| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Unidad de Suelos y Riegos, Centro de Investigación y Tecnología Agroalimentaria (CITA-DGA) and Laboratorio de Agronomía y Medio Ambiente (DGA-CSIC), Apdo. 727, 50080-Zaragoza, Spain
* Corresponding author (disidoro{at}aragon.es)
Received for publication February 23, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: NCR, nitrogen in canal releases • NDN, denitrification • NF, nitrogen fertilization • NGI and NGO, nitrogen in ground water inflows and outflows • NI, nitrogen in irrigation water • NL, nitrogen load in La Violada Gully • NMW, nitrogen in municipal wastewaters • NP, nitrogen in precipitation • NSF, symbiotic fixation • NSR, nitrogen in surface runoff • NU, nitrogen uptake by crops • NV, volatilization • VGW, La Violada Gully watershed • VID, La Violada Irrigation District
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The contribution of N fertilization to NO3 concentration and fluxes in surface waters has also been extensively investigated. In a long-term study, Lucey and Goolsby (1993) found that as much as 25% of the N fertilizer applied in an agricultural basin was exported through its drainage outlet. The seasonal patterns of N fluxes in surface waters have been related to land use and management (Arheimer and Lidén, 2000), flow regimes (Ekholm et al., 2000), or catchment characteristics (Van Herpe and Troch, 2000; Causapé et al., 2004).
Concerns about nitrate contamination in the surface waters of the Ebro River basin (Spain), where our study area is located, are based on the prevalence of irrigated agriculture in this basin (783 900 ha or 9.2% of the total surface; Confederación Hidrográfica del Ebro, 2004a). Within the Ebro River basin, Bellot et al. (1989) found a N export of 65.8 kg N ha–1 yr–1 in La Violada Irrigation District and Basso (1994) found a net nitrate export between 35.5 and 59 kg N ha–1 yr–1 (16-30% of the N applied as fertilizer) in the drainage waters of the 25 000-ha Bardenas Irrigation District.
The objectives of this work are to (i) characterize the N fertilization practices in La Violada gully watershed, (ii) determine the nitrate concentrations and loads exported in La Violada drainage waters, (iii) relate N fertilization practices to nitrate export patterns, and (iv) recommend improvements in present N management inefficiencies with the aim of reducing off-site N pollution.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
A companion paper (Isidoro et al., 2006) described the sampling of the different flow components in La Violada Gully (bypass flows from irrigation ditches, drainage from the irrigated lands, and ground water inflows) and their separation during the irrigation seasons of 1995 and 1996.
Nitrogen Balance
The main N inputs and outputs for VID and VGW were measured or estimated during the 1995 and 1996 hydrological years. The N balance equation may be written as:
 |
Neglecting the errors in the estimation of the terms of the N balance equation, the difference between the N inputs and outputs (
N) represents the change in the organic and mineral N content in the system. As not all the terms of the N balance equation have been measured, N will show the inaccuracies and approximations made in the balance.
While some terms of the N balance equation (NF, NU, and NL) were calculated from extensive data gathered from the study area, others were estimated from available data or from mean literature values. The estimates of NO3-N load in the different (minor) water flows (NI, NP, NSR, NCR, NMW, and NGI) were based on actual measurements or available data on their concentrations and on estimates of the flows (obtained by hydrograph separation, or data available from the Ebro River Basin Authority). For three significant terms, NSF, NV, or NDN, no actual measurements were made in situ. Their estimates are based on simple assumptions or mean values from the literature and they are presented only to show their importance relative to the other terms of the balance.
Finally, the N balance was performed on total N, and thus the different N pools (organic and mineral) in the system and the transformations of N from one to another (i.e., immobilization and mineralization) and between the different N mineral forms were not considered.
Nitrogen Inputs
Fertilization
The N fertilization practices for the irrigated crops grown in VID were determined through interviews with local farmers in 1995 (18 interviews) and 1996 (27 interviews). Farmers were asked about the type of N fertilizers applied (including manure), the amounts per unit area, and the dates of application. We also asked for the number and dates of irrigation, crop yields, and management of crop residues. The crops surveyed were corn (Zea mays L.), alfalfa (Medicago sativa L.), wheat, barley (Hordeum vulgare L.), sunflower (Helianthus annuus L.), rice (Oryza sativa L.), pepper (Capsicum annuum L.), and fruit trees (mainly olive, Olea europaea L., and apple, Malus domestica Borkh.) (Table 1). Due to the low number of answers obtained in each year for some minor crops such as pepper, rice, and fruit trees, both years were coupled together to get average figures. The N fertilization practices were also surveyed for rain-fed fruit trees and winter cereals (mainly barley) grown on the non-irrigated land of VID and on the dryland area of VGW (mainly wheat, mean answer for both years). The amounts of the different N forms applied (nitrate N, ammonia N, amidic N, or organic N) were obtained from fertilizer composition and bulk applications. For each N fertilizer application to each crop, the mean and standard deviation of the amount of N applied, the percentage of farmers that performed that application, and the dates of application were determined (Table 2).
|
|
|
Irrigation (NI), Precipitation (NP), Canal Releases (NCR), and Municipal Wastewater (NMW)
The N in irrigation water was determined as the volume of irrigation water diverted to VID times its N concentration (Confederación Hidrográfica del Ebro, 2004b). Since the historical Ebro River Basin Authority (Confederación Hidrográfica del Ebro) records showed that the NH4-N + NO2–-N concentrations in Los Monegros Canal water represent 22.5% of the NO3-N concentration, the NO3-N concentrations measured in 29 samples of irrigation water taken during the study period (average = 0.72 mg NO3-N L–1) were multiplied by 1.225 to get the total N concentration (average = 0.89 mg N L–1).
The N in precipitation was calculated as the product of the volume of precipitation recorded in a meteorological station located in VID times the average N concentration (0.49 mg L–1) measured by Bellot et al. (1989) in VID.
The N in canal releases was calculated as the product of the volume of canal releases estimated from hydrograph separation and information given by the Ebro River Basin Authority (Isidoro et al., 2004) times the Monegros Canal average N concentration (0.89 mg N L–1).
The N in municipal wastewaters was calculated as the product of the volume of municipal wastewaters, estimated as 80% of the water delivered for urban supplies, times an average N concentration of 50 mg L–1 (Hernández, 1992).
Surface Runoff (NSR) and Ground Water Inflows (NGI)
Nitrogen in surface runoff (NSR) entering VID was estimated from the NO3 concentration measured in two water samples taken in the three main gullies entering the irrigated area (0.62 mg NO3 L–1). The flow peaks due to surface runoff were clearly identified in the D-14 limnigraph and thus the surface runoff volume was estimated through the hydrograph separation method (Isidoro et al., 2006).
The contribution of ground water input to the gully flow could only be estimated for the 1996 irrigation season and was not included in the N balance. Ground water inflow for the 1996 irrigation season was estimated by hydrograph separation techniques as 2.1 x 106 m3, 6.5% of total outflow through D-14 (Isidoro et al., 2006). Nitrate concentration in ground water inflow was determined in water samples taken in Fuente de los Tres Caños spring (average 31.0 mg NO3 L–1 in the 1996 irrigation season). The contribution of ground water inflow during the 1996 irrigation season was used to determine the N yield of the irrigated land in the 1996 irrigation season.
Nitrogen Outputs
Nitrate Exported through La Violada Gully Drainage Waters (NL)
The daily N loads (NL) were calculated from the NO3 concentrations at the time of sampling and the daily mean flows measured at the D-14 monitoring station during the October 1994 to September 1998 period. Daily water samples were taken with an automatic sampler and mean daily flows were provided by the Ebro River Basin Authority.
Although N in drainage waters may be found as NO3–, NH4+, and organic N forms (the latter usually associated with sediments) only NO3– was analyzed in this work. This was not regarded as an important source of uncertainty since most of the N found in drainage waters from agricultural lands is in the NO3– form (Lowrance, 1992). However, the unaccounted organic N exported from La Violada Gully could be of some relevance during high-peak flows, particularly in winter when sediment transport is higher in the gully. The NO3– concentration in the gully is always presented hereafter as NO3 (mg NO3 L–1) and the mass exports in terms of NO3-N (NO3 as nitrogen, Mg NO3-N) just like the other terms of the N balance are presented as N (Mg N).
Nitrogen Uptake by Crops (NU)
Crop N uptake was estimated from the yield data gathered in the field surveys, the surface area of the different crops (Table 1), and the N content measured in crops grown in similar agricultural systems (Domínguez, 1984). Only the N content present in the harvested material was considered in NU because, according to the interviews, farmers incorporated the crop residues into the soil. However, as an unknown part of these residues are burned or stacked and exported from VID, this estimate of NU must be considered as a lower limit for N export in crop matter.
Volatilization (NV), Denitrification (NDN), and Nitrogen in Ground Water Outflows (NGO)
Volatilization losses from applied urea and ammonia fertilizers and from manure can be important in agricultural environments. Meisinger and Randall (1991) considered that NV could be between 0 and 20% of the surface-applied urea and ammonium nitrate, 0 to 40% of the applied ammonium sulfate, and 15 to 30% of the broadcasted solid manure. Following Puckett et al. (1999) we have chosen a mean loss of 10% N for applied urea and ammonia fertilizers and 35% N for applied manure.
Hauck and Tanji (1982) reported denitrification losses ranging from 5 to 20 kg N ha–1 in several agroecosystems in the United States and from 3 to 17 kg N ha–1 found in regional studies. From those data we have chosen a value of 15 kg N ha–1 to characterize NDN in VID and VGW.
Finally, NGO was not incorporated in the N balance because it was considered to be insignificant due to the narrow basin outlet at D-14 and because there is no deep percolation to regional aquifers (Isidoro et al., 2004).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nitrogen fertilization inputs in VID were mainly concentrated in January to April (corn pre-sowing and winter cereals side-dress applications) and in June to July (corn side-dress applications) (Fig. 1). A smaller N peak was also observed in October-November due to pre-sowing applications to winter cereals.
The total N fertilizer inputs were 23% higher in 1996 than in 1995 (Table 3 and Fig. 1) due to the 20% larger corn area (Table 1) and the 16% higher N rates applied to corn in 1996 (Table 3). This is also reflected in the daily distribution of N fertilizer in both 1995 and 1996 irrigation seasons (Fig. 1). Thus, the daily N peaks of corn pre-sowing and, in particular, the two corn side-dress applications were higher in 1996 than 1995. One reason for this higher N peaks could be the higher precipitation in the irrigation season of 1996 (precipitation = 223 mm) than in 1995 (precipitation = 129 mm), which limited the number of dates available for farmers to fertilize.
The total N fertilizer input in VGW was estimated at 2175 Mg in 1995 and 2795 Mg in 1996. The N applied to the irrigated crops in the district accounted for 39% of the total N fertilizer inputs to the basin. During the irrigation season (April-September) almost all the fertilization took place in the district (Fig. 2). During the non-irrigation season (October-March) the wheat fertilization in the rain-fed area of the basin led to two peaks per year corresponding to the pre-plant applications in October 1994 and 1995 and to the wheat's side-dressing in January 1995 and January-February 1996. As the fertilization doses of dryland wheat are known only through one interview and the exact distribution of wheat and barley in the dry land is not well known (as wheat is the dominant crop we assumed 100% wheat), the results of NF for VGW are only approximate.
|
|
|
The daily NO3 concentrations were much more variable during the non-irrigation season than during the irrigation season. The highest seasonal maxima took place in the rainiest seasons, and the concentration peaks generally lagged the precipitation events (Fig. 3a). During the low-precipitation non-irrigation season of 1995, the daily NO3 concentrations were very low because the low precipitation did not leach the residual N present in the soils or the pre-sowing N applied in October in the winter grains (Fig. 2). In the non-irrigation season of 1996 the high precipitation of December 1995 increased the NO3 concentration in late December, and the precipitation during January and February 1996 leached the post-plant N fertilizer applied to winter grains raising the NO3 concentration up to values of 80 mg NO3 L–1 on 18 Feb. 1996 (Fig. 3a). Similar peaks were observed in January-February 1997 and 1998 with even higher maxima (123 mg NO3 L–1 on 19 Feb. 1997 and 114 mg NO3 L–1 on 15 Feb. 1998). The fact that these peaks took place after the side-dress N applications to winter grains and during the high base-flow periods following the January-February precipitations suggested that an important part of the applied N was leached and exported from the basin through La Violada Gully. Thus, the N contribution in ground water inflows (NGI) from the upper dryland was important during the winter seasons.
The NO3 concentration was above the 50 mg NO3 L–1 (11.3 mg NO3-N L–1) European standard (European Union, 1998) several times during the irrigation and non-irrigation seasons, but only in the 1998 irrigation season did it remain above the standard for a long, continuous period (2 mo; Fig. 3a).
The daily N exported from La Violada Gully at D-14 (NL) resulted from a combination of daily flows and NO3 concentration; consequently high NL values were found during the four irrigation seasons, when high flows in the gully originated from drainage in VID. The flows in the June-September period were similar over the four years, so that the higher NL in the 1998 June-September period was due to higher NO3 concentration (Fig. 3a) in that season.
High NL also took place in April of the four irrigation seasons. Irrigating after pre-sowing N applications to corn in April to prepare the soil for sowing is a usual practice in VID. Also, farmers irrigate after sowing to promote corn emergence in these soils, which are prone to crusting (soil crusting in La Violada district is a major problem, hindering emergence and sometimes even forcing a second sowing). These practices caused the N fertilizer losses in April, as they leached the pre-sowing N fertilizer of corn, and should be discouraged when possible.
The daily NL during the irrigation season ranged from 0.3 Mg NO3-N d–1 (in 3 Apr. 1995) to 6.3 Mg NO3-N d–1 (in 22 July 1998), with the seasonal means increasing from 1995 to 1998 following the increases in NO3 concentration [1.2 Mg NO3-N d–1 in 1995 (standard deviation, SD = 0.8), 1.4 Mg NO3-N d–1 in 1996 (SD = 0.9), 1.8 Mg NO3-N d–1 in 1997 (SD = 0.8), and 2.6 Mg NO3-N d–1 in 1998 (SD = 1.7)].
During the non-irrigation season, daily NL was more irregular (in each season and between seasons). On the average, N loads were lower in the non-irrigation season and presented a higher relative variation [0.2 Mg NO3-N d–1 in 1995 (SD = 0.1), 0.7 Mg NO3-N d–1 in 1996 (SD = 0.6), 1.0 Mg NO3-N d–1 in 1997 (SD = 0.7), and 0.6 Mg NO3-N d–1 in 1998 (SD = 0.3)]. The highest mean NL took place in the rainier 1996 and 1997 non-irrigation seasons, and the maximum non-irrigation season daily NL took place during the high base-flow and high NO3 concentration events indicated previously (absolute maximum = 3.1 Mg NO3-N d–1 in 2 Feb. 1996).
Mean monthly 1995-1998 NL values were highest in the summer months (June-August) due to the combination of high flows and NO3 concentrations (Fig. 4). The high flows were due to drainage of excess irrigation applied in VID, whereas the high NO3 concentrations were due to side-dress N applications given to corn. Nitrogen load in July presented the highest variability due to the unusually high value of 1998. July 1998 presented the highest NO3 concentration and NO3 exports. This summer export pattern is due to the leaching of N fertilizers applied to corn and, in particular, to the application of liquid fertilizers (UAN 32) with inefficient irrigations given to corn grown in the shallow soils of the northeastern VID (Playán et al., 2000).
|
Seventy-five percent of the total nitrate load exported by the drainage waters occurred during the irrigation season (Table 5). The yearly NL more than doubled along the study period from 252 Mg NO3-N in 1995 to 579 Mg NO3-N in 1998. No explanations were found for this fact, although the study period is too short to ascertain any increasing tendency in N loads exported by La Violada Gully. The N yield of VGW averaged 21.8 kg NO3-N ha–1 for the four years studied, varying from 12.8 kg NO3-N ha–1 in 1995 to 29.5 kg NO3-N ha–1 in 1998 (Table 5).
|
The NO3-N exported by La Violada Gully waters (NL) during the irrigation seasons was 26% (1995) and 25% (1996) of the applied fertilizer N (NF) applied to the irrigated crops in 1995 and 1996 (Table 3).
The estimation of the monthly nitrate loads from the monthly flows under the present conditions of land-use, crop management, and irrigation practices may be a valuable tool to predict nitrate loads in the absence of concentration data to allocate allowable loads in watershed planning. Monthly NL (Mg NO3-N mo–1) and monthly flows (Q, hm3 mo–1) were significantly correlated (probability < 0.0001) by means of the following equation:
 |
Nitrogen Balance
Based on the information gathered in the 1995 and 1996 hydrological years, we calculated the N balance for VID and VGW (Table 6). Nitrogen fertilization (NF) and symbiotic fixation (NSF) accounted, respectively, for 57 and 65% and 40 and 32% of total inputs in VID, whereas the rest of the inputs were negligible. The N extracted by the crops (NU) and the N exported by La Violada Gully waters at D-14 (NL) accounted, respectively, for 67 and 60% and 19 and 27% of total outputs from VID. Nitrogen uptake (NU) accounts only for the N removed in the cropped matter. Although most farmers incorporated crop residues into the soil, a certain amount of N is lost when residues are burned or packaged and removed from the district, adding to the uncertainty of the results.
|
N) represents the unaccounted terms that include ground water inputs and outputs. The N values were 362 Mg in 1995 and 355 Mg in 1996 in VID, representing around 23% of the average inputs and outputs in both years. We, therefore, considered that the closing of the balance was satisfactory. Higher N values have been found in other studies at the irrigation district level, such as those of McMahon and Woodside (1997), who found a 51% N value in eight agricultural catchments.
The N values were calculated for VGW (thus, NSR was not considered) as 804 Mg in 1995 and 539 Mg in 1996. The lower value in 1996 was due to higher nitrate leaching resulting from higher winter rainfall and extraordinary yields of the rain-fed barley and wheat in that year (higher NL and NU). These results for VGW are only approximate because NF and NU for dryland wheat are based on one single interview and the actual surface of wheat and barley in the dry land was not known.
During the 1996 irrigation season, ground water inflows into VID were estimated by end-member mixing analysis (EMMA) (Isidoro et al., 2006). From the EMMA estimates of ground water inflows and their NO3-N concentration, NGI was estimated as 18.7 Mg or 7.1% of the NL exported at D-14. This reduced contribution of ground water inflows to NL during the 1996 irrigation season was due to the low contribution of ground water inflow to total flow (6.5%; Isidoro et al., 2006) and the NO3 concentration of ground water inflow (30.9 mg NO3 L–1), slightly lower than NO3 concentration in the drainage waters of VID. Discounting from NL the N inputs from outside VID (NGI and NSR), we estimated that the nitrate loading attributable only to irrigation in VID was 246.3 Mg NO3-N during the 1996 irrigation season, or 63.7 kg NO3-N ha–1 for the irrigated land.
Most of the N export took place during the irrigation season and was carried by the drainage waters from the irrigated soils. Thus, the strategies to reduce N loading should focus on reducing drainage flows and NO3-N in the drainage waters. Achieving better on-farm irrigation efficiency would reduce drainage flow (change of the irrigation system and looking for alternatives to irrigation to overcome soil crusting). The adequate timing of irrigation and fertilization practices and adjusting fertilization rates to crop requirements could reduce NO3-N in drainage water.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The time distribution of N loads in La Violada Gully clearly lagged behind N fertilization in VID during the irrigation season and was determined by N fertilization in VGW and rainfall patterns in the non-irrigation season. The highest non-irrigation season loads took place in February of the more rainy seasons (1996 and 1997) following high rainfall events in January-February and also after the side-dress N applications to wheat in the rain-fed area of the basin. So, N loads in the non-irrigation season were mainly determined by the leaching of the wheat N fertilizer by rainfall.
Most of the N export through La Violada Gully (75%) took place during the irrigation season, when flow is dominated by drainage from VID, which unlike rainfall is a controllable factor. Thus, efforts to reduce N exports should focus on reducing drainage and NO3-N concentration in drainage waters. The high N loads in the gully from June to August were caused by important N fertilizer losses from the corn side-dress applications. Also, the usual practice of fertilization followed by irrigation previous to corn sowing (and immediately after, to promote its emergence) was responsible for significant N fertilizer losses in April. Therefore, fertilization practices in VID are responsible for a good deal of the N export through drainage along with timing and low efficiency of irrigation. To reduce N fertilizer losses in La Violada Gully, the doses of N fertilizer applied to corn (especially in the side-dress applications) and other crops should be reduced and the use of liquid fertilizers with the traditional surface irrigation system of the district discouraged.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
