Introduction
The present research is part of a broader study aimed at reducing loads of agricultural pesticides in the environment using vegetative filters (biofilters) such as grassed waterways or natural grasslands. The focus here is on atrazine, a residual herbicide that is a major contributor to poor water quality on the Liverpool Plains, New South Wales (NSW) (Boey and Cooper 1996; Mawhinney 1998). To manage pesticides in the environment effectively, it is important to know their degradation rate. In the case of atrazine, not only does degradation rate determine its effectiveness as a residual herbicide, but it also affects the amount of pesticide available for runoff from fields. Moreover, the effectiveness of biofilters receiving runoff depends on both load reduction by filtration (Patty et al. 1997) and the degradation rate of pesticides retained within the biofilter. This rate determines whether pesticide concentration increases to the point where the pesticide becomes available for desorption and transport with subsequent runoff events, or in the case of residual herbicides such as atrazine, where they become toxic to plants in the biofilter.
The Liverpool Plains is a series of floodplains along the Mooki River, north-western NSW. The Plains are now largely cropped except for some remnant grasslands. Cropping starts from the hillslopes immediately below the surrounding mountain ranges and continues down onto the floodplain. Runoff is most commonly generated after heavy rainfall between August and February, which includes the summer cropping season when atrazine is widely used. Once runoff is exported from the hillslopes, it is transported through the catchment. It may be confined to channels and streams, or it may pass through natural grasslands, where it spreads over areas of lowland relief through a network of small, ephemeral braided streams that are typical of natural floodplains. The runoff from cropped land carries herbicides and other pollutants. Our interest is in the capacity of the natural grasslands to detain and subsequently degrade herbicides such as atrazine.
Herbicide persistence expresses the duration of activity in a soil or plant system. Herbicide persistence is measured by a half-life value (Kookana et al. 1998), i.e. the period for 50% of a herbicide in the soil to degrade. Field studies on atrazine persistence in soils have shown contrasting results. Some show half-lives of about 30-60 days (Sengalevitch et al. 1987; Tomlin 1994), while others report shorter half-lives of 8 days (Barrett et al. 1991; Mathess 1994). Under laboratory incubation, much shorter half-lives of about 5-20 days have been reported (Topp et al. 1995; Pussemier et al. 1997; Vanderheyden et al. 1997), which, in comparison with half-lives observed by Ferris and Haigh (1993) ([t.sub.50] = 4 months, [t.sub.90] = 9 months), illustrates the variability in degradation rates under different conditions. Variation has been related to atrazine adsorption to organic matter and vegetation, soil pH, and atrazine treatment history (Barriuso and Houot 1996; Pussemier et al. 1997; Vanderheyden et al. 1997; Houot et al. 2000). Microbial degradation is the main mechanism for atrazine mineralisation (WSSA 1989; Bollag and Liu 1990). Adaptation and selection of specific atrazine-degrading bacterial communities that may utilise atrazine as a source of C and N for their metabolism has been shown, with total numbers of these degraders increased with regular long-term application of atrazine (Racke and Coats 1990). Atrazine degradation rate increases accordingly (Barriuso and Houot 1996; Pussemier et al. 1997; Vanderheyden et al. 1997). Soil light fraction organic matter and total N are also thought to influence degradation rates of atrazine by providing competing sources of C and N (Sims and Cupples 1999; Yassir et al. 1999).
The aims of this work were to quantify degradation rates in soils from 2 sites in the Liverpool Plains with contrasting agronomic history (cropped land and grassland) using laboratory incubation and glasshouse bioassays, and to determine whether there is any association between atrazine mineralisation dynamics and changes in soil microbial communities identified using terminal-restriction fragment length polymorphism (t-RFLP; Marsh et al. 2000). Effects of selected soil properties, including light fraction organic matter and total N, on atrazine degradation rates were also examined, assuming higher nutrients and organic matter content in grasslands than in cropped land.
Materials and methods
The study comprised 2 stages. During the first exploratory stage, soil for determination of atrazine degradation rates was sampled in February and May 2002. The soil was used for laboratory incubation, identification of microbial communities, and a plant bioassay experiment. In the second stage, confirmation of initial promising results was sought in more detailed incubation and bioassay experiments in 2004.
Sites
The soils of the floodplain are predominantly self-mulching Black Vertosols (Isbell 1996) derived from basalt alluvium. Soil was sampled from sites on the floodplain and footslopes with contrasting histories of herbicide exposure. Samples were taken from: Site 1, an intensively managed cropped field on the footslopes just above the floodplain, receiving regular atrazine applications during the last 10 years; Site 2, a grassland dominated by Stipa aristiglumis on the floodplain, adjacent to, and receiving runoff, from Site 1 (this site will be referred to as upland Stipa); and Site 3, a Stipa grassland situated 2.5 km downstream from Site 2 and presumably intercepting less water, pollutants, and sediments than Site 2 (lowland Stipa).
Soil sampling and handling
Exploratory studies
Soil was sampled from the 0-0.05m layer on 2 occasions. In February 2002, the soil was sampled for an incubation experiment and microbial community analysis by t-RFLP. Total N (Kjeldahl), total C (Walkley and Black 1934), available phosphorus (Olsen et al. 1954), soil pH (Plaster 1992), and bulk density were determined. At each site, a composite sample was obtained by collecting and mixing 5-10 soil subsamples from each of 5 randomly chosen locations situated at about 100-150m from each other, perpendicular to the direction of stream flow.
In May 2002, these sites were re-sampled for more comprehensive analysis of basic soil properties, total N and C (Dumas 1981), and available phosphorus (Colwell 1963), light fraction organic matter determination as percentage of the total organic matter (Gregorich and Ellert 1993), and proportion of organic C in the light fraction component (Haenes 1984). For the re-sampling, a nested sampling design was adopted with replicate samples being taken from different land micro-relief (soil surface and slight depressions in the drainage lines of the braided streams). A total of 48 samples were analysed. A composite sample of soil from each site was used for the exploratory bioassay experiment. Thus, for the exploratory studies, there was no field replication of soil sampling sites (but laboratory and glasshouse treatments were replicated).
Confirmation experiments
At each of the field sites, 3 independent replicates samples were obtained in July 2004. The 3 replicates (subsites) within each site were randomly located at about 450-500-m intervals from each other, perpendicular to the direction of stream flow. Each subsite comprised 10 soil subsamples selected randomly from within an area of …
No comments:
Post a Comment