JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 May 2008
Published in J Environ Qual 37:889-897 (2008)
DOI: 10.2134/jeq2007.0251
© 2008 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Vigiak, O.
Right arrow Articles by Valentin, C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Vigiak, O.
Right arrow Articles by Valentin, C.
Agricola
Right arrow Articles by Vigiak, O.
Right arrow Articles by Valentin, C.
Related Collections
Right arrow Surface Water Quality
Right arrow Best Management Practices
Right arrow Hillslope Analysis
Right arrow Runoff
Right arrow Soil Conservation

TECHNICAL REPORTS

Surface Water Quality

Trapping Efficiencies of Cultivated and Natural Riparian Vegetation of Northern Laos

Olga Vigiaka,*, Olivier Ribolzib, Alain Pierretb, Oloth Sengtaheuanghoungc and Christian Valentind

a International Water Management Inst., IWMI-Laos, P.O. Box 811 Vientiane, Lao PDR (current address: DPI Rutherglen Centre, RMB 1145 Chiltern Valley Road, Rutherglen, VIC 3685, Australia)
b Institut de Recherche pour le Développement, IRD, P.O. Box 5992, Vientiane, Lao PDR, seconded to IWMI
c National Agriculture and Forestry Research Inst., P.O. Box 811, Vientiane, Lao PDR
d Insitut de Recherche pour le Développement, IRD, 32, av H. Varagnat, 93143 Bondy cedex, France, seconded to IWMI

* Corresponding author (olga.vigiak{at}dpi.vic.gov.au).

Received for publication May 21, 2007.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Conclusions
 REFERENCES
 
In northern Laos, intensification of cultivation on sloping land leads to accelerated erosion processes. Management of riparian land may counteract the negative impacts of higher sediment delivery rates on water quality. This study assessed water and sediment concentration trapping efficiencies of riparian vegetation in northern Laos and the effect of cultivation of riparian land on water quality. Runoff flowing in and out of selected riparian sites was monitored by means of open troughs. In 2005, two native grass, two bamboo, and two banana sites were monitored. In 2006, adjacent to steep banana, bamboo, and native grass sites, three upland rice sites were established and monitored. Water trapping efficiency (WTE) and sediment concentration trapping efficiency (SCTE) were calculated on an event basis; means and 95% confidence intervals (CIs) were estimated with a bootstrapping approach. Confidence intervals were large and overlapping among sites. Seepage conditions severely limited trapping efficiency. Native grass resulted in the highest WTE (95% CI, –0.10 to 0.23), which was not significantly different from zero. Banana resulted in the highest SCTE (95% CI, 0.06–0.40). Bamboo had negative WTE and SCTE. Median outflow runoff from rice sites was nine times the inflow. Median outflow sediment concentration from rice sites was two to five times that of their adjacent sites and two to five times the inflow sediment concentration. Although low-tillage banana plantation may reduce sediment concentration of runoff, cultivation of annual crops in riparian land leads to delivery of turbid runoff into the stream, thus severely affecting stream water quality.

Abbreviations: SCTE, sediment concentration trapping efficiency • WTE, water trapping efficiency


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
Materials and Methods
 Results
 Discussion
 Conclusions
 REFERENCES
 
SOUTHEASTERN Asian countries are experiencing a rapid change of land use, entailing the intensification of cultivation on sloping land and the reduction of forest and secondary vegetation cover (Roder et al., 1997; Ziegler et al., 2004). In northern Laos, the traditional shifting cultivation system consists of clearing off a field by slashing and burning natural vegetation at the beginning of the rainy season, cultivating for 1 yr with annual crops (mostly upland rice, Oryza sativa L.), then abandoning the field to natural vegetation regrowth for several years (fallow period) to restore soil fertility. No external inputs are usually applied. In the last two decades, however, restricted access to land coupled with a sustained increase of population has forced farmers to shorten the fallow period from 15 to 20 yr to 3 to 5 yr while protracting the cultivation period from 1 yr to almost 2 yr (Saito et al., 2006; Lestrelin and Giordano, 2007). The reduction of fallow period and the concurrent intensification of cultivation of annual crops cause losses of soil fertility and crop yield (Roder et al., 1997; Saito et al., 2006) and accelerate erosive processes on the slopes, leading to higher sediment delivery rates (Ziegler et al., 2004; Chaplot et al., 2005).

The impact of accelerated erosion from sloping land on sediment concentration of water bodies could be reduced by trapping sediments in the riparian areas. Research conducted in temperate climates showed that riparian land can retain up to 70 to 99% of incoming pollutant loads, and proper management of riparian areas is among the most recommended practices for water quality improvement (Karssies and Prosser, 1999; Dosskey, 2001). However, there is a lack of information about riparian land in tropical environments (Karssies and Prosser, 1999). In the wet tropics, McKergow et al. (2004) reported lower sediment trapping efficiencies (37–46%) and even negative values when exfiltration in the riparian area was observed. In northern Laos, cultivation systems and environmental conditions are different from those in which most riparian trapping efficiency studies have been conducted (e.g., Dosskey, 2001). Moreover, growing demand for fresh vegetables in urban centers attracts farmers to cultivate riparian land, where irrigation is easier. Vegetable gardens are mainly a dry-season activity; however, reaches of headwater catchment streams are also put under cultivation for vegetables or upland rice during the rainy season. The effects brought about by cultivation of riparian land use are largely unknown.

The objectives of this research were to assess (i) the water and sediment concentration trapping efficiencies of riparian vegetation of northern Laos and (ii) the potential effect of cultivation of riparian land on these trapping efficiencies and water quality.


   Materials and Methods
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
 Discussion
 Conclusions
 REFERENCES
 
Study Area
The study was conducted in the Houay Pano catchment (Luang Prabang Province, northern Lao PDR). Rainfall amounts are on average 1259 mm yr–1, more than 90% of which falls during the monsoon season from mid-May to mid-October. The geological substrate consists of argillites, siltstones, and fine-grained sandstones from Permian Upper Carboniferous (Rumpel et al., 2006). Soils prevailing along the slopes are deep (2.5–4.5 m) and clayey Alfisols (Rumpel et al., 2006).

The catchment is representative of the no-input slash and burn system of Southeast Asia, but the fallow periods have been shortened from 10 to 15 yr to 2 to 5 yr (de Rouw et al., 2005; Chaplot et al., 2005). The most common annual crop is upland rice (Oryza sativa L.). The term "upland" refers to rainfed cultivation as different from paddy rice cultivation. Preparation of sowing bed and weeding operations are conducted by hand, often by hoeing. On average, farmers clear the upland rice fields four to five times during the rice cultivation (de Rouw et al., 2002). Hoeing entails disturbance of the top 7 to 10 cm of soil, causes severe tillage erosion, and facilitates the formation of soil seals and crusts, which inhibit water infiltration (Janeau et al., 2003). Tillage can also cause reduction of saturated hydraulic conductivity in the soil profile, which may result in enhanced generation of lateral subsurface flow during rainstorms (Ziegler et al., 2004). Cultivation of annual crops may be repeated a second year, before the field is abandoned to secondary vegetation (fallow). In the first and second year of fallow, saturated hydraulic conductivity is still low, but over time, the growth of secondary vegetation is strong enough to recover infiltration and reduce overland flow generation to levels comparable to undisturbed surfaces (Ziegler et al., 2004).

Recently, parts of the catchment have been taken out of the slash and burn system. Approximately 13% of the catchment area has been converted to perennial crops (mainly banana [Musa sp.]), tree plantations (mainly teak [Tectona grandis L.]), or vegetable gardens, cultivated on a continuous basis (Lestrelin and Giordano, 2007). The cultivation of banana in Houay Pano is a low-input farming activity. Farmers plant bananas, sometimes burn the undergrowth at the onset of the rainy season, and cut single stems carrying casks ready to be sold at the market. They do not till the soil, and vegetation cover remains high throughout the year. Such a low-labor, low-input management system must not be confused with the intensive banana plantations of other tropical areas. Teak plantation is also a low-labor system: After plantation, farmers limit their activities to cultivating the fields for 2 yr with annual crops before the teak crowns become too large. No cover crop is used. Cultivation of vegetables is similar to other annual crops; land is cleared by burning, no drainage system is put in place, no fertilization is applied, and soil preparation and weeding operations during the crop growth are done by manual hoeing.

The 64-ha catchment of Houay Pano feeds a 1200-m-long, second-order perennial stream of irregular topography with an average slope gradient of 0.19 m m–1 (Ribolzi et al., 2005). Riparian areas are mainly of convex or convex-concave morphology, steep (10–130%), and narrow (4–23 m). Riparian soils are mainly Dystrochrepts with redoximorphic features and clay loam topsoil (Rumpel et al., 2006). More than 43% of the Houay Pano riparian land is covered with grass and shrub vegetation dominated by the grass Microstegium ciliatum A. Camus (hereafter referred to as "native grass"). Bamboos (mainly Dendrocalamus sp. and Cephalostachium virgatum) cover 19% of riparian areas. Cultivation of banana extends over 15% of the riparian areas. The remaining riparian areas are covered with forest (15%), cassava (Manihot utilissima Phol., 5%), and elephant grass (Pennisetum purpureum, 3%). For the past few years, patches of riparian land have been cleared of the natural vegetation and cultivated with vegetables (chili, watercress) or upland rice (Fig. 1 ).


Figure 1
View larger version (158K):
[in this window]
[in a new window]

  		Fig. 1. In northern Laos, cultivation of upland (rainfed) rice on steep slopes may be extended to the riparian land of headwater streams.

 
Experimental Setting
Incoming and outgoing flows across riparian sites were monitored using 0.50-m-wide Gerlach troughs (Gerlach, 1967) connected by PVC tube to PVC water buckets. Each site (approx. 20 m wide), characterized by homogeneous vegetation, was instrumented with three troughs placed at the upper rim of the riparian area and three troughs placed at the lower rim. The upper rim was established immediately below the upslope field border; the lower rim was established close to the stream at approximately 1 m above the stream water level to avoid water from the stream entering the buckets. The total width of riparian sites therefore varied from 3 m to almost 12 m (Table 1 ). After each rainfall event, the total volume of runoff collected in the buckets was measured. After energetic stirring, a 1-L sample was collected from each bucket to measure average runoff sediment concentration. The sample was filtered, oven-dried at 105°C for 24 h, and weighed to obtain the sample's sediment amount. Trapping efficiency was calculated for runoff water volumes (water trapping efficiency [WTE]) and for sediment concentration (sediment concentration trapping efficiency [SCTE]) as the portion of inflow trapped between the upper and the lower rim (e.g., McKergow et al., 2004):

Formula 1[1]
where Xin is the water flow amount in liters per linear meter of contour line (L m–1 for WTE) or the average sediment concentration (g L–1 for SCTE) of the three upper troughs (inflow), and Xout is the water flow amount (for WTE) or the average sediment concentration (for SCTE) of the three lower troughs (outflow).


View this table:
[in this window]
[in a new window]

  		Table 1. Characteristics of the riparian sites instrumented with Gerlach troughs, Houay Pano catchment, Lao PDR.

 
Six sites were monitored during the 2005 and 2006 monsoon seasons, but the riparian sites differed among the seasons (Table 1). In 2005, we assessed WTE and SCTE of the most frequent riparian vegetation types occurring in the catchment (i.e., native grass, bamboo, and banana). More than 75% of Houay Pano riparian land is very steep (slope >40%); however, to assess the vegetation performances on different topographic settings, for each vegetation type we selected a gentle sloping site (slope <20%) and a steep site (slope ~60%). Conversely, in 2006 we assessed the effect of cultivation of annual crops on steep riparian land. Sites BB2 and BA2 (Table 1) were kept as representative sites for bamboo and banana in steep conditions. However, the placement of Gerlach troughs changed and defined slightly different areas, so the sites cannot be considered exactly the same in the two seasons and are named 3BB and 2BA, respectively. Adjacent to 3BB, 2BA, and a newly selected site of native grass, we cleared the vegetation and established three upland rice sites. Upland rice was chosen because it is the most frequent annual crop and because its cultivation in riparian land has been observed in the upper reaches of the Houay Pano stream. Moreover, given that vegetable cultivation is similar to upland rice, we considered the effect of upland rice cultivation as representative of annual crop cultivation in general. The four vegetation types differed in canopy and ground cover (Table 2 ) and in species composition (Appendix 1).


View this table:
[in this window]
[in a new window]

  		Table 2. Vegetation characteristics of selected vegetation types, Houay Pano catchment, Lao PDR.

 

View this table:
[in this window]
[in a new window]

  		Appendix 1.

Species composition of vegetation types as assessed on a 3 x 3 m plot at the middle of monsoon season (12 July 2005 and 21 July 2006).

 
Rainfall was measured with a tipping-bucket rain guage located in the middle part of Houay Pano catchment at a distance of approximately 100 m from most sites up to a maximum distance of approximately 500 m for site NG2.

Data Analysis
Previous studies showed that event trapping efficiencies are not normally distributed (McKergow et al., 2004; Sheridan et al., 1999; Lowrance and Sheridan, 2005). The WTE and SCTE distributions we measured could not be satisfactorily described using simple model distributions, so we used a nonparametric statistical approach for the data analysis. We computed means and 95% confidence intervals (CIs) by nonparametric bootstrapping (Efron and Tibshirani, 1993). A 20% trimmed mean was used to estimate the center of symmetry of each distribution. First, we generated 999 independent and identically distributed realizations of the empirical distribution of the actual data (i.e., 999 independent samples of the data, with replacement). This was achieved using the "sample" function of the freeware R statistical package (R Development Core Team, 2005). The sorted values of these samples were used to derive empirical quantiles of the bootstrap approximation of the distribution's center of symmetry, the k-th value in sorted order being equivalent to the k/(nboot + 1) quantile. These quantiles were used to calculate the bootstrap CI with {alpha} = 0.05 by the so-called percentile method (i.e., using the {alpha}/2 and 1 – {alpha}/2 percentiles of the bootstrap distribution to define the interval).


   Results
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
Discussion
 Conclusions
 REFERENCES
 
Some of the meteorological and hydrological data collected in 2005 and 2006 are shown in Table 3 . In 2005, the observations started half-way through the rainy season (end of July) and accounted for only 60% of total rainfall of the season (1100 mm), whereas in 2006, observations started at the beginning of May and covered the whole monsoon season. In terms of number of events, maximum event rainfall amounts, and maximum rainfall intensity, the 2005 and 2006 data collection campaigns were comparable. The distributions of inflow water runoff and inflow sediment concentration were log-normal. The geometric mean of incoming water runoff was higher during the 2006 season than during the 2005 season, but this reflected the different choice of sites rather than differences in catchment hydrological conditions. Student t tests comparing the lognormal distribution of water inflow in site BB2/3BB and site BA2/2BA (i.e., the only sites that were almost identical in the two seasons) indicated no significant differences among the samples. Inflow sediment concentration ranged from 0.03 g L–1 to 16.34 g L–1 but was generally low, with a geometric mean of 1.32 g L–1 across the two seasons.


View this table:
[in this window]
[in a new window]

  		Table 3. Summary of 2005 and 2006 data collection campaigns.

 
The estimated means and 95% CIs of WTE in 2005 and 2006, sorted from the worst to the best performing site, are shown in Fig. 2 (values are reported in Appendix 2). In 2005, WTE was comparable in four sites out of six, with values between –2 and 0. NG1 was the only site with positive WTE (0.20–0.56). The worst performer was BA1, with a mean WTE of –5, meaning that water runoff flowing out of the site was around six times the water runoff inflow. In this site, we repeatedly observed seepage occurrence; in some cases, we continued to collect water in the lower rim buckets even a day or two after the rainfall event. In 2006, WTE was largely negative in all upland rice sites and 3BB; WTE was positive only at sites 2BA and 1NG. In the banana site BA2/2BA, which was monitored throughout the campaign, WTE did not differ between 2005 and 2006, whereas the bamboo site BB2/3BB performed worse during the 2006 rainfall season. During both seasons, native grass sites were among the best performers, either crossing the zero line or being completely positive, which suggest that native grass may be the most effective vegetation to retain surface runoff. However, the 95% CIs show that most sites performed similarly.


Figure 2
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 2. Riparian site water trapping efficiency (WTE) mean and 95% confidence interval for 2005 and 2006, Houay Pano catchment, Lao PDR. Note the different scale of y axes in the two seasons. Site values are reported in Appendix 2.

 

View this table:
[in this window]
[in a new window]

  		Appendix 2.

Riparian sites' water trapping efficiency (WTE) means and 95% confidence intervals (CI). The values are presented in Fig. 2.

 
The estimated means and 95% CIs of SCTE in 2005 and 2006 are shown in Fig. 3 (values are reported in Appendix 3). In 2005, all riparian sites performed similarly, and SCTE ranged from –2 to 0.7. There seemed to be a trend (but not significant at {alpha} = 0.05) related to topographic settings, with the worst SCTE occurring on gentle sloping sites rather than on steep sites. In 2006, upland rice sites exhibited very negative SCTEs. In the banana site BA2/2BA, SCTE was higher in 2006 than in 2005, but in both seasons the site was the best performer, with mainly positive SCTE values. In the bamboo site BB2/3BB, SCTEs were comparable in 2005 and 2006.


Figure 3
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 3. Sediment concentration trapping efficiency (SCTE) site mean and 95% confidence interval in 2005 and 2006, Houay Pano catchment, Lao PDR. Note the different scale of y axes; SCTE y-axis is discontinued below –10 to allow comparison among sites. Site values are reported in Appendix 3.

 

View this table:
[in this window]
[in a new window]

  		Appendix 3.

Riparian sites' sediment concentration trapping efficiency (SCTE) means and 95% confidence intervals (CI). The values are presented in Fig. 3.

 
Means and 95% CIs estimated from all sites per vegetation type (Table 4 ) indicated that native grass had the best WTE, which was, however, not significantly different from zero. In terms of net infiltration, native grass riparian sites did not contribute water to the runoff inflow they received but did not retain water either. Banana and bamboo had mean WTE around –1 (i.e., the runoff outflow was around two times the runoff inflow); in these sites, riparian land contributed water to the runoff outflow as much as the sloping land above it. The mean WTE of upland rice was –8: The surface runoff outflow from rice in riparian land was nine times the runoff inflow received from the slope above it.


View this table:
[in this window]
[in a new window]

  		Table 4. Estimated mean and 95% confidence interval of trapping efficiencies for water (WTE) and sediment concentration (SCTE) for the four vegetation types of riparian land (data from two seasons and all sites per vegetation type), Houay Pano catchment, Lao PDR.

 
Banana sites consistently reduced the sediment concentration of surface runoff, with mean SCTE of 0.28. Native grass and bamboo had slightly negative SCTEs, and outflow sediment concentration was approximately 30 to 40% higher than the inflow sediment concentration. Upland rice sites showed a mean SCTE of –3: The sediment concentration in the outflow was four times higher than the inflow.

The sediment concentration in the outflow is of particular concern because riparian outflow directly reaches the stream and therefore immediately affects stream water turbidity. Inflow sediment concentration in 2006 was homogeneous (geometric mean = 1.6 g L–1) across all sites but two (sites 2BA and 3R), which received more turbid surface runoff (geometric mean = 2.6 g L–1). Median outflow sediment concentration differed significantly among riparian sites (Table 5 ). In the riparian sites converted to upland rice, outflow sediment concentration was two to five times the sediment concentration of their adjacent sites and two to five times the sediment concentration of the inflow.


View this table:
[in this window]
[in a new window]

  		Table 5. Median sediment concentration of the outflow (lower rim) measured at the riparian sites of 2006. Wilcoxon two-sample bilateral tests were conducted on adjacent sites.

 

   Discussion
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
Conclusions
 REFERENCES
 
Inflow and outflow measurements with Gerlach troughs are prone to errors. The area draining into the troughs cannot be easily identified and probably changes from one event to the next. Hillslope overland flow is also highly variable in time and space (e.g., Kirkby, 1988). Because of this, we considered the three 0.50-m Gerlach troughs of each rim as a unique system that effectively monitored 1.5 m of contour line, disregarding the variation observed among the three troughs and referring to water flows in terms of liters per linear meter of contour line. In some cases, buckets were found full, and bucket overflow probably occurred. Full buckets were found in 2005 in site BB1 (three cases, lower rim), BB2 (two cases, lower rim), BA2 (two cases, upper rim), and especially in BA1 (11 cases, lower rim). Despite increasing the number of buckets to collect water for each trough, in 2006 full buckets were observed in sites 2BA (one case, upper rim), 3BB (eight cases, lower rim), 2R (one case, lower rim), 3R (one case, lower rim), and 1R (seven cases, lower rim). The probable overflow means that WTE was underestimated where full buckets were found in the upper rim (site 2BA/BA2) and overestimated where full troughs were found in the lower rim.

Given the high variability of hillslope flows in space and time and the measurement errors, WTE estimations should be considered an explorative assessment. Water trapping efficiency depends on water runoff inflow and outflow, but the origin of the water runoff outflow is difficult to identify, being composed of water runoff inflow, overland flow generated in the riparian land, and return flow that may exfiltrate in the riparian land (seepage). Seepage was frequently observed in Houay Pano riparian land on gentle sloping sites such as BA1 and on steeper areas (Fig. 4 ). The locations of places where return flow inputs to streams are important cannot be identified simply from riparian topography in Houay Pano because of the complex geological structure of the catchment (Ribolzi et al., 2005). Depending on the geological setting of the site and hydrological conditions, which may change from one event to the next, seepage may occur above the riparian land, within it, or not at all. Seepage creates saturation of soils, which inhibits infiltration, reduces soil resistance to detachment and transport, and may trigger landslide movements and stream bank collapses. Under these hydrologic conditions, riparian land cannot effectively trap inflows in situ; instead, these sites may become important sources of runoff water and sediment to the stream (e.g., McKergow et al., 2006). Indeed, trapping efficiencies observed in the Houay Pano catchment agree with the work of McKergow et al. (2004), who found low trapping efficiencies of riparian land and negative trapping when seepage occurred. Ziegler et al. (2004) reported that in Northern Vietnam return flow is a dominant component of catchment hydrology, especially at the footslope of recently abandoned fields. Our study verified that, in at least one site, seepage in riparian land was the common situation. We suspect that return flow is a major water pathway in the Houay Pano catchment as well. Further research should verify this hypothesis.


Figure 4
View larger version (154K):
[in this window]
[in a new window]

  		Fig. 4. The presence of seepage could be observed during the clearing of riparian land. This riparian site was later cultivated with vegetables.

 
The hydrologic conditions of Houay Pano riparian land were probably the main reason why WTE and SCTE values were largely negative. Also contributing to the low trapping efficiencies we observed were the field conditions in which we operated and the analysis of event data instead of annual summaries.

Studies conducted in open field conditions with devices similar to the Gerlach troughs (Sheridan et al., 1999; McKergow et al., 2004; Helmers et al., 2005) reported lower trapping efficiencies (on the order of 15–20% of WTE and 20–60% of SCTE) than plot studies conducted under rainfall simulations. Plot experiments that are conducted under rainfall simulations and applying (often very turbid) runoff inflow work under conditions that are ideal for trapping; these experiments are important to understanding the factors affecting sediment trapping but may lead to an overestimation of trapping efficiency in the open field (Dosskey, 2001).

The gentler the slope and the longer the width of riparian area, the higher the trapping efficiency (Karssies and Prosser, 1999; Dosskey, 2001). In our study, we could not identify critical topographic factors affecting trapping efficiency, but the natural setting of riparian land in the Houay Pano catchment, which is very steep and narrow, is not ideal for trapping water and sediments.

Low SCTE may also be explained by taking into account sediment characteristics because smaller sediment particles have more difficulty settling in the riparian area than larger ones (Karssies and Prosser, 1999; Dosskey, 2001; Syversen and Borch, 2005). Soils along the slopes of the Houay Pano catchment are clayey. Textural analysis of composite samples collected during the 2005 field campaign showed that clay fraction amounted to 45 to 60% of inflow sediments, and 85 to 92% of sediment particles were <20 µm (fine silt and clay). These very fine sediment particles would require much longer and flatter riparian areas to settle.

We chose to estimate mean WTE and SCTE on an event basis. Conversely, almost all literature reports annual trapping efficiency values. The few studies that looked at event trapping efficiencies showed a large scatter in riparian response (Sheridan et al., 1999; McKergow et al., 2004; Wang et al., 2005; Schoonover et al., 2006). Event trapping efficiency depends on many factors that are not well understood, such as antecedent catchment hydrologic conditions and inflow amounts. The relationship between inflow runoff characteristics (i.e., water amount and sediment concentration) and trapping efficiency in the literature is controversial; some studies reported a positive effect, others reported a negative one, and others were inconclusive (e.g., Dosskey, 2001; Gharabaghi et al., 2001; Abu-Zreig et al., 2004). In Houay Pano, WTE was slightly positively correlated to inflow water amount (r2 = 0.13; n = 252; significant at {alpha} = 0.05), and SCTE was correlated to inflow sediment concentration (r2 = 0.40; n = 252; significant at {alpha} = 0.01), so better trapping efficiencies were observed for the most erosive events. Annual estimates would be more influenced by the fewer most erosive events, whereas event-based mean trapping efficiencies reflect the prevalence of lower inflow events across the observation campaign. Our choice of using event values was methodological (i.e., we wanted to make the most use of the statistical information of the dataset), and we considered that the observed variation in riparian performances was valuable information that would have been lost by using annual summaries. The bootstrapping method allowed us to achieve this methodological objective without requiring statistical assumptions about the dataset.

Because of measurement errors and high temporal and spatial variability of the observed trapping efficiency, WTE and SCTE estimations per vegetation type (Table 4) must be taken with care. Estimates for native grass and upland rice, which were derived from three different sites, are probably more accurate than for banana or bamboo. Moreover, in native grass sites, bucket overflow never occurred, which excludes measurement errors on WTE. Upland rice WTE is probably overestimated, especially because of site 1R, but this does not affect the result that upland rice was the worst vegetation for trapping water and sediment. More controversial are the results for banana. We effectively monitored only two sites: BA1 resulted in negative (and probably overestimated) WTE, but its hydrology was heavily affected by seepage occurrence; BA2/2BA performed very well. Overall, banana WTE was not different from bamboo. The same applies to the SCTE estimates: Even if banana was among the best performers in all cases, only site 2BA in 2006 had a significantly higher SCTE than other sites. The good SCTE of banana is probably linked to the maintenance of undergrowth vegetation (Table 2), the low soil disturbance level of the cropping system of banana plantations in Houay Pano, and the high evapotranspiration rate, which reduces soil moisture.

Bamboo sites were sources of water and sediment to the stream. The low undergrowth cover of bamboo sites (Table 2) probably limited the sediment retention. Our results differ from those reported by Schoonover et al. (2006), who found that the bamboo species giant cane (Arundinaria gigantea Chapm.) is an effective filter of water and sediment already at a riparian width of 3.3 m. Although some species of bamboo are probably more effective than others, the Schoonover et al. (2006) study area had high infiltration rates in riparian land, which is very different from Houay Pano conditions. We could not find other studies on the effectiveness of bamboo in trapping sediments. In Southeast Asia, bamboos are multipurpose species whose products are important for household consumption and market (de Beer and McDermott, 1996; Belcher, 1998); therefore, further research should assess their effect on soil and water conservation. Although banana cultivation in Houay Pano does not negatively affect the riparian filtering of sediment and seems to reduce sediment concentration in the outflow, cultivation of upland rice in riparian areas results in a very clear detriment to water quality in the stream.

The sampling method we used for measuring average sediment concentration underestimated the real sediment concentration of water runoff. The magnitude of the error depended on (i) lag time between the end of stirring and the sample collection, (ii) the settling velocity of sediments, (iii) the vigor exercised while stirring the water, and (iv) the utensils used for sampling (Bagarello and Ferro, 1998; Ciesiolka et al., 2006). Although this error may have led to underestimating sediment concentrations by a factor of two to four (Bagarello and Ferro, 1998; Ciesiolka et al., 2006), we believe that this did not affect the estimation of SCTE because the same error applies uniformly to the upper and lower rim troughs. However, the median sediment concentrations reported in Table 5 probably underestimated the sediment concentration of runoff water exiting riparian areas and directly entering into the stream. Moreover, sediment concentrations were event averages; peak sediment concentrations were probably much higher. Gentry et al. (2006) found a strong correlation between fecal bacteria contamination and water turbidity. Higher sediment concentration in the streams may thus negatively affect population livelihood and health. In northern Laos, less than 40% of households have access to safe water, and the link between land degradation, contaminated water, and poverty is undeniable (Dasgupta et al., 2005). Given that most rural populations rely on surface water for household consumption, the higher contamination risk brought about by cultivation of annual crops in riparian land may negatively affect human health.


   Conclusions
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Conclusions
REFERENCES
 
Trapping efficiencies of riparian vegetation measured in the Houay Pano catchment were largely negative. The natural settings of riparian land in this steep, narrow, and clayey headwater catchment limit the possibility of trapping sediment and pollutants in situ. Moreover, in Houay Pano, like in other humid and wet tropic environments (Ziegler et al., 2004; McKergow et al., 2004; Sidle et al., 2006), seepage in riparian land is commonly observed. Seepage inhibits infiltration and reduces soil resistance to detachment and transport. In these hydrologic conditions, the negative impacts on water quality resulting from the more intense use of sloping land cannot be counteracted by interventions limited to the riparian areas. On the contrary, under such circumstances, riparian land is more susceptible to erosion processes. Therefore, proper management of riparian land cannot replace proper management of sloping land, and proper management of riparian land is also necessary to avoid erosion taking place in the immediate proximity of streams.

In northern Laos, riparian land offers important opportunities for income-generating activities for the rural population. Relatively gentler slopes and the presence of water for irrigation make riparian land particularly apt for the cultivation of vegetable gardens from which produce can fetch high prices on the market. However, because of its proximity to streams, riparian land use heavily affects water quality. Native grass was the best vegetation in terms of infiltrating runoff inflows, thus reducing sediment mass delivered to the stream. However, it can generate income only if grazed, which would increase the risks of water contamination by fecal bacteria. Although bamboos, whether naturally occurring or planted, are important sources of nontimber forest products, they were not effective in reducing sediment pollution to the streams. Conversely, bamboo sites were sources of water and sediment. This study indicates that more research is needed to assess the effect of bamboo on soil and water conservation. Cultivation of banana offers good opportunities: Casks are a profitable cash product, and, when plantations are managed with little tillage while maintaining good ground cover, banana may reduce surface runoff sediment concentrations. On the contrary, the cultivation of upland rice on riparian land led to two- to five-fold increases in sediment concentration of surface runoff flowing to streams.

Our study addressed only one aspect of the relationship between riparian vegetation and water quality (i.e., the efficiency of trapping sediments flowing from the slopes via nonconcentrated water flows). Other ecological aspects may be equally or more important. Given the importance of subsurface water flows highlighted during the study, filtering of dissolved nutrients and prevention of stream bank collapses may be important aspects to be addressed by further research to thoroughly assess the influence of riparian vegetation on water quality.


   ACKNOWLEDGMENTS
 
The research was conducted within a post-doc fellowship issued by International Water Management Inst. of Colombo. We thank the colleagues of the National Agricultural Forest Research Inst. of Lao PDR (NAFRI) for their useful assistance during the study. The research was part of the Management of Soil Erosion Consortium (MSEC) program. Dr. Anneke de Rouw (IRD-Laos) and Dr. J.F. Maxwell (Herbarium, Faculty of Science, Chiang Mai Univ.) assisted in the identification of the species reported in Appendix 1. The work of Mr. Inpaeng DuangVong, Ms. Nora van Breusegem, Mr. Khankham Aphaivong, Mr. Rudolf van der Helm, and the collaboration of all MSEC field assistants were essential to this research and are gratefully acknowledged. We also thank the review team of the journal, whose comments helped improve the manuscript.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Conclusions
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Conclusions
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Vigiak, O.
Right arrow Articles by Valentin, C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Vigiak, O.
Right arrow Articles by Valentin, C.
Agricola
Right arrow Articles by Vigiak, O.
Right arrow Articles by Valentin, C.
Related Collections
Right arrow Surface Water Quality
Right arrow Best Management Practices
Right arrow Hillslope Analysis
Right arrow Runoff
Right arrow Soil Conservation

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome