probably also partly caused by the fact that sediments are more likely to be trapped in the papyrus swamps. Also earlier studies pointed out that wetlands may act as a buffer for sediment transfers (e.g. Jordan et al., 2003; Kelderman et al., 2007). As Fig. 5 shows, both the Rwamunena and Koga catchment tend to have lower SSC concentrations compared to the catchments with moderately and severely encroached wetlands. The lack of a clear correlation between SSC and the corresponding runoff discharge at all stations (Table 3), indicates that this is not merely a result of the lower runoff discharge in these catchments with intact wetlands. In fact, the lack of clear trends between Q and SSC might be partly attributable to the presence of wetlands that buffer the transfer of sediments through the catchments (even in the case that swamps are severely encroached). Nonetheless, this remains uncertain, as also other studies in tropical catchments observed that SSC showed no correlation with Q (e.g. Lootens and Kishimbi, 1986). Another mechanism that may explain why catchments with minimally encroached wetlands have overall lower SSC values (Fig. 5), is that these wetlands significantly reduce the sediment connectivity between important sediment sources and the river network (MWE, 1995; Hamels, 2011). Field observations in the study area clearly indicate that many gullies developed on the hill slopes that are currently used for grazing (upper parts) and the production of bananas (lower sections). Several of these gullies directly deliver fine sediments and even large boulders to the river channel at reaches where the wetlands were encroached (Fig. 7), while this was not observed at reaches with intact wetlands. Hence, the wetland degradation also induces a change in the texture of the sediments transported by river flows. Also other sediment point sources, such as river bank collapses or bank degradation at cattle drinking places caused by intense trampling, were observed frequently at encroached reaches. Nonetheless, it is noteworthy that the differences in SSC between catchments with encroached or intact wetlands at their outlet are fairly limited and perhaps smaller than expected. Caution is also required when interpreting these data as runoff and sediment dynamics may strongly vary from year to year (e.g. Vanmaercke et al., 2012) while our results are based on only one hydrological year of observations. More research, based on longer observation series and where also the sediment input into the wetlands is quantified, is required to fully understand these sediment dynamics. 5.2. Temporal variation in runoff and sediment concentrations It is noteworthy that, although catchments with moderately and severely encroached wetlands showed a clear variability in monthly runoff, this variability is not clearly correlated to the corresponding monthly rainfall (Figs. 3 and 4). This prolonged delay implies infiltration and exfiltration and a large importance of subsurface flow in the studied catchments, which might further help explaining the lack of a correlation between SSC and runoff discharge (Table 3). Nevertheless, also surface runoff contributes to the total river discharges, especially near the end of the second wet season. At that time, saturation occurs over vast areas, small basins are filled and overland flow will occur. As discussed in the previous section, catchments with minimally encroached wetlands showed a much smaller seasonal variability (Figs. 3 and 4), which is likely at least partly attributable to their higher evapotranspiration rates. However, also on shorter time scales (i.e. the 30 minute intervals of our continuous flow depth measurements), we noticed that the catchments with intact wetlands near their outlet

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

763

Fig. 6. Runoff depth and sediment yield (SY) relations with catchment area, % cropland area and mean catchment slope (°) (Table 1), symbols are shaded according to the encroachment status of the wetlands in the river reaches upstream of the gauging station (black for severe encroachment, gray for moderate encroachment, white for minimal encroachment).

show a more moderate flow regime compared to the catchments with encroached wetlands (Fig. 8). A measure to express the variability in runoff discharge and the relative importance of flood events is VW2, i.e. the share of the total runoff volume that is transported by the 2% highest runoff discharges (e.g. Moatar et al., 2006). VW2 values typically range between 5 and 20% for rivers of comparable size in Europe or

the USA (Moatar et al., 2006), but can exceed 40% in rivers with a strong flash flood regime, such as in the semi-arid Ethiopian highlands (e.g. Vanmaercke et al., 2010; Zenebe et al., 2013). The VW2 values for our seven subcatchments range between 4 and 14%, indicating that the tributaries of the Upper Rwizi are characterized by a relatively constant flow regime.

764

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

Fig. 7. a) Gully channel draining directly into river Rwizi upstream of Ndaija without the presence of any buffering papyrus vegetation. b) Boulders and debris deposited in river Rwizi at the outlet of the gully channel indicating a high sediment connectivity between the hill slope and the river network (August, 2010).

Whereas it can normally be expected that VW2 decreases with increasing catchment area (e.g. Moatar et al., 2006), we observed an increasing trend with catchment area for the seven studied subcatchments of the Upper Rwizi (Fig. 9). As the two smallest catchments on these graph are those with nearly intact wetlands and the encroachment rate tends to increase with catchment area (Fig. 9), this clearly indicates that the wetlands play an important role in buffering peak flow discharges and moderating the flow regime of the studied rivers. Unfortunately, too few samples could be collected to allow for a detailed temporal analysis of the suspended sediment concentrations in relation to the state of the papyrus wetlands in the considered catchments. Nonetheless, we could observe that SSC values are generally larger during the long dry season compared to values for the wet seasons, which indicate that SSC is strongly supply-controlled (Table 2). Several processes may explain these higher concentrations. Firstly, the relative water shortage during the dry season forces cattle to drink at the river. As argued in other studies (e.g. Trimble and Mendel, 1995), cattle can act as an important geomorphic agent by loosening the soil material, trampling river banks and destroying riparian vegetation. However, due to the presence of the impenetrable papyrus swamps, cattle cannot reach the stream at RWA and KOG and is forced to drink on the edges of the swamps, resulting in a minimal soil disturbance. During field surveys, we mainly observed cattle impacts at river reaches with encroached wetlands, with 10 or more possible sediment point sources per km channel length, while in river reaches with intact

Fig. 8. Cumulative runoff discharge plotted against cumulative time, both calculated based on 30 minute interval data, sorted in a descending order and expressed as a percentage of the total annual runoff discharge/time. Note the more gradual runoff release where wetlands underwent minimal encroachment.

wetlands this was limited to 0–3 point sources. In addition, it is common practice in the study area to burn the vegetation over large areas during the dry season in order to stimulate the growth of new grass for cattle grazing. This process leaves the soil bare and prone to water erosion (e.g. Inbar et al., 1998; Shakesby and Doerr, 2006) and, in particular, gully erosion as the topographic threshold for gully incision is drastically lowered after burning (Torri and Poesen, 2014). Such gullies are not only a significant sediment source, but also increase the sediment connectivity between the uplands and the river system (Poesen et al., 2003). Furthermore, the combustion of the soil organic matter in the top soil can damage the soil structure and increase its erodibility (e.g. Shakesby and Doerr, 2006; Lyon and O'Connor, 2008). 5.3. The importance of the papyrus wetlands Several studies have already shown that papyrus wetlands provide valuable ecosystem services, such as regulating nutrient cycles and reducing eutrophication risks (e.g. Kelderman et al., 2007; Kansiime et al., 2007; Mitsch et al., 2010), improving the water quality by acting as a filter (e.g. Brix, 1994; Mnaya and Wolanski, 2002), sequestering large amounts of carbon (e.g. Saunders et al., 2007; Mitsch et al., 2010), maintaining biodiversity by acting as an important habitat (e.g. Brix, 1994; Gichuki et al., 2001; Mnaya and Wolanski, 2002; Maclean et al., 2003; Kansiime et al., 2007; Keddy, 2010), providing a nursery ground for fish (e.g. Mnaya and Wolanski, 2002), acting as fishery grounds (e.g. Gichuki et al., 2001; Kansiime et al., 2007) and supplying raw materials for fuel, handcarfts and roofs (e.g. Gichuki et al., 2001; Maclean et al., 2003; Kansiime et al., 2007; Keddy, 2010). As this study indicates, papyrus wetlands also buffer runoff and function as sediment traps,

Fig. 9. The fraction of the total annual runoff that is transported by the 2% highest peak runoff discharges VW2, versus catchment area, with indication of the encroachment status of the papyrus wetlands.

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

765

reducing the sediment connectivity in these catchments, and, hence, mitigating the impacts of hill slope soil erosion on downstream water bodies. As discussed in the previous section, papyrus swamps may also lower flood risks and related impacts by buffering peak runoff discharges. Due to these benefits, papyrus wetlands play a key role when facing the hydrological and ecological challenges of Lake Victoria discussed in the Introduction section. Protecting and, if possible, restoring wetlands could be an important strategy in limiting the sediment, nutrient and pollutant input into Lake Victoria and reducing the impacts of floods. Nonetheless, the large population pressure and the resulting food production demands make this a difficult task. Moreover, the benefits of wetlands are often underestimated by governments in the Lake Victoria basin (Swallow et al., 2003). More insights in the runoff and sediment transport reducing effects of wetlands in relation to their spatial extent and their location within the catchment could help address this task, as it would allow prioritizing the wetlands that urgently need protection or restoration. The data collected in the framework of this study are insufficient to fully obtain such understanding. Nonetheless, our results allow for some first recommendations. For example, due to practical constraints, it was impossible to map the extent of all wetland reaches in the Upper Rwizi catchment. We therefore focused mainly on the status of the wetlands 5 km upstream of the catchment outlet, as we hypothesized that these wetlands were most relevant for buffering runoff and sediments (see Section 2). Since our results indicated that the status of papyrus wetlands in these final reaches indeed have an impact on runoff and sediment dynamics, this hypothesis appears to be confirmed. Hence, this also suggests that wetlands near the downstream reaches of the river (i.e. close to the outlet) should be prioritized for protection or restoration.

services, their protection (and where possible) restoration is highly recommended from an ecological point of view. Nonetheless, this may be challenging due the large demand for agricultural land and the increasing population pressure within the Lake Victoria basin. Further research may contribute to addressing this challenge by trying to better understand how the location, spatial extent or other wetland characteristics influence their runoff and sediment transport reducing capacity, as this would allow us to prioritize wetlands that should be protected. Our first-guess recommendation based on this study would be to mainly focus protection efforts on wetlands in the downstream reaches of a catchment.

6. Conclusions

Asselman, N.E.M., 2000. Fitting and interpretation of sediment rating curves. J. Hydrol. 234 (3), 228–248. Bartram, J., Ballance, R. (Eds.), 1996. Water Quality Monitoring: A Practical Guide to the Design and Implementation of Freshwater Quality Studies and Monitoring Programmes. CRC Press. Boar, R.R., Harper, D.M., 2002. Magnetic susceptibilities of lake sediment and soils on the shoreline of Lake Naivasha, Kenya. Lake Naivasha, Kenya. Springer, Netherlands, pp. 81–88. Bolwig, S., 2002. Land use change and soil degradation in the southwestern highlands of Uganda. SCRIP Phase II Report Submitted to the USAID Uganda Mission. International Food Policy Research Institute, Washington, DC COMPETE Presidential Conference on Export Competitiveness. Brix, H., 1994. Functions of macrophytes in constructed wetlands. Water Sci. Technol. 29 (4), 71–78. Campbell, L.M., Hecky, R.E., Muggide, R., Dixon, D.G., Ramlal, P.S., 2003. Variation and distribution of total mercury in water, sediment and soil from northern Lake Victoria, East Africa. Biogeochemistry 65 (2), 195–211. Chenery, E., 1960. An Introduction to the Soils of the Uganda Protectorate. Department of Agriculture, Kawanda Agricultural Research Station, Kampala (80 pp.). Cohen, M.J., Shepherd, K.D., Walsh, M.G., 2005. Empirical reformulation of the universal soil loss equation for erosion risk assessment in a tropical watershed. Geoderma 124 (3), 235–252. Cooper, J.R., Gilliam, J.W., Daniels, R.B., Robarge, W.P., 1987. Riparian areas as filters for agricultural sediment. Soil Sci. Soc. Am. J. 51 (2), 416–420. De Vente, J., Poesen, J., Verstraeten, G., Govers, G., Vanmaercke, M., Van Rompaey, A., BoixFayos, C., 2013. Predicting soil erosion and sediment yield at regional scales: where do we stand? Earth Sci. Rev. 127, 16–29. Douglas, I., Alam, K., Maghenda, M., Mcdonnell, Y., McLean, L., Campbell, J., 2008. Unjust waters: climate change, flooding and the urban poor in Africa. Environ. Urban. 20 (1), 187–205. FAO, 2005. Local climate estimator (New LocClim 1.06). FAO, Rome. Available on: http:// www.fao.org/nr/climpag/pub/en3_051002_en.asp. Gichuki, J., Guebas, F.D., Mugo, J., Rabuor, C.O., Triest, L., Dehairs, F., 2001. Species inventory and the local uses of the plants and fishes of the Lower Sondu Miriu wetland of Lake Victoria, Kenya. Hydrobiologia 458 (1–3), 99–106. Gray, J.R., Simões, F.J., 2008. Estimating sediment discharge. Sedimentation Engineering— Processes, Measurements, Modeling, and Practice, Manual 110 pp. 1067–1088. Hamels, K., 2011. Topographical Thresholds and Impacts of Gully Erosion in Southwestern Uganda. (Master thesis), KU Leuven, Belgium (108 pp.). Harrop, J.F., 1960. The soils of the Western Province of Uganda. Mem. Res. Div. 1 (6). Hecky, R.E., 1993. The eutrophication of Lake Victoria. Verh. Int. Ver. Limnol. 25, 39–48. Inbar, M., Tamir, M.I., Wittenberg, L., 1998. Runoff and erosion processes after a forest fire in Mount Carmel, a Mediterranean area. Geomorphology 24 (1), 17–33. Jordan, T.E., Whigham, D.F., Hofmockel, K.H., Pittek, M.A., 2003. Nutrient and sediment removal by a restored wetland receiving agricultural runoff. J. Environ. Qual. 32 (4), 1534–1547.

The Lake Victoria basin faces important hydrological and ecological challenges that pose a significant threat to the livelihood of its growing population. Several of these challenges relate to the increased sediment input and changing runoff dynamics of the rivers draining into the Lake and demand for adequate catchment management strategies. Developing such strategies requires an understanding of the magnitude and dynamics of catchment runoff and sediment export in relation to their characteristics (e.g. land use patterns). This study contributed to a better understanding by providing runoff and sediment export data for seven tributaries (ranging in size between 358 and 2121 km2) of the Rwizi catchment. These are among the first measured data available for the Lake Victoria basin and, to our knowledge, the first measurements of this kind for Uganda. Our results indicate that the presence of papyrus wetlands in the downstream reaches of the catchments can significantly reduce runoff depths, temporal variations in runoff discharge, sediment concentrations and, consequently, sediment export. The impacts of wetlands appear to be the largest on runoff depths, while the impact on sediment concentrations was significant but fairly limited. A large part of the observed differences in runoff depth between catchments with intact wetlands and catchments with (severely) encroached wetlands is most likely attributable to the retention of runoff in combination with the high evapotranspiration rates of papyrus wetlands. However, the observed differences were larger than could be expected from evapotranspiration alone. This suggests that other mechanisms, such as percolation, also play a role. Additional research is needed to investigate this further in more detail. Although our results are subject to some uncertainty and are insufficient to disentangle the role of the various factors that may affect runoff and sediment yield, they clearly illustrate that papyrus wetlands could play a crucial role in catchment management strategies that aim at reducing the sediment input into Lake Victoria. As previous studies demonstrated that these wetlands also provide various other ecosystem

Acknowledgments The authors would like to thank VLIR-UOS for financing the VLIR-OIRIPAVIC project, which made this study possible. Our thanks go also to the IRO-KU Leuven for supporting the fieldwork campaign of N. Ryken in Uganda and towards Koen Hamels for his assistance during the fieldwork campaign. M. Vanmaercke acknowledges his research grant from the Research Foundation Flanders (FWO), Belgium. We further thank Fernando A.L. Pacheco and an anonymous reviewer for their constructive comments on an earlier draft of this manuscript. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2014.12.048. References

766

N. Ryken et al. / Science of the Total Environment 511 (2015) 756–766

Kansiime, F., Saunders, M.J., Loiselle, S.A., 2007. Functioning and dynamics of wetland vegetation of Lake Victoria: an overview. Wetl. Ecol. Manag. 15 (6), 443–451. Keddy, P.A., 2010. Wetland Ecology: Principles and Conservation. Cambridge University Press. Kelderman, P., Kansiime, F., Tola, M.A., van Dam, A., 2007. The role of sediments for phosphorus retention in the Kirinya wetland (Uganda). Wetl. Ecol. Manag. 15 (6), 481–488. Khan, Z., Midega, C., Pittchar, J., Pickett, J., Bruce, T., 2011. Push–pull technology: a conservation agriculture approach for integrated management of insect pests, weeds and soil health in Africa: UK government's Foresight Food and Farming Futures project. Int. J. Agric. Sustain. 9 (1), 162–170. Kitaka, N., Harper, D.M., Mavuti, K.M., 2002. Phosphorus inputs to Lake Naivasha, Kenya, from its catchment and the trophic state of the lake. Lake Naivasha, Kenya. Springer, Netherlands, pp. 73–80. Knapen, A., Kitutu, M.G., Poesen, J., Breugelmans, W., Deckers, J., Muwanga, A., 2006. Landslides in a densely populated county at the footslopes of Mount Elgon (Uganda): characteristics and causal factors. Geomorphology 73 (1), 149–165. Lehman, J.T., 1996. Pelagic food webs of the East African great lakes. The Limnology, Climatology and Paleoclimatology of the East African Lakespp. 281–301. Lootens, M., Kishimbi, Y., 1986. Some aspects of water and sediment discharge in the upstream section of the Kafubu River (Shaba-Zaire). Geogr. Ann. A 383–392. Lufafa, A., Tenywa, M.M., Isabirye, M., Majaliwa, M.J.G., Woomer, P.L., 2003. Prediction of soil erosion in a Lake Victoria basin catchment using a GIS-based Universal Soil Loss model. Agric. Syst. 76 (3), 883–894. Lwasa, S., 2010. Adapting urban areas in Africa to climate change: the case of Kampala. Curr. Opin. Environ. Sustain. 2 (3), 166–171. Lyon, J.P., O'Connor, J.P., 2008. Smoke on the water: can riverine fish populations recover following a catastrophic fire‐related sediment slug? Austral Ecol. 33 (6), 794–806. Maclean, I.M., Tinch, R., Hassall, M., Boar, R.R., 2003. Towards optimal use of tropical wetlands: an economic valuation of goods derived from papyrus swamps in southwest Uganda. Environmental Change and Management Working Paper, p. 10. Maetens, W., Vanmaercke, M., Poesen, J., Jankauskas, B., Jankauskien, G., Ionita, I., 2012. Effects of land use on annual runoff and soil loss in Europe and the Mediterranean: a meta-analysis of plot data. Prog. Phys. Geogr. 36, 599–653. Mitsch, W.J., Nahlik, A., Wolski, P., Bernal, B., Zhang, L., Ramberg, L., 2010. Tropical wetlands: seasonal hydrologic pulsing, carbon sequestration, and methane emissions. Wetl. Ecol. Manag. 18 (5), 573–586. Mnaya, B., Wolanski, E., 2002. Water circulation and fish larvae recruitment in papyrus wetlands, Rubondo Island, Lake Victoria. Wetl. Ecol. Manag. 10 (2), 131–141. Moatar, F., Person, G., Meybeck, M., Coynel, A., Etcheber, H., Crouzet, P., 2006. The influence of contrasting suspended particulate matter transport regimes on the bias and precision of flux estimates. Sci. Total Environ. 370 (2), 515–531. Moliere, D.R., Evans, K.G., Saynor, M.J., Erskine, W.D., 2004. Estimation of suspended sediment loads in a seasonal stream in the wet‐dry tropics, Northern Territory, Australia. Hydrol. Process. 18 (3), 531–544. Morrison, E.H., Harper, D.M., 2009. Ecohydrological principles to underpin the restoration of Cyperus papyrus at Lake Naivasha, Kenya. Ecohydrol. Hydrobiol. 9 (1), 83–97. Muyodi, F.J., Bugenyi, F.W., Hecky, R.E., 2010. Experiences and lessons learned from interventions in the Lake Victoria Basin: the Lake Victoria environmental management project. Lakes Reserv. Res. Manag. 15 (2), 77–88. Mwamburi, J., 2003. Variations in trace elements in bottom sediments of major rivers in Lake Victoria's basin, Kenya. Lakes Reserv. Res. Manag. 8, 5–13. Mwanuzi, F., Aalderink, H., Mdamo, L., 2003. Simulation of pollution buffering capacity of wetlands fringing the Lake Victoria. Environ. Int. 29 (1), 95–103. MWE, 1995. National Policy for the Conservation and Management of Wetland Resources. Ministry of Water and Environment, Kampala (16 pp.). Nakiyemba, W.A., Isabirye, M., Poesen, J., Maertens, M., Deckers, J., Mathijs, E., 2013. Decentralised governance of wetland resources in the Lake Victoria Basin of Uganda. Nat. Resour. 2013 (4), 55–64. Nash, J., Sutcliffe, J.V., 1970. River flow forecasting through conceptual models part I—a discussion of principles. J. Hydrol. 10 (3), 282–290. Ntiba, M.J., Kudoja, W.M., Mukasa, C.T., 2001. Management issues in the Lake Victoria watershed. Lakes Reserv. Res. Manag. 6 (3), 211–216. Odada, E.O., Olago, D.O., Kulindwa, K., Ntiba, M., Wandiga, S., 2004. Mitigation of environmental problems in Lake Victoria, East Africa: causal chain and policy options analyses. Ambio 33 (1), 13–23. Ogutu-Ohwayo, R., Hecky, R.E., Cohen, A.S., Kaufman, L., 1997. Human impacts on the African great lakes. Environ. Biol. Fish 50 (2), 117–131.

Phillips, J.M., Webb, B.W., Walling, D.E., Leeks, G.J.L., 1999. Estimating the suspended sediment loads of rivers in the LOIS study area using infrequent samples. Hydrol. Process. 13 (7), 1035–1050. Poesen, J., Nachtergaele, J., Verstraeten, G., Valentin, C., 2003. Gully erosion and environmental change: importance and research needs. Catena 50 (2–4), 91–133. Rwakakamba, T.M., 2009. How effective are Uganda's environmental policies? A case study of water resources in 4 districts, with recommendations on how to do better. Mt. Res. Dev. 29 (2), 121–127. Saunders, M.J., Jones, M.B., Kansiime, F., 2007. Carbon and water cycles in tropical papyrus wetlands. Wetl. Ecol. Manag. 15 (6), 489–498. Shakesby, R.A., Doerr, S.H., 2006. Wildfire as a hydrological and geomorphological agent. Earth Sci. Rev. 74 (3), 269–307. Swallow, B., Okono, A., Ong, C., Place, F., 2003. TransVic: improved land management across the Lake Victoria Basin. Research Towards Integrated Natural Resources Management: Examples of Research Problems, Approaches and Partnerships in Action in the CGIAR. CGIAR Interim Science Council: Centre Directors' Committee on Integrated Natural Resources Management, Rome, Italy, pp. 65–78. Swallow, B.M., Sang, J.K., Nyabenge, M., Bundotich, D.K., Duraiappah, A.K., Yatich, T.B., 2009. Tradeoffs, synergies and traps among ecosystem services in the Lake Victoria basin of East Africa. Environ. Sci. Pol. 12 (4), 504–519. Syvitski, J.P., Milliman, J.D., 2007. Geology, geography, and humans battle for dominance over the delivery of fluvial sediment to the coastal ocean. J. Geol. 115 (1), 1–19. Torri, D., Poesen, J., 2014. A review of topographic threshold conditions for gully head development in different environments. Earth Sci. Rev. 130, 73–85. Trimble, S.W., Mendel, A.C., 1995. The cow as a geomorphic agent—a critical review. Geomorphology 13 (1), 233–253. Turowski, J.M., Rickenmann, D., Dadson, S.J., 2010. The partitioning of the total sediment load of a river into suspended load and bedload: a review of empirical data. Sedimentology 57 (4), 1126–1146. Van Dam, A.A., Dardona, A., Kelderman, P., Kansiime, F., 2007. A simulation model for nitrogen retention in a papyrus wetland near Lake Victoria, Uganda (East Africa). Wetl. Ecol. Manag. 15 (6), 469–480. Vanmaercke, M., Zenebe, A., Poesen, J., Nyssen, J., Verstraeten, G., Deckers, J., 2010. Sediment dynamics and the role of flash floods in sediment export from medium-sized catchments: a case study from the semi-arid tropical highlands in northern Ethiopia. J. Soils Sediments 10 (4), 611–627. Vanmaercke, M., Poesen, J., Radoane, M., Govers, G., Ocakoglu, F., Arabkhedri, M., 2012. How long should we measure? An exploration of factors controlling the interannual variation of catchment sediment yield. J. Soils Sediments 12 (4), 603–619. Vanmaercke, M., Poesen, J., Broeckx, J., Nyssen, J., 2014a. Sediment yield in Africa. EarthSci. Rev. 136, 350–368. Vanmaercke, M., Kettner, A.J., Van Den Eeckhaut, M., Poesen, J., Mamaliga, A., Verstraeten, G., Govers, G., 2014b. Moderate seismic activity affects contemporary sediment yields. Prog. Phys. Geogr. 38 (2), 145–172. Vanonckelen, S., 2009. Assessment of Land Use and Land Use Dynamics in the UpperRwizi River Catchment, Lake Victoria, Southwestern Uganda. (Master thesis), K.U. Leuven, Belgium (125 pp.). Verschuren, D., Johnson, T.C., Kling, H.J., Edgington, D.N., Leavitt, P.R., Brown, E.T., Hecky, R.E., 2002. History and timing of human impact on Lake Victoria, East Africa. Proc. R. Soc. Lond. B Biol. Sci. 269 (1488), 289–294. Walling, D.E., 1977. Assessing the accuracy of suspended sediment rating curves for a small basin. Water Resour. Res. 13 (3), 531–538. Walling, D.E., Webb, B.W., 1996. Erosion and Sediment Yield: A Global Overview 236. IAHS Publications-Series of Proceedings and Reports-Intern Assoc Hydrological Sciences, pp. 3–20. Witte, F., Goldschmidt, T., Wanink, J., van Oijen, M., Goudswaard, K., Witte-Maas, E., Bouton, N., 1992. The destruction of an endemic species flock: quantitative data on the decline of the haplochromine cichlids of Lake Victoria. Environ. Biol. Fish. 34, 1–28. World Meteorological Organization (WMO), 1994. Guide to hydrological practices. 5th ed. WMO-No. 168World Meteorological Organization, Geneva. Zenebe, A., Vanmaercke, M., Poesen, J., Verstraeten, G., Haregeweyn, N., Haile, M., Nyssen, J., 2013. Spatial and temporal variability of river flows in the degraded semi-arid tropical mountains of northern Ethiopia. Z. Geomorphol. 57 (2), 143–169.

Impact of papyrus wetland encroachment on spatial and temporal variabilities of stream flow and sediment export from wet tropical catchments.

During the past decades, land use change in the Lake Victoria basin has significantly increased the sediment fluxes to the lake. These sediments as we...
1MB Sizes 0 Downloads 8 Views