Using weeds to reduce pest insect numbers in organic vegetable crops - a desk study (OF0329)

Summary

At present, weeds and pest insects are two of the major constraints to the production of high quality organic vegetable crops, primarily because the range of control measures available to UK growers is extremely limited. Several studies have shown that the numbers of pest insects found on crop plants are reduced considerably when the crops are allowed to become weedy. However, if weeds are not removed by a certain stage of crop growth, they start to compete with the crop plants and yield begins to decline. Warwick HRI already has models that are used independently to predict both the onset of weed competition and the timing of pest insect attacks. However, in practice, growers have to manage weeds and pest insects at the same time and hence the decisions they make often involve compromises. Therefore, it would seem a natural progression to combine the existing models to show how reductions in crop yield, from altered weeding practices, could be offset by improvement in crop quality from reduced pest numbers. The aim of this project was to demonstrate how the existing crop protection models could be used to optimise weed and pest control in organically-grown vegetable crops. The strategy could be very useful, as the range of crop protection interventions available to organic growers is strictly limited. Vegetable brassicas were chosen as the experimental crop, because their pest and weed models have been well validated, and because Brassica crops account for approximately 20% of the total organic vegetables grown in the UK. However, the research also demonstrates how the system could be adapted for other crop/pest combinations. The first objective was to quantify the plant ‘architecture’ needed from weed populations to reduce pest insect infestations in cabbage crops. Three non-host weed plants were used in the study; Chenopodium album (fat hen), Stellaria media (common chickweed) and Tripleurospermum inodorum (scentless mayweed). These species were selected because they are common weeds in field vegetable crops, reduce colonisation by the cabbage root fly (Delia radicum) and have contrasting architecture (spread and height). The treatments combined weed species, planting times, plant sizes and plant densities to examine the impact on pest colonisation of cabbage. The different plant sizes and plant densities were achieved using weed plants that were raised in modules and then planted around cabbage plants. The three test insect species were Brevicoryne brassicae (cabbage aphid), Delia radicum (cabbage root fly) and Plutella xylostella. (diamond-back moth). The trial demonstrated that, for D. radicum and B. brassicae, the greatest reductions in crop colonisation were produced by a high density of large weeds. The effects on P. xylostella were less clear-cut. The second objective was to combine weed and pest insect models to quantify the interactions
between crop growth, weed growth and pest insect colonisation. The initial analysis of the data obtained in Objective 1 used a log-linear model to relate pest counts to various assessments of crop (cabbage) and weed plant size (weed fresh weight, weed dry weight, weed height, weed diameter, crop height and number of crop leaves). Whilst these regressions suggested a number of relationships, they often accounted for only a small percentage of the variance. In this second objective, further statistical analyses were done to consider relationships between the numbers of pests per plant and various crop and weed parameters. The ratio of weed biomass to crop biomass (dry weight) provided the strongest relationship with pest numbers, and is also a measurement that can be obtained easily in the field. Using this relationship, critical values were calculated for each pest : weed combination. Maintaining the weed : crop dry weight ratio above this critical value should provide protection against pest colonisation. For example, a weed : crop dry weight ratio of 32.7 was required for S. media to “control” D. radicum. The final objective was to determine the impact of different weeding strategies on 1) yield reduction through weed competition and 2) pest colonisation. A field trial was designed to test these strategies using S. media to ‘control’ D. radicum on cabbage in organic growing conditions. Three ‘weed’ strategies were devised using the crop-weed competition model in combination with the D. radicum forecasting model. They were intended to give: 1) a ratio always greater than the ‘critical value’ (the greatest pest protection, but also the greatest weed competition), 2) a ratio close to the ‘critical value’ (good pest protection, but less crop yield loss than Treatment 1 due to weed competition), and 3) a ratio always below the ‘critical value’ (the least pest protection, but also the least yield loss due to weed competition). Different weed : crop ratios were achieved by transplanting module-raised S. media of different ages into the cabbage plots. The strategies were timed to coincide with periods of either low or high pest incidence in the field, predicted using the D. radicum forecast, and the weeds were removed 4 weeks after transplanting, when the cabbage plants should have been sufficiently wellestablished to withstand a certain amount of root damage. Unfortunately, the growth patterns of some of the weed treatments deviated from predictions made by the crop-weed competition model, so that the ‘strategies’ did not achieve the precise range of weed : crop ratios that had been intended. Despite these deviations, the percentage reduction in cabbage weight at harvest, due to the presence of weeds, was not dissimilar to that predicted by the model. Damage by D. radicum was relatively low. None of the cabbage plants wilted or died, although all suffered larval feeding damage to their roots. However, where the critical ratio was achieved, weed presence had a distinct effect on root damage and the highest density of weeds reduced root damage by the greatest amount when compared with the weed-free control treatment. At harvest, cabbage fresh weight was negatively correlated with the weed : crop ratio, due to competition from the greater biomass of weeds. Thus despite practical difficulties in achieving the intended weed : crop ratios using transplanted weeds, the trial confirmed the value of a ‘green background’ in reducing colonisation by D. radicum. It also confirmed the negative effect of weed competition on crop yield and demonstrated the ‘trade-off’ between reductions in pest damage and yield. Although the weeds were removed after the first 4 weeks of cabbage growth, the high density of weeds required over this period caused a significant reduction in crop yield.
Scientific conclusions
1. The current crop-weed competition model is driven purely by radiation interception and this study has highlighted some possible model limitations, in particular, the need to take account of crop and weed growth in nutrient deficient soils and in response to moisture stress. The model also need to be able to cope with multi-species populations of natural weed flora (i.e. parameters need to be obtained for a wider range of species).
2. The crop-weed competition model provides an adequate prediction of competition based on average plant density. Modification of the model to account for the spatial proximity of the plants and their growth plasticity would enable more realistic and practical strategies to be evaluated (e.g. only having weeds within a small radius of the crop plant)
3. Pest models currently provide a robust prediction of the timing of pest infestations, but predictions of pest abundance and subsequent damage would be needed to fully evaluate the trade-off between yield loss through plant competition and the reduction in pest damage.
Practical conclusions
1. The trade-off between crop yield and pest control is clearly illustrated by the study. Yield loss (up to 30%) due to competition may be tolerable as an alternative to severe pest damage, in situations where infestation levels are high. Variety choice may have compensatory benefits, however, the impact of varietal choice on competition needs to be explored.
2. The study has shown that the strategy of allowing weed presence for a limited period whilst maintaining a weed : crop ratio (total dry weighs per unit area) above a threshold can provide some protection against pest damage. The threshold is dependent on the pest and weed species involved.
3. In practice, planting into a background of natural flora is probably the most practical way of achieving this protection. However, the weeds would need to be well established before the cabbage was transplanted to achieve the required weed : crop ratio. Further information is needed on the development of natural floras, and on the effect of different weed species compositions.
4. Weeds in close proximity to the crop do reduce pest colonisation, as seen in other studies. Hence a lower total number of weeds could potentially achieve the same protective effect, providing they are close to the crop plant. Further information is needed on the spatial characteristics of plant competition to enable more realistic and practical strategies to be evaluated.