Measures that describe weed spatial patterns at different levels of resolution and their applications for patch spraying of weeds

If weeds occur in aggregated spatial patterns, it is possible to reduce herbicide use by spraying only weed patches. The reduction in herbicide use will then depend on the spatial resolution of the patch sprayer and the weed-free area at that level of resolution. Three distance measures are presented that describe the weed spatial pattern at different levels of resolution. They give information on aspects of pattern that are relevant to patch spraying. The distance measures were applied to a spatial pattern of Galium aparine L. seedlings recorded in an area of 18.0 m × 32.4 m. In this area, the herbicide use of an idealized patch sprayer that detects and sprays all weeds with a spatial resolution of 1.0 m would be 41% of the amount required for a whole-field application. Spraying with a finer spatial resolution of 0.5 m would give a further 26% reduction in herbicide use.     

Long-term modelling of weed control strategies: analysis of threshold-based options for weed species with contrasted competitive abilities

A simple life-cycle-based demography model was adapted for two contrasting weed species ( Alopecurus myosuroides and Poa annua ). This model included a seed production function that accounted for population self-regulation through weed:weed interactions. The A. myosuroides version of the model was tested with field data. Long-term simulations of population demography were then performed to investigate the relationship between weed control strategies based on density thresholds and both the frequency of herbicide use and the long-term economic profitability. This study confirms that threshold-based weed management strategies are more cost-effective than spraying every year and may allow important reductions in herbicide use. However, after the first transient years of either systematically spraying or withholding herbicide, the long-term spraying frequency was insensitive to threshold values between 0.01 and 100 plants m⁻². The highest long-term profitability was obtained for the lowest threshold tested, and the profitability decreased rapidly when the threshold was raised above 4–6 and 10–20 plants m⁻² for A. myosuroides and P. annua respectively. The study thus indicates that the exact threshold value is of little importance for the long-term reliance of the system on herbicide, provided that it is reasonably low. For species with low competitive ability, high thresholds may be used in some cropping systems to reduce the spraying frequency for environmental considerations, but those options would also reduce the profitability if no compensatory measures were taken.     

Simulating the effects of weed spatial pattern and resolution of mapping and spraying on economics of site-specific management

The economic benefits of using site‐specific weed management (SSWM) are related to the proportion of the field that is weed‐infested, the number of weed patches and the spatial resolution of sampling and spraying technologies. In this paper we simulate different combinations of these factors using parameter values obtained for Avena sterilis ssp. ludoviciana growing in Spanish winter barley crops. The profitability of SSWM systems increased as the proportion of the field infested by this weed decreased and when patch distribution was more concentrated. Under most of the conditions tested, positive net returns for SSWM were obtained when the weed‐infested area was smaller than 30%. The highest net return occurred using a 12 m × 12 m mapping and spraying resolution. The critical parameter that determined the economic viability of patch mapping and spraying resolution was the technology costs. The site specific strategy was economically superior to the standard strategy (overall herbicide application) in most cases. However, the differential between the two strategies decreased when the number of patches and the resolution of mapping and spraying increased, such that the highest net returns were obtained with a single patch covering 14% of the field and using a 12‐m mapping and spraying resolution; whereas the worst net returns were obtained for all patch numbers when 64% of the field was infested and a 3‐m mapping and spraying resolution was used.     

Precision farming for weed management: techniques

Site-specific weed control techniques have gained interest in the precision farming community over the last years. Managing weeds on a subfield level requires measuring the varying density of weeds within a field. Decision models aid in the selection and adjustment of the treatments, depending on the weed infestation. The weed control can be done either with herbicides or mechanically. A site-specific herbicide application technology can save large amounts of herbicides. Mechanical weed control techniques adapting to the weed situation in the field are applicable to a wide spectrum of crops. Site-specific techniques for the detection and management of weeds are presented. A system for the discrimination of different weed species and crops from images is described, which generates weed maps automatically. Models for the yield effect of weeds are developed and applied in on-farm-research experimental setups. Economic weed thresholds are derived and used for a herbicide application with a patch sprayer.     

Effect of environmental conditions on Alopecurus myosuroides germination. II. Effect of moisture conditions and storage length

Summary More effective control of Alopecurus myosuroides with a minimum amount of herbicide may be achievable by using models describing the impact of cropping system effects on the weed's demography. In order to develop the submodel relating the weed seedbank to the emerged seedlings, a series of laboratory experiments was set up to analyse and quangify the effects on A. myosuroides germination of length of seed storage in dry conditions, soil water potential and dryness during imbibition. The base soil water potential for germination was estimated at −1.5 MPa. A four-parameter non-linear equation depending on accumulated day–degrees was fitted to the observed germination data, resulting in the estimation of parameters illustrating germination proportion and rate. Germination proportion and rate increased with seed storage length, and were increased by short and decreased by long dry periods occurring after imbibition. These results showed that, in contrast to crop species, A. myosuroides germination cannot be expressed simply as a function of hydrothermal time, but also requires the determination of the proportion of non-dormant seeds as a function of seed age and environmental conditions.     

A decision algorithm for patch spraying

It has been established that weeds are spatially aggregated with a spatially varying composition of weed species within agricultural fields. Site‐specific spraying therefore requires a decision method that includes the spatial variation of the weed composition and density. A computerized decision method that estimates an economic optimal herbicide dose according to site‐specific weed composition and density is presented in this paper. The method was termed a ‘decision algorithm for patch spraying’ (DAPS) and was evaluated in a 5‐year experiment, in Denmark. DAPS consists of a competition model, a herbicide dose–response model and an algorithm that estimates the economically optimal doses. The experiment was designed to compare herbicide treatments with DAPS recommendations and the Danish decision support system PC‐Plant Protection. The results did not show any significant grain yield difference between DAPS and PC‐Plant Protection; however, the recommended herbicide doses were significantly lower when using DAPS than PC‐Plant Protection in all years. The main difference between the two decision models is that DAPS integrates crop–weed competition and estimates the net return as a continuous function of herbicide dose. The hypothesis tested is that the benefit of using lower herbicide doses recommended by DAPS would disappear after a few years because weed density will increase and thus require higher doses. However, the results of weed counting every year did not confirm this hypothesis.     

A risk-qualified approach to calculate locally varying herbicide application rates

Weed competition can decrease crop yield and profit. Herbicides are applied to reduce weed populations, minimize crop loss and maximize profit. Traditional practice is to apply herbicides at a uniform rate over an entire field. Complete knowledge of the weed distribution and appropriate instrumentation on the spraying equipment would allow the farm manager to apply the 'correct' locally varying herbicide application rate. The locally variable rate would be greater in areas of high weed density and less where there are few weeds. A locally varying treatment would have both economic and environmental advantages. A major challenge facing farm managers is the unavoidable uncertainty in the spatial distribution of weeds in any particular field. This uncertainty in weed distribution influences the optimal locally varying herbicide rate. A mathematical model is presented to calculate the optimal herbicide application rate using geostatistical models of uncertainty in weed density combined with principles from decision making. Weed data from a 34-ha field near Saskatoon, Saskatchewan, Canada, illustrate the application of these tools. Weed control was achieved with a significant reduction in total herbicide use.     

背负式静电喷雾器静电喷施对2种玉米田除草剂的减量效应

为了明确背负式静电喷雾器静电喷施对玉米田除草剂的减量效应,以土壤处理除草剂40%乙·莠SE和茎叶处理除草剂30%苯唑草酮SC+90%莠去津WG+专用助剂作为参试除草剂,测定了背负式静电喷雾器在不同施药剂量下的静电喷施效果。结果表明:背负式静电喷雾器静电喷施对40%乙·莠SE和30%苯唑草酮SC+90%莠去津WG+专用助剂均有明显的减量效应。当40%乙·莠SE和30%苯唑草酮SC+90%莠去津WG+专用助剂中90%莠去津WG的施用量分别降至2 850 mL/hm 2和825 g/hm 2时,较各自推荐施用量分别减少用药量36.67%和21.43%,对玉米田主要杂草仍有理想的控制效果,株防效超过64%,鲜重防效超过95%;并可最大限度地改善玉米茎、穗部性状,减轻对玉米产量的影响,较人工除草分别减产1.96%和0.29%。可见,背负式静电喷雾器在玉米田土壤处理除草剂和茎叶处理除草剂喷施中具有较大的应用价值。     

Improved management of Avena ludoviciana and Phalaris paradoxa with more densely sown wheat and less herbicide

Summary The effectiveness of crop competition for better weed control and reducing herbicide rates was determined for Avena ludoviciana and Phalaris paradoxa . Four experiments, previously broadcast with seeds of the two weeds in separate plots, were sown with three wheat densities, and emerged weeds were treated with four herbicide doses (0–100% of recommended rate). The measured crop and weed traits were first analysed across experiments for treatment effects. Grain yield and weed seed production data were then analysed using cubic smoothing splines to model the response surfaces. Although herbicide rate for both weeds and crop density for P. paradoxa had significant linear effects on yield, there was a significant non-linearity of the response surface. Similarly, herbicide rate and crop density had significant linear effects on weed seed production, and there was significant non-linearity of the response surface that differed for the weed species. Maximum crop yield and reduction in seed production of P. paradoxa was achieved with approximately 80 wheat plants m⁻² and weeds treated with 100% herbicide rate. For A. ludoviciana , this was 130 wheat plants m⁻² applied with 75% herbicide rate. Alternatively, these benefits were achieved by increasing crop density to 150 plants m⁻² applied with 50% herbicide rate. At high crop density, application of the 100% herbicide rate tended to reduce yield, particularly with the A. ludoviciana herbicide, and this impacted adversely on the suppression of weed seed production. Thus, more competitive wheat crops have the potential for improving weed control and reducing herbicide rates.     

Variation in seed dormancy and light sensitivity in Alopecurus myosuroides and Apera spica-venti

Variation in seed dormancy and light sensitivity was studied in Alopecurus myosuroides and Apera spica-venti . Seeds were collected from different populations, at different dates and from plants emerging in autumn or spring, and used in four experiments. In the first experiment, initial dormancy was investigated in light and darkness. In Expt 2, buried seeds were exhumed on 16 occasions, from September 1997 to March 2000, and germinated in light, in darkness and after a 5-s light exposure. In Expt 3, emergence was recorded for seeds sown in pots outdoors. In Expt 4, stratified seeds of A. myosuroides only were exposed to photon irradiance ranging from 0.1 to 25 600 μmol m⁻². Variation was high among seed collections, but both species showed winter annual dormancy patterns. Apera spica-venti germinated to high percentages in autumn but negligibly in spring. Alopecurus myosuroides germinated less in spring when tested in darkness and after a short light exposure and emerged poorly in spring, which reflected photo-desensitisation during cold stratification. We conclude that the peak of emergence in A. myosuroides , and to some extent in A. spica-venti , is largely regulated by exposure to light interacting with low-level dormancy. This offers valuable information regarding optimal timing of weed control measures.     

Characterizing the spatial pattern of Abutilon theophrasti seedling patches

Spatial biology of weed populations is the study of weed patches and their relevant patch-level processes. In this context, a patch was defined as an area in which individuals are aggregated into discrete subdivided populations. Four Abutilon theophrasti seedling patches in two continuous maize production fields were surveyed using a contiguous grid of quadrats between 1995 and 1997. Surveyed area was dependent on patch size and ranged from 96 m² to 1134 m². Within each area, all seedlings were counted in each 1 m × 0.75 m quadrat in June, just before post-emergence weed control, and in mid-July after all weed control practices were completed. The spatial pattern observed in the seedling distribution maps was single or multiple focal points of high seedling density that decreased with distance from the focal point. Two-directional correlograms corroborated this visual observation, such that A. theophrasti seedling density in neighbouring quadrats was spatially autocorrelated, and correlation strength decreased with distance separating quadrats. Autocorrelation coefficients decreased at a greater rate across crop rows than parallel to crop rows. Visually, patch shape was elliptical and oriented in the direction of field traffic. Factors affecting patch-level processes of spatial aggregation, stability and edge dynamics were considered.     

Practical experiences with a system for site-specific weed control in arable crops using real-time image analysis and GPS-controlled patch spraying

Information on temporal and spatial variation in weed seedling populations within agricultural fields is very important for weed population assessment and management. Primarily, spatial information allows a potential reduction in herbicide use, when post‐emergent herbicides are only applied to field sections with high weed infestation levels. This paper presents a system for site‐specific weed control in sugar beet, maize, winter wheat, winter barley, winter rape and spring barley. The system includes on‐line weed detection using digital image analysis, computer‐based decision making and Global Positioning System‐controlled patch spraying. In a 2‐year study, herbicide use with this map‐based approach was reduced in winter cereals by 6–81% for herbicides against broad leaved weeds and 20–79% for grass weed herbicides. Highest savings were achieved in cereals followed by sugar beet, maize and winter rape. The efficacy of weed control varied from 85% to 98%, indicating that site‐specific weed management will not result in higher infestation levels in the following crops.     

Effect of environmental conditions on Alopecurus myosuroides germination. I. Effect of temperature and light

Summary Control of Alopecurus myosuroides with a minimum amount of herbicide may be improved by models describing the impact of cropping system effects on the weed's demography. In order to develop the submodel relating the weed seedbank to the emerged seedlings, a series of laboratory experiments was set up to analyse and quangify the effects on A. myosuroides germination of temperature, storage location, seed age and light incidence for various seed populations. The minimum temperature requirement for germination was estimated at 0 °C, and it did not depend on seed population or weight. In the subsequent experiments, a four-parameter non-linear equation depending on accumulated day-degrees was fitted to the observed germination data, resulting in the estimation of parameters quangifying germination proportion and rate. Storage location had little influence on germination. Germination proportion and rate increased with seed age and were higher in light than in darkness. The later the germinating seeds were submitted to light, the less was their germination stimulated. These observations were synthesized by a series of equations quangifying the relative variation in germination parameters as a function of storage and light conditions, independent of seed population.     

Real‐time variable‐rate herbicide application for weed control in carrots

A camera sensor for precision weed control in arable fields has been developed at the Leibniz Institute for Agricultural Engineering. For herbicide spraying in carrots, the sensor was positioned at the front three‐point linkage of a tractor and was operated between carrot rows. In field trials in 2 years, real‐time (online) technology in which sensing and spraying were performed in one step was evaluated in comparison with a conventional uniform spraying application. The spray volume was linearly adjusted to the camera‐detected weed coverage level from a minimum of 200 L ha^?1 if no weeds were present to a maximum of 400 L ha^?1. The herbicide savings were 30% in 2007 and 34%, 43% and 36% for each of three applications in 2010. There were no significant differences between the camera‐based and conventional spraying approaches with regard to yield parameters, total carrot weight and weight of marketable carrots. Regarding the weed control efficiency of the camera‐based spraying procedure in the 2010 experiment, no trend was observed between the weed coverage and the application rate of the previous spraying.     

Patch spraying: future role of electronics in limiting pesticide use

Developments relating to the control of application equipment can deliver improvements in pesticide use by better matching applications to target requirements. This may have components relating to the spatial distribution of a weed, pest or disease or methods by which the target, particularly a crop canopy, can be described with respect to a given application. Changes in application can relate to the dose and/or volume applied, but may also concern the way in which a treatment is delivered in terms of parameters such as spray trajectory angle and droplet size distribution. For many weed species there is evidence of patchy distributions in field situations. Studies have shown that savings of typically up to 40% in herbicide use can be achieved by adopting patch spraying approaches in such situations. Weed patch detection is key to the performance of such patch spraying systems. In widely spaced rowcrops such as vegetables, there is considerable scope for developing fully automated detection systems based on image analysis, and for the development of accurate guidance systems that apply pesticides only to the crop row. In crops with a relatively high plant density, weed detection in the medium/short term is likely to be based on manual discrimination. The costs of labour for manual weed patch mapping have been estimated at less than 1.50 ha(-1) pounds sterling. Potential savings in pesticide use can also be made if applications are matched to crop canopy structure. This is most important in bush and tree crops where savings of up to 75% in pesticide use could be achieved. In crops such as cereals, studies have shown that savings in fungicide use may be possible, particularly at earlier stages of growth by adjusting spray delivery to measured canopy characteristics. Key components of the performance of application systems concern the ability to deliver over a dose rate range of more than 3:1 while maintaining control of variables such as delivery trajectory angle and spray quality. Traceability and the effective monitoring of applications is likely to be a major driver influencing the uptake of more sophisticated control systems. Methods of labelling pesticides with systems that can be read by the application unit will be an important step in the development of recording and data handling systems that will operate safely with the minimum of operator input and enable the environmental advantages of targeted pesticide application to be monitored.     

Comparison of sampling methodologies for site-specific management of Avena sterilis

The ability to manage weed infestations in a spatially precise manner requires efficient and accurate methods of mapping weed distributions. A study was conducted to compare four different ground-based methods for collecting georeferenced information on infestations of Avena sterilis in winter wheat and barley. Sampling was performed at harvest by scoring panicle density, either from the ground or from a combine, by counting the number of panicle contacts with a stick moved horizontally over the crop canopy by an observer walking through the field, and by sampling A. sterilis seed rain on the ground. No significant differences were observed among the populations estimated by the four methods. A partial budget analysis of the in-season costs and benefits of spraying patches using these methods showed that visual scoring from the combine was the most appropriate method for the creation of weed management maps to be used for patch spraying in the following season. As a large variety of spatial patterns may be found in fields, the recommended sampling method might be field-specific and optimality should be verified for general use.     

Assessment of management options for Salsola australis in south-west Australia by transition matrix modelling

A matrix model of the life cycle of Salsola australis was constructed, based on population ecology data collected from the district of Lake Grace, Western Australia. The model was used to assess potential control strategies for this summer annual weed within the Western Australian broad acre grain cropping system. The population growth rate (λ) of S. australis in the absence of weed control strategies was 1.49 and was virtually unaffected by the dormant seedbank. However, λ increased to 8.21 if it was assumed that a constant number of seed immigrated into the area in question from neighbouring populations of S. australis , through farm-scale seed dispersal. As a result, effective weed management depended on reducing seed dispersal. The model determined that burning all senesced, mobile plants in late autumn, combined with herbicide control of the largest cohorts of S. australis in summer and autumn, reduced population growth rate to 0.79. This control strategy resulted in a 66.1% chance of the population becoming extinct over 25 years. Management strategies are proposed based on the results of the models and further research is required to validate their effectiveness and practicality in the field.     

Effect of patch size on seed removal by harvester ants

The harvester ant Messor barbarus can be responsible for high weed seed losses in dry land cereals in Spain. Because weeds occur in patches, harvester ants have to be able to find and exploit patches. However, seed patches can differ in size and may, therefore, differ in the probability of being discovered and exploited. Here, 90 patches varying in size from 0.25 to 9 m² were created in three 50 × 50 m subareas in a cereal field. Oat seeds were sown as weed seed surrogates in the patches at 2000 seeds m^?2. After 24 h, those remaining were collected and the exploitation rate (the percentage of seeds removed per patch discovered by ants) was estimated. Harvester ant nests and the location of the seed patches were georeferenced and used to estimate distances between them. The patch encounter rate (the proportion of patches discovered by the ants) decreased slightly, but significantly, with decreasing patch size, though not the exploitation rate, which was lowest in the smallest patches (78–94%) and highest in the largest (99–100%). Seed patches that were not found or partially exploited were mostly located in subareas with a lower ant nest density or a longer distance away from the nearest nest than seed patches that were fully exploited. The results of this study indicate that the interaction between the spatial distribution of ant nests and the patchy distribution of seeds can create opportunities for seeds to be subjected to lower levels of predation.     

Modelling herbicide dose and weed density effects on crop:weed competition

The effects of a range of herbicide doses on crop:weed competition were investigated by measuring crop yield and weed seed production. Weed competitivity of wheat was greater in cv. Spark than in cv. Avalon, and decreased with increasing herbicide dose, being well described by the standard dose–response curve. A combined model was then developed by incorporating the standard dose–response curve into the rectangular hyperbola competition model to describe the effects of plant density of a model weed, Brassica napus L., and a herbicide, metsulfuron‐methyl, on crop yield and weed seed production. The model developed in this study was used to describe crop yield and weed seed production, and to estimate the herbicide dose required to restrict crop yield loss caused by weeds and weed seed production to an acceptable level. At the acceptable yield loss of 5% and the weed density of 200 B. napus plants m^–2, the model recommends 0.9 g a.i. metsulfuron‐methyl ha^–1 in Avalon and 2.0 g a.i. in Spark.     

Response of Sorghum halepense demographic processes to plant density and rimsulfuron dose in maize

In spatially heterogeneous weed infestations, variable dose technologies could be used to minimise herbicide use; high doses could be applied to reduce high‐density patches and low doses to maintain weed populations in low‐density portions of a field. To assess the potential short‐ and long‐term effects of variable herbicide dose and site‐specific management, the major weed demographic processes were described and parameterised in this study. Various doses of rimsulfuron (from 0 to 12.5 g a.i. ha^?1) were applied to different densities of Sorghum halepense (0–100 plants m^?2). Contrary to similar studies with other weed species, higher herbicide efficacy was not observed at low densities, suggesting that the same rimsulfuron dose should be applied regardless of the S. halepense density. The highest percentage of control was obtained with the full rimsulfuron dose. However, it did not guarantee a decrease of the infestation in the following season in the field areas where the initial S. halepense density was lower than 60 plants m^?2. Reduced doses of rimsulfuron to control S. halepense cannot be recommended based on our results.