Introduction

Wheat is an important cereal crop besides rice in Bangladesh and plays an important role in attaining food and nutritional security (Hossain et al. 2019). During 2018–19, 1.15 million tons of wheat were produced from 0.33 million ha that could meet only 20% of the national requirement (Barma et al. 2019). On the other hand, the demand for wheat has been increasing at the rate of 13% per annum due to rapid changes in dietary habit, increase in socio-economic status and per capita income, rapid growth of fast food restaurant, and the growth of branded bakery and biscuit industries, etc. (Barma et al. 2019). Due to the decrease in wheat area by 15% in 2018–19 compared to the previous year, wheat production in Bangladesh also reduced by about 12%. Though there is decrease in wheat area, there has been a significant increase in its average yield per ha (national average: 3.49 t/ha) in Bangladesh due to the use of high yielding, disease-resistant and stress-tolerant varieties (Barma et al. 2019). Wheat production however is also constrained by several management factors of which severe weed infestation and lack of appropriate weed control measures is considered to be the most important one limiting wheat yield (Khaliq et al. 2013b). Fahad et al. (2015) and Jabran et al. (2017) reported that more than 20% yield loss in wheat in Bangladesh is occurred due to weed infestation. The degree of wheat yield losses due to weeds depends on many factors such as the availability of farmers’ resources, weed species and density, time of emergence of crop and weeds, crop growth stage, and the duration of weed interference to wheat crop (Estorninos et al. 2005; Hussain et al. 2015).

Farmers in Bangladesh are usually reluctant to control weeds in their wheat field and those who control mostly weed manually. However, manual weeding these day has been impractical due to increasing labour shortage as well as labour wages. The chemical weed control using herbicide is the easiest, cheapest, reliable and a timely measure. However, lack of knowledge in selection of appropriate herbicide, farmers’ inadequate skill in its use, and its excessive use have adverse effects on anaimal and human health; such inappropriate practices can harm the environment and develop resistance biotypes supremacy of minor weeds due to their resistance against herbicide (Baghestani et al. 2007; Khaliq et al. 2013a; Chauhan et al. 2015; Ahmed et al. 2019). Therefore, to reduce the dependency on herbicides, alternative, low costing and environmentally friendly weed control options need to be explored as the sustainable weed management approach (Khaliq et al. 2013c; Ahmed et al. 2020).

Mechanical weed management options are found to be more economical than manual weeding and more sustainable than chemical methods with herbicides (Subudhi 2004; Cloutier et al. 2007; Jabran et al. 2012; Gongotchame et al. 2014; Narwariya et al. 2016). The tractor/power tiller driven rotary type power weeder (power weeding) are used to control weeds for low-land and upland row crops (other than wheat) in developed countries (Olaoye and Adekanye 2011; Hossain et al. 2011) while power tiller driven shovel type power weeder are used for upland row crops (rice, wheat, maize, etc.) in many developing countries including Bangladesh (Matin et al. 2010; Hoque et al. 2010). Considering the pertinent issues of labor shortage and labour wages for manual weeding and development of herbicide resistance and environmental hazards associated with chemical weeding, Bangladesh Agricultural Research Institute (BARI) developed a two-wheeled power-tiller multi-row weeder which can weed 6 rows at a time in wheat and upland row crops and reduce the weeding cost by reducing the labor requirement (Hossen et al. 2019). However, the performance of multi-row weeder needs to be evaluated for its recommendation for the farmers’ fields. In this context, an experiment was undertaken to investigate the effect of different weed control methods in wheat production and find out the most economically viable and environmentally friendly weed control method for sustainable wheat production in Bangladesh. We hypothesized that the multi-row power weeder would be more economically viable and more sustainable than other weed control methods in terms of reducing weeding cost and increasing wheat yield.

Material and Methods

Experimental Duration and Location

The experiment was conducted at the Regional Wheat Research Centre (RWRC; 23° 59′ N, 90° 24′ E; 13.1 m MSL), Bangladesh Agricultural Research Institute, Gazipur-Joydebpur in agro-ecological zone 28 (AEZ–28) (Modhupur Tract) during November to March in consecutive two years (2012–13 and 2013–14).

Soil Characteristics

Soil in the AEZ-28 is weakly acidic, organic matter content is low like others AEZs while total nitrogen (N) and phosphorus (P) are very low. Potassium (K), sulfur (S), boron (B) and zinc (Zn) contents are below critical levels. Overall, with the exception of P, all nutrients were deficient (Jahan et al. 2018; Hossain et al. 2018).

Meteorological Information

The climate of the area is subtropical with highly variable rainfall, which during the months of November to May ranges from 4050 mm. The temperature during this period also varies greatly, with maximum daily temperature ranging from 20–35oC and minimum temperature from 10–15oC. Weekly maximum (MaxT) and minimum temperature (MinT), sunshine (hours) and total rainfall (mm) during the growing seasons in two years for the experimental location is presented in Fig. 1.

Fig. 1
figure 1

Meteorological information, particularly maximum (MaxT) and minimum temperature (MinT), sunshine (hours) and total rainfall (mm) during both the growing seasons in the experimental location

Full size image

Treatments and Design

The trial considered five weed control methods such as i) control plot (weedy check), ii) one hand weeding (1 HW) at 25 days after sowing (DAS), iii) mechanical weeding by using a BARI dry-land weeder (BARI weeder) at 25 DAS, iv) mechanical weeding by using power tiller operated weeder (PTOW) at 25 DAS, and v) chemical weed control. Treatment details are presented in Table 1. All treatments were arranged in a randomized block design and repeated three times. Each treatment plot size was 5 m × 2.4 m (12 m2).

Table 1 Treatment details of the trial
Full size table

Wheat Variety and Characteristics

The seeds of wheat variety ‘BARI Gom 26’ were collected from Bangladesh Wheat and Maize Research Institute (BWMRI), Dinajpur, Bangladesh. It is a high-yielding popular variety released in 2010. The yield potential of the variety ranges from 3.5 to 5 t ha−1 (Barma et al. 2019). It can tolerate terminal heat stress caused by late sowing andis resistant to Bipolaris leaf blight, leaf rust and stem rust race (Ug 99).

Land Preparation and Application of Fertilizers

The experimental field was deep ploughed and cross-ploughed with a two-wheel power tiller to obtain good tilth required for higher crop yield. All weeds and stubble were removed from the experimental field before the seed sowing. The soil was treated with Furadan (Carbofuran) 5G at the rate of 8 kg ha−1 (marketed by FMC International S.A. Bangladesh Ltd.) to protect the young plants against insect attack. Fertilizer was applied as recommended by BWMRI: 110-27-40-20–1 kg ha−1 of N, P, K, S, and B, respectively, through urea, triple superphosphate, muriate of potash, gypsum and boric acid respectively. Two-thirds of N and full amounts of other fertilizers were applied as basal doses during final land preparation. The remaining (1/3) N was applied immediately after the first irrigation during the crown root initiation stage (17–21 DAS).

Seed Treatment and Seed Sowing

Before sowing, seeds were treated with Provax-200 WP (marketed by Hossain Enterprise CC Bangladesh Ltd., an agrochemical company, in association with Chemtura Corp., USA), which is a carboxin and thiram containing fungicide to control fungi in the soil at the seedling stage (Hossain et al. 2013). In 2012, seeds were sown on 26 November and in 2013 on 28 November. Seeds were sown manually at the rate of 120 kg seeds ha−1 with a spacing of 20 cm and depth of 4–5 cm by making specific narrow furrows with an iron rod. After sowing, seeds were covered with soil and slightly pressed by hand. Special care was taken to protect seeds from birds.

Irrigation and Weed Management

The first irrigation was performed at 17–21 DAS, the second at 53–56 DAS during the panicle initiation stage, and the last one at 78–81 DAS at the grain-filling stage. A proper drainage system was developed to drain off excess water. A total of 0.0251 m or 2.51 cm of irrigation water was applied in each season. Weeding was performed as per treatments.

Harvesting and Post-harvest Operation

In both years, crop was harvested at full maturity in the last week of March, when leaves and stems became yellowish in color. The central position of each plot area was harvested for grain and biomass yield and data were converted to t ha−1. The harvested plants were tied into bundles, transferred to a threshing floor, and sun-dried by spreading out evenly on the threshing floor. Seeds were separated from the chaff by a mechanical thresher and then cleaned, dried and weighed.

Data Collection and Their Procedure

Weed Data

Weed density and biomass were recorded at 20 days after the imposition of treatments. Weed samples were collected from the randomly placed 0.5 m × 0.5 m quadrate from three spots of each plot. After collection, the total weed number was counted, oven-dried at 70 °C for 72 h, and dry weight recorded. Weed control efficiency (WCE %) and weed control index (WCI %) were estimated (Kumar et al. 2015) from total number of weeds and dry weed biomass as per the following equation:

$$\mathrm{WCE}(\mathrm{\% })=\frac{\left(\begin{aligned} & \text{Weed drymatter in weedy plot}- \\ & \text{Weed drymatter in treated plot}\end{aligned} \right)}{\text{Weed drymatter in weedy plot}}\times 100$$
(1)

Crop Data

For recording yield contributing characters of wheat, plant height (cm), spike length (cm), spikes m−2 (no.), grains spike−1, and 1000-grain weight (TGW, g) data were collected from the ten randomly selected plants from each plot. Grain and biomass yield were recorded from an area of 8 m2 (4 m × 2 m) from each plot and grains oven dried. One thousand grains were counted from the yield area of each plot and the grain weight (g) recorded with an electrical balance. Grain yield were converted to 12% moisture content (Hellevang 1995):

$$\mathrm{Y}\left(M_{2}\right)=\frac{100-M_{1}}{100-M_{2}}\times 10^{-6}\times \mathrm{Y}\left(M_{1}\right)$$
(2)

where, Y (M2) = grain weight with 12% moisture, Y (M1) = grain weight with actual moisture %, M1 = actual moisture %, and M2 = expected moisture %.

Grain yield (GY) and straw yield (SY) together were regarded as the biological yield of wheat. The biological yield was calculated with the following formula:

$$\text{Biological yield}=\mathrm{GY}+\mathrm{SY}$$
(3)

Harvest index (HI) denotes the ratio of GY to biological yield and was calculated with the following formula (Gardner et al. 1985):

$$\mathrm{HI}\left(\mathrm{\% }\right)=\frac{\mathrm{GY}}{\,\text{Biological yield}}\times 100$$
(4)

Economic Analysis

Economic analysis was performed to determine the efficiency of different weed control methods. For this, only the variable weeding cost was was considered. Other management costs remained same for all treatments and hence were not included in the analysis. The amount of commercial products of herbicides required for one hectare was calculated and the cost was estimated based on their market price. The number of labours for herbicide spraying and mechanical weeding were counted and labor wage was based on an 8‑hour work a day. The market price of wheat was determined during the years of the experiment and used for calculating the gross return. Benefit-cost ratio (BCR) and marginal benefit-cost ratio (MBCR) were calculated with the following formula:

$$\mathrm{BCR}=\frac{\text{Gross return}}{\text{Total cost}}$$
(5)
$$\text{MBCR}=\frac{\text{Gross Return}_{(\text{Specific Management})}-\text{Gross Return}_{\left(\text{Control}\right)}}{\text{Variable Cost}_{(\text{Specific Management})}-\text{Variable Cost}_{\left(\text{Control}\right)}}$$
(6)

Statistical Analysis

Data for both years were analyzed separately using a R package (Core Team R 2013). Since there were significant differencse between years, data were presented year-wise separately. Treatment means were separated using the least significant difference (LSD) at the 5% level of significance. Weed density and biomass data were subjected to square root transformation (√(x plus 0.5) before analyses); but since the relationships did not improve much original data were used for final analysis.

Results and Discussion

Effect of Weed Control Methods on Weed Incidence and Weed Control Efficiency

Major weed species found in the experimental plots were Chapra/Goosegrass (Eleusine indica Gaertn.), Shama/Barnyardgrass (Echinochloa crus-galli (L.) P. Beauv.), Subuj shiyal-leza/Green foxtail (Setaria viridis (L.) P. Beauv.), Mutha/Purple nutsedge (Cyperus rotundus L.), Bothua/Lambs quarter (Chenopodium album L.), Bon begun/Black night-shade (Solanum nigrum L.), Bishkatali/water pepper (Polygonum hydropiper L.), Bon mushur/Wild lentil (Vicia sativa L.), Shakenotae/Green amaranth (Amaranthus viridis L.), and Karpet agacha/Green carpetweed (Mullugo verticillata L.). There was higher weed infestation in 2012–13 (first year) than in 2013–14 (second year). Though the most dominant weed species in the second year was Cyperus rotundus it was comparatively less in the first year. Weed density (number m−2) at 45 DAS (after 20 days of treatment imposition) was significantly influenced by the weed control method in the first year, but not in the second year (Table S1 and Figs. 2 and 3).

Fig. 2
figure 2

Weeds m−2 of wheat as affected by weed control methods. SD ± in each treatment were calculated from three replications

Full size image
Fig. 3
figure 3

Weed dry biomass m−2 of wheat as affected by weed control methods. SD ± in each treatment were calculated from three replications

Full size image

Cyperus rotundus is a persistent, prolific, and the worst weed species in the world. This weed is very difficult to control by manual or mechanical weeding or even by application of a pre-emergence herbicide (Ahmed and Chauhan 2014) because it can spread and survive even following the destruction of its aerial parts (Horowitz 1972). This weed can propagate through both tubers and rhizomes. The rhizome and tuber function both as storage and reproductive organs. When rhizomes elongate, they form tubers and ensure the translocation of nutrients and assimilates between above ground and underground parts (Horowitz 1972). In 2012–13, the highest weed density was recorded in weedy check (1209 weeds m−2); all weed control treatments had significantly lower weed density than the weedy check. Among the weed control methods, one hand weeding (HW) had the lowest weed density (51% less than the weedy check) followed by the herbicide treatment (31% less). Weed density may not be the right parameter to measure the weed control efficiency (WCE) because many times a weed control treatment helps retardation of the weed growth but weeds do not fully die or eradicate (Ahmed and Chauhan 2014; Chauhan et al. 2015). In crop-weed competition when weed growth is retarded due to the application of any control measure, the crop may get a growth advantage.

Weed biomass was significantly influenced by the weed control method in both years. In both years, compared to the weedy check treatment, different weed control methods reduced weed biomass by 35–75% (Table 1S and Fig. 3). In both years, the lowest weed biomass was found with one HW which was 61–75% lower than the weedy check. Among the weed control methods, both BARI weeder and PTOW plots had similar weed biomass and WCE (Fig. 4), but the latter was significantly lower than with one HW or with herbicide. In terms of WCE, manual weeding was the best option but currently it is not an economically viable option (Ahmed and Chauhan 2014, 2015). Manual weeding is the non-chemical and ecologically sound weed control method that provides the best clean and thorough weeding but is only good for resource-poor farmers where labour is available at low wages. In Bangladesh, a significant number of farmers still rely on manual weeding due to lack of farmer’s knowledge and unavailability of herbicide information, lack of farmer’s skill on herbicide application and lack of suitable mechanical weeders, and hence farmers spend a lot of money on manual weeding. In the current study, both types of mechanical weeders performed similarly in terms of WCE but compared with the weedy check they reduced weed biomass by only 35–44%, indicating that significant numbers of weeds were not controlled by the mechanical weeders. No doubt mechanical weeding needs less time and causes less drudgery than the manual weeding, it has some limitations also. To control weeds mechanically using a mechanical weeder, plants must be in straight rows and soil should be moist before weeding. Often, it becomes difficult to remove weeds within crop rows, the cut weeds above root system can re-establish, and the improper use of weeders can damage the crops.

Fig. 4
figure 4

WCE (%) and WCI (%) in different weed control methods of wheat

Full size image

Yield and Yield Contributing Characters of Wheat

Among the yield contributing characters, only the number of spikes m−2 in 2012–13 (first year) was significantly affected by the weed control method; other parameters were not influenced significantly in any year (Table S2 and Figs. 5, 6 and 7). The highest number of spikes m−2 was obtained from the one HW plot and was statistically different from all other treatments; it could be due to the effect of higher WCE. The lowest spikes m−2 was observed in the weedy check due to higher weed infestation and suppressing the growth and development of the growing plants.

Fig. 5
figure 5

Plant height (cm) and spike length (cm) of wheat as influenced by different weed control methods

Full size image
Fig. 6
figure 6

Spikes m−2 and grains spike−1 of wheat as influenced by different weed control methods

Full size image
Fig. 7
figure 7

1000-grain weight and harvest index of wheat as influenced by different weed control methods

Full size image

Grain yield in both years was influenced significantly by the weed control method with almost similar yield trends (Table 2). The highest grain yield (4.07–4.35 t ha−1) was recorded from one HW which was 25–35% higher than the weedy check and 23–25, 11–25, and 9–20%, higher respectively than the PTOW, herbicide, and BARI dry land weeder. The higher grain yield with one HW was due to significantly higher number of spikes m−2 and slightly higher grains spike−1 (Fig. 6). Compared with weedy check, the BARI dry land weeder had only 15% higher yield indicating that in controlling weeds it did not perform well which was also evident from WCE. In the current study, there was no complete weed-free treatment; however, the highest yield obtaining HW treatment had similar grain yield to the weed-free yield of some previous studies using the same variety (Ahmed et al. 2019, 2020). Grain yield data indicate that one HW around 25 DAS in wheat is enough to obtain a similar yield to weed-free condition yield. In Bangladesh farmers mostly practice one HW and they usually perform that late (30–50 DAS). Grain yield with herbicide data in the current study indicate that if farmers apply only one post-emergence herbicide without any additional hand weeding, they will lose some yield. In situations where farmers would like to control weeds in wheat fields effectively by using only herbicide they will need to apply both pre-and post-emergence herbicides.

Table 2 Wheat straw biomass and grain yield as influenced by different weed control methods
Full size table

Economic Analysis

The highest weeding cost (107 US$ ha−1) was required for the one HW treatment while lowest (23 US$ ha−1) for the herbicide treatment (Table 3). The mechanical method, PTOW and BARI weeder had similar weeding cost and slightly higher than the herbicide treatment. The fixed cost was similar for all the weed control treatments, therefore, due to higher weeding cost (only variable cost) for the one HW treatment, the total production cost was also higher (Table 3). Although the highest gross return was recorded for one HW due to higher production cost, the highest BCR (1.51) and MBCR (10.44) were recorded for PTOW. The total production cost for the herbicide treatment was slightly lower than the PTOW treatment but due to lower grain yield in the herbicide treatment, the BCR and MBCR were also lower than for the PTOW treatment. One HW treatment had similar BCR to, but lower MBCR, than the herbicide treatment. Many previous studies have reported that the application of herbicide would be the best option to reduce weed control cost and increase farmers’ net return in wheat production (Marwat et al. 2006; Safdar et al. 2011; Shehzad et al. 2012). However, the major concerns in using herbicides are environmental pollution and human and animal health hazards, which would discourage the widespread use of herbicides in wheat production in Bangladesh. Mechanical weeding would probably be the best option for farmers as it reduces farmers’ weed management cost without environmental and human health hazards but it has limitations too. Integrated weed management approaches with use of mechanical weeder and optimum use of herbicides would be required for best would control, reduce weeding cost and increase farmer’s yield and profit by resource-rich farmers with labor constraints but manual weeding would still remain an option for the resource-poor farmers with abundant family labour.

Table 3 Economic analysis (hiring basis) as influenced by different weed control methods
Full size table

Conclusion

The results of the present study revealed that one hand weeding at 25 DAS was the best in terms of lower weed density (m−2), lower weed biomass (m−2), and higher weed control efficiency (WCE%). Consequently, one hand weeding at 25 DAS also resulted in highest grain yield (increased 28% as compared with control) of wheat followed by the PTOW, herbicide, and BARI weeder treatments, while the lowest grain yield was obtained from the weedy check treatment. The highest BCR and MBCR (1.51 and 10.44 respectively) were found with PTOW followed by herbicide, hand weeding, and BARI weeder. Considering the negative effect of herbicides on the environment and human and animal health hazards and also labour crisis at the peak period of weed control in wheat, farmers can use power tiller operated multi rows weeder to reduce the weeding cost as well as to increase yield with the higher net return. However, manual weeding would still remain an option for the resource-poor farmers with abundant family labour in Bangladesh.