Agricultural production worldwide has been intensified in recent decades. It is characterized by high productivity, high inputs (i.e. pesticides, fertilizers and water), increased mechanization (Le Féon et al., 2010) and more simplified cropping sequences (Stoate et al., 2001) where the crop choice is often market driven (Vasileiadis et al., 2011). However, the long-term sustainability and environmental effects of the intensification of agricultural systems have raised public concerns in terms of negative consequences (Matson et al., 1997).

Intensive tillage operations (i.e. with soil inversion) have been found to adversely affect soil structure and cause excessive breakdown of aggregates, leading to soil erosion in higher rainfall areas (Chauhan et al., 2012), but also to have a negative impact on environmental quality by accelerating soil carbon loss and greenhouse gas emissions (Reicosky & Allmaras, 2003). Such concerns provided an incentive to investigate tillage systems that minimize negative impacts on the environment without reducing crop yields. These systems can be broadly termed conservation tillage (CT) systems and involve the reduction or suppression of primary tillage operations such as plowing, disking or chiseling (Chauhan et al., 2012). CT is also stated as a tool in the general principles of integrated pest management (IPM) as documented in the Directive 2009/128/EC (EC, 2009). However, what are the implications switching from conventional to reduced or no tillage systems in terms of weed seedbank dynamics? Inversion tillage through moldboard plowing reduces weed populations by burying a large proportion of freshly shed seeds deep in the soil profile (Vasileiadis et al., 2007). Conversely, chisel plowing retains them nearer the surface facilitating germination and emergence, thereby increasing the requirement of management inputs to control them (Ball, 1992; Vasileiadis et al., 2007). Other studies indicate that weeds increase after several years of reduced tillage because of increased seed accumulation at the soil surface, the lack of disruption of underground vegetative reproductive organs (Cardina et al., 1991; Bàrberi et al., 1998) and changes in timing of weed emergence compared with inversion tillage (Bullied et al., 2003). Indeed, the timing of weed seedling emergence in the field varies depending on, individual species germination requirements, seed burial depth, soil disturbance (Roberts & Feast, 1972) and environmental conditions (Forcella, 1992). Seeds in CT systems remain near the soil surface, rather than being deeply buried, thus are more likely to emerge and be depleted from the seedbank due to exposure to favorable conditions for germination (Morris et al., 2010; Chauhan et al., 2012).

These prerequisites call for investigation on the timing of seedling emergence for key weed species when implementing CT systems, but also for the construction-calibration of decision support systems (DSS) for efficient weed management decisions in these systems. By developing such DSS, a decision-making process that will determine ‘‘if’’, ‘‘how’’ and ‘‘when’’ weed control is needed can be established, thus optimizing the time of weed control while reducing redundant herbicide applications or mechanical interventions (Masin et al., 2011; Vasileiadis et al., 2011). There is considerable information on emergence timing of cool temperate weed species (e.g. Roberts & Boddrell, 1983; Froud-Williams et al., 1984; Ogg & Dawson, 1984; Egley & Williams, 1991), whereas comparable information is limited for species of more Mediterranean distribution, especially under CT system conditions. Such research is thus necessary for specific geographic locations subject to different environmental conditions, soil types and CT conditions. Hence, this study aimed at investigating the seedling emergence dynamics of barnyardgrass (Echinochloa crus-galli (L.) Beauv. ECHCG) and jimsonweed (Datura stramonium L. DATST), two common annual weed species prevalent in spring-sown arable cropping systems of northern Greece, in the case of a switch from conventional tillage to reduced or no tillage conditions and from 5 and 10 cm depth scenarios, given that 1) seeds shed the previous year would be distributed in the top 5-10 cm soil layer after a transitional soil cultivation using reduced tillage implements (i.e. 10 cm working depth of these implements) and 2) seeds of most weeds would germinate and emerge from the top 3 cm of the soil layer anyway.

Total barnyardgrass seedling emergence after two years showed only a significant effect of burial depth (p<0.001) (Table 2). Emergence was significantly greater when seeds were buried at 5 cm depth, attaining 96% of emergence compared to 78% from 10 cm burial depth. Soil disturbance resulted in 89% emergence, whereas 85% emergence was determined when soil was not disturbed. The interaction between factors showed no significant difference and total emergence of this weed species ranged from 97% at 5 cm depth with soil disturbance to 75% when at 10 cm depth without disturbance. The % emergence of this weed species was stable across years from the different depths and tillage regimes (Fig. 1).

Jimsonweed total seedling emergence (sum of emergence after two years) was significantly affected by soil disturbance (p<0.01) and by the interaction of the two factors (p<0.05). The emergence percentages were lower compared to barnyardgrass, with 32% from 5 cm depth and 37% from 10 cm depth. Soil disturbance resulted in 41% emergence, whereas 28% emerged when soil was not disturbed. The significant interaction showed higher emergence from an initial 10 cm burial depth when subject to disturbance (48%) and lower from 10 cm without disturbance (26%) (Table 2). Analysis of the dynamics of emergence across years identified a significantly higher emergence from 10 cm depth with soil disturbance in Year 2 (p<0.001) compared to all other treatments (Fig. 1).

Emergence timing of barnyardgrass did not vary for the different burial depths by soil disturbance regimes, as in both years of the study emergence started in April and lasted till the end of May (Fig. 2). Jimsonweed showed a bimodal pattern, with the first emergence starting end of April till mid-May, in both years of the study irrespective of burial depth and soil disturbance regime, and the second ranging from mid-June to mid-August from 10 cm depth and from mid-July to mid-August from 5 cm depth, irrespective of soil disturbance in both cases (Fig. 2).

Previous trials under laboratory conditions in a silt-loam soil (23% sand, 52% silt, 25% clay) have shown no emergence of barnyardgrass from 10 cm depth without disturbance (Benvenuti et al., 2001), which is not in agreement with the current study. However, this disparity could be explained by differences in soil type, environmental conditions (i.e. real field conditions vs. controlled conditions), bulk density and the fact that seeds were buried within a few days of collection avoiding any physiological changes that might occur during dry storage and allowing them to overwinter and break dormancy naturally. In fact, sandy soils as in this study have greater air permeability, which is also important in limiting hypoxia and removing germination-inhibiting volatile metabolites, thus enhancing seed germination (Benvenuti, 2003).

On the other, the decrease in germination of jimsonweed with increasing burial depth and no soil disturbance as identified in this study may be linked to poor gas exchange in the environment surrounding the buried seeds (Benvenuti, 2003), the lack of a light trigger (Benvenuti & Macchia, 1998) and the general ability of this weed species to build a persistent seedbank as a consequence of its survival strategy based on seed dormancy and longevity (Reisman-Berman et al., 1991). Weaver & Warwick (1984) reported that jimsonweed seeds may remain viable for 39 years in the soil, whereas Toole & Brown (1946) similarly reported 91% of jimsonweed emergence after 38 years, thus confirming the persistence of this weed species. Presumably in the current investigation, soil disturbance re-distributed jimsonweed seeds from the initial 10 cm burial depth closer to the soil surface resulting in this significant interaction. This increased stimulation of jimsonweed germination after soil disturbance was possibly due to the loss of innate (primary) dormancy as in most summer annual species following exposure to low temperatures over winter (Murdoch & Ellis, 2000) and exposure to light prior to or during disturbance (Benvenuti & Macchia, 1998).

In terms of emergence timing, the spring emergence flush of barnyardgrass and jimsonweed, during both years, was due both to the gradually warming soil in April through May that provided the temperatures required for seed germination of these species and the positive water balance as spring rainfall was adequate in both years (Table 1). The fact that only mature seeds were used in this study may also have influenced seed germination early in the season as after-ripening of seeds is known to decrease the level of primary dormancy (Murdoch & Ellis, 2000). On the other, the later midseason emergence flush of jimsonweed, during both years, indicates that some of the seeds possibly required a longer exposure to higher temperatures in order to germinate.

These results suggest that if a farmer makes the change from conventional to reduced or no tillage, considering only the seeds shed the previous year and distributed in the top 5-10 cm soil layer, after a transitional soil cultivation using reduced tillage implements and under similar climatic and soil conditions, he/she should expect a minimum of 75% emergence of barnyardgrass and 26% emergence of jimsonweed (i.e. from the worst case scenario of 10 cm depth without disturbance) in his/her field in the first two growing seasons and subsequent emergence of those seeds that didn’t suffer fatal germination in the following years. Previous research on barnyardgrass and jimsonweed determined that fatal germination (i.e. the seed germinates, but the seedling dies before reaching the soil surface; Davis & Renner, 2007) of these weeds was approximately 25% and 17%, respectively, from a depth of 12 cm in a silt-loam textured soil (Benvenuti et al., 2001). In fact, Gardarin et al. (2010) reported that the % of fatal germination is proportional to the soil structure (i.e. less in sandy soils) and greater for monocotyledonous than dicotyledonous species. Similarly, Martinková & Honěk (2013) reported that a shallow tillage that brings barnyardgrass seeds at ≥ 10 cm depths could cause approximately 10–20% seed mortality for by fatal germination.

As CT systems do not exhume weed seeds from the deeper soil profile nearer to the soil surface where conditions are favorable for germination, the farmer could eventually deplete barnyardgrass and jimsonweed from the seedbank by controlling these weeds during the growing season and avoiding any escapes that will shed new seeds. This statement can also be valid for the latter weed species that has a very persistent seedbank, but over a longer time span. This agrees with Popay et al. (1994), who reported fewer weeds emerging in uncultivated and shallow cultivated plots compared to deep cultivated plots over the initial 7-year period, stating that seeds under the CT conditions were probably near to the surface and germinated without being replaced due to weed control. But even if weed escapes happen, in an integrated weed management context, the increase of the seedbank from these weeds could be buffered by crop rotation, the specific crop practices implemented and a higher diversity level of the system (i.e. inclusion of winter and summer crops, cover crops) (Melander et al., 2013; Vasileiadis et al., 2015). A good example comes from a field experiment adjacent to the cylinder experiment (i.e. same soil conditions but different years), where after two years of continuous cotton, cotton-sugar beet rotation and continuous tobacco, barnyardgrass seedbank and population density was significantly reduced by the crop specific management practices (i.e. pre-transplant herbicide and seedbed operations) in late established tobacco (i.e. transplanted end of May) compared to the other spring-sown crops (Vasileiadis et al., 2007, 2012).

Under similar soil and climatic conditions and reduced or no tillage system conditions, this study identified mid-May as the best time for post-emergence herbicide application against both weed species. With careful choice of crop in the rotation in terms of competitiveness (e.g. sowing time, crop density and early canopy closure) and mechanical weeding in reduced tillage systems only (e.g. inter-row hoeing in corn; Zea mays L.) the later emerged jimsonweed could suffer competition and be suppressed by the already established crop and its related interventions, not causing any yield loss and having a reduced seed production.

The results obtained are of practical relevance for farmers willing to change from conventional to reduced or no tillage systems that retain weed seeds in the top 10 cm soil horizon. This study provides knowledge on the emergence dynamics of two important and troublesome weed species from various depths and soil disturbance regimes, and the timing of their emergence that allows for more sustainable weed management decisions in CT systems, making the best use of all principles of integrated weed management and maintaining weed populations at economically acceptable levels.