JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

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 Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
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 Maître, V.
Right arrow Articles by Guenat, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Maître, V.
Right arrow Articles by Guenat, C.
Agricola
Right arrow Articles by Maître, V.
Right arrow Articles by Guenat, C.
Related Collections
Right arrow Water Quality
Right arrow Wetlands and Aquatic Processes
Right arrow Nutrients
Right arrow Ecosystem Management
Right arrow Nitrogen
Published in J. Environ. Qual. 34:707-716 (2005).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Wetlands and Aquatic Processes

A Methodology to Estimate the Denitrifying Capacity of a Riparian Wetland

Véronique Maîtrea, Anne-Claude Cosandeyb, Aurèle Parriauxa and Claire Guenatb,*

a Ecole Polytechnique Fédérale de Lausanne (EPFL), GEOLEP, GR B 1 383 (Bâtiment GC), Station no. 18, CH-1015 Lausanne, Switzerland
b Ecole Polytechnique Fédérale de Lausanne (EPFL), LPE, GR B 1 423 (Bâtiment GR), Station no. 2, CH-1015 Lausanne, Switzerland

* Corresponding author (claire.guenat{at}epfl.ch)

Received for publication May 10, 2004.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Numerous studies have shown that riparian wetlands can play an important role in reducing nitrate concentrations before the ground water discharges into streams. Denitrification has been identified as an important process for this removal. Several approaches have been proposed to predict the denitrifying removal capacity of a riparian wetland, but no widely used tool exists to precisely quantify this capacity at the landscape scale. We propose such a methodology based on modeling the spatial variation of soil–water interactions in the entire riparian wetland. Mean values of denitrification enzyme activity (DEA) within three soil-denitrifying classes were 604, 212, and 24 ng N g–1 h–1 for Classes 3, 2, and 1, respectively. The study area, having a ground surface of about 15000 m2, was underlain by an aquifer with a calculated volume of 60000 m3, less than 10000 m3 of which corresponded to active denitrifying horizons (Classes 2 and 3). By volume, approximately 30% of Class 3 and 70% of Class 2 were interacting with ground water. The denitrifying removal capacity of our wetland was calculated to be about 1.8 kg N m–2 yr–1. The calculated denitrifying capacity of our site was less than expected. This is due to the fact that not all ground water interacts with the horizons having the highest denitrifying capacity. Thus, we show that whatever the system is, specific local pedological and hydrogeological conditions and their interactions are paramount in controlling the denitrification process.

Abbreviations: DEA, denitrification enzyme activity • DNT, in situ denitrification • OC, organic carbon • TDBN, total dissolved bacterial nitrogen


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
RIPARIAN AREAS are the most important buffer zones between terrestrial and aquatic ecosystems (Blinn and Kilgore, 2001). In agricultural catchments, there are high nitrate loading rates to ground water from fertilizers, and numerous studies have shown that these areas can play an important role in reducing these nitrate loads in ground water before ground water discharges into streams (Cooper, 1990; Gilliam, 1994; Haycock and Pinay, 1993; Jacobs and Gilliam, 1985; Jordan et al., 1993; Peterjohn and Correll, 1984). Denitrification, which requires anaerobic conditions and specific microbiological populations, has been identified as an important process for nitrate removal in riparian areas. Because wetlands are defined as ecosystems characterized by hydric soils and plant and animal species adapted or partially adapted to life in saturated conditions (National Research Council, 1995), they are often considered as having a high ground water nitrate removal capacity (Hauer and Smith, 1998), but each type of wetland does not possess the same denitrifying removal capacity (Groffman et al., 1996a). Because this removal capacity is due to the specific soil–water interactions that occur in these environments, the effectiveness of the removal is strongly linked both to the proportion of ground water moving within the denitrifying active layers of the wetland (Lowrance, 1997; Maître et al., 2003) and to the denitrifying capacity of these active layers. Thus, it is difficult to predict the relevance of riparian wetlands in removing nitrate from ground water by denitrification. However, several approaches have been proposed at different scales. Some methods exist to quantify ground water denitrification in discrete locations (i.e., a transect or a ground surface of one square meter to a few hundred square meters) of riparian areas, and these rely on either laboratory microcosm studies (Groffman et al., 1996b; Schipper et al., 1993; Pinay et al., 1993) or direct in situ tracer tests (Addy et al., 2002; Sabater et al., 2003; Starr et al., 1996). As the denitrification process is highly variable spatially, these measurements should be extensive to extrapolate to the landscape scale. This implies intense and expensive monitoring. Moreover, riparian zones are complex environments that are spatially heterogeneous in both horizontal and vertical dimensions with respect to hydrology, soil characteristics, and biochemical processes. Thus, a difficulty in extrapolating from localized observations to the landscape scale is understanding the three-dimensional extent of interaction between nitrate-enriched ground water and zones of high removal capacity (Gold et al., 2001; Schipper et al., 1993; Flite et al., 2001; Hill et al., 2000; Cey et al., 1999).

Broader approaches have proposed various wetland classifications to identify riparian sites having high capacity for ground water nitrate removal. These classifications were based on geomorphology (Quinn et al., 2001), combined cartography of geomorphology and hydric soil status (Rosenblatt et al., 2001; Gold et al., 2001), hydrologic setting (Baker et al., 2001), or soil texture (Pinay et al., 2000). But these classifications are qualitative or semiquantitative. Recently, the method proposed by Ducros and Joyce (2003) underscores the effectiveness of riparian areas as buffers. This method was based on field criteria where each criterion was subjectively assigned an ordinal score. The authors concluded that validation of this method is needed, suggesting that monitoring of water quality field data be used for comparative assessment of their evaluation tool. Thus, no widely used tool exists to precisely quantify and predict the denitrifying removal capacity of a riparian wetland at the landscape scale, whereas there is an increasing trend for using and optimizing this natural property of riparian zones for the protection of surface water quality in many countries (Blinn and Kilgore, 2001; Tytherleigh, 1997; Cooper et al., 1997; Lin et al., 2002). Without a tool to help resource managers to identify riparian wetlands having high natural denitrifying removal capacity, there is still a great risk that mistakes will be made, the net result being nitrate-laden ground water moving into surface waters. In this paper, we propose a tool to predict the efficiency of a riparian wetland in attenuating the nitrate load in ground water. The aim is to quantify its denitrifying removal capacity at the landscape scale. Our methodology requires the spatial variation of the soil–water interactions over an entire riparian wetland to be modeled.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Site
The study site is located at the foot of the Jura mountains in Switzerland and belongs to the Swiss Plateau geographical unit (46°36' N; 6°24' E). The site elevation is 660 m above mean sea level with a mean annual precipitation of 1280 mm (1961–1990; Swiss Meteorological Institute data). Recent glacial deposits contain a regional ground water flow whose outlet generates a wetland.

This site was selected because it provides favorable conditions for denitrification in the ground water system as evidenced by the following. First, the riparian area consists of two parts, a sloped area of 3% decline, and an adjacent flat riparian zone. The study site essentially corresponds to the flat area (Fig. 1a) . Indeed, a flat riparian area (Burt et al., 2002) was proposed as indicative of denitrifying conditions. Second, the riparian area is partly forested and partly covered by an extensive meadow. The vegetation includes the wetland indicator plants European alder [Alnus glutinosa (L.) Gaertn.], European ash (Fraxinus excelsior L.), and bird cherry (Prunus padus L.) in the forest and meadowsweet [Filipendula ulmaria (L.) Maxim.], reed canary grass (Phalaris arundinacea L.), and common reed [Phragmites australis (Cav.) Trin. ex Steud.] in the meadow. Third, the area contains hydric soils [hydric soil designation being linked to both redoximorphic features and features related to accumulation of organic matter (Soil Survey Staff, 1993 in Rosenblatt et al., 2001)], the presence of which has been used as an indicator of effective denitrification (Simmons et al., 1992; Nelson et al., 1995; Gold et al., 1998). The spatial distribution of the soil (Fig. 1b) is the result of a complex spatial arrangement of nine horizons (Table 1). The soils, described in detail in Cosandey (2001a) and Cosandey et al. (2003a), were attributed to three references (reductisol, histosol, organosol) and three transitions (brunisol-reductisol, reductisol-brunisol, reductisol-histosol), which are briefly described in Table 2 according to the French classification (Association Française pour l'Etude des Sols, 1998). The equivalents according to the FAO (Food and Agriculture Organization of the United Nations, 1989) soil classification are Gleysol, Histosol, Leptosol, gleyic Cambisol, mollic Gleysol, and terric Histosol, respectively.



View larger version (46K):
[in this window]
[in a new window]
  		Fig. 1. Map of the study site showing (a) major physical constraints, slope, altitude, land use, and (b) soil type.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Characteristics of the soil horizons and their classification into the three soil-denitrifying classes.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Definition of soil types (ADc and ADs could occur in each soil type).

 
Soil Survey
Denitrification enzyme activity (DEA), organic carbon (OC), and total dissolved bacterial nitrogen (TDBN) were measured twice, once in the winter of 1999 (December, 54 samples) and once in the summer of 1999 (August, 39 samples). These data were obtained from 18 sampling profiles in the context of a European project on nitrogen removal in riparian zones (Pinay and Burt, 2001). Within each profile the sampling depths were chosen so that each sample was representative of one homogeneous soil horizon. The methodology is detailed in Cosandey et al. (2003b). Denitrification enzyme activity was measured for each sample using the Smith and Tiedje (1979) procedure, with an amendment of nitrate and carbon (30 g of fresh soil in 30 mL of a solution containing 4 mg glucose C + 10 µg KNO3–N per g fresh soil). Total dissolved bacterial N, representing the total microbial pool, was estimated using a simplified version of the Jenkinson and Powlson (1976) fumigation procedure. Total carbon content was measured by means of 1000°C combustion under oxygen flow using a carbon analyzer (Casumat 8-Adge; Wösthöff, Bochum, Germany). Mineral carbon content was determined using the same analyzer after addition of phosphoric acid. Organic carbon was calculated from the difference between those two results.

Cosandey et al. (2003a) applied the concept of functional horizons (Wopereis et al., 1993) to the denitrification process and were able to classify the horizons into three soil-denitrifying classes, according to the distribution of their DEA. For these three soil-denitrifying classes defined in Table 1, average values of DEA, OC, and TDBN were calculated and are reported in Table 3.


View this table:
[in this window]
[in a new window]
  		Table 3. Summary results of denitrification enzyme activity (DEA), total dissolved bacterial nitrogen (TDBN), and organic carbon (OC) for each soil-denitrifying class.

 
Aquifer Formation Characteristics and Ground Water Monitoring
Hydrogeological investigations, described in detail in a previous work (Maître et al., 2003), indicated that the aquifer formation [i.e., a rock unit that is sufficiently permeable so as to supply water to wells (Domenico and Schwartz, 1998)] consists of both fluvioglacial material (Dx, Dl) and overlying organic (An, A) and organomineral (S, G, ADc, ADs) horizons. This hydrogeological unit fills an old U-shaped valley formed on a till (compacted silty-clayed formation, Dm, that corresponds to the aquitard). The ground water flow is oriented northeast (azimuth 45°) (Fig. 2) . The general gradient is about 0.5% and increases near the river. This hydraulic condition prevails from November to July. The "low water" period is short, lasting only from July to October. Except for these annual "low water" periods, water table position is relatively stable. Moreover, the amplitude of the water table variation is weak compared with the total thickness of the saturated zone (see Fig. 4). Thus, in our site, it is relevant to work on annual hydraulic condition. April 2000 was chosen as representative of this annual condition.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 2. Map of the study site showing major physical constraints, location of the piezometers, and contours of the ground water flow (drawn for April 2000).

 


View larger version (32K):
[in this window]
[in a new window]
  		Fig. 4. Water table position from September 1998 to October 2000 in relation to the soil-denitrifying classes for selected piezometers. Numbers indicate soil-denitrifying classes. See Fig. 2 for the location of the piezometers.

 
Soil–Water Interactions: Definition and Cartography
Definition
In this paper, soil–water interactions are defined as the three-dimensional relationship between the three soil-denitrifying classes and the ground water flow within our riparian wetland. They correspond in vertical extent to the thickness of each soil-denitrifying class interacting with ground water, except for the zone influenced by capillary movement.

Cartography
The riparian wetland, representing a ground surface of about 15000 m2, was mapped in the following steps. First, a map of the bottom elevation of the aquifer formation (that corresponds to the elevation value of the top of the till aquitard [Dm]) was hand-drawn by interpolation from lithological observations made during piezometer installations completed by bore hole observations where the aquifer formation is the deepest (p5 and p2 in Fig. 2). Second, the thickness of the aquifer formation (i.e., An, A, S, G, ADc, ADs, Dx, and Dl horizons) was recorded from hand auger observations at each point along a grid (varying from 10 to 20 m). Third, the water table position was interpolated at each point of the pedological grid from the piezometric map. The thickness of the saturated zone (i.e., the part of the aquifer formation interacting with ground water, except for the zone influenced by capillary movement) was estimated from the elevation values of the water table and of the bottom of the aquifer formation. For each point of the pedological grid, horizons of the same soil-denitrifying class were grouped together, so that the profile of a particular class could be determined. Then, the spatial pattern of the soil-denitrifying classes was combined with the spatial pattern of the water table position using functions of a GIS tool (ArcView GIS 3.2; ESRI, 2005) following the procedure detailed in Cosandey et al. (2003a). This methodology enabled maps of the thickness of each soil-denitrifying class interacting with ground water to be drawn. Then, both the total volume and volume of each soil-denitrifying class interacting with ground water were calculated.

Calculation of Denitrifying Removal Capacity
For each horizon, the denitrifying removal capacity (kg N yr–1) was estimated by multiplying the volume interacting with ground water (m3), the bulk density (Mg m–3), and the characteristic DEA value of the horizon based on its soil-denitrifying class (ng N g–1 h–1), extended to one year. Because our aim was to calculate the denitrifying removal capacity (integrated over one year), we used the DEA value corresponding to the mean between winter and summer values (Table 3). Then, the denitrifying removal capacity of the mapped wetland was obtained by summing the denitrifying removal capacity of all horizons. Finally, considering that the ground surface above the calculated volumes of horizons is 15000 m2, the denitrifying removal capacity was also reported per unit of ground surface by dividing the results obtained in kg N yr–1 by this surface.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil Denitrification Rates
A summary of soil results is given in Table 3. Mean values of DEA within the soil-denitrifying classes were 604, 212, and 24 ng N g–1 h–1 for Classes 3, 2, and 1, respectively. Class 3, corresponding to the amalgamation of surface soil horizons, presented not only the highest DEA value, but also the highest OC and TDBN values (for example, for TDBN compare 191 mg N kg–1 for Class 3 with 6 mg N kg–1 for Class 1). The active denitrifying horizons corresponded to the horizons of Classes 3 and 2.

Spatial Variation of Soil–Water Interactions
As shown in Fig. 3a and 4 , the thickness of the saturated zone varied considerably within the riparian meadow, from more than 5 m in the middle of the wetland (p40, p26) to less than 1 m near the river (p23, p14). Thus, the thickness of the saturated zone was the greatest (about 8 m) in the middle of the wetland and decreased to the east and north (to less than 1 m), which corresponded to the direction of ground water flow.



View larger version (43K):
[in this window]
[in a new window]
  		Fig. 3. Map of the study site showing (a) thickness of the saturated zone and (b) water table position below the ground surface. The grid is 100 x 100 m.

 
As shown in Fig. 3b, the water level position ranged from 60 to 80 cm (to the south) to less than 20 cm below the ground surface (near the river and to the west). The total thicknesses of both Class 3 (p33, 0.75 m; p26, 0.25 m) and Class 2 (p26, 0.6 m; absent in p14, p23, p33) were highly variable (Fig. 4). Moreover, the thickness of each soil-denitrifying class interacting with ground water varied considerably throughout the wetland. Thus, in the middle of the meadow (p40), only the upper part of the ground water (i.e., a few centimeters) flowed through Class 3 whereas the thickness of Class 1 interacting with ground water was very high. Near the river, on the other hand, almost the total thickness of the denitrifying horizons interacted with ground water. The three-dimensional soil–water relationship is illustrated in Fig. 5 and 6 . From these maps, it is possible to compare the total thickness of a soil-denitrifying class with the thickness interacting with ground water (Fig. 5a and 5b for Class 3; Fig. 6a and 6b for Class 2). In the northeast, near the forest–meadow boundary, a large portion of Class 3 interacted with ground water. In the south, except in the western part, the water level was below Class 3, even where the thickness of this class was large. Almost all of Class 3 interacted with ground water near the river. Class 2 was not present in the north part of the study site and its thickness was greatest in the southeast part. Where present, almost the total thickness of this soil-denitrifying class was interacting with ground water. Total volumes of each soil-denitrifying class were compared with the volume of each interacting with ground water (Table 4). In fact, below the study area, representing a ground surface of about 15000 m2, the volume of the aquifer formation was calculated to be 60000 m3, less than 10000 m3 of which corresponded to active denitrifying horizons (Classes 2 and 3). Approximately 30%, by volume, of Class 3 and 70% of Class 2 were interacting with ground water. The volume of Class 3 interacting with ground water represented only 2.5% of volume of all soil-denitrifying classes interacting with ground water. More than 93% of total volume of all soil-denitrifying classes interacting with ground water was Class 1.



View larger version (43K):
[in this window]
[in a new window]
  		Fig. 5. Map of the study site showing (a) total thickness of the soil-denitrifying Class 3 and (b) thickness of Class 3 interacting with ground water. The grid is 100 x 100 m.

 


View larger version (46K):
[in this window]
[in a new window]
  		Fig. 6. Map of the study site showing (a) total thickness of the soil-denitrifying Class 2 and (b) thickness of Class 2 interacting with ground water. The grid is 100 x 100 m.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Calculation of the total volume and the volume interacting with ground water for each soil-denitrifying class and estimation of class contribution to the denitrifying removal capacity within the riparian wetland. This calculation was based on annual denitrification enzyme activity (DEA) mean values for the date selected as representative of an average annual hydraulic condition (April 2000).

 
Denitrifying Removal Capacity
The volume of Class 2 interacting with ground water (2339 m3) was almost twice the volume of Class 3 interacting with ground water (1423 m3). Despite the different DEA values of Class 2 and Class 3 they had similar denitrifying removal capacities (361 and 305 g N m–2 yr–1, respectively). In the same way, whereas the DEA value for Class 1 seemed negligible in comparison with that of Classes 2 and 3, its denitrifying removal capacity was much higher (1135 g N m–2 yr–1). Indeed, the ground water flow was located mainly in the Dx horizon.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Spatial Variation of Soil Denitrification Rates
Based on DEA measurements, our calculated denitrifying removal capacity represents an index of denitrifying enzyme densities and not actual denitrification rate at the time of sampling. In a previous study (Cosandey et al., 2003b), we have measured the springtime rates of both DEA and in situ denitrification (DNT) in our horizons. The average DEA ranged from 73 to 1232 ng N g–1 h–1 and the average DNT ranged from 4 to 36 ng N g–1 h–1. Nevertheless, the DEA is very helpful for the distribution of rates within horizons because it represents a potential capacity, which could occur if all conditions were optimal (Groffman et al., 1992; Schnabel et al., 1996). Moreover, DEA varies very little over the course of the year relative to DNT (Pinay et al., 2000). Thus, it is more adapted to our aim of calculating the annual integrated denitrifying removal capacity. Our results show that the potential denitrification rate varies to a great extent within our wetland, depending on the horizons. It was very low for all horizons of Class 1, where there was a lack of bacterial population. Similar results were found by Burt et al. (1999) and Ambus and Lowrance (1991), who found that there was no significant evidence of deep denitrification for, respectively, <0.4 m in different geomorphological units of a riparian zone in England, and subsoil samples taken from the top of the aquifer in Georgia, USA. Other authors, who considered the soil horizon as observation scale, found similar patterns. Bohlen et al. (2001) measured DEA varying from 760 to 1908 ng N g–1 h–1 in humic horizons and from 46 to 366 ng N g–1 h–1 in deeper mineral horizons in north-central New Hampshire. Flite et al. (2001), considering soil horizons in their sampling scheme, also found that denitrification may not extend to the entire riparian zone (Pennsylvania) because of mineral horizons not harboring bacterial populations capable of denitrification (DEA of 200 ng N g–1 h–1 in the A horizon; N2O production below detection value in the C horizon).

Denitrifying Removal Capacity of the Wetland
The denitrifying removal capacity of our wetland was calculated to be about 1.8 kg N m–2 yr–1, based on DEA measurements. In recent literature, some authors extrapolate results from discrete locations to an area. The comparison of our results with their data is difficult because these authors generally do not work on measured soil volumes interacting with ground water.

For example, Pinay et al. (1993) and Hanson et al. (1994) averaged subsurface measurements (0.10 and 0.15 m, respectively) without considering the water table level. Pinay et al. (1993) calculated the denitrifying capacity of two French riparian zones to be as high as 972 and 1018 kg N m–2 yr–1, based on DEA measurements. Hanson et al. (1994) estimated annual denitrification (based on DNT measurements) ranging from <5 to 40 kg N m–2 yr–1 in the USA. Groffman et al. (1996b) and Ashby et al. (1998) expressed the denitrifying capacity taking account of the thickness of the active soil. The denitrifying capacities of their riparian zones are lower (respectively, 6 kg N m–2 yr–1 and ranging from 1.43 to 0.28 kg N m–2 yr–1) than the other two studies mentioned but remain considerably higher than our results because they are based on DNT measurements, which were very low in our site (Cosandey et al., 2003b).

The validity of our results depends on the hydrological information about the riparian area. Indeed, the efficiencies of each soil-denitrifying class in removing nitrate from ground water can be compared only if the ground water flux occurring through each soil-denitrifying class is equivalent. Because the soil-denitrifying classes were defined only for the aquifer horizons, working directly on the horizons containing the ground water flow is fundamental to our methodology. Moreover, saturated hydraulic conductivities measured by piezometers using pumping tests (Maître et al., 2003) were similar for the organic horizons (2.5 x 10–5 to 3 x 10–4 m s–1) constituting Class 3 and the fluvio-glacial material Dx (8 x 10–5 to 3.7 x 10–4 m s–1) constituting a great proportion of Class 1 (Table 4).

Our results, similar to those of Addy et al. (1999), Ambus and Lowrance (1991), Hill et al. (2000), and Willems et al. (1997), highlight the importance of the soil–water interactions within a riparian wetland in controlling ground water denitrification. The denitrifying removal capacity of the riparian area could be supposed to be very high based on the existence of organic horizons (Class 3) alone. However, the ground water interacts mainly with the mineral horizons within the wetland. The ability of stream riparian zones to effectively remove nitrate has been found to depend on whether the denitrifying activity occurs in areas where the most ground water transport occurs (Lowrance, 1997; Hedin et al., 1998). However, the high denitrifying activity of the nonsaturated Class 3 can potentially increase ground water quality by reducing nitrate leaching. On the other hand, most of the ground water flows through Class 1, which presents a low but nonzero DEA, and this corresponds in our wetland to the class with the highest denitrifying removal capacity.

Another nitrate removal process is plant uptake. In our site, the mowing of the meadow followed by a removal of the plant biomass constitutes an efficient process of nitrogen removal, but the annual removal of nitrogen by denitrification was measured to be higher than the nitrogen removal by plant uptake (Cosandey et al., 2001b), whatever the pedo-hydrogeological settings are. Actually, for two different sampling stations, we obtained a removal by in situ denitrification (which is lower than DEA) that reached 0.042 and 0.055 kg N m–2 yr–1, whereas the removal by plant uptake was 0.011 and 0.013 kg N m–2 yr–1, respectively.

Interest for Management Tool
Our aim was to use attributes that are easy to measure and to map at the landscape scale to estimate riparian nitrogen removal capacity of riparian wetlands. We propose to extrapolate measurements of soil denitrification rates at discrete locations to entire riparian wetlands. This methodology takes the spatial variability of the soil, of the ground water, and of the soil–water interactions into consideration, and produces an estimate of the annual denitrifying removal capacity; it does not aim to study the intra-annual variations. This methodology is composed of the following steps:

Our results show that making predictions of denitrifying removal capacity is very difficult. Our site presents some simple indicators for efficient denitrification, such as flat topography, vegetation typical of a wetland, high DEA values measured in surface horizons, and hydric soils. Nevertheless, the calculated denitrifying removal capacity was lower than expected. This is due to the fact that not all the denitrifying horizons interact with ground water. In recent literature, more and more studies highlight the importance of understanding ground water flow to estimate the nitrate removal capacity of riparian wetlands (Cey et al., 1999; Devito et al., 2000; Simpkins et al., 2002). For example, Baker et al. (2001), on the basis of repeated observations of six general riparian hydrological types throughout lower-peninsula Michigan, showed that in any of these types, water may reach the river without being significantly influenced by riparian processes. The need to understand the relationship between ground water flow path and active denitrifying layers was also shown by Hill et al. (2000) and Gold et al. (2001). Our methodology combines both hydrogeological and soil factors, which are key factors that determine the ability of a riparian wetland to remove nitrate. Moreover, the variation of soil–water interactions is taken into account both vertically and laterally.

Generally, soil cartography is based on field observations and corresponds to a descriptive approach. With such an approach, horizons are identified on the basis of differences in texture, structure, color, and presence of mottles and other, mainly visual, characteristics, that do not necessarily reflect actual processes. We applied the concept of functional horizons, which is highly appropriate to quantify a process such as the denitrification. It provides a suitable basis for extrapolation of denitrifying data to the landscape scale.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Many guidelines for management of riparian zones, focusing on water quality protection, have been drawn up by landscape managers. These guidelines generally comprise various basic components, one of which is a minimum width (Blinn and Kilgore, 2001). In agreement with Devito et al. (2000) and Crow et al. (2000), our opinion is that soil–water interactions are more important than width of vegetated strips in influencing the ability of a riparian wetland to remove nitrate. We showed that specific local pedological and hydrogeological conditions remain predominant factors in controlling the denitrifying removal capacity. As already emphasized (Nelson et al., 1995; Rosenblatt et al., 2001), it is necessary to utilize high-resolution, large-scale soil and ground water maps simultaneously to guide the preservation and management of riparian areas.


   ACKNOWLEDGMENTS
 
These results were obtained from the European project NICOLAS (Nitrogen Control by Landscape Structures in Agricultural Environments), which was funded by the Environment and Climate Programme ENV4-CT-97-0395.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2005 34: 403-407. [Full Text]  




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 Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
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 Maître, V.
Right arrow Articles by Guenat, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Maître, V.
Right arrow Articles by Guenat, C.
Agricola
Right arrow Articles by Maître, V.
Right arrow Articles by Guenat, C.
Related Collections
Right arrow Water Quality
Right arrow Wetlands and Aquatic Processes
Right arrow Nutrients
Right arrow Ecosystem Management
Right arrow Nitrogen

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