Leaching Losses: Materials and Methods

The experiment was located within an existing long term N, P and K fertilizer response trial on oil palm in Tawau, Sabah, Malaysia. Tenera (Dura × Pisifera) oil palms were planted in the field in 1982 in a triangular pattern with a planting distance of 9.2×9.2×9.2 m. The site was undulating to rolling with slope of 0-6°. The soil type was mapped as Kumansi soil series (Typic Hapludults) or Haplic Acrisols under the FAO classification. The experimental design of the fertilizer response trial comprised a 3x3x2 factorial combination of N, P and K arranged in a Randomized Complete
Block Design (RCBD) with 3 replicates. The plot size was 30 palms. To achieve the objectives of this study, 4 treatments from each replicate were selected as follows: N0P0K0, N0P2K1, N1P2K1 and N1P2K0. The N and K fertilizers were applied in the palm circle, about 1-2 m from the palm stem. Annually, 3 rounds of N and 2 rounds of K were applied. P fertilizer was broadcast evenly in the inter row areas and over the frond heaps once a year. The N source was Ammonium Chloride (AC) which has an N concentration of 25%. The N rates were 0 kg AC palm-1 year-1(N0), 3.75 kg AC-1 palm year-1 (N1) and 7.5 kg AC palm-1 year-1 (N2). The K source was Muriate of Potash (MOP) (K = 49.8%) which was applied at 0 kg MOP palm-1 year-1 (K0) and 4.5 kg MOP palm-1 year-1 (K1). Jordanian Rock Phosphate (JRP) was used as a source of P with application rates of 4.0 kg JRP palm-1 year-1 (P2) and 0 kg JRP palm-1 year-1 (P0).

The soil water sampler was a standard vacuum lysimeter (Soil Moisture Equipment Corp., Santa Barbara, Canada; Model: 1900L), which wascommonly used for studying pore liquid sampling from vadose zone in order to measure leaching losses[18,33]. It is also called a suction lysimeter. It comprised a Polyvinyl Chloride (PVC) tube with an external diameter of 4.85 cm and a porous ceramic cup with length of 5.8 cm mounted to one end. To install the lysimeter, a vertical hole was drilled with a soil auger, which had a diameter similar to that of the lysimeter.
To optimise the contact surface between the suction ceramic cup and the soil, a small amount of slurry with the soil material from the auger was made and poured back into the hole before inserting the suction lysimeter. Once the lysimeter was located at the correct depth, the hole was backfilled with the same soil material from the auger. The soil was then stamped firmly around the lysimeter to prevent surface water
from running down the cored hole. In order to study the downward movement of N and K nutrients, the lysimeters were installed at depths of 30, 60 and 120 cm in each treatment.

The monitoring well was constructed based on the principle described by Bouwer[19]. It comprised a PVC pipe with an external diameter of 6.0 cm. Due to soil and geological constraint, monitoring wells were only installed in 6 plots with treatments comprising N0P0K0, N0P2K1, N1P2K1 and N1P2K0, N2P2K1. The water samples from both lysimeter and monitoring well were collected at 15 day intervals starting from the first fertilizer application on 23/9/2008. The samplings were carried out for a period of 5 months from October 2008 to February 2009 (150 days) or a total of 10 samplings, which basically covered the entire monsoon period in Sabah.

All water samples were collected and stored in a narrow mouth polyethylene bottle with cap[34]. In this study, water samples collected from the monitoring well was referred to as groundwater[19,20] and that from the lysimeter at 120 cm depth as leachate[35]. The samples were preserved with 2-4 mL of chloroform per litre of water to retard bacteria decomposition prior to laboratory analysis. Samples were then analysed immediately upon arrival at the laboratory or were refrigerated at 4°C if analysis cannot be carried out on the same day[34]. Samples were filtered and analysed for ammonium-N (NH4-N) using automated phenate method, nitrate N (NO3-N) + nitrite N (NO2-N) and NO2-N using automated hydrazine reduction method on a Bran and Luebbe AutoAnalyzer 3[34]. Potassium (K) concentration was analysed using flame photometric method with a Sherwood flame photometer[34]. The concentration of total inorganic N was calculated as the sum of N concentration in each form of inorganic N(NH4-N + NO3-N + NO2-N) in the sample. NO3-N concentration was obtained by subtracting NO3-N + NO2-N with NO2-N.

The experiment was carried out using a split plot design where fertilizer treatments were randomised in the main plots and depths of lysimeter in the sub plots. To test treatment effects on leaching over time, analysis of variance was performed using GenStat statistical software by treating times of sampling as repeated measures. The effect of fertilizer treatments on groundwater was examined by regarding the fertilizer x time as a factorial combination without replication and the higher interaction was used as an error term.