Pd/a Crsp Eighteenth Annual Technical Report

When fish are recovered from ponds, the effluent is often drained, presenting both an environmental challenge and an agricultural opportunity. The effects of irrigation with pond effluent and its interaction with applied fertilizer were assessed in a field experiment using French bean (Phaseolus vulgaris) and kale (Brassica oleracea) over two growing seasons near Sagana, Kenya. Fresh and dry matter yields of the crops were recorded at harvest, and samples were collected for determination of tissue nutrient concentration. In the first season, French bean fresh pod yield differed significantly (P = 0.05) among treatments. Plots receiving canal water and fertilizer at recommended rates had the highest yield (9.1 t fresh pod ha-1), while those receiving no fertilizer or irrigation had the lowest yield (1.3 t fresh pod ha-1). In the second season, significant differences (P = 0.05) were observed among treatments in fresh bean pod and fresh kale leaf yields. The highest (4.4 t ha-1) fresh pod yield was observed in pond-effluent-irrigated and fertilized plots, while the lowest (1.3 t ha-1) was observed in non-irrigated, unfertilized plots. The highest fresh kale leaf yield (11.5 t ha-1) was obtained using irrigation with canal water combined with fertilizer application, while the lowest (4.2 t ha-1) was observed in non-irrigated, unfertilized plots. Low nutrient status in the pond water, together with inadequate water supplied to some crops due to emitter clogging, was responsible for low yields in treatments where pond water was substituted for canal water. Pond water from the Sagana Fish Farm supplied low amounts of nitrogen (N) and phosphorus (P) for crops, indicating that recommended rates of mineral fertilizers should be used when pond water is used for irrigation. In the second experiment, the effectiveness of two types of soil occurring at Sagana, Kenya— a vertisol (black clay soil) and a cambisol (red clay soil)—in retaining nutrients from pond effluent was investigated. A laboratory experiment was conducted with soil columns containing red or black clay soil. Pond effluent application rates of 31, 81, and 161 mm d-1 were tested on both soils. Both soils retained over 60% of total phosphorus from pond effluents, with red clay soil retaining 27% more phosphorus than black clay soil. At the high effluent loading rate, low % N removal was observed in both soils. Total nitrogen removal efficiency declined with time after 21 days at the high rate, and after that time no nitrogen removal was observed where red clay soil was used. Black clay soil was more enriched by nitrogen than red clay soil, while phosphorus enrichment was higher in red clay soil than in black clay soil. It appears that land application can remove substantial amounts of phosphorus and nitrogen from pond effluent. EIGHTEENTH ANNUAL TECHNICAL REPORT 70 water combined with 75 to 100% of the nitrogen requirement as fertilizer were comparable with treatments irrigated with aquacultural effluents combined with 25 to 50% of the nitrogen requirement. These results imply that the application of 150 to 225 kg N ha-1 for well water irrigation and 75 to 160 kg N ha-1 for aquaculture effluent irrigation containing 40 mg l-1 is sufficient for optimum grain yield and WUE. Similar results were obtained by Al-Jaloud et al. (1993). When pond effluents are applied in arid and semiarid environments, greater crop returns may be obtained through more efficient application methods. In Kenya, where farm ponds can also serve as water reservoirs for irrigation, drip irrigation could be profitable. Drip irrigation is a technique whereby water and fertilizers can be placed directly over the root zone through use of emitters that are calibrated for low flow rates. Drip irrigation appears most promising when water and fertilizer application is split into several events over a cropping season. Little work has been conducted in East Africa on the use of fish pond effluent as a source of irrigation water for high-value crops. A study was undertaken to determine the effects of irrigation with polyculture tilapia (Tilapia aureus) and African catfish (Clarias gariepinus) pond water on yield of French bean (Phaseolus vulgaris) and kale (Brassica oleracea). Specific objectives of the study were to: 1) Evaluate pond effluents as a source of irrigation water for French bean and kale; 2) Assess the ability of pond effluents to supply nitrogen and phosphorus to French bean and kale; and 3) Determine the effectiveness of two soil types from the Sagana, Kenya, area to retain nutrients from fish pond effluent. METHODS AND MATERIALS The project was conducted between October 1998 and September 1999 at the Department of Fisheries Fish Farm at Sagana in central Kenya. The farm lies at an elevation of 1,231 m above sea level. Rainfall at the farm ranges from 1,332 to 1,612 mm yr-1, and daily average air temperatures range from 16.3 to 26.9°C. Water supply to the farm was from the Ragati River. Separate studies were conducted in this project. The first study was conducted to investigate the potential benefits of applying pond effluents to crops and the effects of the effluents on the yield of French bean and kale. The second investigation examined the potential use of land to purify fish pond effluents. Soils at the farm are “black cotton soils” (vertisols) of volcanic origin. Table 1 shows the pH, nutrient, and other chemical concentrations (Hue and Evans, 1986), bulk density, and hydraulic conductivity (Klute et al., 1986) of the vertisol (black clay soil) and cambisol (red clay soil) used in both the experiments at the start of the trials. Water analyses for both experiments were done using standard methods. French Bean and Kale Field Experiment The experiment was conducted during two growing seasons. The first season started in October 1998 and ended in February 1999. The second season started in June and ended in September 1999. For both runs of the experiment, one of the fish ponds on the Sagana Fish Farm was selected to supply effluent. The pond was fertilized with 8 kg P ha-1 as diammonium phosphate during a 17-wk period prior to stocking. The pond was then stocked with tilapia and African catfish. Subsequently, the pond received 20 kg N ha-1 wk-1 and 8 kg P ha-1 for the 17-wk grow-out periods for both runs of the experiment. First Growing Season Eighteen field plots measuring 10 × 6 m were prepared on land previously under star grass (Digituara scalarum). Plots were hand-tilled and hand-harrowed sufficiently for planting French bean. In October 1998, plots were planted with French bean (var. Samantha) at a spacing of 0.6 × 0.1 m. Bean plants were sprayed with Antracol® and Ripcord® at a rate of 80 l ha-1 at 14-d intervals for pest and disease control. The experimental design was an incomplete factorial arranged as a randomized complete block with six treatments replicated three times. Treatments consisted of: • Non-irrigated, unfertilized (–I –F); • Non-irrigated, fertilized (–I +F); • Drip-irrigated with canal water, unfertilized (+I –F); • Drip-irrigated with canal water, fertilized (+I +F); • Drip-irrigated with fish pond effluent, unfertilized (+P –F); and • Drip-irrigated with equal parts canal and pond water, unfertilized (+IP –F). At planting, diammonium phosphate (DAP) (200 kg ha-1) was applied to treatments receiving fertilizer. These treatments received an additional 200 kg ha-1 of calcium nitrate as top dressing after bean emergence. Plots receiving irrigation water were fitted with garden drip irrigation systems. A 10-l distribution bucket suspended on a post held water (canal or pond) to irrigate individual plots receiving irrigation treatments. Plots receiving water via drip irrigation were fitted with a F1 0.75-in filter (Lego Inc., Israel) that was intended to remove particulate matter. Drip-irrigated treatments received 0.33 mm water d-1 over a growing season of 74 days. French bean harvest began 46 days after planting and continued for 28 days. Fresh and dry weight of bean pods and total biomass dry weight (dry pods, leaves, and stems) were recorded. Twenty-one days after transplanting, leaf samples were picked for nutrient analysis. After 76 days, the bean crop was uprooted and the above-ground biomass dried out on polythene sheets for biomass yield records. Second Growing Season Twenty-four plots measuring 5 × 6 m were prepared on the previous season’s experimental site, with an additional twelve plots prepared on an adjacent uncultivated portion under star grass. The land was hand-tilled and hand-harrowed to the recommended tilth for French bean and kale. The experiment consisted of twelve treatments arranged as a 2 (crops; French bean and kale) × 2 (fertilization; 0 and recommended rates) × 3 (drip irrigation; 0, canal water, and pond water) factorial laid in randomized complete block design with three replicates. Treatments were: • No irrigation and no P or N for kale (K: –I, –F); • No irrigation and no P or N for bean (B: –I, –F); • No irrigation plus 78 kg N ha-1 in splits of 48 kg N ha-1 at EFFLUENTS AND POLLUTION RESEARCH 71 planting time and 30 kg N ha-1 four weeks after planting and 54 kg P ha-1 for kale (K: –I, +F); • No irrigation plus 36 kg N and 40 kg P ha-1 for bean (B: –I, +F); • Drip irrigation at 2.3 mm d-1 with canal water and no P or N for kale (K: +IC, –F); • Drip irrigation at 2.3 mm d-1 with canal water and no P or N for bean (B: +IC, –F); • Drip irrigation at 2.3 mm d -1 with water from the canal and application of 78 kg N ha-1 in splits of 48 kg N ha-1 at planting time and 30 kg N ha-1 four weeks after planting and 54 kg P ha-1 for kale (K: +IC, +F); • Drip irrigation at 2.3 mm d-1 with water from the canal and application of 36 kg N and 40 kg P ha-1 for bean (B: +IC, +F); • Drip irrigation at 2.3 mm d-1 with pond effluent providing 6.3 kg N ha-1 and 2.6 kg P ha-1 for kale (K: +IP, –F); • Drip irrigation at 2.3 mm d-1 with pond effluent providing 6.3 kg N ha-1 and 2.6 kg P ha-1 for bean (B: +IP, –F); • Drip irrigation at 2.3 mm d-1 with pond effluent providing 6.3 kg N ha-1 and 2.6 kg P ha-1 with application of 78 kg N ha-1 in splits of 48 kg N ha-1 at planting time and 30 kg N ha-1 four weeks after planting and 54 kg P ha-1 for kale (K: +IP, +F); and • Drip irrigation at 2.3 mm d-1 with pond effluent providing 6.3 kg N ha-1 and 2.6 kg P ha-1 with application of 36 kg N and 40 kg P ha-1 for bean (B: +IP, +F). Kale seedlings were transplanted to the field on 9 June 1999 at a spacing of 0.9 × 0.3 m and watered for seven days using a watering can to facilitate establishment. The crop was sprayed with Dimethoate® and Antracol® every two weeks to protect it against insect and fungal attacks, respectively. French bean seeds were direct seeded on 12 June 1999 at a row spacing of 0.6 m and line spacing of 0.1 m. Other practices were done as in the first season. Plots receiving irrigation water were fitted with garden drip irrigation systems. Water for drip irrigation was lifted to a 70-l distribution barrel in each irrigated plot using a pedal pump and applied daily at 1100 h. Plots under French bean receiving drip irrigation water were fitted with an Alkal 1.5-in filter (which was bigger and more efficient than F1 0.75-in used in the first season), while those under kale were fitted with F1 0.75-in filters used in the first season. Kale harvesting began 22 days after transplanting and continued for 42 days by removal of the lowest three leaves per plant every four days. The weight of fresh kale leaves harvested was measured at the desired harvesting period from each of the plots receiving different treatments and summed over the harvest period. Samples for leaf nutrient analyses were collected from the second uppermost leaf of kale just before the final harvesting period. Harvest ceased 80 days after transplanting, and dry matter yield for kale was determined for all treatments as the sum of leaf dry matter over the course of the experiment. Second-season French bean harvest began 52 days after planting and continued for 28 days. The weight of fresh French bean pods were measured at the desired harvesting period from each of the plots receiving different treatments. Samples for leaf nutrient analyses were collected for French bean from the third uppermost leaf during flowering. After 81 days the bean crop was uprooted and the above-ground biomass dried out on polythene sheets for biomass yield records. Total aboveground dry matter yield for French bean was determined from all the treatments at the end of the experiment. Plant Tissue Analysis For both growing seasons, plant tissue samples for nutrient analyses were oven-dried at 65°C for 24 hours, hand-crushed using a mortar and pestle, and kept in plastic cans for analysis in laboratories at the Department of Agronomy and Soils at Auburn University. Total nitrogen in plant tissue was determined by dry combustion with a LECO CHN-600 analyzer (LECO Corp., St. Joseph, Michigan) (Hue and Evans, 1986). Phosphorus, potassium, iron, manganese, and aluminum in plant tissue were measured by dry ashing, followed by dissolution in 1 M hydrochloric acid, followed by determination with a Jarrell-Ash inductively coupled argon plasma (ICAP) spectroscope (ICAP 9000, Thermo Jarrell Ash, Franklin, Massachusetts) (Hue and Evans, 1986). Statistical Analysis For both growing seasons, analyses of variance were performed to determine variation in French bean and kale (second season only) fresh and dry biomass and leaf nutrient concentrations owing to treatments. Pond Effluent Nutrient Removal by Soil Experiment The experiment, conducted in the laboratory at Sagana Farm, was designed with soil columns set up to filter and retain pollutants from fish pond effluents. Soil columns are commonly used in solute transport and nutrient leaching experiments. Columns simulating a soil profile allow easy access to through-flow water and hence their adoption in this study. Two soil samples were obtained, one from an uncultivated field under star grass for the vertisol (black clay soil) and the other batch from a field previously cultivated with soybean (Glycine max) for the cambisol (red clay soil) by excavation to a depth of 45 cm using a soil auger. For the two types of soils, samples were taken from 0–15 cm, 15–30 cm, and 30–45 cm depths and maintained as individual samples. The soil samples were air-dried in the laboratory, crushed, and sieved through a 2-mm mesh screen. Dry bulk density, hydraulic conductivity, and initial concentration of total nitrogen and extractable phosphorus were determined from subsamples taken from each of the soil types as previously described. In the same fields, undisturbed samples were obtained using the core ring method (Klute et al., 1986). Four sites were selected randomly on the experimental site. Mini pits were dug to a depth of 0.50 m and steps demarcated at 0–0.15 m, 0.15–0.3 m, and 0.3–0.5 m. Three cores were placed at random on each step and driven into the soil using a core driver. The cores were then dug using a sharp knife, wrapped in aluminum foil, and taken to the laboratory for analysis. Three portions of pipe (10-cm diameter) were used to simulate a soil profile of three layers with depths of 0–15 cm, 15–30 cm, and 30–45 cm. Each portion of 15-cm length was filled with soil taken from a particular soil layer. Based on the determined bulk densities (Table 1), 1.56 kg of the red clay soil and 1.93 kg of the black cotton soil from the 30–45 cm soil layer were EIGHTEENTH ANNUAL TECHNICAL REPORT 72 pushed down into the lowest portion. The second 15-cm portion of the pipe was fitted on the top side of the first portion already filled with soil from the 30–45 cm soil layer and fixed by duct tape; then 1.48 kg of the red clay soil and 1.84 kg of the black cotton soil from the 15–30 cm soil layer were packed into this second portion of the pipe. A third portion of the pipe, which was longer (22-cm depth) to hold pond effluent, was fitted on the top side of the system and fixed using the same procedure. Red clay soil (1.51 kg) and black cotton soil (1.63 kg) from 0–15 cm soil layer were packed into the portion to a depth of 15 cm and compacted by shaking so as to attain the bulk density of the field soils in the same horizon. The three portions fixed together formed an individual soil column filter, which was mounted on a collection pan. Pond water to be purified was then collected from the pond receiving 20 kg N ha-1 wk-1, containing on average 5.18 mg l-1 N and 0.68 mg l-1 P, and passed through the soil column filters at varying depths of irrigation, which served as the treatments. Four treatments were administered at the depths of water corresponding to varying loading rates of pond effluent to land as shown in Table 2. Three replicates of the soil column filters were arranged on the laboratory floor in a completely randomized design. At the end of experiment, soil was retrieved from column filters at the three 15-cm depths, prepared, and analyzed for total nitrogen and extractable phosphorus. The through-flow water from soil columns was collected on Tuesdays and Fridays and stored at 4°C for chemical analysis. Using standard methods, pond effluents and through-flow water were analyzed for total nitrogen and phosphorus. The difference between the concentration of total nitrogen and phosphorus in pond effluents and the through-flow water obtained from the soil column gave the estimated nutrients retained from the pond effluents by the soil columns. Percent nutrient removal (% NR) from effluents was calculated as: %NR = (conc. X – conc. F) × 100/conc. X where %NR = % nutrient removal, conc. X = nutrient concentration in the pond effluent, and conc. F = nutrient concentration in the filtrate. Change in soil nutrient content after application of effluents was determined using the following equation: dX = X1 – X2 where dX = change in nutrient X content in the soil, X1 = concentration of nutrient X in the soil at 0 level of effluent application, and X2 = concentration of nutrient X in the soil at a given irrigation depth used to apply pond effluent. Data on soil nutrient content from the columns were entered into a spreadsheet with the various treatments in rows and the nutrient levels in columns. Analysis of variance was performed using the general linear model of Statgraphics to compare the mean effects of water application rates on nutrient retention by the soil. The means were separated using the least significant difference procedure. RESULTS AND DISCUSSION French Bean and Kale Field Experiment The two sources of irrigation water used in this study differed in nitrogen, phosphorus, and total suspended solids (TSS) concentrations (Table 3). Total nitrogen and phosphorus concentrations were higher in pond water than canal water. After filling the pond with canal water and subsequent fertilization, nitrogen and phosphorus concentrations in pond Treatment Irrigation Intensity (mm d) Application Period (d) Water per Land (m m) Pond:Land 1 0 0 0 0:0 2 31 32 1 1:1 3 81 62 5 5:1 4 161 62 10 10:1 Source Season 1 Season 2 Total N Total P TSS Total N Total P TSS