|
Effects of Mechanical Weed Control in Spring Cereals – Flora, Fauna and Economy 4 DiscussionThe present project has focused on two major questions that have often led to discussions when chemical and mechanical weed control have been compared. One of these was an indication from a preceding project which shed light on the phase of conversion from conventional farming to organic farming, one of the changes being replacing chemical weed control with mechanical weed control (Navntoft et al. 2003). The indication of interest was that mechanical weed control might have adverse effects on beneficial soil-dwelling arthropods. These beneficials prey upon pest species providing ecosystem services of economic importance for the farmers (Östman 2003) and are also serving as prey for birds. The other debate area was the fate of bird nests during the operations of mechanical weed control. Both for the direct effects on invertebrate fauna and on bird nests as well as for the indirect effects (like, e.g., loss of bird food) more damage was anticipated after several than after few harrowings. By comparing two, three and four weed harrowings, and in addition including other weed control results in the modelling, the project results show effects on wild plants, the crop, the yield, arthropods and ground-nesting birds. 4.1 Weeds and arthropodsMore than two weed harrowings had significant effects on flora and arthropod fauna in the experimental fields. The weed biomass decreased 43% after three weed harrowings when compared to two harrowings and as much as 48% after four harrowings compared to two. Along with the change in biomass, also a marked difference in weed species occurrence was seen at the high treatment level compared to only two harrowings. The findings from the phenological studies with respect to the impact on flowering frequency and seedling frequency, respectively, may be explained in two ways. Firstly, there might be a sublethal effct on the plants due to harrowing (much like the effects of low dosages of herbicides); when the weed plant once has been covered by soil, the chance diminishes for this plant to reach maturity, make flowers and set seeds. Secondly, more seedlings may be provoked to germinate by the harrowing, but they fail to reach maturity due the late time for germination. Only the latter explanation is substantiated by the analysis of the field data obtained in this project. The observations seem to indicate, that higher harrowing frequency reduces the number of weed plants by delaying the development of surviving plants, thus reducing flower and seed production. High weeding frequency thereby results in lower above ground weed biomass as well. Weed harrowing had a significant negative effect on numbers of the arthropod predator complex Agonum spp., Bembidion spp., Tachyporus spp. and Linyphiidae. Shortly after the third harrowing the total density of these predators was 35% higher in plots where only two harrowings had been carried out, and shortly after the fourth harrowing this difference was 21%. The spiders Linyphiidae were the most vulnerable to weed harrowing showing a direct negative response to treatment after both third and fourth harrowing, probably as the result of a lethal effect on the mechanical operation as supported by findings of Thorbek and Bilde (2004). The rove beetles Tachyporus spp. were also vulnerable to weed harrowing, responding significantly and directly following the third harrowing. The ground beetles Agonum spp. and Bembidion spp. were the least sensitive to weed harrowing since they did not respond significantly directly to weed harrowing. The third weed harrowing was more detrimental to the polyphagous predators than the fourth harrowing. Non-crop vegetation cover may provide an attractive foraging and resting site, which is particularly valuable prior to canopy closure (Lee & Landis 2002). This means that the arthropods may be more vulnerable to early weed harrowings carried out when the canopy is relatively open, which might explain the stronger effect of the third weed harrowing. We do not know, however, how much of the negative effect of weed harrowing after the fourth harrowing that can actually be explained by carry-over effects of the third harrowing. The findings of Thorbek & Bilde (2004) indicated that recolonization of spiders following weed harrowing eliminated the direct negative effects of the treatment within seven days. Also for carabids and staphylinids there were indications that immigration following weed harrowing compensated for direct negative effects of harrowing within a week (Thorbek & Bilde 2004). Weeds were very important for the presence of beneficial arthropods, and weed harrowing was therefore also indirectly harmful to the beneficials. The negative effects of weed harrowing could be explained by the change in plant biomass since there was a highly significant, positive relationship between vegetation biomass (weeds and crop) and arthropod densities, with weeds having the highest positive effect on arthropod densities. An extra 1 g biomass of weed per sample could increase the density of the predator complex with up to 8 %. The importance of the vegetation was further emphasised by the statistical analysis of the effect of four harrowings because inclusion of the covariates ‘weed’ and ‘crop’ biomasses replaced the treatment effect of weed harrowing. The three predator taxa Agonum spp., Linyphiidae and Tachyporus spp. all responded positively to a higher vegetation biomass. For Bembidion spp. a significantly increase as a result of higher weed biomass was only found when data from two harrowing plots was analysed separately; in fact, a significant negative relationship between Bembidion densities and crop biomass was revealed. The contrary response of Bembidion spp. is in line with the findings of Mitchell (1963), who found that the common B. lampros had a preference for sparse vegetation (maybe because of a need for warmer soil). However, this response of B. lampros is not typical of the predatory ground beetle complex of the field taken as a whole (Speight & Lawton 1976). Experiments with A. dorsalis and B. lampros (field arena experiments) gave further evidence that indirect effects of weed harrowing on ground beetles is species dependant. A. dorsalis had a preference for un-harrowed ground, which is in line with its positive relationship with weed biomass, whereas B. lampros had no preference for harrowed or un-harrowed ground, a result that is also in line with the observations from the larger field experiment. The probability of finding the ground beetles on the weed harrowed halves decreased during the day (as shown by the significant, negative parameter ‘time’ in Table 3.13). We do not have a bulletproof explanation, but it may be taken as an indication that the beetles had an increasing preference for undisturbed microhabitats during the day, maybe because of increasing temperatures. Honek (1997) revealed that epigaeic predators preferred shaded control stands at air temperatures of 18-25°C, while at temperatures below 16°C the catches on shaded and bare ground surface were similar. This may explain the relatively evenly distribution between weed harrowed and undisturbed soil found in the field-frame experiment on 5 June, which was a rather cool day with showers, and the strong preference for un-harrowed soil revealed for A. dorsalis on 11 June, which was warm and sunny. The significant, positive relationship between beneficial arthropods and weed biomass made it possible to draw up a non-linear model which may be useful in the implementation of farming strategies that favour enhanced predation in spring cereal fields. Other investigations have also revealed a positive effect of weeds on beneficial arthropods. Speight & Lawton (1976) demonstrated that beetles exerted a higher predation pressure on artificially placed fruit fly pupae (Drosophila) in weedy and presumably more humid fields. They observed a positive correlation between ground beetles and Poa annua L. and found that the underlying mechanisms were probably complex, but that is was likely that the weeds protect the predators from extremes of climate, i.e. insolation during the day and desiccation both during the day and at night. Rivard (1966) found higher catches of carabids in area of higher humidity and Powell et al. 2004 found further indications that there is a negative relationship between the majority of the epigael invertebrates and soil moisture in the summer. These findings might suggest that our results were mainly due to an arthropod – humidity relationship rather than an arthropod – weed relationship. We carried out analyses (Sørensen similarity and weighted Ellenberg indices, section 3.1.1.3) in order to test this possibility. Weed species composition reflects differences in ecological growth conditions e.g. differences in soil humidity. Differences in weed species composition should therefore unveil underlying growth conditions. There was however no indications of a different weed species composition as a result of more humid soil (e.g. in lower parts of the fields) as tested by weed species index calculations, and the positive relationship between weed biomass and important arthropods can be regarded as undisturbed this factor. This means that positive effects of the weed itself (e.g. an improved microclima) most likely is an important variable for the occurrence of generalist arthropod predators. In addition to microclimatic considerations, it is also possible that there is an indirect effect operating via the abundance of natural prey, which may be more common in the denser patches of weed (Speight & Lawton 1976). Potts & Vickerman (1974) found a positive relationship between the numbers of predatory Coleoptera in cereal fields and the abundance of macroscopical Isotomidae (Collembola). Also Hawthorne & Hassall (1995) demonstrated that carabid density was positively correlated with the density of the prey items Collembola and aphids, besides being positively correlated with the vegetation cover. Pearce & Zalucki (2006) however, found that predator aggregation did not correlate consistently with pest aggregation, plant damage or predation rate. Collembola can be inhibited by dry microclimates (Basedow 1994). Mechanical weed control may therefore indirectly have a negative impact on the complex of polyphagous predators through a negative effect on alternative prey. However, Odderskær et al. (2006) found no negative effects of weed harrowing on densities of Collembola in organic spring wheat. This may be taken as an indication that effects of weed harrowing on alternative prey may not have influenced the distribution of the polyphagous predators in the present experiment. The distance to perennial vegetation e.g. field margings is important for the dispersal and distribution of generalist arthropod predators in arable fields (Coombes & Sotherton 1986), a result that was also found in this experiment, based on data following the fourth harrowing. However, following the third harrowing the effect of distance to perennial vegetation was insignificant, even the estimates of the effects of distance were in most cases negative indicating that predator densities also in this case decreased at increased distance to field margins. Generalist arthropod predators are especially important early in the growth season when specialist enemies are not yet present (Ekbom et al. 1992) These predators act as natural enemies of crop pests (Sunderland 1975), and their presence in the field early in the season can reduce the build up of pests such as the bird cherry-oat aphid (Rhopalosipum padi L.) in spring cereals (Ekbom et al. 1992). The population build up of R. padi in spring cereals usually starts late May and early June (L.M. Hansen pers. comm.). It is therefore important that the number of arthropod predators is as high as possible at that time for high natural control of aphids. To ensure a time period of one week for recolonization of predators (Thorbek & Bilde 2004), harrowing should not take place later than mid May. This should allow the population of predators to recover before aphid build up. In all cases the 2nd weed harrowing was carried out by mid May in this experiment (Table 2.1). The present experiments only concern effects on adult individuals of polyphagous predators. Soil tilling may also affect juvenile stages (larvae and pupae) negatively. Juveniles occurring in the soil during the time of weed harrowing are often offspring of “autumn-breeders” that will emerge as adult beneficials in the field later in the season. It is also important that densities of these predators remain high later in the growth season, when they may impact pests such as Grain Aphids (Sitobion avenae (F.)) (Holland & Thomas 1997). The effect, if any, of weed harrowing on the juvenile beneficials is unknown. For the promotion of early-occurring beneficial arthropods, and based upon the results of the present experiment, we suggest that weed harrowing is carried out early and limited to an absolute minimum and that more weed should be tolerated in the fields; the modelling indicates up to approximately 15 - 20 g weed biomass per m² in spring wheat should be allowed based on data collected in late May and early June (Figure 3.9). 4.2 Birds4.2.1 SkylarkThe number of hatchlings and the number of chicks leaving the nest were two to three times higher in plots exposed to two weed harrowings than in plots exposed to four harrowings. The significant difference in Skylark productivity between treatments was due to a higher percentage of nests being successful in “2 harrowings” plots than in “4 harrowings” plots (65% vs. 28%), whereas neither the number of nests per ha nor the number of chicks per successful nest differed between treatments. Only 5% of the Skylark nests found in “2 harrowings” plots were destroyed by weed harrowing whereas this was the case in 26% of the nests in “4 harrowings” plots. In the latter plots, an additional 4% of the nests were exposed to harrowing but survived with a reduced number of eggs/young. Weed harrowing was generally destructive to the nests affected: 83% of the nests and 87% of the eggs or nestlings that were subject to harrowing were lost. In a study where artificial Skylark nests with eggs of Coturnix chinensis were exposed to weed harrowing, Odderskær et al. (2006) found that 72% of the nests were destroyed by a single harrowing. In total, only 16% of the monitored Skylark nests were exposed to weed harrowing because most harrowings were performed before the number of Skylark nests peaked. The first post-emergence harrowing (weed harrowing no. 2) affected only 3 nests, whereas no. 3 and 4 affected 5 and 10 nests, respectively, although they were performed on just half of the area. Generally speaking, harrowings performed 37 days or less after sowing affected very few Skylark nests while the harrowings performed 40 days or more after sowing clearly put Skylark nests at risk. In the study of Odderskær et al. (2006), where the last harrowings were carried out 33-38 days after sowing, only two natural Skylark nests with eggs (and 3 without eggs) out of 92 nesting attempts were affected by weed harrowing. Odderskær et al. concluded that the direct, negative effect of weed harrowing on the reproductive output was insignificant, which is generally in accordance with the results from the “2 harrowings” plots in the present study. In “2 harrowings” as well as in “4 harrowings” plots, predation was the most important cause of nest losses. Contrary to Lapwings and Oystercatchers, Skylarks have no active defence against predators and rely solely on their camouflage. Predation rates were particularly high early in the season with an estimated 68% of nests established before 21 May being predated. Later in the season, the frequency of predation was significantly higher in plots that were subject to four times weed harrowing than in plots that were harrowed only twice; the estimated predation rates were 57% and 25%, respectively. The risk of predation seems to be highly dependent on the vegetation cover. Early in the season, ground cover is sparse and the nests are fairly easily visible. Later, the higher risk of predation in “4 harrowings” plots suggests that the less developed weed cover caused by the extra harrowings makes the nests more visible to predators, probably aided by changes to the crop structure for a few days after harrowing. The main predators were surely Crows, especially early in the season, but, interestingly, Marsh Harrier turned out to be an important nest predator in late season in some fields. Being adapted to foraging in reedbeds, Marsh Harriers seem able to localize the Skylark nests even in high vegetation, provided the ground cover is not too dense. Overall, the difference in nest success rate (and thus in productivity) between treatments was due to the combined effects of a higher direct nest mortality caused by weed harrowings and a higher frequency of nest predation, probably due to the less developed vegetation cover, in “4 harrowings” plots. The impact of weed harrowing on Skylark breeding success in spring-sown wheat thus depends strongly on the timing of the harrowing and the number of harrowings performed:
It must be stressed that these recommendations for the timing of weed harrowing are based solely on results from spring-sown wheat and that they should not uncritically be extrapolated to other spring-sown cereal crops where phenology and growth pattern may be different. The negligible effects of weed harrowing on natural Skylark nests found in the study of Odderskær et al. (2006) are fully consistent with the above recommendations. Using a modelling approach, Odderskær et al. extrapolated their results from spring-sown wheat to spring barley and found that weed harrowings performed before 20 May have no significant impact on the Skylark population. Their recommendation was that weed harrowing in spring cereals should be carried out no later than 30 days efter sowing, which is probably on the safe side. Chick survival after leaving the nest was not monitored in the present study and has indeed been very little studied (Donald 2004). Thus, the above discussion and recommendations are based solely on the number of Skylark nests producing young leaving the nest and do not take into account that the prospects of these young may differ during the season. In many bird species, early young have a better first year survival and a greater lifetime reproduction than late young. If this is the case also in Skylarks, the negative effects of early weed harrowings on the population may well be underestimated. 4.2.2 Lapwing and OystercatcherWeed harrowing was the most frequent cause of failure of Lapwing nests in the study. Harrowing of nests was generally destructive: only 15% of the harrowed nests survived with all eggs intact, and 66% of the eggs were destroyed. On fields where weed harrowing was carried out, 40% of all recorded nesting attempts completely failed because of harrowing and a further 25% of the nests were damaged but survived with one or more viable eggs still being incubated. The mean number of hatchlings per nest was 0.63 in fields with weed harrowing, compared to 2.35 in similar fields where weed harrowing was not performed and 2.80 in reference areas with perennial grassland. Few Oystercatcher nests were found, but the available data suggest that weed harrowing is at least as damaging to this species. None of the five Oystercatcher nests found in plots that were subject to weed harrowing was successful, whereas the two nests placed in untreated plots produced two hatchlings each. There was no significant difference in Lapwing breeding success between plots that were subject to only one weed harrowing after emergence of the crop and plots where several post-emergence weed harrowings were carried out. Lapwings breed early in the season – among others because their anti-predator strategy depends on the incubating bird being able to see the predator at some distance – so the major damage is caused by the first weed harrowings. In fields where weed harrowings were carried out, only 6 out of 40 nests were established after the first post-emergence weed harrowing (several of them probably as replacement clutches), and only one of these nests was successful while 4 were predated. Thus, most of the Lapwings whose nests were destroyed by post-emergence weed harrowing did not produce replacement clutches (at least not inside the study fields), and for those who did, the success rate was low. Ettrup & Bak (1985) also state that the success rate of late clutches is poor. However, if successfully hatched, late chicks survive at least as well as early chicks (Klomp & Speek 1971). Perhaps to compensate for the poor prospects of late replacement clutches, Lapwings and Oystercatchers are much less prone to abandoning a damaged nest than Skylarks are. As a rule, incubation of a harrowed nest was resumed as long as one or more eggs were intact or only slightly damaged, even if the eggs were scattered (up to at least 30 cm) outside the nest. In several cases where the old nest was destroyed, a new nest scrape was made and the surviving eggs were rolled into the new nest. Interestingly, at least 4 eggs with small holes or cracks after weed harrowing proved able to hatch 4 to 24 days after being damaged. From the outset, the emphasis of the study was on effects of weed harrowing performed after the emergence of the crop (because these are the harrowings affecting Skylark breeding success). Hence, the study only to a minor extent elucidates the impact of pre-emergence harrowings on Lapwing breeding success. Data from 2004-05 indicate that at least on some fields, active Lapwing and Oystercatcher nests occurred already at the time when the pre-emergence harrowings were performed, but the frequency of such early nests is unknown. In 2006, the vast majority of Lapwing nests (= 80%) were established before 20 April in undisturbed areas or within 12 days after sowing in cereal fields and would thus be vulnerable to pre-emergence harrowings, but the time of nesting varied greatly between fields (cf. Figure 3.13). Thus, the data do not provide any clear indication of the frequency of Lapwing breeding attempts that (potentially) collide with pre-emergence weed harrowings. Even if nesting attempts are destroyed by pre-emergence harrowings, laying of replacement clutches with fair chances of survival is possible at this stage, because the height of the crop is still suitable for Lapwings. Galbraith (1988) found that first clutches and replacement clutches did not differ significantly in size, so early replacement clutches may – at least in theory – be as productive as first clutches. Unfortunately, data are too sparse to allow the determination of a cut-off date, after which the probability of success is significantly reduced. On the Zealand study fields, few nests were established after 5 May, and only one of these (placed in an area where crop growth was poor) was successful. By contrast, on the Krovang field at Kalø, successful Lapwing and Oystercatcher nests were established throughout the first 3 weeks of May 2005, whereas in 2004 and on Keglehøj, all nests established after the first few days of May failed. Other farming activities (ploughing, harrowing/sowing and rolling) were completely destructive to all nests present on the field. Ten nests were destroyed by these activities on a field left untreated until 18 May, but most or all of the pairs affected probably laid replacement clutches. As a rule, re-laying is possible after ploughing and sowing, whereas rolling may be more problematic – and comparable to weed harrowing – in this respect. Ettrup & Bak (1985) state that farming activities in spring is most years do not significantly reduce the number of chicks produced because the lost clutches are replaced. However, they also infer that this may not be the case in wet and cold years, where farming activities are delayed and take place over a longer period. Humid patches, where cultivation takes place later than in the rest of the field, represent a particular problem. Such areas with sparse or no vegetation may be the only areas suitable for nesting after mid-May, when the crop is too high elsewhere. Galbraith (1988) describes how newly cultivated areas in late May were quickly colonized by Lapwings whose previous breeding attempt had failed. However, experiences from Vibygård in 2005 indicate that if the soil and vegetation structure of these humid patches is already suitable for Lapwings, late cultivation may be disastrous, destroying the nests at a time where successful re-laying is no longer possible. Really wet, uncultivated patches with a high water table and little vegetation may be of great value as a feeding habitat for the chicks. Nest predation was less important than in Skylarks with an estimated 23% of Lapwing clutches being lost to predators. This is in good accordance with other studies (compiled by Trolliet 2000) where predation losses varied between 9 and 50% with an average of 23% of clutches. The relatively low incidence of nest predation is probably related to an effective, active defence against predators. This predator defence is particularly effective where Lapwings breed semi-colonially, and the collapse of the colony structure (e.g. following weed harrowing of the area) may increase the risk of predation of surviving nests or single replacement clutches. For example, at Oremandsgård in 2004, not even two abandoned nests were predated while a colony of at least 6 pairs was active. However, after the colony had been destroyed by weed harrowing, the single surviving nest and one replacement clutch were predated within a week. In the solitary breeding Oystercatcher, the estimated predation rate (46%) was twice as high as in the Lapwing. Partly compensating for these egg losses, the survival of Oystercatcher hatchlings is better, probably because the chicks (uniquely among European waders) are fed by the parents (Cramp & Simmons 1983). Various Lapwing studies compiled by Hudson et al. (1994) indicate that an annual production of 0.80 to 0.97 fledged young per pair is necessary to prevent Lapwing populations from declining. Lapwing chicks are unable to fly until 35-40 days old (Cramp & Simmons 1983), and mortality rates of unfledged chicks are huge; four studies cited by Trolliet (2000) reported mortality rates between 58 and 93% (average 75%). Thus, each female must produce at least 3.2 hatchlings per year to maintain population size. In Table 3.18, the mean number of hatchlings per nest and per successful nest represent the minimum and maximum productivity per female (and year), respectively. On the fields exposed to weed harrowing, where nest losses occurred late and all but one of the supposed replacement clutches were unsuccessful, annual productivity per female was surely close to the recorded mean number of hatchlings per nest and far too low to be sustainable. On fields that were untreated after crop emergence and on perennial grassland, the productivity was probably close to the level needed in order to keep the population stable. Working in Scotland, Galbraith (1988) found that Lapwing chick survival was higher on grassland than on arable land and that survival in arable areas was positively related to the proximity of pasture. During their first days of life, most “arable” chicks in his study moved from their natal field to nearby pasture to find suitable feeding conditions. In years where weather conditions were favourable for crop growth, survival of chicks at the arable sites was as low as 7-15% , almost certainly due to food shortage (Galbraith 1988). Such survival rates may well be typical for conventional arable farmland and suggest that even if the hatching success is high, the number of fledglings produced is too low to maintain population size. Ettrup (2002) states that the Lapwing population in Danish farmland is maintained by immigration from meadows, but this has been questioned because breeding success on meadows is often poor (H. Olsen pers. comm.). Organic fields generally hold higher amounts of insects and other birds’ food items than conventional fields (e.g. Hald & Reddersen 1990, Brooks et al. 1995, Navntoft et al. 2003) and are probably superior to conventional fields as foraging habitats for Lapwing chicks. This may be important for chick survival; in one year of Galbraith’s (1988) study, adverse weather retarded crop growth and retained suitable feeding conditions for Lapwings in the fields for a longer period. This enabled many “arable” chicks to stay and feed in their natal field, increasing chick survival to 31%. Furthermore, organic farms more frequently hold a mosaic of spring-sown fields and pasture that is favourable to Lapwings. Although conventional (sprayed) fields may seem superior to organic (weed harrowed) fields as a Lapwing breeding habitat, the benefits of omitting weed harrowing may to a great extent be offset by less favourable conditions for chick survival. Therefore, the optimum solution is not to replace mechanical weed control with pesticide sprayings in Lapwing areas but to confine weed harrowing and other soil treatments to the period when Lapwings are still able to produce a replacement clutch with good chances of success. The available information is insufficient to allow the determination of a definite cut-off date, but the critical date may be around 1 May or few days after emergence of the crop. The preservation of wet patches with low or sparse vegetation in (or adjacent to) fields with breeding Lapwing may also increase the breeding success by improving chick survival. 4.3 Weed control4.3.1 The economic objectiveThere are lots of reasons why the weed most be controlled in spring cereals and other field crops. The weeds compete with the crop, weed plants can make it difficult to harvest the crop, uncontrolled weed will produce more seed, weed seed can cause higher water content in the grain, and weed seed can be expensive to remove from the grain. Also uncontrolled weed and more weed seed will dynamically increase the weed problem, from year to year and from crop to crop. In some cases weed species causing problems in one crop are most effectively (effect) or efficiently (costs) controlled in another crop. For instance some broad-leaved weed species causing problems in sugar beets are most effectively and efficiently controlled in spring cereals. And some species like Lamium spp. and Polygonum spp. are weakly and moderately competitive, respectively, in spring barley but moderately and weekly competitive in winter wheat (Rasmussen et al. 1997). So even if specific weed species or weeds in general are not a problem in spring cereals, the farmers may still have good reasons to control the weed also in the spring cereals. In the actual analyses the farmers objective is to control the weed effectively (a specific reduction) and efficiently (at the lowest possible costs). It is normally considered that a 70-80% reduction in weed biomass will be sufficient to prevent the weed from causing the above mentioned direct, indirect and dynamic problems and costs. In the Crop Protection Online (PVO) decision support system (Jørgensen et al. 2007 and Bøjer & Rydahl 2007) problematic weed species have individual, density based reduction thresholds. In spring barley the most problematic weed species like Galium aparine are, regardless of the density, to be reduced by at least 80% of biomass and in case of the highest density by at least 95%. More inferior weed species like Viola arvensis are not to be reduced at low densities but should be reduced by at least 75% at the highest densities. By using these individual thresholds the weed is found to be effectively controlled as well in a short-run as in a long-run perspective. The economic objective in the Crop Protection Online (PVO) system is to recommend herbicide solutions that will satisfy the individual thresholds at the lowest cost possible. Actually the recommendation is chosen independent of crop yield, crop prices and yield gain from controlling the weed. The herbicide-weed response parameters and the individual thresholds used in the Crop Protection Online (PVO) system have been established by using field trial data including detailed registrations of weed species density and biomass. Such detailed registrations are however not included in the available field trials used by Landsforsøgene (Petersen 2002) to asses effects of mechanical weeding and to compare effects of herbicide and weed harrowing in spring cereals. In the Landsforsøgene the weeding effect is systematically expressed in terms of net yield and in some cases also in terms of the total weed density (not individual species) before the first and after the last treatment[1]. In the Landsforsøgene context, the implicit economic objective is to optimize the short-run net yield. In this way the trials are perfect for comparing the efficiency of effective and robust, but not specific, solutions for weed control. As long as the tested solutions are believed to control the weed effectively in a long-run perspective, there is no conflict between the short-run economic objective and a long-run dynamic objective. However, if we are not sure about the effectiveness of the solutions, as is the case of mechanical weeding, the Landsforsøgene with their implicit economic short-run objective has some critical shortcomings. It is found that the selected trials are the best available data to asses the timing and weed control effect of strategies involving more than one or two post emergence harrowings in spring barley. However, due to missing information on harrowing intensity, missing information on weed biomass, and missing treatments with separate pre-emergence harrowings, treatments with separate harrowings performed at different periods, and treatments with different varieties and crop density, the selected trials are not the ideal background for selecting and recommending efficient mechanical weeding strategies. But these trials and the actual analysis might still be a good (and the best available) setting to evaluate and recommend mechanical weed control strategies involving more than just one or two post emergence harrowings. The following conclusions should be assessed in the light of all these shortcomings, limitations and possibilities. 4.3.2 Six weed harrowing trialsIn general spring cereals suppress the weed very well, and the analysis (see 3.3.1) of six Landsforsøg weed harrowing trials has shown that weed harrowing in most cases is not profitable in spring barley. An increasing number of harrowings did not increase the crop yield in spring barley. Weed harrowing is however an effective tool to control the weed in spring cereals. The higher the number of harrowings, the higher is also the weed reduction and the yield damage. Using a combination of one pre- and one or more post-emergence harrowings, the pre-emergence harrowing does not seem to affect the weed density (in general also confirmed by J. Rasmussen (pers. comm.)). The weed control effect of harrowing differs from trial to trial, but within each trial the effects from additional harrowings are almost constant. In some cases, one post-emergence harrowing will reduce the weed density by 70%, in other cases three post-emergence harrowings are needed to give a 50% reduction. The variation in the weed density reduction can neither be explained by region nor by weed density. Timing is normally considered to be a critical (but not verified) parameter in mechanical weeding. It is however found that the relative weeding effect of additional harrowings (five harrowings with one week intervals) is almost constant within each of the selected trials. This indicates that the timing of (at least the additional) harrowings after all is not important. Also Rasmussen & Nørremark (2007) found that the weeding effect in general was unaffected by the timing. It is well known that weeding often fails to be profitable, at least in a short-run perspective. According to Jørgensen et al. (2007) the average net yield gain from using herbicides is negative. Also Rasmussen (1993) stated that crop yield gain was often not apparent in experiments with post-emergence harrowing: “It is however difficult to reveal the causes, because quantitative approaches were not used. If crop yield gain fails to appear, the failure may be caused by low competition from weeds, or may be a result of a balance between crop damage caused by harrowing and yield gain”. Rasmussen used a modelling approach (Rasmussen 1991) to clarify such doubts. 4.3.2.1 Weed retardation, density, and biomassIn the absence of registrations on weed biomass, the analyses presented in section 3.3.1 and 4.3.2.1 have focused on the weed density. Analyses of five-year herbicide trials in spring barley (Jørgensen et al. 2007) indicate that the relative reduction in weed density raised to the power of 0.6 is a good approximation for the corresponding reduction in weed biomass. That the reduction in weed density and the reduction in biomass are not the same can be explained by the fact that the herbicides not just kill some weed but also retard the surviving weed, and also that different weed species may have varying potential biomass and varying sensitivity to herbicides. Also in the case of mechanical weed control there are differences in the weed species’ potential biomass and sensitivity to mechanical weed control, and the surviving weed will also to some extent be retarded. However, the difference in weed retardation between the two weed control methods has apparently not yet been reported. As mentioned before, a 70%-75% reduction in the weed biomass is normally considered to be an acceptable result and the implicit objective for mechanical as well as mechanical weeding in cereals. If the weed is retarded to the same extent by harrowing as by herbicides, a 55% weed density reduction in needed to accomplish this objective. If the harrowing does not at all retard the surviving weed plants, a 70% reduction in weed density is required to accomplish the objective. According to Table 3.22, a strategy with two or three post emergence harrowings will reduce the weed density by 54% and 74%, respectively. If the harrowed weed is retarded to the same extend as by herbicides, two post emergence harrowings are needed to accomplish the 70% objective; otherwise three post emergence harrowings are needed. Also according to Table 3.22, the second and third post emergence harrowing on average took place on 2nd and 12th of May, respectively, corresponding to 14 and 21 days after sowing, and the latest of these harrowings were performed on 17th and 23rd of May. In case a higher, acceptable reduction in the weed biomass can only be achieved by increasing the number of one week interval harrowings, as implied in the selected trials (Petersen 2002), an effective weed harrowing will be costly and occasionally in conflict with Skylark breeding. 4.3.2.2 Weeding intensityAn increased reduction in weed density and biomass might also be achieved by intensifying the individual harrowings. Unfortunately, information on the intensity of harrowings is missing in the selected field trials. It is however well known (e.g. Rasmussen 1991, Rasmussen 1993, Rasmussen et al. 2004) that the weed control effect and the crop damage definitely depend on the harrowing intensity. It is also established that two or more, less intensive harrowings may equal the effect of one very intensive harrowing. In that case, differences in harrowing intensity could explain the significant variation in weed control effect between trials (Figures 3.15 and 3.16). The harrowing intensity is often measured in terms of the relative crop soil cover (CSC) after harrowing. Normally a 20% CSC is recommended. LandbrugsInfo (Petersen 2006) recommends a 10% to 20% CSC, and in Rasmussen (1993) and Rasmussen & Nørremark (2007) even higher CSC values were successfully applied. On average, a 20% CSC from a single post emergence harrowing will reduce the weed biomass by at least 70% (e.g. Rasmussen & Nørremark 2007, figure 2). Analysis of the 32 selected Landsforsøg trials involving post emergence harrowing (see 2.4.1.2) indicates that the average CSC used in these trials was 10%, varying from less than 2% to more than 50%. The cereal type and variety, vigour, and crop density are important parts of a mechanical weed control strategy (e.g. Rasmussen 1993, Rasmussen & Rasmussen 2000, Gundersen et al. 2006). The higher the initial crop density, the more intensive harrowing can be performed without yield loss, and a weed competitive variety will reduce the need for a high crop density and for weed control. Also differences in these parameters could help explaining the huge variation in the weed control effect from trial to trial. But variety, vigour and crop density were not included in the selected trials. Because one intensive post emergence harrowing is more efficient than two or more, less intensive harrowings, and because the effect is not influenced by timing, a strategy with one single (but intensive) post-emergence harrowing carried out between 7 and 30 days after sowing, and no later than 20th May, appears to be an effective (biomass reduction), efficient (costs) and Skylark-friendly mechanical weed control strategy. Actually the farmers are advised to perform the first and most efficient harrowing as an early post emergence harrowing, and the later the harrowing is performed, the more Skylark nests are damaged (cf. sections 4.2 and 4.4). ). This means that there doesn’t need to be conflict between an effective and Skylark friendly mechanical weed control. As mentioned weed control is needed not just to optimize the actual net yield, but also to reduce harvest problems and long-run dynamic problems. It is normally considered, that there is a direct linear correlation between weed biomass production and the weed seed production, and thus the long-run dynamic weed problem. It is found that one intense harrowing is as effective as two or more less intensive harrowings to control the weed, and that timing of (at least) the additional harrowings is of limited importance to the weed effect and thus also of limited importance to the weed seed production. However, it has been found, section 3.1.1.4, that weed surviving two times of harrowing is more likely to flower than weed surviving four times of harrowing. In seems that the flowering is postponed in case of four harrowings. It is however hard to say whether or not it is the extra intensity or the later harrowing that causes the postponed flowering. 4.3.3 More field trials and a modelling approachBecause the mechanical weed control is accomplished most efficiently by using a few intensive post-emergence harrowings, and because the timing of at least the additional harrowings obviously is of limited importance, the investigations on mechanical weeding can be expanded to field trials with just one or more harrowings applied liberally from sowing to late May and early June. Unbalanced data from these extra field trials (different design and strategies) that also include chemical weed control have been used to construct new chemical, mechanical and mixed weeding strategies by using the complex of weeding models established (see 2.5.1.1). By using this model approach (as also suggested by Rasmussen 1993) mechanical and chemical weeding have been analysed (see 3.3.2) in a broader context than normally used in the Landsforsøgene. As an example the model is used to estimate the weed and yield effect of a higher (or lower) herbicide dose or more (or less) intense harrowing than observed in the actual field trials. In this way data from many different unbalanced trials can be transformed and treated as if they were carried out on the same locations (weed density and weather conditions) and with the same design and treatments (herbicide doses, harrowing intensity and timing etc.). The extended analysis has shown that a 0.3 TFI herbicide low dose strategy is probably the most (economically) efficient weed control strategy in conventionally grown spring barley. It has also been shown that a single, but intensive, post-emergence harrowing (resulting in at least 20% crop soil cover) is the most cost-efficient mechanical weed control strategy. The intensive post-emergence harrowing is more effective and just as stable, but more expensive (0.7 hkg or 60 DKK per ha) than the most efficient 0.3 TFI herbicide strategy. In case the conventional farmer needs a weed control effect beyond the normally required 70% weed biomass reduction, a low herbicide dose followed up by one post-emergence harrowing is maybe the most (economically) efficient strategy. It is also found that the timing of at least the additional post-emergence harrowings, except for the sometimes fatal too early harrowing in the one to two leaf stage of the crop, has no significant effect on neither weed biomass nor crop yield. In case the conventional farmer wants to reduce the use of herbicides in spring barley, the most efficient strategy is to reduce the herbicide dose. When the herbicide dose becomes too low to effectively control the weed, the herbicides should be followed up by a single post-emergence harrowing applied at least 7 to 14 days after herbicide spraying. In this way, herbicide use can be reduced from a TFI about 0.3 to a TFI below 0.1 at a modest extra cost (0.5 hkg or 40 DKK per ha). A weed harrowing 14 days after spraying occurs on average in mid-May or in some cases in late May or early June. However, it should be borne in mind that harrowing performed more than 35 days after sowing or later than 20 May have a significant, negative impact on Skylark productivity (section 4.2.1). Also seed density and barley variety should be considered in the analysis. Using a higher seed density or other barley varieties could be costly, but a weed competitive variety and a more vigorous and dense seed might reduce the weed control costs due to better weed competition and less sensibility to harrowing (see Rasmussen & Rasmussen 2000). In case the spring barley has to be grown without herbicides, these parameters are of great importance. Also some of the variations in the local (ex post) parameter estimates might not be stochastic. In that case, it will be possible to design more intelligent strategies and decision support systems, involving seed varieties and seed density etc. and utilising the farmers’ expectations and knowledge on yield potential, weed density, weed species etc. The economic and environmental potential for such more intelligent strategies and decision support systems has not been analysed. The established complex of weeding models has proved useful in the process of connecting and accessing the different weeding aspects, trials and strategies. In Rasmussen & Nørremark (2007) the need for more detailed and specific field trials and a modelling framework is recommended: “It is, however important that the research in sub-items related to physical weed control are connected in modelling frameworks that secure the practical aspects and perspectives are not lost in reductionism”. The actual analysis using the established complex of weeding models (see 3.3.2.2) has however shown that the variation, and especially the co-variation, in the herbicide, harrowing, and timing effects, as well as the variation in the general growth conditions, are essential parameters in that process. In order to understand the weed control mechanisms and to assess the value of pre-emergence harrowing, as is the case in 3.3.2.2, it is important to know how its effect is correlated to the weed growth conditions, the post-emergence harrowing effect, the herbicide effect, the weed density etc. These parameters must be measured simultaneously in many trials over a span of years to reveal the required correlations. A model is needed to estimate and quantify the parameters and their correlations (as in 3.3.2.2), but the correlations cannot be estimated by using a model and piecewise data from specific field data. 4.3.4 Herbicides versus weed harrowingDuval (1997) compared mechanical and chemical weed control in cereals. The advantages and disadvantages of the two technologies are summed up in Table 4.1. Duval (1997) mentions that an increased seeding rate and deeper planted cereal seed will compensate for or reduce losses during mechanical weeding. Table 4.1. Advantages and disadvantages of mechanical weed control according to Duval (1997).
Many of these advantages and disadvantages are verified, or are not included, in the actual economic analysis. The timing problem mentioned by Duval (1997) is however in open conflict with the actual analysis and with Rasmussen & Nørremark (2007). Also the reduced costs and the less effective weed control mentioned for mechanical weeding divert from the actual findings. More intensive harrowing might level these differences. Unfortunately, Duval (1997) provides no references to support the statements concerning timing and harrowing intensity. Stony fields are of course a problem in case of mechanical weeding, but the large-area less-time advantage might not be a problem in the case of Danish agriculture. The width, quality and speed of the harrowing equipment have improved alongside the improvements in capacity and width of the spraying equipment. The widest harrows are now as wide as the widest sprayers, the sprayers are susceptible to wind, and the harrows are even faster than the sprayers. 4.3.5 Weed harrowing and timingThe farm economic study of this project is based on findings from literature and has for the first time ever comprised most field trials on weed harrowing and the weed - crop relations in spring cereal available and relevant from 1999 till to day from all over Denmark. In this way the farm economic results, if any, can easily be generalized to all parts of Danish agriculture and covers different weeding strategies and growth conditions in spring barley. It is found that efficient weed harrowing in spring cereals must be carried out by using one or two intensive post emergence weed harrowings. The intensity is here measured in terms of relative crop soil cover. It is found that at least a 20% crop soil cover is needed to effectively control the weed and that the required crop soil cover can be achieved by one intensive or repeated less intensive harrowings. It is also found that the timing of the additional weed harrowings is unimportant to the weeding effect. Thus, from an economic point of view, one intensive post emergence weed harrowing is the most efficient harrowing strategy. The results from a recent Danish field trial (Rasmussen & Nørremark, 2007) have implied that also timing of the first and only, intense weed harrowing is unimportant to the weeding effect (the selectivity, but not necessarily the crop damage is constant over time). In that case, from a weeding point of view, a single intensive weed harrowing liberally performed in the period from 7 days to 21 days after sowing might be an efficient, flexible, and competitive alternative to herbicides. Traditionally weed harrowing and other alternatives to herbicides are considered to be less reliable, less flexible and less effective than using herbicides. Some times weed harrowing works, and some times it doesn’t work. When it works, it is traditionally considered, that timing of the harrowings has been perfect, and if it doesn’t work, it is considered, that the timing was wrong. It is however found in this project that the weeding effect of harrowing varies from trial to trial (as for herbicides), but that the weeding effect seems to be independent of timing. And it is found, that the variation in the weeding effect is all most the same or maybe even lower for weed harrowing than for herbicides. The net yield gain is however lower and less stable for weed harrowing than for using herbicides. Traditionally the farmers using mechanical weeding in spring cereals are advised to perform a pre emergence harrowing and at least one, early post emergence harrowing. And traditionally it is considered, that it is difficult for the farmer to find time to perform theses two operations within a short span of time and at the same time (by chance) get a perfect timing. Because nobody knows if the timing was right in the first place, the farmer might be encouraged to perform more harrowings to hit the right timing. The more harrowings, the better is the chances. Following this advises and logic, the farmer will however end up with a costly mechanical weed control, not at all competitive to herbicides, and with maximum damage to the arthropods and the Skylarks. Contrary to the traditional understanding of weed harrowing, it is found in this project, that the pre emergence harrowing is inefficient, that timing of the remaining post emergence harrowing from a weed control point is unimportant, and that a single intensive weed harrowing may be as effective and reliable as using herbicides to control the weed. Consequently weed harrowing, using the right strategies, can be a more efficient, flexible and reliable alternative to herbicides than normally considered, but from an economic point of view a low herbicide dose is still the most efficient and reliable strategy. 4.3.6 Weed harrowing, herbicides and biodiversityThe main purpose of the mechanical and chemical weeding in spring cereals is to control the weed. Although the most efficient weed harrowing strategies (a single or a few intensive post emergence harrowings) are found to be efficient and as effective, flexible and reliable to control the weed as herbicides, they are not as efficient as the most efficient herbicide strategies. It also appears that a single, early and intensive post emergence weed harrowing may result in a higher flowering ratio in the surviving weed and may also result in fewer ground-dwelling arthropod predators at a time when they are needed by the farmer to control aphids or by the Skylarks to feed their young. Even if it turns out that timing and weed flowering is not a problem from a weed control point of view, timing may still be an important issue from a crop damage, and therefore also from an economic, point of view. One intensive weed harrowing performed less than 35 days after sowing and no later than 20th of May is an effective, flexible and reliable mean to control the weed, doing little harm to the Skylarks. Probably due to timing, intensity and frequency related crop damages the net costs from weed harrowing are more varying and in general a little higher (0.7 hkg per ha or 11 € per ha) than using herbicides. In order to improve the efficiency of weed harrowing and if possible to do better than the herbicides, a more reliable and less crop damaging strategy with respect to timing, intensity and frequency has to be found. It has been implied that weed harrowing one to two weeks after sowing is less harmful to the crop. It is hard to say whether or not such a strategy by nature always will be in conflict with the arthropods being valuable for Skylarks, Lapwings and pest control. In the worst case, from a flora and bird perspective, Skylark friendly weed harrowing is as harmful as using herbicides and from a Lapwing perspective the Skylark-friendly weed harrowing will almost eliminate the production of young for the whole season. The control of root weed and pests has not been analysed in this project. The arthropods and their timing, is indeed related to the pest problem, and from organic farming we know, that also root weed can be controlled by using mechanical weeding. But without doubt, herbicides like glyphosat and MCPA and insecticides are the fare more efficient and reliable means to control root weed and pests also in spring cereals. However, insecticide use is clearly more damaging to birds’ food resources than any pest control measure used in pesticide-free farming systems. Consequently an efficient and Skylark-friendly weed harrowing in spring barley will save the herbicides used to control the dicot and seed-grass weed in spring cereals for a net cost of 2.3€ per ha. Such a strategy might be almost harmless to the Skylarks and effectively control the dicot and seed-grass weed. But in most cases, such an efficient weed harrowing will be as harmful to the weed flora and weed seed production as using low dosage herbicides and even more harmful to the early ground-dwelling arthropods. It has also been found, that the farmer by using a lower, less efficient herbicide dose or a more extensive weed harrowing than the optimal, for a net cost of 3€ per ha, can “produce” more weed and enough weed to obtain the optimal conditions for the arthropods. Performed the right way, both the use of herbicides and the weed harrowing can be more Lapwing, Skylark, arthropods and weed friendly for a low additional cost. However the long-run costs of such more weed and arthropods friendly strategies has not been analysed. Thus from an economic point of view (and thus in conventional practice) low doses of herbicides turns out to be a more convenient and efficient mean to control the weed than weed harrowing. And the efficient use of low dosage herbicides is friendlier to Lapwing, Skylark and ground-dwelling beneficial arthropods than a frequent use of weed harrowing. However, in a broader perspective the use of herbicides may also be related to a reduction in the frequency of highly weed dependant insects like butterflies, flowing weed etc., and related to an increased risk of herbicide drift, human health and reproduction problems, and point pollutions etc. In that broader perspective weed harrowing, may still be a reliable, efficient and environmental better alternative to the herbicides. 4.4 General discussionThe present project has followed up upon two areas of debate. One of these is an indication from a preceding project, which shed light on the phase of conversion from conventional farming to organic farming, one of the changes being replacing chemical weed control with mechanical weed control (Navntoft et al. 2003). The indication of interest was that mechanical weeding might have adverse effects on beneficial epigaeical arthropods (species which prey upon pest species) that also serve as prey for birds. The other area of debate was the fate of bird nests during the operations of mechanical weed control. Both for the direct effects on invertebrate fauna and bird nests and for the indirect effects (like, e.g., loss of birds’ food items) more damage was anticipated at an increasing number of weed harrowings. By comparing two, three and four times weed harrowing, and in addition including other weed control options in modelling, the project results show effects on wild plants, insects and several birds as well as on crop and on yield. In contrast to earlier results of reducing herbicide dosages by 50% and 75%, which created significant increases in the number of weed species, there was no significant difference between the numbers of weed species found after 2, 3 and 4 times harrowing, respectively. The weed occurrence, however, showed a fairly clear effect of an increased number of harrowings (Figure 3.1), and the weed biomass decreased up to 48% after four weed harrowings and 43% after three (Table 3.1). An interesting ecological question is whether weed harrowing has both direct and indirect effects on epigaeical fauna. For Linyphiidae spiders, Thorbek & Bilde (2004) found an immediate decrease due to harrowing. The present project confirms that result for spiders and for selected epigaeic beetles. The project also shows a very interesting and firm, positive correlation between vegetation biomass and the density of the investigated polyphagous predators (Figure 3.9). This correlation proved to be strongest for weed biomass. The general indicating is that vegetation plays a major role for arthropods on the soil surface. Therefore one effect of harrowing is negative effects on the beneficial arthropod complex, both directly and indirectly through interruption of the positive influence of vegetation. Therefore harrowing can already on this background be seen as a field operation with potential damage on organisms across the food web: plants directly, epigaeic arthropods directly + indirectly and birds indirectly through food depletion (+ directly through nest mortality). The overall result of comparing different numbers of harrowing is that more than two times harrowing will reduce selected epigaical predators by 20 - 35 % at least in the short term. The food aspect is in general of importance to the bird potential. Previous work has proved that the epigaeic arthropods involved here are important as food for Skylark and several other birds (Elmegaard et al. 1999, Moreby & Stoate 2001, Navntoft et al. 2003), and a retrospective study in UK has shown a correlation between the decline of insects in agricultural land and the decreasing abundance of birds (Benton et al. 2002). The possible influence of reduced food amounts on Skylark populations was not studied in the present project, but the direct effects of harrowing on breeding success came out very clearly. The fate of Skylark nests was a question raised from the outset but at several of the involved farms it became obvious that Lapwing nests should also be included. For the Skylark more than two times harrowing proved very damaging to nests and thus to the production of offspring. Two times harrowing allowed a nest success more than two times higher (65%) than four times harrowing (only 28% success), and in terms of the number of fledglings per ha the similar proportion was almost 3:1 (1.14/ha versus 0.40/ha). This reduction of offspring numbers includes crushed and buried eggs etc. as well as increased predation. Such a massive effect of course raises the question whether weed harrowing is at all compatible with successful Skylark breeding. The project results, however, include the answer, which is proper timing. Pre-emergence harrowing can in this respect be neglected because Skylark nests are not established until after the emergence of the crop, and if harrowing is finished no later than 35-37 days after sowing and never later than 20th May the reduction of Skylark reproduction will be negligible. For the Lapwing the timing is even more crucial. This species is highly vulnerable to weed harrowing. More than half of all nests examined were completely unsuccessful, mainly due to farming operations, and clutch size was reduced by harrowing in many of the remaining nests. Overall, the number of hatchlings per nest was reduced by 73% in fields subject to post-emergence weed harrowing compared to similar cereal fields where post-emergence harrowings were not performed. In relation to farming practice a further problem is that already the first weed harrowings can cause a massive mortality if not carried out very early (no later than a few days after crop emergence). The project results indicate that intensive agriculture on Lapwing rich areas has to be carefully considered in the light of the present vulnerability of this bird species. This is especially important because breeding success in the other main habitat of the species, meadows, is also poor. Viewed from an agricultural perspective, weed harrowing must have some damaging effect on weeds – otherwise this control operation is meaningless. However, the interesting question is whether one or two times harrowing is sufficient for control and whether there are other positive or negative effects on yield. Within the latter category is the disturbance of polyphagous predators. If this live buffer against, e.g., aphids in cereals is removed it can cause a loss of 3 hkg/ha in spring barley attacked by aphids (Östman et al. 2003). In relation to the more central question of general efficiency, a re-analysis of a long series of farm placed experiments (Landsforsøgene) is important. This shows that in two thirds of all cases two times weed harrowing is sufficient in terms of control efficiency, and the harrowing is competitive with chemical weed control. Other results further point at the intensity, measured as soil coverage of the crop after the mechanical operation, as being very important. Thus 20% crop soil cover is the most appropriate, and at this harrowing intensity level one operation may be sufficient and two will be as good as and often better than a chemical treatment, if measured by yield at harvest time. In terms of net economy the picture is, however, somewhat different. Then the best option (found by re-analysis of data on weed treatment from the Danish Agricultural Advisory Service and the use of modelled scenarios) is one herbicide treatment plus one weed harrowing. Further analysis of data from organic fields with spring barley shows that one, not too late, post-emergence harrowing has on average a cost of 0.5 - 1 hkg/ha more than one herbicide treatment. Interestingly, the more and more used pre-emergence harrowing seems to be an option with a highly variable outcome, although it sometimes may be useful in combination with a later treatment. Thus it appears that already at present the farmers make use of weed control strategies, which are in terms of net yield more “risky” than the most efficient mechanical weed control strategies implying a net loss of 0.5 - 1 hkg/ha compared with a chemical strategy. Overall, the results point at the fact that several weed control strategies are possible, depending on farm practice and incentives. For instance, it will be rather easy to ensure successful breeding of Skylark under the use of mechanical weed control in spring cereals (cf. above). However, it has to be underlined that the highly cost-efficient use of a combination of (low-dose) herbicides and mechanical weed control – which possibly includes insecticide use – is probably far from optimal to Skylarks. This is partly because harrowing carried out at least 7-14 days after herbicide application (as recommended in section 4.3.2) will invariably destroy a significant number of nests, partly because pesticide use reduces the number of offspring produced (Odderskær et al. 1997). Conservation of the Lapwing in Danish farmland represents the greatest challenge, however, because of its unfavourable conservation status and vulnerability to harrowings performed later than a few days after crop emergence. Only a very skilled manager can carry out the harrowing(s) in such a way that sufficient weed control is attained and Lapwing breeding success is ensured. Overall, the demands of skills and managerial incentive is also a cloudy area of potential, combining respect to biodiversity with the request to income. Footnotes[1] In some cases, but not systematically, density for a few weed species after the last treatment or total weed coverage after harvest might be registered.
|
||||||||||||||||||||||||||