Effect of herbicides on weed dynamics, soil fertility, energetics and productivity of upland rice in an acidic subtropical soil of the Eastern Himalayas, India

ABSTRACT Weed management in upland rice under subtropical climate with high rainfall is challenging. The diverse weed flora in upland rice ecosystem emerges in several flushes, necessitating sequential application of herbicides. A 4-year study conducted in Meghalaya (950 m above sea level), India indicated that sequential application of Cyhalofop butyl [2-{4-(4-cyano-2-fluorophenoxy) phenoxy} propionic acid, butyl ester (R)], a grass weed killer high efficacy low volume herbicide applied 80 g ha−1 at 25 days after sowing (DAS) and 2,4-D, a broadleaf weed killer herbicide applied 0.75 kg ha−1 at 35 DAS was effective for weed control and produced significantly higher grain yield (3572 kg ha−1) of rice with the highest weed control efficiencies than other treatments. Application of Pretilachlor followed by (fb) 2,4-D or Fenoxaprop-p-ethyl fb 2,4-D yielded significantly higher chlrophyll (chl) a, chl b and total leaf chl content compared with other herbicides at 90 DAS. The highest net energy (111,443 MJ ha−1), energy profitability and benefit:cost ratio was recorded with sequential application of Cyhalofop butyl and 2,4-D. Thus, sequential application of Cyhalofop butyl and 2,4-D could provide a sustainable weed management option in upland rice under high rainfall conditions in the Eastern Himalayas, India.


Introduction
Rice (Oryza sativa L.) is a staple food for more than half of the world's population. Approximately, 90% of the world's rice is produced and consumed in Asia (FAO-Food and Agriculture Organisation 2014). In India, rice is grown in 44 million hectares with a total production of 155.7 Tg of rice. Rice is grown under different ecologies in India of which irrigated ecology accounts over 22 million ha, i.e. 49.5 of total acreage. In the Eastern Himalayan region of India, rice is mostly cultivated under rainfed ecosystem mostly by transplanting in puddled field. In recent times, progressive shift has been witnessed towards direct seeding mainly due to shortage of labour and water (Mallikarjun et al. 2014). In India, approximately 6.1 million hectares are under upland rainfed rice, which is about 13.5% of the total area under rice (Kar et al. 2003). Direct seeding minimizes approximately 30% of the total water requirement and reduces cost involved in puddling, uprooting and transplanting of seedlings. To sustain present food self-sufficiency and meet future requirements, India has to increase its rice productivity by 3% per annum. In the hilly ecosystems of Eastern Himalayas, rice is grown in well drained upland situation with the onset of monsoon. Under such ecosystem, weed emerges in several flushes and often along with the rice and competes with the crop from the early growth stages Kamoshita et al. 2010). Having inherent competitive ability, weeds reduce rice growth and productivity. Many of the weeds mimic to rice seedlings and thus, making it difficult to control manually at the early stage. The low productivity of rice in such ecology is attributed mainly to heavy weed pressure (Rao et al. 2007;Sanusan et al. 2010). Yield loss due to weed problems in upland rice varies from 40% to 100% depending upon the ecosystem, management practices, nature and degree of infestation of weed, etc. (Oerke and Dehne 2004;). Worldwide 1800 weed species are reported to exist in rice ecologies (Rao et al. 2007). The adoption of need-based weed management practices is therefore, important in order to minimize the negative effect of the principal weeds on crop growth and productivity. Weed management is therefore, a major challenge in direct seeded rice (DSR) Chauhan 2012;Chauhan and Opena 2012). Manual weeding is expensive, backbreaking and often constrained by scarcity of labour coupled with extreme weather conditions (Kumar and Ladha 2011). Weed control through herbicide is gaining popularity for its cost effective timely weed control and less labour needs as well as flexibility involved in herbicide applications. However, single application of either pre-or post-emergence herbicides often fail to provide satisfactory weed control due to their narrow spectrum of weed control and variation in weed emergence time (Gibson et al. 2002;Yadava et al. 2009). Early emergence of weeds along with crop seedlings due to favourable soil conditions and their rapid growth result in severe competition for nutrients, space, light, etc. Evaluation of physiological attributes at different stages of a crop under varied weed management measures gives an idea on the initial crop establishment as well as their effects on crop yields. Moreover, leaf pigment content provides valuable information on the physiological status of plants (Ustin et al. 2009).
Variation in leaf pigments (chlorophyll and carotene) may be influenced by genetic makeup, stress levels, environmental conditions and management practices. Leaf area and the associated leaf indices are also important growth parameters influencing photosynthetic efficiency of crop and eventually crop yield under different weed ontrol measures (Channappagoudar et al. 2013a). Appropriate leaf structure and its thickness (LT) are pivotal for leaf and plant functioning and associated to species' strategies for adequate resource acquisition and transport (Vile et al. 2005). LT is often used as a tool to screen species or cultivars for their productivity (White and Montes 2005) and the ratio of leaf mass to surface area may act as surrogate for LT (Wright and Westoby 2002). But other validated measures of leaf traits like specific leaf area (ratio of leaf area to leaf dry biomass) and leaf dry matter content (ratio of leaf dry mass to saturated fresh mass or leaf water content) can also be used as a good estimate of leaf volume (Garnier et al. 1999;Vile et al. 2005) and contribution to photosynthesis and biomass production under stress condition. Leaf area has been found to be positively correlated with total chlorophyll, biological yield and grain yield.
Inefficient weed control in DSR usually leads to shift in weed flora towards difficult to control weeds (Singh et al. 2013). Repeated use of herbicide with similar mode of action also results in weed shift towards more persistent perennials and builds up herbicide residues in soils and consumable products (Kathiresan 2001). Butachlor, Pretilachlor, 2,4-D, etc., are some of the commonly used herbicides in rice culture. However, it is imperative to evaluate the new generation low dose high efficacy herbicides for sustainable weed management in upland rice. Cyhalofop butyl, Fenoxapropp-ethyl, etc., are low volume and high efficacy herbicides and may provide efficient and sustainable weed control particularly under high rainfall situation. The hypotheses tested was that new generation herbicides (like Cyhalofop butyl, grass weed killer) used at low doses along with 2,4-D may be highly effective in managing weeds in upland rice. The present study was therefore, undertaken with the objectives to identify efficient combinations of herbicides to develop an economic, energy efficient and sustainable weed management practice for DSR in upland ecosystem of the Eastern Himalayas, India.

Study site
A field experiment was conducted at experimental farm, Indian Council of Agricultural Research (ICAR) Complex for North Eastern Hill (NEH) Region, Umiam, Meghalaya, India for four consecutive years of 2012-2015. The farm is located at 25°30' N latitude and 91° 51' E longitude with an elevation of 950 m above sea level. The site receives an average annual rainfall of 2548.5 mm. The minimum and maximum temperature varied between 16.6°C (October) and 29°C (July) during the crop cycle. Four year mean monthly weather parameter and monthly rainfall during the experimental period are depicted in Suppl. Figure 1. The soil of the experimental field was a well-drained silty loam classified as Typic Paleudalf and was low in available P (18.5 kg ha −1 ), medium in N (252.6 kg ha −1 ) and high in K (206.5 kg ha −1 ). Organic carbon content and pH of the soils were 2.33% and 4.98, respectively.
The experiment with seven treatments (Table 1) was laid out in a randomized block design with three replications. Under weedy check, no weed control was done. The herbicides as per treatment were applied with a spray volume of 450 litres of water ha −1 using flat fan nozzle. The rice crop was sown in line during the month of May in all the years with a row to row spacing of 30 cm using a manually drawn 3 tyne furrow opener. The crop received a recommended fertilizer dose of 80:60:40 kg N, P 2 O 5 and K 2 O ha −1 , respectively, along with 5 Mg ha −1 farmyard manure. The N, P and K were supplied through urea, single super phosphate and muriate of potash, respectively. Half of N and full doses of P and K were applied at the time of transplanting and remaining half of N was applied in two equal splits at tillering and panicle initiation stage. The observations on various growth, yield components and yield were recorded at harvest.

Plant sampling and analysis
Five random plant samples were collected from the earmarked area of each plot during harvesting leaving two border lines. Plant samples were kept in an oven at 60°C till constant weights were obtained and oven-dried plant samples were used for determination of N, P and K content. Data on weeds (density, dry weight) were recorded at 30, 60, 90 and 120 days after sowing (DAS) from randomly placed quadrants (0.5 × 0.5 m) on each treatment. Weeds were uprooted gently, the roots were washed and their counts recorded. Weeds were oven dried at 60°C for 48 h after sun drying for recording dry weight. The dry weight of weeds was expressed as g per 0.25 m 2 and data were converted into g m −2 . Major weed species of experimental site are listed in the Table 2. The weed control efficiency (WCE) was worked out through the formula suggested by Mani et al. (1973) and expressed in percentage: where WCE = weed control efficiency (%); DWC = weed dry weight of control plot; DWT = weed dry weight of control plot. The weed index (WI) expresses the competition offered by weeds measured by per cent reduction in yield owing to their presence in the field (Gill and Vijay Kumar 1969). WI was calculated by using following formulaL where WI = weed index (%); YWF = yield from weed free check; YTP = yield from treatment plot.

Measurement of leaf characteristics
The fully expanded matured leaf (the fourth leaf) from the top was used for leaf thickness (LT) and specific leaf weight (SLW) measurements. The LT was measured using absolute digital verniercalliper (Mitutoyo Corp., Japan) at broadest part of the leaf excluding major veins with accuracy of ±0.01 mm and expressed in µm. This measurement was made by pressing the calliper gently to avoid over estimation and injury to leaf. For determining SLW, selected leaves from different treatments were assessed for leaf area using leaf area meter and respective dry weight of the leaf after drying in an oven at 60°C to constant weight. SLW was computed using the formula given below: where SLW = specific leaf weight (g cm −2 ), LW = leaf weight; LA = leaf area.

Estimation of leaf pigments
The contents of chlorophyll, carotenoid and anthocyanin pigment levels in fresh and matured leaves were measured by acetone extraction method. Known weight of leaf tissue (0.5 g) was fully homogenized in 15 ml of 80% acetone:20% water solution using a pre-cooled pestle and mortar at 4°C. A pinch of CaCO 3 was added to extraction solution during grinding of the samples to neutralize any plant acids which might be liberated during grinding. After thorough grinding, the extracted solution was filtered through filter paper (Whatman No. 42) in volumetric flask and the volume of the extract made to 25 ml using 80% acetone:20% water solution. Each sample was placed in a cuvette and absorbance was recorded at 645, 663 and 480 nm through UV visible spectrophotometer (UV-2100). The absorbance values were substituted in the following formulae to Carotenoids ¼ ðA480 þ 0:114Þ � ðA663 À 0:638Þ � A645 � ðV=WÞ � ð1=1000Þ; Chl a, Chl b and Chl a + b are in mg g −1 .
Here V refers to the total volume of the extract and W refers to weight of the tissue taken for pigment measurements and A663, A645 and A480 are the optical absorbance values recorded by UV-2100 at 663, 645 and 480 nm, respectively. Besides this SPAD chlorophyll meter reading (SCMR), an indication of total chlorophyll content in leaf was measured at different stages using a hand held SPAD meter (Konica Minolta Inc, Japan).

Plant nutrient uptake
The nutrient (N, P and K) content in rice (grain and straw) was determined at harvest following standard procedures. The total N was analysed through micro-Kjeldal method, while total P and K were determined using sulphuric-nitric-percholoric acid digest (Prasad et al. 2006). The nutrient content in grain/straw was multiplied with their respective biomass production to obtain uptake in kg ha −1 .

Economics
The cost of cultivation and returns were taken into account for calculating economics of treatments.
The gross returns were considered as total income from the produce of grain and straw yield based on prevailing market price (INR is Indian rupees and 1 US $ = 73 INR). The net return, benefit-cost ratio and economic efficiency were calculated with the help of following formulae: where GR = gross return $ ha −1 ; RY = return from seed yield. ii.
where NR = net return $ ha −1 ; TC = total cost of production, GR = gross return. iii.
where B:C ratio = benefit cost ratio; GR = gross return; TC = total cost of production, where EE = economic efficiency $ ha −1 day −1 ; NR = net return; TD = total duration of crop (in days).

Energetics
Energy input and output were calculated by converting the various inputs used viz. labour, organic manures, paddy straw and output i.e. grain and straw into energy units (MJ) as shown in Supplementary Table 1. Input energy: Energy equivalents for all inputs were summed to provide an estimate for total energy input as suggested by Devasenapathy et al. (2009) and Tuti et al. (2012).
Output energy: Biomass crop yield is total yield of grain and by-product (straw). Energy output from the product (grain) was calculated by multiplying production and its corresponding energy equivalent. Energy outputs from by-product were estimated by multiplying amount of by-product and its corresponding energy equivalent.

Statistical analysis
The analysis of variance (Panse and Sukhatme 1978) was used to analyse the data. The significance of different sources of variations were tested by error mean square of Fisher Snedecor's 'F' test at 5% probability level (P = 0.05). In the summary tables of the results, the standard error of the mean (SEm ±) and least significant difference (LSD) were provided to compare the means.

Weed dynamics
The major weed species (Table 2) observed in the experimental field were broadleaved -Borreria hispida, Ageratum conizoides, Spilanthes spp., Drimaria cordata, Erecthritisva lerianifolia, grassy weeds; Paspulumum congugatum, Paspulum longifolium, Digiteria sanguinalis, Eleusin eindica and sedges -Cyperus rotundus and Cyperus iria. Broadleaved were the major (62%) weed followed by grasses (28%) and sedges (10%) under weedy check (Table 3). The grassy weeds emerged in the initial stage of crop growth followed by broadleaved and sedges. Weed management measures significantly reduced the weed population compared with unweeded check. Application of pre-and postemergence herbicides reduced the weed density and weed dry weight. Minimum weed counts (Table 4) were observed (at 30, 60 and 90 DAS, respectively) with the application of 80 g ha −1 Cyhalofop butyl at 25 DAS and 0.75 kg ha −1 2,4 D at 35 DAS followed by the treatment receiving sequential application of 1.5 kg ha −1 Butachlor at 3 DAS and 2,4-D at 25 DAS. Cyhalofop-butyl is a new generation herbicide registered under the group aryloxyphenoxy propionate which controls a wide range of grass weed species in different growth stages. The herbicide Cyhalofop-butyl halts activity of acetyl-coenzyme A carboxylase (ACCase), by blocking fatty acid biosynthesis, thus prevents formation of lipid and secondary metabolites in susceptible plants (Kaundun 2014). Tolerant rice inactivates the esterase producing a lack of functionality thus, minimizes the conversion of Cyhalofop-butyl to its active form Cyhalofop acid . Hindrance in fatty acid production subsequently inhibits cell division. In rice plant, Cyhalofop-butyl is metabolized into inactive form diacid. In grasses, the herbicidal efficacy of Cyhalofop-butyl is due to the biologically active monoacid metabolite (Antralina et al. 2015). The effective control of grassy weeds by post emergence grass killer i.e. Cyhalofop butyl or Butachlor (pre-emergence) followed by control of broadleaved weeds through 2,4-D might have resulted in minimum weed density in present study as was previously indicated by other researchers Singh et al. 2017). Maximum weed density at 30, 60 and 90 DAS, respectively was observed under weedy check where no weed control was done, while weed free check registered the minimum weed density. In general, the weed dry weight was found higher at 60 DAS compared with at 30 and 90 DAS in all the treatments. The higher availability of soil moisture along with favourable temperature regime might have increased the growth rate of weeds (Das et al. 2015). The weed dry weight (pooled data over 4 years) as influenced by weed management practices indicated a highly significant variation among the treatments at different growing stages (Table 4). Minimum weed dry weight was recorded under weed free check followed by the treatment receiving 80 g ha −1 Cyhalofop butyl fb 0.75 kg ha −1 2,4-D. This is owing to control of weeds with sequential application of herbicide effective against grassy and broadleaved weeds. As expected, weedy check registered the maximum weed dry weight. High rainfall with favourable temperature regime results in emergence and rapid growth of weed and smothered the crop completely under weedy check. Our study showed complete crop failure under weedy check. The highest WCE at different crop stages were obtained with application of 80 g ha −1 Cyhalofop butyl and 0.75 kg ha −1 2,4-D and was comparable to weed free treatment (Table 4). Effective control of wide range of weeds following sequential application of grass and broadleaved weed killer herbicides resulted lower weed dry matter accumulation and higher WCE ). The lowest WI (Figure 1.) was consistently obtained from the treatment that received post emergence application of 80 g ha −1 Cyhalofop butyl at 25 DAS fb 0.75 kg ha −1 2,4-D at 35 DAS indicating the minimum reduction in grain yield due to efficient control of weeds under this treatment followed by the treatment receiving 60 g ha −1 Fenoxaprop (15 DAS) fb 0.75 kg ha −1 2,4-D at 25 DAS. The higher WI was associated with weedy check. The WI with either Butachlor 1.5 kg ha −1 or 0.75 kg ha −1 Pretilachlor fb 2,4-D 0.75 kg ha −1 was at par and comparable to mechanical weeding (20 and 40 DAS). The higher bio-efficacy of Cyhalofop-butyl than other herbicides in rice is widely reported ).

Physiological parameters
In the present study, the leaf chlorophyll content assessed through SCMR and chlorophyll extraction indicated that rice grown with different weed management practices recorded higher SCMR and increased total chlorophyll at all stages of crop growth compared with unweeded check (Table 5). However, the effect was more pronounced at later stages of crop growth (60 and 90 DAS). All the weed management practices were not significantly different in respect to leaf chlorophyll content up to 60 DAS. Application of Pretilachlor fb 2,4-D or Fenoxaprop-p-ethyl fb 2,4-D were not significantly different from each other and resulted in significantly higher chl a, chl b and total chlorophyll at 90 DAS compared with other weed management practices. The low chlorophyll content in rice plants under weedy check compared with treatments with weed management practices might be due to exhaustion of nutrients especially N (which is integral part of chlorophyll) through robust weed growth at critical stages of crop growth (Moran et al. 2000). Weeds being important biotic stress factors in agro-ecosystem can induce relentless stress and competition affecting various metabolic processes in roots and individual leaves resulting in both reduced nutrient uptake and loss of chlorophyll and a change in its distribution pattern (Barton 2000). In addition, chlorophyll and carotenoids contents also varied with microclimatic conditions of crop growth (Shaikh and Dongare 2008). The specific leaf weight (SLW) and leaf thickness (LT), which are inextricably linked and dependent on each other varied significantly among different weed control measures. Both of the parameters were lowest under weedy check compared with weed management practices evaluated. However, no significant differences in SLW and LT of rice plants were recorded among the weed management practices. Higher SLW was recorded with the application of Butachlor fb 2,4-D, whereas, Fenoxaprop fb 2,4-D resulted in higher values of LT at 60 and 90 DAS as compared with other weed management treatments. Leaf temperature of rice plants grown with different weed control measures also varied significantly ( Table 5). The rice plants grown with Pretichlor fb 2,4-D and weed free up to 60 DAS had significantly lower leaf temperature compared with other methods of weed control. The higher leaf temperature, however, was recorded under Butachlor fb 2,4-D and Fenoxaprop fb 2,4 D. As leaf characteristics are the important growth parameters for maintaining photosynthetic efficiency of crop and eventually crop yield, their decrease in weedy check over different weed control measures indicates the impact of weeds on reduced crop growth and decreased leaf area and leaf mass (Varshney et al. 2012;Channappagoudar et al. 2013a). As proper leaf structure, leaf volume and other leaf traits are necessary for biomass production and crop performance, increased SLW and LT in crop due to adoption of the weed management practices ensured increased photosynthesis, increased biomass and there by enhanced crop yield (Vile et al. 2005;White and Montes 2005). Positive correlation among leaf traits with total chlorophyll, biological yield and grain yield under different weed management practices was reported earlier (Chandola et al. 2015). Higher LT and leaf dry biomass lead to greater number of leaf mesophyll cells with higher chlorophyll content under different weed management practices, which ultimately leads to greater and efficient use of photon flux density and results in enhanced photosynthetic production (Manzoor and Goutam 2014). Higher dry matter accumulation in rice under weed management practices is also an indicative of the better utilization of resources in absence of weed growth (Channappagoudar et al. 2013b).

Crop productivity
Aerobic soil conditions and tillage in upland rice along with favourable alternate wetting and drying cycles favours germination and growth of highly competitive weeds like grasses, sedges and broadleaved weed causing substantial yield los in rice (Singh 2012). In present study, unattended weed resulted in complete crop failure under weedy check ( Table 6). The vigorous weed growth of diverse species composition resulted in serious crop weed competition from the early vegetative stage of the crop (Suppl. Figure 2) for growth factors and thus, resulted in complete suppression of the crop   1.81 fb, followed by; DAS, days after sowing; SEm, standard error of the mean; LSD, least significant difference. *The means within the same column with different letters are significantly different. (Jayadeva et al. 2011;Singh et al. 2014). Application of 80 g ha −1 Cyhalofop butyl at 25 DAS fb 0.75 kg ha −1 2,4-D at 35 DAS gave consistently and significantly higher grain yield (Suppl. Figure 3) of rice (Table 6), which was not significantly different with the treatment receiving mechanical weeding with grubber at 20 and 40 DAS. The effective and timely control of weeds with Cyhalofop butyl fb 2,4-D have reduced crop weed competition during the critical crop growth stages and provided better crop growth environment, which ultimately increased the yield components and yield (Menon et al. 2014;Singh et al. 2017). Further, the yield obtained with either 1.5 kg ha −1 Butachlor fb 0.75 kg ha −1 2,4-D or 0.75 kg ha −1 Pretilachlor fb 0.75 kg ha −1 2,4-D were comparable (Table 6), which might be due to similar mode of action of these herbicides. As expected, the significantly higher rice yield was obtained under weed free treatment (up to 60 DAS) mainly due to minimum or no competition between crop and weed during the critical period of crop weed competition. The yield obtained with post emergence application of 60 g ha −1 Fenoxaprop fb 0.75 kg ha −1 2,4-D was comparable with that of mechanical weeding (20 and 40 DAS). Straw yield followed the similar trend as that of grain yield ( Table 6). The highest harvest index (40.3) was with weed free treatment followed by 60 g ha −1 Cyhalofop butyl fb 0.75 kg ha −1 2,4-D (38.92).

Nutrient uptake
Pooled data (Figure 2) revealed that weed management measures significantly influenced the crop nutrient uptake. The total N, P and K uptake by the crop ranged from 109 to 140, from 13.7 to 19.7 and from 65 to 82.2 kg ha −1 , respectively. Maximum NPK uptake was obtained under weed free situations owing to minimum crop weed competition. No uptake of nutrient by crop at harvest could be recorded under weedy check due to complete smothering of crop by the weeds. Application of Butachlor or Pretilachlor fb 2,4-D were not significantly different from each other but resulted in significantly higher uptake, which might be due to minimum crop weed competition from early crop growth stages. Butachlor or pretilachlor effectively controls the grass weeds emerged earlier and 2,4-D controls the broadleaved weeds emerged at later growth stages of the crop and results in minimum crop weed competition, which leads to higher nutrient uptake (Rajkhowa et al. 2001). However, all the weed control practices resulted in increased nutrient uptake by the crop due to satisfactory control of weeds. The uptake (NPK) by the weeds ranged from 13.5 to 50.8, from 14.8 to  40.1 and from 2.8 to 13.5 kg ha −1 , respectively. Weedy check registered maximum nutrient uptake. All the weed control practices significantly reduced the nutrient uptake by weeds indicating the necessity of weed control to reduce the loss of nutrients (NPK) through weeds (Gowda et al. 2009;Babar and Velayutham 2012). Among the herbicides, post emergence application of 60 g ha −1 Fenoxapro-p-ethyl followed by 2,4-D 0.75 kg ha −1 recorded the least nutrient uptake and remained at par with weed free treatment. The least nutrient removal by weeds was obtained with weed free treatment. Efficient and timely weed control associated with herbicide applications might have resulted in reduced weed growth and ultimately led to reduced nutrient uptake (Sanjay et al. 2006;Mandal et al. 2011).

Economics
The weed free treatment resulted in the highest net return ($322.4 ha −1 ), B:C ratio (1.78 ha −1 day −1 ) as well as economic efficiency ($2.39 ha −1 day −1 ) followed by the treatment that received 80 g ha −1 Cyhalofop butyl at 25 DAS fb 0.75 kg ha −1 2,4-D at 35 DAS with net return ($317.5 ha −1 ), B:C ratio (1.87) and economic efficiency ($2.35 ha −1 day −1 ). These were comparable to the use of Fenoxapropp-ethyl and 2,4-D for suppression of weeds. Efficient and timely weed control provided better growth environment to the crop for increased economic yield (Table 7). Further, reduction in cost of weeding with herbicides also resulted in higher economic efficiency (Jadhav et al. 2010;Dewangan et al. 2011).

Energetic
Maximum input energy (Table 7) was recorded with Butachlor fb 2,4-D (12,024.2 MJ ha −1 ) treatment followed by Pretilachlor fb 2,4-D (11,919.48 MJ ha −1 ) and minimum in weedy check (11,103.8 MJ ha −1 ). Whereas, the highest output energy was recorded with weed free up to 60 DAS (126,864.92 MJ ha −1 ) followed by Cyhalofop butyl fb 2,4-D (12, 3114.03 MJ ha −1 ). The higher output: input ratio was observed with sequential application of pre-and post-emergence herbicides. Complete crop loss under weedy check due to smothering effect of weeds resulted in zero output energy. Application of Cyhalofop butyl fb 2,4-D resulted in higher net energy (111,443.83 MJ ha −1 ), energy productivity (0.31 MJ ha −1 ) and energy profitability (9.53) and was comparable to weed free treatment. Similar report on efficacy of Cyhalofop butyl have been previously reported (Kaur and Singh 2016;Kumar et al. 2016).

Conclusions
The data supports the following conclusions • Weed flora in upland rice ecosystem are complex consists of grass, broadleaf and sedges. The highest grain yields were obtained under weed free conditions whereas, there was complete crop loss under weedy check, indicating need for adoption of effective weed management programme for upland rice in the study agro-ecosystem. • Sequential application of 80 g ha −1 Cyhalopfopbutyl (25 DAS) fb 0.75 kg ha −1 2,4-D (35 DAS) provided efficient and profitable control of diverse weeds in upland rice with higher net energy, energy productivity and energy profitability. • Alternatively, pre emergence application of either 1.5 kg ha −1 Butachlor or 0.75 kg ha −1 Pretilachlor followed by post emergence application of 0.75 kg ha −1 2,4-D were also found effective for efficient control of weeds in upland ecosystem of the Eastern Himalayas.