Predictive models of weed population dynamics

Many of the challenges faced by weed ecologists can be met only by the capability to predict the responses of weed populations to changes in their environment or management. In spite of this, a review of papers published in Weed Research suggests that weed ecologists are remarkably reluctant to produce detailed, quantitative predictions. This may result from uncertainty in the accuracy of predictions and indeed, a variety of reasons have been put forward to suggest that the potential utility of weed models may be limited in this regard. In this study, we review the applications to which weed models have been put. Focusing on predictive population modelling, we highlight several limitations that can lead to failures of this approach and we discuss the likely prospects for weed population modelling. We make three points regarding the future of weed modelling. First, owing to prohibitive data requirements, the development of highly mechanistic models that attempt to make detailed predictions of weed population numbers is unlikely to be very successful. Second, data collection for developing weed models needs to be rethought. Weed models are most commonly compromised by a lack of spatial and temporal replication, preventing modellers from measuring parameter variability and error effectively and limiting assessments of model uncertainty. Finally, the utility of models needs to be better appreciated; models are key tools in making long range predictions of how management will affect weed populations, but, we estimate, they are used in only a small fraction of studies. Without the further development of models for weed population dynamics, our ability to predict long-term dynamics will be restricted.     

Modelling the population dynamics of Papaver rhoeas under various weed management systems in a Mediterranean climate

The demography of the annual dicotyledonous weed Papaver rhoeas and the efficacy of different management practices were studied during three consecutive years in winter cereals in the north‐east of Spain. These data were used to estimate the parameters of a weed life cycle model that was used to describe the population dynamics of this species and to predict the effect of various control strategies and integrated weed management (IWM) scenarios. Without control, the annual rate of increase was 40 (λ_t), and the minimum control level required to maintain the population stable was 99% of the emerged plants. The annual application of post‐emergence and/or pre‐emergence herbicides did not prevent the growth of the population. Using various cultural control tactics (delayed seeding, harrowing and fallow) resulted in different trends in the overall population depending on the techniques and combinations analysed. Simulations showed that delayed seeding, fallow and pre‐emergence herbicides are the best techniques to employ in IWM programmes, always using a combination of these and other more common practices (i.e. post‐emergence herbicides). Sensitivity analysis indicated interaction between the parameters and that the model was especially sensitive to seed losses and also to fecundity, seedling survivorship and emergence. The study shows that new strategies should be sought to control these parameters. To develop IWM programmes for P. rhoeas, the combination of two or more control strategies is required.     

Long-term management of Striga hermonthica: strategy evaluation with a spatio-temporal population model

The parasitic weed Striga hermonthica poses a serious threat to cereal production in sub-Saharan Africa. Striga hermonthica seedbanks are long-lived; therefore, long-term effects of control strategies on the seedbank only emerge after several years. We developed a spatially explicit, stochastic model to study the effectiveness of control strategies in preventing invasion of S. hermonthica into previously uninfested fields and in reducing established infestations. Spatial expansion of S. hermonthica and decrease in millet yield in a field was slower, on average, when stochasticity of attachment of seedlings to the host was included and compared to the deterministic model. The spatial patterns of emerged S. hermonthica plants 4–7 years after point inoculation (e.g. seeds in a dung patch) in the spatial-stochastic model resembled the distribution typically observed in farmers' fields. Sensitivity analysis showed that only three out of eight life cycle parameters were of minor importance for seedbank dynamics and millet yield. Weeding and intercropping millet with sesame or cowpea reduced the seedbank in the long term, but rotations of millet with trap crops did not. High seedbank replenishment during years of millet monoculture was not sufficiently offset by seedbank depletion in years of trap crop cultivation. Insight from simulations can be employed in a participatory learning context with farmers to have an impact on S. hermonthica control in practice.     

Non-chemical weed control on traffic islands: a comparison of the efficacy of five weed control techniques

The efficacy of five non‐chemical weed control methods for reducing weed cover on traffic islands was investigated in the growing season of 2004. Three trial sites were divided into six treatment areas which were treated with either flame, steam, hot air, hot water, brushes or left untreated. The treatments were carried out at regular intervals throughout the growing season. The percentage weed cover was measured every second week using a 75 × 75 cm quadratic frame with 100 squares. In the control areas, a rapid increase in weed cover was observed, whereas all treatments reduced weed cover. Hot water was the most effective method, although not significantly better than hot air or steam. Hot air treatment was more effective than brushing, whereas hot water was more effective than both flaming and brushing. The doses that were used were relatively high (150–355 kg ha^?1), partly because of the irregular shape of the traffic islands and the treatment intervals were quite short in comparison with those in similar studies. However, the treatments could keep down the weeds only to a certain extent. The present knowledge of the efficacy of various weed control methods, as well as an increase in our knowledge of adequate treatment intervals, supports an optimisation of hard‐surface weed control. Data and experience gained from these trials were used to develop further calibrated application studies.     

An evaluation of four crop : weed competition models using a common data set

Summary To date, several crop : weed competition models have been developed. Developers of the various models were invited to compare model performance using a common data set. The data set consisted of wheat and Lolium rigidum grown in monoculture and mixtures under dryland and irrigated conditions. Results from four crop : weed competition models are presented: almanac , apsim , cropsim and intercom . For all models, deviations between observed and predicted values for monoculture wheat were only slightly lower than for wheat grown in competition with L. rigidum , even though the workshop participants had access to monoculture data while parameterizing models. Much of the error in simulating competition outcome was associated with difficulties in accurately simulating growth of individual species. Relatively simple competition algorithms were capable of accounting for the majority of the competition response. Increasing model complexity did not appear to dramatically improve model accuracy. Comparison of specific competition processes, such as radiation interception, was very difficult since the effects of these processes within each model could not be isolated. Algorithms for competition processes need to be modularised in such a way that exchange, evaluation and comparison across models is facilitated.     

Predicting weed emergence: a review of approaches and future challenges

Summary In relatively short-term studies it is often difficult to identify the factors that are important in determining the patterns of emergence for different weed species. When results from longer-term studies are averaged over years, they demonstrate that some weed species follow characteristic, and potentially predictable, patterns of annual emergence. In recent years, our understanding of the seasonal changes in dormancy and germination behaviour, and the interaction of these processes with the environment, has advanced considerably. In particular, the capacity for computer-based statistical analysis and the ability to handle large datasets has made it possible to use this information in a predictive way. Currently, there are several approaches to developing predictive weed emergence modelling. Some researchers have taken an empirical approach, seeking to identify correlations between environmental variables and observed emergence patterns. Others have taken a reductionist approach, subdividing the emergence process into its component stages of dormancy, germination and pre-emergence growth, to work towards achieving an eventual understanding of the physiological processes involved. Despite recent advances, a number of major challenges remain (e.g. population variability, dormancy and the quality of the input data) that must be overcome before these emergence models can be implemented in practice. The level of complexity and degree of parameterization incorporated within such models must also be addressed in relation to the intended use of the model.     

Response of Arabidopsis thaliana to 22 ALS inhibitors: baseline toxicity and cross-resistance of csr1-1 and csr1-2 resistant mutants

The baseline toxicity of 22 acetolactate synthase (ALS)-inhibiting herbicides and the cross-resistance patterns of chlorsulfuron- and imazapyr-resistant (R) lines on these 22 ALS-inhibiting herbicides were investigated using the model species Arabidopsis thaliana. The 22 herbicides consisted of 18 sulfonylureas (SU), three imidazolinones (IMI) and one triazolopyrimidine (TP). The ED₅₀ values (doses of herbicides required to reduce dry matter by 50%) of the post-emergence-treated Col and Ler susceptible (S) lines ranged from 22 to 4822 mg ha⁻¹ and from 17 to 3143 mg ha⁻¹ respectively. The csr1-1 chlorsulfuron-resistant line (substitution of Pro₁₉₇ to Ser) conferred a high resistance to the only TP tested as well as to nine SU herbicides (R:S ratio ≥30), a low resistance to two SU herbicides (R:S≥5 and <30) and little or no resistance to the three IMI and seven other SU herbicides (R:S <5). This result contradicts the expectation that an ALS mutation selected by an SU herbicide confers high cross-resistance to other SU herbicides. We found that the efficacy of specific ALS inhibitors was different for different species and therefore could not be predicted from our results with A. thaliana; however, the cross-resistance patterns in A. thaliana were highly correlated with cross-resistance patterns in unrelated species with the same resistance mutation. These results have implications for resistance management.     

地膜覆盖和除草剂对夏玉米田杂草及玉米生长发育的影响

研究了地膜覆盖和除草剂对夏玉米田杂草和玉米生长发育的影响.结果表明,在玉米3～5叶期施药,药后33 d,杂草数量降至最低,56％2甲4氯钠可湿性粉剂处理的杂草减退率和相对株防效分别比40 g/L烟嘧磺隆油分散制剂提高了11.71百分点和12.65百分点.揭膜后,白色地膜覆盖下杂草的平均鲜重显著低于对照,比对照减少73.25％,相对鲜重防效分别是黑色地膜覆盖、56％2甲4氯钠可湿性粉剂、40 g/L烟嘧磺隆油分散制剂处理的1.12、1.15、1.73倍.地膜覆盖出苗率高且整齐一致,平均株高和叶数比对照高,对玉米的穗粒重、百粒重及其产量有明显的增效作用,以白色地膜覆盖下玉米产量最高,增产率比黑色地膜覆盖的高1.23百分点,比40g/L烟嘧磺隆油分散制剂处理高16.09百分点;而施用56％2甲4氯钠可湿性粉剂严重影响了玉米产量的形成,其产量比对照降低3.09％.     

Secondary dispersal, spatial dynamics and effects of herbicides on reproductive capacity of a recently introduced population of Bromus sterilis in an arable field

An arable field was subdivided and subjected to either deep inversion ploughing or non‐inversion cultivation after viable seeds of Bromus sterilis had been sown into oilseed rape stubble. After sowing in isolated plots distributed within the field, sequences of cropping treatments for the establishment of two successive winter wheat crops were applied. Each subfield was split into an uphill and a downhill direction for soil cultivation. The field had a 10° slope. In the season following seed introduction, 2.6% of the introduced seeds had successfully germinated and established in the non‐inversion cultivation regime, when no effective graminicide was applied. Ploughing eradicated B. sterilis. Using differential global positioning system (DGPS) mapping of the whole field population, emerged plants were observed up to 8.7 m (uphill treatment) and 21.3 m (downhill treatment) of their initial source. The median distance seeds were transported was 2.3 m uphill and 4.8 m downhill. Post‐emergence application of the herbicide propoxycarbazone slightly reduced weed density and seed weight, and almost halved weed seed production. Application of fenoxaprop‐P‐ethyl was followed by higher density of plants, tillers and seeds of B. sterilis. Seed viability was unaffected by herbicide use. Thus, in the second wheat crop following seed rain, the weed population was dispersed more widely in the field, such that 20–30% of seeds were dispersed more than 5 m distance from the first year's foci of infestation. The relevance of soil cultivation to secondary dispersal of B. sterilis is discussed.     

Effects of host plant genotype and seedbank density on Striga reproduction

Prevention of seed input to the seedbank of Striga hermonthica‐infested fields is an important objective of Striga management. In three consecutive years of field experimentation in Mali, Striga reproduction was studied for 10 sorghum genotypes at infestation levels ranging from 30 000 to 200 000 seeds m^?2. Host resistance was identified as an important determinant of Striga reproduction, with the most resistant genotypes (N13, IS9830 and SRN39) reducing Striga reproduction by 70–93% compared with the most susceptible genotype (CK60‐B). Seedbank density had a significant effect on Striga seed production. Higher seedbank density resulted in more Striga plants, which led to increased intra‐specific competition and consequently a reduced level of reproduction per plant. For the most susceptible sorghum genotypes, density dependence also occurred in the earlier belowground stages. Striga reproduction continued beyond harvest. At the high infestation level just 8% of the total reproduction was realised after harvest, whereas at the low infestation level 39% was attained after harvest. Even though host‐plant genotype plays a significant role in Striga reproduction, calculations indicated that only at very low infestation levels the use of the most resistant genotype was able to lower the Striga seedbank.