Importance of a Wet Season Cover Crop to Nitrate Recovery in Tropical Vegetable Production
S.L. Bithell, N. Hartley, C.C. Martin, M. Hearnden, and S.H. SmithDepartment of Resources, Northern Territory Government, GPO Box 3000, Darwin, Northern Territory, Australia / G. Owens
Northern Territory Agricultural Association, PO Box 2243, Katherine, Northern Territory, Australia.
Keywords: forage sorghum, Sorghum bicolor, nitrate nitrogen, ammonium nitrogen, mineralisation, leaching, catch crops
Abstract
The maintenance of adequate soil nutrient levels for vegetable production in the Northern Territory (NT), Australia is challenging due to low fertility soils and a tropical wet dry climate where approximately 2m of rain falls each wet season. Large scale vegetable cropping usually occurs over the dry season. This study investigated the level of nitrate recovered by a forage sorghum (Sorghum bicolor) cover crop sown at the start of the wet season in December in comparison to plots of unmanaged weeds or herbicide fallow plots (with and without fertiliser). Soils were sampled over a four month period to a depth of 1.8m at 30 cm intervals, and analysed for nitrate and ammonium nitrogen. By the end of the wet season in late March, forage sorghum plots had the lowest nitrate-N concentrations (0.92 mg/kg) due to plant uptake. The nitrate-N concentrations of deeper soil under unmanaged weeds indicated that the weeds were less efficient at recovering mineralized nitrate from a depth of about 1m than forage sorghum at one month after establishment. Fallow plots showed a significant increase in nitrate-N concentrations due to progressive mineralisation (4.2 mg/kg final value March). This information is useful for growers as it indicates the level of nitrate-N mineralisation over the wet season and the capability of an effective cover crop to recover this from depth for later use in vegetable production.
INTRODUCTION
Melon production is an important horticultural industry in the Northern Territory (NT). They are grown in the dry season with relatively high inputs of water and fertiliser. The farm-gate annual value of this industry was estimated to be $20-40 million/year (Anon 2010). During the wet season, melon growers establish rotational cover crops, predominately forage sorghum (Sorghum bicolor) or pearl millet (Pennisetum typhoides), to protect against soil erosion, to provide a break from melon pests and disease, to recover nutrients lost deep in the soil profile during the melon season and to prevent the leaching of nitrogen mineralised from the organic fraction of soil during the wet season. Leguminous cover crops are avoided by vegetable growers in the NT due to the associated buildup of pest nematode populations. A number of common soil types used in horticulture in the NT such as sandy red Kandosols (eg. Blains) have a high sand, low clay and low organic matter contents, these properties contribute to high water infiltration rates and low water holding capacities (Slatyer 1954; Day 1977). Melon crops are commonly grown on sandy soils in the NT. Hence, the potential for nutrient leaching during the wet season is high unless the nutrients can be intercepted and recovered by a cover crop.
The terms cover crop and green manure are often used interchangeably. However the term green manure is specifically used to refer to crops that are not grown to reproductive maturity and are cultivated into the soil to improve soil fertility and quality (Rosolem et al. 2005; Cherr et al. 2006). Green manures are often legume based in order to improve nitrogen fertility content. In contrast, cover crops, as implied by the term cover are sown to provide a protective cover across the soil, which is important to reduce erosion in high rainfall areas. Cover crops and green manures also improve other aspects of soil quality during their growth or after incorporation such as the soil organic matter content and the water and nutrient holding capacity of the soil. A specific form of a green manure called a ‘catch crop’ is grown to recover or retain nutrients already present at a site. Catch crops are an important means of recovering surplus nutrients (eg. nitrogen) remaining after a crop or nutrients that are released from the soil or crop residues.
The recovery of nutrients that remain in the soil after melon crops is important in the NT. A survey of melon farms in 1989 found that on some farms high levels of nutrients were lost through the soil profile during the growing season (Smith 1991a). At that time, fertilisers were side dressed and overhead sprinklers were used for irrigation (Smith 1991a), while modern farms use soluble fertilisers applied through irrigation systems (fertigation) as standard practice. Smith (1991b) demonstrated that appropriate practices substantially reduced nitrate losses, resulting in minimal nitrate-N remaining after the crop. Blackburn et al. (1987) also established that appropriate irrigation schedules with trickle tape can minimise the loss of water and nutrients through the profile. Even so, there has been no recent evaluation of potential nutrient loss or recovery in, or following, melon crops.
The recovery of nutrients that are released during the wet season is also an important aspect to successful cover cropping in the NT. The majority of conversion of organic nitrogen to mineral nitrogen in the soil (mineralisation) occurs during the high rainfall period of the wet season in the NT (Norman 1966). With no cover crop, there is the potential for significant leaching of nitrogen to deep layers in the soil. In some tropical systems a pulse of mineralisation occurs at the start of the wet season (Hagedorn et al. 1997). There is limited understanding of the timing of mineralisation in relation to cover cropping under NT conditions.
Identifying sustainable cover crop options, including species and management practices, was identified as a high research priority by melon growers at an industry research planning meeting in the NT in November 2009 (P. Stephens pers. comm.). Limited cover crop species comparisons in the NT have established that millet has a superior root system to sorghum in terms of the depth of rooting and the recovery of nutrients (Wetselaar and Norman 1960). Millet crop residues can be difficult to incorporate and manage due to the often thick stems of mature millet plants, making millet unpopular with some vegetable growers. Prior NT based studies were predominately carried out on Tippera soils (loamy sand) which have higher silt content than the Blains and Cockatoo types (sands or sandy loams). Sandier soils have been identified as being more prone to nutrient losses through leaching, due to their naturally high permeability. There has been no comparison between cover crop species on the sandier soils (Blain or Cockatoo types) in the NT.
This wet season study sought to ascertain the level of nitrogen mineralised during the wet on sandy soils and compare the ability of the two cover crops to recover nitrogen on a sandy loam soil.
MATERIALS AND METHODS
Treatments
A field experiment with four treatments (1. Forage sorghum ‘Jumbo’, 2. Katherine Pearl millet, later re-classified as weed treatment, 3. Chemical fallow – no fertiliser, 4. Chemical fallow plus fertiliser) was drilled on 17 December 2009 on a Blain soil at Berrimah Farm, Darwin. The drill was passed through all treatments. The forage sorghum (thousand seed weight (TSW) of 24.7g) was drilled at 12.6 kg/ha, Katherine Pearl millet (TSW 10.1 g) was drilled at 14.3 kg/ha. Seed was germination tested and sowing rates were corrected for non-viable seed and an emergence of 90% of viable seed, effective emergence target populations for the respective species were 40 and 70 seedlings/m2.
The area, which had been in pasture, was sprayed off with glyphosate, mown and residues were removed prior to light surface cultivation of the area. The experiment was a randomised complete block design, with four blocks and each plot measuring 16x4 m. Neighboring plots had a 1m gap which alternated with a 10m gap, to allow tractor access from the 16m side of each plot. Each plot was divided into four, 4x4m sub-plots with a sub-plot randomly allocated to each of four sampling times. Seed was drilled with a tractor mounted drill in 15 cm row spacings with granulated fertiliser (13.5% N (ammonium form), 15% P, 12.5% K and 1.2 % S.) banded behind the seed at 53 kg/ha. Treatment 3 received no fertiliser. Because of poor establishment the millet plots were re-sown by hand on 24 December 2009 with another seed source. This was also unsuccessful so plots were re-classified as a ‘weed treatment’ with weeds allowed to grow and any millet removed by hand weeding.
Pre and post-trial soil sampling in the melon rooting depth zone
Soil samples were collected on 10 of December 2009 (pre-treatment) at depths of 0-150mm and 150-300mm using a hand auger (interior diameter 48.9 mm). Eight samples from each depth were collected from each plot and bulked. Samples were placed in a forced air oven (65○C) on the day of sampling and dried for 48 h. For the pre-treatment sampling, samples from each treatment were bulked again to provide four samples from each depth. Samples were analysed by CSBP Laboratories, Lake Pibra, Western Australia for total carbon and carbon to nitrogen ratio. A LECO combustion analyser combusted samples at 950˚C and flushed with oxygen. The generated gases were collected and measured on both an infrared detector and a thermal conductivity cell for total carbon and total nitrogen respectively.
Pre and in-trial soil sampling to 1.8m
Soil samples were collected on four occasions at one month intervals. Each plot was sampled from a randomly allocated subplot. Depth samples were collected using a tractor mounted hydraulic auger (interior diameter 47.7 mm). Before sowing three blocks (12 plots) were sampled on 11 December 2009 until heavy rain (347mm from 11 to 15 of December) and wet soil conditions delayed the completion of sampling until 16 December. The trial was drilled the following day. This pre-sowing sampling was from depths of 0-10, 30-40, 60-70, 90-100cm, 1.4-1.5m, and 1.7-1.8m from a single sample point in each plot, giving 96 samples.
The three remaining samplings each occurred over a two day period on: 18-19 January, 15-16 February and 22-23 March 2010 (in-trial sampling). The January, February and March samples, were taken from depths of 0-30, 30-60, 60-90, 90-120 cm, 1.2-1.5m and 1.5-1.8m from two sample points in each sub-plot and then bulked and dried. After drying, a 300g sub-sample was sent to CSBP for analysis of nitrate and ammonium-N. Soil nitrate nitrogen and ammonium-N were extracted with a 1M potassium chloride solution for 1 hour at 25°C. After dilution the resulting soil solution was measured on a Lachat Flow Injection Analyzer. The concentration of ammonium-N was measured colorimetrically at 420nm using the indo-phenol blue reaction. The nitrate was reduced to nitrite through a copperised-cadmium column and measured colorimetrically at 520nm.
Plant sampling
Two 0.25m2 quadrats per forage plot were used to sample the biomass of weeds and forage sorghum and assess forage populations at 42, 77 and 102 days after sowing (DAS). Plants were removed by cutting at ground level and then dried in a forced air oven at 65○C for 48 h before weighing. Weed species were identified in four quadrats per plot of the weed treatment and percentage cover of each species estimated.
A mixed effects two factor ANOVA (treatments and depth) repeated in time (month) model by S+ (Anon 2008) was used for data analysis. For plots, December means were plotted for reference but excluded from the models, for effects of interest, pair-wise means comparisons were calculated using Tukey’s LSD.
RESULTS
Rainfall during the months of December, January, April and May were higher than the long term monthly averages for Darwin (1941-2010) (Figure 1), this contributed to an above average rainfall over the wet season of this study. Monthly rainfalls recorded adjacent to the study site were 57 and 36 mm lower than at the Darwin site in December and January, but 17 mm higher during both March and April (Table 1).
Pre-trial samples collected from the melon rooting zone (0-150 mm and 150-300 mm) had total carbon levels of 2.2 % (standard error of the mean (SEM) = 0.11) and 1.4 % (SEM=0.09) and C:N ratios of 19 (SEM=1.2) and 16 (SEM=2.2), respectively.
Plant growth
Forage sorghum populations from 42 DAS had densities of 34-45 plants m-2 (Table 2). The dry matter (DM) of forage by the end of the trial averaged 15.04 t/ha, weed levels in forage plots were substantial at 2.47 t/ha, overall biomass in weed only plots was 7.21 t/ha. Leguminous weeds recorded in plots were Alysicarpus vaginealis, Macroptilium gracile, Chamaechrista rotundifolia, Stylosanthes humilis, and Crotalaria goreensis; broadleaf and grasses were Oldenlandia corymbosa, Corchorus trilocularis, Boerhavia dominii, Sida acuta, Cyperus rotundus, Cynodon dactylon, Urochloa mosambicrnsis. Legume weeds contributed to about 30% of the weed spectrum on the basis of percentage cover.
Pre and in-trial soil sampling to 1.8m
Nitrate-N values were low for samples collected in December before sowing with slightly higher concentrations in the surface profile (Figure 2) Respective concentrations from the six sample profiles (0-10, 30-40, 60-70, 90-100cm, 1.4-1.5m, and 1.7-1.8m) were 1.44, 1.06, 1.03, 0.72, 0.56, and 0.75 mg/kg. Figure 3 shows ammonium-N concentrations also trended to higher values for samples in the top three profiles, values were 1.44, 1.06, 1.03, 0.72, 0.56, and 0.75 for the respective depths of 0-10, 30-40, 60-70, 90-100cm, 1.4-1.5m, and 1.7-1.8m.