| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Agriculture and Agri-Food Canada, Lethbridge Research Centre, Lethbridge, AB, Canada
b National Water Research Institute, Saskatoon, SK, Canada
c Prairie Farm Rehabilitation Administration, Canada-Saskatchewan Irrigation Diversification Centre, Outlook, SK, Canada
d National Water Research Institute, Saskatoon, SK, Canada
* Corresponding author (allan.cessna{at}ec.gc.ca)
Received for publication August 1, 2000.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: SSRID#1, South Saskatchewan River Irrigation District #1
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Irrigation runoff occurs with flood irrigation in which excess irrigation water is allowed to run off the lower end of the field to ensure adequate irrigation. As excess irrigation water leaves an irrigated field within an irrigation district, the runoff generally enters a system of drainage ditches, which carries cumulative runoff from several flood-irrigated fields to some type of receiving water.
In Saskatchewan, the largest irrigation district is the South Saskatchewan River Irrigation District #1 (SSRID#1), which encompasses 20600 ha on the east side of the South Saskatchewan River near the town of Outlook. Approximately 15000 ha are currently irrigated with water originating from the river. Of the irrigated area, approximately 11000 ha are irrigated by sprinkler systems, with the remainder irrigated using various flood-irrigation methods. Drainage water from the irrigation district is currently returned to the river.
Runoff water from flood-irrigated fields or treated irrigation canals can be unsuitable for downstream irrigation of crops due to contamination with herbicides (Jame et al., 1999). As well, nutrients and pesticides entering receiving waters via agricultural runoff may endanger freshwater aquatic wildlife or render the water unsafe for human or animal consumption. This study addresses the concern that drainage water from SSRID#1 entering the South Saskatchewan River may be detrimental to the quality of the river water with potential implications for downstream water use.
The objectives of the study were to (i) determine, over three growing seasons (1994 to 1996), inputs of several herbicides and plant nutrients (nitrogen [N] and phosphorus [P]) into the South Saskatchewan River via flood-irrigation runoff from the SSRID#1; (ii) relate these amounts to herbicide and fertilizer use on flood-irrigated fields within the irrigation district; and (iii) investigate the remedial effect of passage of the drainage water through a natural wetland.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Water Sampling and Flow Measurement
Drainage water samples were collected from the two drainage ditches during the 1994 through 1996 growing seasons. Each year, automated water samplers (Sigma Streamline 800 SL portable liquid sampler with integral flow meter; American Sigma, Medina, NY) were installed on the 1C and 9A drainage ditches for both water sample collection and for flow measurement.
Sampling Sites
In each year, a single automated sampler was installed on the 9A drainage ditch at a point (9A-River) just prior to the drainage water entering the South Saskatchewan River (Fig. 1). Three automated samplers were installed on the 1C drainage ditch each year. One sampler was installed downstream of the wetland at a point (1C-River) just prior to the drainage water entering the river and a second sampler was installed just upstream of the wetland (1C-Wetland); however, no flow measurements were made at the latter sampling site. The positioning of these two samplers permitted an assessment of possible remedial effects, with respect to herbicides and plant nutrients, of passing irrigation drainage water through a wetland. In 1994, the third sampler was installed just downstream of the 1C drainage area, 1C-South (94). In 1995 and 1996, this sampler was installed further upstream (1C-South) to better sample the drainage water from the relatively large flood-irrigated area near the Broderick Reservoir.
In 1995 and 1996, excess irrigation water was similarly sampled at a point (Canal) on the main supply canal just upstream of its juncture with the 1C drainage ditch (Fig. 1). In 1996, irrigation water was also sampled from the main supply canal for the 1C drainage area at a point (Reservoir) just downstream of the Broderick Reservoir for a 2-wk period. No flow measurements were made at the Canal or Reservoir sampling sites in 1996. The irrigation water was sampled to determine both the herbicide and nutrient content of the water applied to the flood-irrigated fields.
At each sampling site, daily water samples were collected from early June, when flood irrigation generally commenced, until early October each year. The automated water samplers were programmed to collect a subsample of drainage water every hour so that each daily 1-L sample of drainage water consisted of 24 subsamples. After shaking to suspend any sediments, approximately 250 mL of each daily sample was poured off for nutrient analysis, and the remainder of the sample was stabilized (pH < 1) for herbicide analysis by the addition of 5 mL of concentrated sulfuric acid. The integrated flow meters were programmed to record the flow of drainage water past the sampling point every hour and daily flows were determined as the mean of 24 hourly readings.
Farmer Survey
All farmers who flood-irrigated cropland within the study area were surveyed at the end of each growing season. Farmers were asked to provide the following information for each flood-irrigated field: crops grown, herbicides and fertilizers applied, rates and methods of application for each fertilizer (kg ha-1) and herbicide (g ha-1), area treated with each fertilizer and herbicide (ha), and timing and amounts (mm) of irrigations. This information was collected so that the relationship between the net outflow of herbicides and nutrients in the return flow water to the South Saskatchewan River and the total amounts applied to flood-irrigated fields within the 1C and 9A drainage areas of the SSRID#1 could be investigated.
Geographic Information System (GIS) Calculations and Map Generation
For each of the three years, land use, herbicide, fertilizer, and irrigation data collected during the farmer survey were imported into Arc/Info (ESRI, Redlands, CA) and merged with spatial data showing the quarter sections (64.75 ha) within the irrigation district. The production data for each quarter section were broken into between one and four parcels depending on individual practices. The resulting Arc/Info files (one for each year) were imported into ArcView for GIS queries and map generation. Although there was generally more than one management within a quarter section, most of the spatial analysis was conducted at the quarter-section scale.
Herbicide and Nutrient Analysis
Prior to herbicide extraction, the water samples, which contained small and varying amounts of sediment, were filtered under reduced pressure through a Buchner funnel equipped with a glass fiber filter paper to remove sediments. An aliquot (500 mL) of each sample was extracted with diethyl ether and the concentrated extract methylated with diazomethane prior to Florisil column cleanup as described previously (Cessna et al., 1985). Gas chromatographic (GC) quantification and confirmation of herbicide [ethalfluralin (N-ethyl-
,,-trifluoro-N-(2-methylallyl)-2,6-dinitroaniline); trifluralin (,,-trifluoro-2,6-dintiro-N,N-dipropyl-p-toluidine); triallate (S-(2,2,-3-trichloroallyl) diisopropylthiocarbamate); clopyralid (3,6-dichloro-2-pyridinecarboxylic acid) (1996 only); mecoprop ((RS)-2(4-chloro-o-tolyloxy)propionic acid); MCPA (4-chloro-o-tolyloxyacetic acid); dicamba (3,6-dichloro-o-anisic acid); bromoxynil (3,5-dibromo-4-hydroxyphenyl cyanide); diclofop ((RS)-2-[4-(2,4-dichlorophenoxy)phenoxy]propionic acid); and 2,4-D ((2,4-dichlorophenoxy)acetic acid)] residues in the drainage water extracts were carried out using a Hewlett–Packard (Avondale, PA) Model 5970B mass selective detector (MSD), which was operated in the selected ion monitoring mode. The GC–MSD system and operating conditions, as well as the criteria for confirmation using peak area ratios of characteristic fragment ions for each herbicide or methyl ester, have been described previously (Cessna et al., 1997). Sediments retained on the glass fiber filters following sample filtration were not analyzed.
The 250-mL subsamples poured off for nutrient analysis were analyzed for nitrate and nitrite, ammonia, total P, and orthophosphate (ortho P) using standard colorimetric methods (Cessna et al., 1994). In 1996, samples were not analyzed for ammonia.
Calculation of Herbicide and Nutrient Transport
Total amounts of the various herbicides (g) and N and P (kg) transported in the drainage water per day past sampling points on the 1C or 9A drainage ditches were determined as a product of the herbicide (µg L-1) or nutrient (mg L-1) concentration in the drainage water sample and the total volume per day (L d-1) of drainage water that flowed past the sampling point. By summing the amounts of nutrients and herbicides transported each day, the total amounts transported to the river over each growing season were calculated. Since each daily water sample consisted of 24 subsamples collected at hourly intervals, it was considered that herbicide concentrations detected in a sample would be representative of the average herbicide concentrations in the drainage water during that day. Similarly, since the daily flow in each drainage ditch, used to calculate the total volume of flow per day, was the mean of 24 measurements taken at hourly intervals, it was considered that the calculated total volume per day was representative of the actual total volume per day.
Nutrients
Three samples per week (Tuesday, Thursday, and Sunday) were analyzed for nutrient content, and nutrient concentrations in the unanalyzed samples were linearly interpolated. Since both nutrients in their various forms were present in all drainage water samples and since nutrient concentrations did not vary markedly between samples, it was considered that this determination of total nutrient transport would have been representative of actual nutrient transport in the drainage water.
Herbicides
Initially, three samples per week were also analyzed for herbicide content. However, unlike nutrient concentrations, herbicide concentrations could vary markedly from day to day. Thus, when any herbicide was detected in a water sample, samples from the previous and following days were also analyzed. Herbicide concentrations in remaining unanalyzed water samples were linearly interpolated. Trace (>0.01 to <0.05 µg L-1) concentrations and zero (when concentrations were <0.01 µg L-1) were used for interpolation. However, trace concentrations in the daily drainage water were not included in the total amounts transported to the river. Since the majority of water samples were analyzed, it was considered that this determination of total transport of each herbicide would have been representative of actual herbicide transport in the drainage water.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Not all of the land available for flood irrigation in the 1C and 9A drainage areas was irrigated each year. During the 3-yr study, the portion of total available land in both drainage areas that was actually flood-irrigated varied from 46 to 60%. However, a higher proportion (72 to 88%) of the 9A drainage area was flood-irrigated each year compared with the 1C drainage area (37 to 50%).
Herbicide Use
In 1994 and 1995, herbicides were applied to approximately 80% of the land available for flood irrigation in both the 1C and 9A drainage areas, but to only 64% in 1996. Generally, a smaller portion of the flood-irrigated land in the 9A drainage area was treated with herbicides compared with the 1C drainage area.
Nineteen herbicides were applied to flood-irrigated land drained by the 1C and 9A drainage ditches during the 1994 to 1996 growing seasons. Five herbicides were applied in all three growing seasons: MCPA; fenoxaprop((±)-2-[4-(6-chloro-1,3-benzoxazol-2-yloxy)phenoxy]propionic acid); ethalfluralin; bromoxynil; and clopyralid. Eight herbicides were applied in two of the growing seasons: thifensulfuron (3-(4-methoxy-6-methyl-1,3,5-triazin-2-ylcarbamoylsulfamoyl)thiophen-2-carboxylic acid); 2,4-D; dicamba; glyphosate (N-(phosphonomethyl)glycine); fluazifop ((R)-2-[4-(5-trifluoromethyl-2-pyridyloxy)phenoxy]propionic acid); tralkoxydim (2-[1-(ethoxyimino)propyl]-3-hydroxy-5-mesitylcyclohex-2-enone); tribenuron (2-[4-methoxy-6-methyl-1,3,5-triazin-2-yl(methyl)carbamoylsulfamoyl]benzoic acid); and clodinafop ((R)-2-[4-(5-chloro-3-fluoropyridin-2yloxy)phenoxy]propionic acid). Six herbicides were applied in only one growing season: 2,4-DB (4-(2,4-dichlorophenoxy)butyric acid); glufosinate (4-[hydroxy(methyl)phosphinoyl]-DL-homoalanine); trifluralin; imazamethabenz ((±)-2-(4-isopropyl-4-methyl-5-oxo-2-imidazolin-2-yl)-p-toluic acid); quizalofop (2-[4-(6-chloro-2-quinoxalinyloxy)phenoxy]propionic acid), and mecoprop.
Only six herbicides (ethalfluralin > MCPA > bromoxynil > 2,4-D > fenoxaprop > glyphosate) were applied in amounts greater than 100 kg over the 3-yr period and these six herbicides accounted for 89.6% of the total amount of herbicide applied. The eight herbicides that were applied to flood-irrigated land and monitored for in the drainage water collectively accounted for 87.5% of herbicide applied.
Each year, approximately 1100 kg of herbicides were applied to approximately 2200 ha of flood-irrigated land, so that, on average, 0.5 kg of herbicide were applied to each herbicide-treated hectare. Approximately 23% of the flood-irrigated land received no herbicide applications. The majority of this land, used for forage production, was within the 1C drainage area.
Fertilizer Use
In each of the three growing seasons, not all of the flood-irrigated land received fertilizer applications. The portion of the flood-irrigated land in both the 1C and 9A drainage areas that was fertilized with either N or P during the 3-yr period varied from 75 to 85%. Fertilizer applications in the 3-yr period totaled 222 Mg N and 69 Mg P. Average application rates were around 60 kg N ha-1 and 20 kg P ha-1 but rates as high as 200 kg N ha-1 and 50 kg P ha-1 were reported. Urea was the most common source of N but anhydrous ammonia and ammonium nitrate were also used.
Most of the fertilizer in the study area was side-banded either during or just prior to seeding. Broadcasting accounted for approximately 30% of N and P applications, whereas 25% of P and 6% of N were seed-placed (Table 1). Fertilizer placement can be a factor controlling nutrient transport in surface runoff. Banded and seed-placed applications are generally placed below the soil–runoff interaction zone, whereas a portion of broadcast applications, which are placed on the soil surface, may remain within this zone even after incorporation.
|
Drainage Water Flows
The variation in daily flows (m3 h-1) illustrated (Fig. 2) for the 1C-South, 1C-River, and 9A-River sampling sites from early June to late September 1995 is typical of that observed in other years. The area of flood-irrigated land drained by the 9A drainage ditch was three times less than that drained by the 1C drainage ditch and this is reflected in the much lower flows in the 9A ditch. Increased flows in the drainage ditches generally corresponded to periods of low rainfall. Conversely, flows generally deceased following significant rainfalls and occasionally ceased completely in the 9A drainage ditch. Flows also decreased toward the end of August when crops no longer required irrigation.
|
Herbicide Transport
Herbicides present in the runoff–soil interaction zone would include preemergence soil-incorporated herbicides and postemergence herbicides that penetrated the crop–weed canopy and deposited on the top few millimeters of soil. Losses of such herbicides in surface runoff, as a percent of the amount applied, are determined not only by the physical–chemical properties of the herbicides but also by the interval between herbicide application and surface runoff. Properties affecting transport in surface runoff include water solubility, vapor pressure, susceptibility to photolysis, persistence in soil, and adsorption potential for different soil types and soil–water regimes.
Ethalfluralin, trifluralin, and triallate have relatively low water solubilities, sorb strongly to soil, and have relatively long field half-lives (Table 2). Due to their relatively high vapor pressures, one of the main routes of dissipation of these herbicides from soil is vapor loss to the atmosphere (Grover et al., 1988). The dinitroaniline herbicides, trifluralin and ethalfluralin, are susceptible to photodegradation either when present on the soil surface or in runoff water (Cessna and Muir, 1991). Such properties would predict not only low mobility in soil and relatively low concentrations of these herbicides in surface runoff, but also that they would not be readily leached from the runoff–soil interaction zone. In contrast, the postemergence herbicides have much higher water solubilities, lower soil sorption, and shorter field half-lives. These properties would predict that a portion of these herbicides may leach, due to rainfall and irrigation, below the runoff–soil interaction zone, and that maximum concentrations of these herbicides would be expected in surface runoff shortly after their application.
|
Unexpected detections of herbicides not reported in the farmer survey could result from the susceptibility of individual field drains, the 1C and 9A drainage ditches, and the wetland to application drift or rainfall runoff from adjacent dryland fields, ditch bank weed control, or farmer error.
Irrigation Water
There were no detections of ethalfluralin and trifluralin in the excess irrigation water in either 1995 or 1996. Trace concentrations of dicamba, mecoprop, bromoxynil, triallate, and diclofop were detected in only a few samples each year. Infrequent trace concentrations of clopyralid were detected in 1996. In both years, more than half the samples had trace concentrations of MCPA with concentrations >0.05 µg L-1 being detected only in the fall of 1995. In contrast, essentially all samples in both years contained 2,4-D, with concentrations >0.05 µg L-1 occurring in nearly all samples in 1995 and approximately half the samples in 1996. Similar results were obtained from irrigation water samples collected on the main supply canal just downstream of the Broderick reservoir, except that no clopyralid was detected in 1996. These results suggest that MCPA and 2,4-D were probably already present in the irrigation water pumped from Lake Diefenbaker to the Broderick Reservoir and that herbicides may also have entered the irrigation water as it passed through the 1C drainage area in the main supply canal.
Drainage Water
Preemergence-Applied Herbicides
The application of preemergence-applied herbicides in the SSRID#1 generally occurs in May prior to seeding, but applications may also be made in the fall. Applied directly to the soil, these herbicides are incorporated into the soil either during application or immediately following application:
Ethalfluralin. Although large amounts of ethalfluralin were applied to large areas in both drainage areas in 1994 and 1996 and the 1C drainage area in 1995 (Table 3), it was not detected in any of the daily drainage water samples collected from any of the sampling sites in any year. The lack of ethalfluralin input to the river probably reflects its limited mobility in soil (Table 2) as well as dissipation by photodegradation and vapor loss to the atmosphere.
|
|
Postemergence-Applied Herbicides
In general, postemergence-applied herbicides were detected in the June–July drainage water samples and these detections probably reflect the first irrigation runoff water following spring application of these herbicides:
Clopyralid. The greatest input to the river occurred with clopyralid (0.31% of amount applied, Table 4) reflecting its much longer field half-life (Table 2). Applied only to the 1C drainage area in 1996 (Table 3) (the only year in which the herbicide was monitored), a total of 56 g of clopyralid entered the river, with the maximum concentration being 0.50 µg L-1 (Table 4). Infrequent trace concentrations of clopyralid detected in the 9A drainage water may reflect applications made in that drainage area the previous growing season.
MCPA. Smaller portions (approximately 0.06%, Table 4) of MCPA and mecoprop entered the river, reflecting their shorter field half-lives (Table 2). Applied in greatest amounts to the greatest area, MCPA was the only herbicide applied to both drainage areas in all three years (Table 3). Quantifiable amounts of MCPA entered the river in all three years from the 1C drainage area, but only from the 9A drainage area in 1995, when the largest amount of MCPA was applied within that drainage area (Table 4). Only trace concentrations of MCPA were detected in the 9A drainage water in 1994 and 1996. In September 1994, quantifiable concentrations were detected in the 1C drainage water, possibly reflecting fall application of the herbicide. The maximum concentration of MCPA detected in the drainage water was 2.8 µg L-1, which exceeded the water quality guideline for aquatic life and for irrigation water (Table 5). Some of the trace concentrations of MCPA detected in the drainage water may reflect the presence of the herbicide in the irrigation water. The amount of MCPA estimated to enter the river during the 3-yr period was 505 g.
|
Dicamba. Lower amounts (0.035%, Table 4) of dicamba applied entered the river, which may reflect greater leaching from the runoff–soil interaction zone, relative to MCPA and mecoprop, due to its higher water solubility and lower soil sorption (Table 2). Applied to both drainage areas in 1995 and 1996 in relatively small amounts (Table 3), infrequent trace concentrations entered the river from the 9A drainage area in both years and from the 1C drainage area in 1996. During 1994 and 1995, 12 g of dicamba were estimated to have entered the river. The maximum concentration of dicamba was 0.46 µg L-1, which exceeded the water quality guideline for irrigation water (Table 5).
Bromoxynil. Bromoxynil represented the lowest input (0.019%, Table 4) to the river, probably as a consequence of its lower water solubility, stronger soil sorption (Table 2), and its susceptibility to photodegradation. Applied to the study area in all three years, the total amount of bromoxynil applied was less than half that of MCPA (Table 3). Generally, only infrequent trace concentrations entered the river. A single concentration > 0.05 µg L-1 entered the river in 1996 when a relatively large amount of bromoxynil was applied to the 1C drainage area. However, unexpected concentrations entered the river in 1994 and 1995 from the 9A drainage area and included the maximum concentration of bromoxynil (0.32 µg L-1) detected in the drainage water. A total of 49 g of bromoxynil was estimated to have entered the river during the 3-yr period.
2,4-D. With properties similar to those for the other phenoxy herbicides (MCPA and mecoprop), the much higher input (1.4%, Table 4) of 2,4-D is an anomaly. In addition to 2,4-D in the drainage water due to application to flood-irrigated fields, this much larger percent transport was probably due to the presence of significant concentrations of this herbicide in the irrigation water and the possible spraying of a portion of the 9A drainage system. Concentrations of 2,4-D in the excess irrigation water were generally <0.2 µg L-1, with an average concentration of 0.074 µg L-1 over the 3-yr period. It was estimated that the irrigation water (using the average 2,4-D concentration and the volume of irrigation water over the 3-yr period) and spraying of the 9A drainage system would have contributed approximately 51 and 15%, respectively, of the amount of 2,4-D that entered the river. Thus, although applied to 1C and 9A drainage areas in 1995 and 1996 in relatively small amounts (Table 3), concentrations of 2,4-D >0.05 µg L-1 very frequently entered the river from both drainage areas in all three years. For extended periods of time, concentrations of 2,4-D in the drainage water exceeded the irrigation water guideline (Table 5). Maximum concentrations of 2,4-D were 0.70 and 0.34 µg L-1 in 1995 and 1996, respectively. In the fall of 1994, a maximum concentration of 75.7 µg L-1, which exceeded the aquatic life guideline, was detected at the 9A sampling site, suggesting that a drainage ditch from an individual field or the 9A drainage ditch had been directly oversprayed with 2,4-D. It was estimated that a total of 2616 g of 2,4-D entered the river over the three growing seasons (Table 4).
Diclofop. Unexpected trace concentrations of diclofop infrequently entered the river from the 9A drainage area in 1994 and 1995, but not in 1996. Some trace and quantifiable concentrations were detected in 1C drainage water in 1994 and 1996 such that a total of 16 g of diclofop were estimated to have entered the river (Table 4). The maximum concentration of diclofop detected was 0.09 mg L-1, which was less than any water quality guidelines (Table 5).
Percent Loss
The percent of applied amounts of the eight herbicides that entered the river from the irrigation district were generally less than edge-of-field losses reported for some of the same herbicides in runoff water from individual fields within the irrigation district (Cessna et al., 1994, 1996). Such edge-of-field losses ranged from 0.07 to 1.0% of applied amounts, depending on the average field half-life of the herbicide and the length of the interval between herbicide application and the first irrigation. Larger edge-of-field losses would be expected because herbicide concentrations in edge-of-field drainage water would frequently be diluted to trace concentrations by drainage water from untreated fields or other dissipation mechanisms, and trace concentrations were not included in the calculation of the percent losses from the irrigation district (Table 4). Consequently, the percent losses from the irrigation district are somewhat of an underestimation.
Effect of Wetland
Drainage water samples were collected upstream and downstream of the wetland on the 1C drainage ditch in order to investigate whether the passage of the drainage water through a natural wetland would have provided any remedial effect with respect to amounts of herbicides entering the river. Contaminant removal from water can result, in part, via uptake by aquatic plants. Constructed wetlands, used for remedial purposes, are generally designed to maximize the residence time of the contaminated water in the wetland and to enhance contact with vegetative growth (Peterson, 1998). Longer residence times also enhance other routes of herbicide dissipation from water, such as photodegradation, volatilization, microbial degradation, etc.
In the present study, vegetative growth, such as cattails and grasses, occurred mainly along the outer perimeter of the wetland. This resulted in a relatively wide vegetation-free channel, from the inlet of the 1C drainage ditch to its outlet to the river, that offered relatively unimpeded flow to the incoming drainage water. Consequently, based on the time interval between detection of pulses of herbicide in drainage water at the 1C-Wetland and 1C-River sampling sites, the residence time of the drainage water in the wetland was estimated to be of the order of 2 d.
The most realistic assessment of a possible remedial effect by the wetland was to consider those herbicides that were detected several times in concentrations >0.05 µg L-1 in samples collected at the 1C-Wetland sampling site. These included dicamba, MCPA, and 2,4-D in 1994, MCPA and 2,4-D in 1995, and MCPA in 1996. Since flow measurements were not taken at the 1C-Wetland sampling site, calculations of amounts entering the wetland were made using flow data from the 1C-River sampling site, assuming that 2 d were required for the water to flow between the 1C-Wetland and 1C-River sampling sites.
The lack of a consistent remedial effect by the natural wetland (Table 6) may reflect the short residence time of the herbicides in the wetland. Because the half-lives of dicamba (
40 d, Scifres et al., 1973) and 2,4-D (>64 d [Robson, 1968]; >170 d [Erne, 1963]) in ditch and pond water are considerably longer, such a short residence time may not have permitted sufficient dissipation to consistently show differences between the amounts of these herbicides entering and leaving the wetland. Apparent dissipation may have occurred due to dilution of herbicide concentrations to trace levels during passage of the drainage water through the wetland, because trace concentrations were not used in calculating amounts leaving the wetland and entering the river. In addition, dicamba, 2,4-D, and MCPA, which have a history of long use within the irrigation district, were probably also present in the sediments of the wetland. Thus, under some environmental conditions, release of sorbed herbicides from wetland sediments to the water column may have played a role in determining amounts of the various herbicides that entered the river. Similarly, herbicide inputs from dryland crop production adjacent to the wetland may have also been a contributing factor.
|
The effect of drainage water from SSRID#1 on the quality of the river water with respect to increased herbicide concentrations would be related to the daily loading (g d-1) of the various herbicides to the river and the corresponding average flow (L d-1) of the river. Generally, worst case scenarios would be represented by maximum daily herbicide loadings to the river. In the present study, maximum daily herbicide loadings to the river could be of the order of 2.5 times the total determined for the 1C plus the 9A drainage ditches (Table 7) because the 1C and 9A drainage ditches account for approximately 40% of the flood-irrigated area within the irrigation district.
|
Assuming that the mean concentration of 2,4-D in the excess irrigation water (0.074 µg L-1) was representative of that in the river water, the increase in river concentration (0.012 µg L-1, Table 7), due to the maximum loading of 2,4-D to the river water, was equivalent to a 16% increase. Such an increase would increase some trace concentrations to >0.05 µg L-1 and concentrations between 0.09 and 0.1 µg L-1 to >0.1 µg L-1, the water quality guideline for irrigation water (Table 5). However, daily loadings of 2,4-D greater than 50 g d-1 occurred only five times during the 3-yr study, so that increases in the river concentration due to the drainage water inflow were generally <7%.
Nutrient Transport
Aliquots of drainage water for total P and ammonia analysis were not filtered to remove sediments. Thus, estimates of transport of these nutrients in the drainage water would have included residues contained in the sediments. In contrast, aliquots for ortho P and nitrate and nitrite analysis were filtered. Because of the high water solubility of nitrate salts, only transport of ortho P in the drainage water may have been somewhat underestimated by removing the sediment through filtration.
Irrigation Water
Excess irrigation water from the Canal sampling site had very low concentrations of ortho P and nitrate, with 20 and 73% of the samples containing less than their respective limits of quantification. In the determination of the geometric means of the ortho P and nitrate concentrations, all concentrations less than the limit of quantification (0.002 mg P L-1 and 0.010 mg N L-1, respectively) were assumed to be equal to the limit of quantification. Concentrations of total P and ammonia in all excess irrigation water samples were greater than their respective limits of quantification of 0.002 mg P L-1 and 0.005 mg N L-1. Due to resource constraints, ammonia concentrations were not determined in any of the 1996 samples.
Drainage Water
Geometric means were used in the analysis of nutrient concentrations measured in the 1C and 9A drainage water (Table 8) because occasional extremely high nutrient concentrations obscured the overall trends if arithmetic means were used. Drainage water samples collected at the 1C-Wetland and 9A-River sampling sites (Table 8) clearly show that as the flood-irrigation water passed over fertilized cropland, nutrient concentrations increased due to interaction of the irrigation water with the runoff–soil interaction zone. Relative to samples collected at the Canal sampling site, concentrations of total P in the drainage water at these sampling sites increased two to five times, whereas those for ortho P increased by two to fourteen times. Nitrate concentrations in the drainage water were also substantially greater, by up to a factor of four, with the exception of the 9A drainage water in 1995, when the mean nitrate concentration showed little increase. In contrast, ammonia concentrations increased by less than a factor of two in the 1C drainage water and, in the 9A drainage water, were relatively unchanged from those detected in the excess irrigation water.
|
There was considerable temporal variation in the concentrations of all nutrients in both drains. The variation typically consisted of sharp increases in concentration that lasted for approximately 1-wk intervals. In 1994, these peak nutrient concentrations were found throughout the year but, in 1995 and 1996, the peaks were only found early and late in the season. Mean concentrations of total P at the 1C-River and 9A-River sampling sites were similar but concentrations of ortho P were greater at the 9A-River site, whereas nitrate and ammonia concentrations were greater at the 1C-River site (Table 8). The higher mean ortho P concentrations in 9A drainage water may be due a greater proportion of broadcast fertilizer applications within the 9A drainage area. Since P is relatively immobile in soil, it may have remained near the soil surface in the soil–runoff interaction zone, resulting in higher concentrations in runoff. The effect of broadcasting would not be seen with nitrogen applications because they tend to be leached out of the soil–runoff interaction zone and deeper into the soil profile.
Percent Loss
Although fertilizer applications are only one of a number of sources of nutrients in the drainage areas, it is still interesting to view amounts of nutrients leaving in the drainage water in terms of amounts of nutrients added as fertilizer. Since ammonia was not measured in 1996, amounts of N and P entering the river in 1994 and 1995 were used to calculate the proportion of the amounts applied. Totals of 161437 kg of N and 50579 kg of P were applied as fertilizer to the 1C and 9A drainage areas during 1994 and 1995 (Table 1). Amounts equivalent to 1.9% of applied N (3024 kg) and 2.2% of applied P (1103 kg) entered the river through the drainage ditches. However, effluent from the 1C drainage ditch corresponded to 2.4% of applied N and 2.8% of applied P, whereas corresponding values for the 9A drainage ditch were 0.3 and 1.0% P, respectively. The difference between the two drainage areas may have been partly due to nutrient contributions from the wetland on the 1C drainage ditch. When the calculations were repeated using concentrations measured for the 1C-Wetland site, only 0.6% of applied N and 1.6% of applied P were estimated to have entered the river in the drainage water.
Effect of Wetland
In general, all nutrient concentrations tended to increase due to passage of the drainage water through the wetland (Table 8). However, the increase was greatest in 1994, less in 1995, and, in 1996, passage of the drainage water through the wetland had little or no effect on the nutrient concentrations. These inconsistent changes in nutrient concentrations on passage of the drainage water through the wetland probably arose from the dynamics of the wetland ecosystem, which may act as either a source or sink of nutrients, depending on environmental conditions in the wetland and surrounding area. For example, increased nutrient concentrations may result from the release of nutrients from sediments and aquatic plants within the wetland or from ground water discharge into the wetland. Conversely, decreases in nutrient concentration after passage through the wetland may reflect nutrient uptake by aquatic plants or deposition on the wetland floor. In the present study, grazing cattle, which had direct access to both the ditch and wetland, were another source of nutrients between the 1C-Wetland and 1C-River sampling sites. The comparison of nutrient concentrations at 1C-Wetland and 1C-River sampling sites indicates that nutrients entering the river from the irrigation district would be derived only in part from flood-irrigation runoff occurring during the growing season.
Effect of Drainage Water on River Water Quality
Although the water of the South Saskatchewan River would have been subject to upstream inputs of both urban runoff and surface runoff from agricultural land, concentrations of the nutrients N and P in the irrigation water from Lake Diefenbaker were low (Table 8). As with the herbicides, the effect of the drainage water on quality of the river water, with respect to increased N and P concentrations, would be related to the daily loading (kg d-1) of the various forms of these nutrients to the river and the corresponding average flow (L d-1) of the river.
Even without considering the dilution effect due to the difference in flow between the drainage ditches and that in the river, nitrate concentrations in the drainage water entering the river were well below the Canadian Drinking Water Quality Guideline of 10 mg N L-1 (Canadian Council of Resource and Environment Ministers, 1987). The maximum daily loading of nitrate (21.3 kg d-1) to the river water during the 3-yr study would have only increased the nitrate concentration from 0.010 to 0.0104 mg N L-1; an increase of only 4% (Table 9). Similarly, average ammonia concentrations in the drainage water entering the river were also well below (by approximately one order of magnitude) water quality objectives, in this case, provincial guidelines for surface water (Saskatchewan Environment and Resource Management, 1995). However, the maximum daily loading (120 kg d-1) to the river occurred in August when the river flow was greatly reduced. Consequently, the ammonia concentration in the river water for that 24-h period increased by 41%, but still remained well below the provincial guidelines. Such increases would have been infrequent because ammonia loadings exceeding 10 kg d-1 occurred only four times during the 3-yr study.
|
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nitrogen and P were detected in all drainage water samples. Total input to the South Saskatchewan River during the 1994 and 1995 growing seasons corresponded to 1.9% of N applied to flood-irrigated fields and 2.2% of P. A portion of the nutrients entering the river originated in the wetland. Concentrations of P in the drainage water ranged from 0.02 to >1.5 mg L-1, and even the irrigation water, which originated from the South Saskatchewan River, exceeded 0.012 mg L-1, the guideline proposed for flowing waters in Alberta. Average ammonia concentrations in the drainage water were well below the Saskatchewan water quality objectives. Nitrate concentrations in the drainage water were always within Canadian water quality guidelines. Due to the large flows in the river, inputs of nutrients did not significantly increase nutrient concentrations in the river water and river concentrations of nitrate and ammonia remained well below water quality guidelines.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Tian, T. C. Granato, R. I. Pietz, C. R. Carlson, and Z. Abedin Effect of Long-Term Application of Biosolids for Land Reclamation on Surface Water Chemistry J. Environ. Qual., January 3, 2006; 35(1): 101 - 113. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
