Climate-smart agriculture practices influence weed density and diversity in cereal-based agri-food systems of western Indo-Gangetic plains

Climate-smart agriculture (CSA)-based management practices are getting popular across South-Asia as an alternative to the conventional system for particular weed suppression, resources conservation and environmental quality. An 8-year study (2012–2013 to 2019–2020) was conducted to understand the shift in weed density and diversity under different CSA-based management practices called scenarios (Sc). These Sc involved: Sc1, conventional tillage (CT)-based rice–wheat system with flood irrigation (farmers’ practice); Sc2, CT-rice, zero tillage (ZT)-wheat–mungbean with flood irrigation (partial CA-based); Sc3, ZT rice–wheat–mungbean with flood irrigation (partial CSA-based rice); Sc4, ZT maize–wheat–mungbean with flood irrigation (partial CSA-based maize); Sc5, ZT rice–wheat–mungbean with subsurface drip irrigation (full CSA-based rice); and Sc6, ZT maize–wheat–mungbean with subsurface drip irrigation (full CSA-based maize). The most abundant weed species were P. minor > A. arvensis > M. indicus > C. album and were favored by farmers’ practice. However, CSA-based management practices suppressed these species and favored S. nigrum and R. dentatus and the effect of CSAPs was more evident in the long-term. Maximum total weed density was observed for Sc1, while minimum value was recorded under full CSA-based maize systems, where seven weed-species vanished, and P. minor density declined to 0.33 instead of 25.93 plant m−2 after 8-years of continuous cultivation. Full CSA-based maize–wheat system could be a promising alternative for the conveniently managed rice–wheat system in weed suppression in north-west India.

www.nature.com/scientificreports/ perennials weed species (C. arvensis and C. arvense) emerged only under CSA-based rice systems (Sc3 and Sc5). After 8 years, S. nigrum and R. dentatus recorded the highest density for all CA/CSA scenarios. The effects of CSA-based management practices were greatest on two abounded species, in which all CSAscenarios, once adopted) after 1 year) the emergence of C. album was eliminated, meanwhile, the density of M. indicus was gradually reduced to be zero under Sc4-6, after 8 years of conversion to CSA (Fig. 2). In Sc4, the density of P. minor was lowered down by 32.3, 68.0 and 93.17%, after 1, 4 and 8 years, respectively, than farmers' practice (Figs. 1,2). In Sc6 (full CSA-based rice system), 7 species were eliminated and the density of P. minor and A. arvensis was reduced to be less than one plant m −2 instead of 25.93 ± 2.69 and 2.33 ± 0.73 (in the first year), respectively. However, the density of S. nigrum was markedly increased under all CA/CSA scenarios (Fig. 2). For example, Sc3 recorded 6.0, 4.27, and 2.73-times higher density of S. nigrum compared to Sc1 after 1, 4 and 8 years, respectively.
Maximum total weed density (TWD; no m −2 ) was recorded in the first year (Table 2). During 2016-2017 and 2019-2020, all CA/CSA-based scenarios significantly reduced TWD, with Sc5 and Sc6 having the highest suppression effect. Minimum TWD was observed in 2019-2020 in which the TWD was lower than Sc1 by 39.17, 39.72, 53.8, 65.0 and 71.7% for Sc2, Sc3, Sc4 Sc5 and Sc6, respectively. The effect of CSA-based management practices on TWD was cumulative and was maximum in the long-term. For example, following Sc2, TWD was reduced by 11.72, 28.47 and 39.17 after 1, 4 and 8 years, respectively. Similarly, TWD in Sc4 was lower than Sc1 by 5.92, 39.97 and 53.83 after 1, 4 and 8 years, respectively. Fig. 3 showed that weed biomass (dry weight, g m −2 ) was affected by CSA-based management practices and the effect varied between years and species. Initially (in 2012-2013), maximum weed biomass was observed for P. minor followed by A. arvensis, M. indicus, R. dentatus and C. didymus. Overall, CSA-based management practices distinctly reduced the biomass of four abounded weed species (P. minor, A. arvensis, M. indicus, and C. didymus), with full CSA-practices (Sc5 and Sc6) having the greatest effect (Fig. 3). However, the biomass of S. nigrum and M. denticulata was increased and having the greatest values with partial CSA-practices (Sc3 and Sc4).

Diversity indices (Shannon, Richness and Evenness). The Shannon Diversity Index and Evenness
showed that different management scenarios had a non-significant effect on weeds diversity across years (Table 2). However, when considering the richness of species (Richness Dmg), CSA-based scenarios demonstrated a significant effect across years (Table 2). In 2012-2013 (after 1 year), a non-significant effect was recorded due to

Abundant weed species across the years.
Phalaris minor. In general, under Sc1 the density of P. minor showed almost the same density across years (2012-2013 to 2019-2020) (Fig. 4). While a downward trend was observed owing to all CSA-based management practices over time. P. minor was favored under Sc1, while after 3 years (in 2014-2015), all CSA-based scenarios significantly suppressed its density (with no significant differences among scenarios). Similar trends were observed in 2015-2016 (Fig. 4). After which, (after 4 years), the density of P. minor was evidently reduced due to all CA/CSA-based practices, with Sc5 and Sc6 having the utmost effect.
Rumex dentatus. The density of R. dentatus under Sc1 found to be almost constant (ranged from 5 to 6 plants m −2 ) during the 8 years of study (Fig. 5). However, the density was non-significantly differed from Solanum nigrum. In general CSA-based management practices significantly increased the density of S.
nigrum and was found to be abundant after 8 years of conversion to CSA (Fig. 5). Interestingly, the density of S. nigrum was maximum under the maize-based partial CSA system (Sc4       Relative weed density. Relative weed density was affected by different crop management practices. Generally, P. minor, A. arvensis, M. indicus, R. dentatus and S. nigrum recorded the highest relative density (RD) across the years (Fig. 7). In 2012, P. minor recorded the highest RD for all CSA-based scenarios (~ 25 to 35%), followed by A. arvensis (~ 25%) and M. indicus (~ 20%). Similarly, in 2017, P. minor recorded the highest RD for all scenarios, in particular Sc1 (~ 45%). CSA-based management practices increased the RD of both R. dentatus and S. nigrum, at the expense of P. minor and M. indicus. In 2020, only in Sc1, P. minor was the most abounded species, in which its RD exceeded 50%. However, all CSA-based scenarios intensely reduced its RD to be less than 5% in full CSA-based scenarios (Sc5 and Sc6), irrespective of cropping systems. The reduction in the RD of P. minor, was coupled with an increase in the RD of R. dentatus and S. nigrum, in particular for Sc4-6. In brief, CSA-practices reduced the RD of P. minor over time to be less than 5% in Sc5 and Sc6 in 2019-2020, instead of Vertical bars indicate ± SE of mean. Sc1 conventional rice-wheat system with flood irrigation (FI), Sc2 conservation agriculture (CA)-based rice-wheat-mungbean system with FI, Sc3 partial CSA-based rice-wheat-mungbean system with FI, Sc4 partial CSA-based maize-wheatmungbean system with FI, Sc5 full CSA-based rice-wheat-mungbean system with subsurface drip irrigation (SDI), Sc6 full CSA-based maize-wheat-mungbean system with SDI. The figure has been generate using Microsoft Excell office professional Plus (2010) (Fig. 7). Since CSA-practices not only suppressed P. minor but also favored both R. dentatus and S. nigrum, in which the RD of both species was maximum (more than 80-90%) in 2019-2020 under CSA, particularly Sc4, Sc5 and Sc6 instead of 5% for S. nigrum and 10% for R. dentatus in 2012-2013.
Contrast analysis. The contrast effects of tillage (CT vs ZT), irrigation (flood vs SDI) and systems (RW vs MW system) on P. minor were significant (Table S2) Table S2). Similarly, the contrast effects of tillage (CT vs ZT), irrigation (flood vs SDI) and systems (RW vs MW system) on broadleaf weed density were significant (Table S2) Table S2). Overall, the TWD was also significantly influenced by tillage (CT vs ZT), irrigation (flood vs SDI) and systems (RW vs MW system) ( Table 1, Table S2).

Discussion
Weed abundance was described by determining the density, biomass, and relative density of individual species.
In the farmer's practices (Sc1), weed density, biomass and diversity almost remained the same across the years (2012-2013 to 2019-2020), therefore the observed shift in weed composition was attributed to CSA-based management practices, e.g., tillage, crop residues mulch, cropping system, precision water and nutrients management and layering of these practices. Therefore, the effects of CSA-management practices on weed diversity and density will be discussed as per the influencing drivers/factors.

Tillage.
The observed lower infestation of major weeds under ZT might be attributed to seedbank depletion over time. In ZT, weed seeds accumulate (60-90%) near the soil surface and are more vulnerable to predation, e.g., insects "especially ants", rodents, birds and other organisms 40 . The higher mortality and loss viability, due to dryness and severe weather variability, lead to more unviable seeds and seedbank depletion over time under ZT 21,25 . The increased number, diversity, and activity of seed-consuming organisms, observed under ZT-fields 41 complement this opinion. In the current study, the long-term adaption of ZT might contributed to exhausting or depleting seedbank in the soil. In a review article, Nichols et al. 21 concluded that seedbanks rapidly decreases under ZT than CT. Thus, seedbank could be exhausted after 5 years of implementing ZT 25 . Moreover, in tilled soils weed seedlings emerge from deeper soil layers 42 , which makes the emerged weeds weaker. Similar to our results, several studies have shown that ZT had lower weed infestation of major weeds than CT 25,43 . However, www.nature.com/scientificreports/ several reports showed higher weed density in ZT with respect to CT even after several years 14,22 , that indicating that other management aspects together with ZT are important.
Crop residue management. Mulching crop residues might be a compelling reason for the observed reduction in weed density and biomass. In the RW system, retaining crop residues on the soil surface can help in weeds suppression by imposing a physical barrier to emerging seeds, restricting the growth of germinated ones by limiting light availability [44][45][46] and by potential allelopathic effects 28,47 . Additionally, surface residue decreases and buffer soil temperature, thus the germination of some weed species' is reduced due to the reduced fluctuations in soil temperature 48 . Under limited light availability conditions, emerged seedlings become etiolated and weak as germinated seeds search for light 49 . In the IGP, India, in RW cropping system (under ZT) mulching rice residue of 5 t ha −1 promoted predation of RW weeds, including P. minor and decreased weed density up to 76% 29 . Similarly, retaining rice crop residues of 5.0 and 7.5 t ha −1 reduced the total weed biomass by 23.4-30.3 and 35.5-44.1%, respectively 10 . In the current study, the average rice/maize residue load (2012-2012 to 2019-2020) was 9.69, 6.97, 8.50, 7.14 and 8.57 for Sc2, Sc3, Sc4, Sc5 and Sc6 respectively, implying that the rice and maize residues were enough to help in weed suppression. Rice and wheat residues have been identified as exhibiting genetically controlled allelopathy which could be browbeaten for weed suppression 50 . Interestingly, since ZT Efficient crop rotation. RW cropping system is the main cropping system, in the IGP-India, therefore weed is a major problem 29 . In the present study, the inclusion of mungbean to the RW cropping system (all tested CA/CSA-scenarios) might contribute to weed suppression. Similar results were reported by Shahzad et al. 28 and Malik et al. 46 in which they found that the inclusion of sorghum, in the wheat-based cropping system may help managing weeds. In the present study, the RW cropping system had the maximum weed diversity, TWD and TWB across years, while maize-based cropping (Sc4 and Sc6), not only showed the least density and biomass, but also had the minimum weeds diversity (Fig. 2), confirming the findings of a review article 21 and a metaanalysis 30 which reported that crop rotation distinctly reduces weed densities. Each crop smears certain restrictions on weed flora, which promote the growth of some weeds and suppress others 28,51 . Rotating crops prevents one weed from being repeatedly successful, thus avoiding its establishment 9,31 . Consequently, each crop can act as a filter, allowing only weeds that can adapt with this sort of crop management 31,51 . In the current study, the RW cropping system favored P. minor and A. arvensis (Figs. 2, 3). However, adapting CSA-based management practices; shifted weed flora toward more broadleaf (e.g., S. nigrum and R. dentatus) in all cropping systems and perennial weeds (C. arvensis and C. arvense) only in ZT-rice-based cropping system. The presence of perennial weeds in CA has been previously reported 9,21 . Interestingly, using different crops also increases predation pressure 33,52 and promote predation of distinct weed seeds 34 . Therefore, rotating crops could increase the diversity of consumed seeds, thus reduce intensification.

Precise water management. Precise water management (through subsurface drip irrigation; SDI) is a
promising technology for altering weed density and weed suppression effects 53 . The positive effect of SDI might be due to adjusting water and fertilizer placement. Since under ZT weed seeds accumulate near and/or on the soil surface, weeds germination is dramatically reduced because SDI leaves the few top cms of the soil dry 54,55 . Additionally, because wheat can germinate in drier environments than can various weeds 56 , sowing under drier conditions (SDI) can reduce weed emergence especially, moisture-loving weeds like P. minor 57 . Interestingly, placement of water and fertilizer using SDI can discourage weed growth rates 21,58 , as SDI better deliver water and nutrients where it will most benefit crops, not weeds 53,55 . Shrestha et al. 54   www.nature.com/scientificreports/ management practices might be due to surface soil strength in ZT (after rice harvest) which obstruct weeds emergence, higher predation of weed seeds and low light availability; all of which weaken germination of many weed species not only P. minor 60 . On the other side, the higher density and biomass of both S. nigrum and R. dentatus, under all CA/CSA-scenarios, might be due to the vast majority of their seeds accumulate near to the soil surface where they can better germinate 10,14 . In contrast, under CT, during tillage, seeds of S. nigrum and R. dentatus are buried and their emergence is markedly decreased due to their high sensitivity to seeding depth 14,61 . The germination of S. nigrum was 93.1 and 4.7% at a seeding depth of 1 and 4 cm, respectively, while its emergence was inhibited at 8 cm seeding depth 61 . In the present study, the effect of CSA-based management practices are additive. Using the same tillage practices (ZT), crop residue retention and cropping system, but with different water/nutrients application (flooding vs SDI), led to more significant weed suppression. Precise water management (through SDI) is promising CSA-practice in weed suppression and showed the superiority of Sc6 over Sc4 and Sc5 over Sc3. The combined-additive effect of CSA-practices on weed flora composition and infestation levels was more evident in the long-term. Similar to our observations, a non-significant difference in total weed density (TWD) and TW biomass (B) in the first year was observed by other researchers 22,62 . After 4 years, a significant reduction in both TWD and TWB was observed. This suppression might be ascribed to weed seedbanks depletion 21,25 . About 4-10 years are required to reach the weed population's equilibrium 63 and it was observed in our study as well with respect to the TWD and TWB. The modern planting technology (Combine harvesting with super SMS followed by Happy Seeder sowing in wheat) have made wheat sowing successfully possible under heavy rice residues (8-10 t ha −1 ) and eased the use of residues for weed management without any negative effect on Figure 7. Effects of climate-smart agriculture-based management practices on relative weed density (%) under different scenarios. Sc1 conventional rice-wheat system with flood irrigation (FI), Sc2 conservation agriculture (CA)-based rice-wheat-mungbean system with FI, Sc3 partial CSA-based rice-wheat-mungbean system with FI, Sc4 partial CSA-based maize-wheat-mungbean system with FI, Sc5 full CSA-based rice-wheat-mungbean system with subsurface drip irrigation (SDI), Sc6 full CSA-based maize-wheat-mungbean system with SDI. The figure has been generate using Microsoft Excell office professional Plus (2010), version (14.0.4734.1000).  46,64 . In this study, ZT under residues retention conditions and integrated with pulse crop (mungbean) and SDI would evidently deplete the weed seedbank.
The collective use of all CSA practices could offer additive benefits, and the potential of weed infestation is higher if only one management practice is applied. Our results imply that layering of CSA-based crop management practices; ZT, crop rotation, crop residue retention and SDI are synergistic means of weed control. Minimum TWD was observed for Sc6, followed by Sc5 and Sc4. Indicating the importance of SDI as CSA practices to minimize weed effects. In terms of TWB, Sc6 recorded minimum TWB, followed by Sc4 and Sc5, implying the importance of the cropping system (replacing rice with maize). This might be due to that weed density does not always represent weed biomass (more feasible when emphasizing crop-weed competition) 21 . The superiority of CSA-based maize systems could be a promising alternative for the RW cropping system, where less weed diversity, density and biomass were recorded. This promotes a big opportunity for weed suppression in IGP, south Asia.

Conclusion
Crop management activities like tillage, crop residue, crop rotations, water application and nutrients management affect weed diversity and composition. Weed community responses to long-term climate-smart agriculture (CSA) management practices in cereal-based agri-food systems of western IGP was investigated. The most abundant species were P. minor, A. arvensis, M. indicus and C. album and were favored by farmers' practice. However, CSA-based management practices markedly reduced total weeds density and biomass and shifted weed flora towards broadleaf weed species i.e., S. nigrum and R. dentatus. The effect of CSA-based management practices on weed flora composition and infestation levels was additive and more evident in the long-term. Implementing long-term CSA-practices might lead to weed seedbank depletion due to encouraging weed predation factors. The superiority of Sc4 over Sc3 and Sc6 over Sc5 "in weed suppression" implies the importance of crop rotation (replacing rice with maize). Similarly, the superiority of Sc6 over Sc4 and Sc5 over Sc3 signifying the role of subsurface drip irrigation (SDI) is promising CSA practice in weed suppression. Our results indicate that layering of CSA management practices was found to be synergistic means of weed control. In conclusion, full CSA-based maize-wheat-mungbean system could be a promising alternative for the rice-wheat system for better weed management in western IGP. Besides the measurement of weed density and biomass, in long-term experiments future studies should also focus on changes of seedbank across years to get more in-depth clarity on weed behavior and dynamics.

Materials and methods
Site description. A study was initiated in 2012-2013 at ICAR-Central Soil Salinity Research Institute, Karnal, (29° 70′ N, 76° 95′ E), India to understand the shift in weed flora in cereal-based cropping systems under different climate-smart agriculture (CSA)-based management practices. During the rainy season, rice and maize crops were grown followed by wheat in the winter season as per the treatment protocols. The site has a semi-arid and sub-tropical climate, with hot and dry to wet summers and cold dry winters. The average annual rainfall (40-year period) was 765 mm with mean maximum temperature was 37.7 °C in June, whereas the minimum temperature was 6.4 °C in January. The soil of the experimental field is silty loam in texture, low in organic carbon (0.45%) with neutral pH. Before conducting this experiment, rice-wheat (RW) cropping system was being practiced under conventional tillage (CT)-based management system for 30 years. (1) conventional-till (CT) rice-CT wheat (Sc1; farmers' practice; CT); (2) CT rice-Zero tillage (ZT) wheat-ZT mungbean with flood irrigation (Sc2; partial CA-based rice system); (3) ZT rice-ZT wheat-ZT mungbean with flood irrigation (Sc3; partial CSA-based rice system); (4) ZT maize-ZT wheat-ZT mungbean with flood irrigation (Sc4; partial CSA-based maize system); (5) ZT rice-ZT wheat-ZT mungbean with SDI (Sc5; full CSA-based rice system); and (6) ZT maize-ZT wheat-ZT mungbean with SDI (Sc6; full CSA-based maize system). The scenarios were arranged in a complete randomized block design, using three replicates. Details of the tested scenarios including drivers of change, crop rotations, tillage, crop establishment method, and residue and water management practices are given in Table 1. Basically, Sc3 and Sc4 were based on CA practices, in which irrigation water and N application were not precisely managed and called it partial climate-smart agriculture (CSA). However, in Sc5 and Sc6, irrigation water and N-was precisely applied SDI and called full CSA. In all scenarios, best crop management practices were applied except Sc1, in which the traditional practices of farmers were followed (Table 1). We used four systems for better understanding, conventional rice-wheat system (Sc1), partial CA-based rice-wheat-mungbean system (Sc2), partial CSA-based rice/maize system (mean of Sc3 and Sc4) and full CSA-based rice/maize system (mean of Sc5 and Sc6). The CSA-based systems include the Sc3 and Sc5 for rice and Sc4 and Sc6 for maize systems.
Crop residue management. In the farmers' practice (Sc1) all the crop residues were removed from the ground level. In rice-based systems (Sc2, Sc3 and Sc5), full (100%) residue of rice was retained on the soil surface www.nature.com/scientificreports/ in wheat crop. However, in maize-based systems (Sc4 and Sc6), full (100%) resides of maize were retained for the first 3 years and for the remaining years partial (65%) maize residue was retained in the wheat crop. The incorporation or retention of the crop residue was depended on the biomass production of the previous crops that varied from 1 year to another years. In different scenarios, the total amount of rice crop residue ranged from 5.30 to 12.70 Mg ha −1 , while for maize it was 5.80-13.10 Mg ha −1 in 8 years of study (Table S3). The average annual crop residue load was ranged from 6.97 to 9.69 Mg ha −1 across the different management scenarios (Table S3).
Crop management. In Sc1, both rice and wheat were conventionally established following farmers' practice. In CT rice plots, dry-tillage (two harrowing and two cultivators followed by wooden planking), and wettillage (puddling; two harrowing and one planking) were done before rice transplanting. In CT wheat, two passes of harrowing and cultivator each followed by planking was done prior to sowing. Under Sc1, rice was manually transplanted in puddled fields with 25-30 days old seedlings, while wheat was planted by manual broadcasting in tilled soil. In Sc2, after wet-tillage (three passes of harrowing and one planking in standing mungbean crop after picking), rice was manually transplanted in a random geometry (20 × 15 cm). Under ZT conditions (Sc3-Sc6) rice, wheat and mungbean were planted in rows (22.5 cm apart) using Happy Seeder with an inclined plate seed metering mechanism. However, maize was seeded by Happy Seeder at a row spacing of 67.5 cm. Rice (hybrid Arize 6129) was seeded with a seed rate of 10 and 20 kg ha −1 in CT and ZT plots, respectively. However, wheat (HD 2967) was seeded at a seed rate of 120 kg ha −1 in CT plots and 100 kg ha − Weed management. Before seeding, no herbicides were applied in conventional-till (CT) plots i.e., sc1 (farmer's practice), while glyphosate @ 900 g a.i. ha −1 was applied prior to sowing in zero-till (ZT) plots (Sc2-Sc6). To control grassy and broadleaf weeds, a mix solution of clodinafop-ethyl + metsulfuron @60 + 4 g a.i. ha −1 or sulfosulfuron + metsulfuron @32 g a.i. ha −1 were used from the year 2012-2013. However, from the year 2016-2017 onwards, pinoxaden (5EC)-axil@1000 ml/ha + metsulfuron methyl 20% WG-algrip@20 g a.i. ha −1 were used at 35 days after sowing as and when required (Table S4). Before spraying of herbicides in wheat crop, four demarked areas (quadrate of 1.0 m 2 each) in each plot from where samples were taken (after 45 DAS) was covered with polythene. Every year the sampling location was different to know the real diversity and density of weed species.
Weed observations. Authors confirm that experiment on weeds/plant species in the present study complies with the Institute guidelines. Weed samples were taken at 45 days after sowing (DAS) of the wheat crop, when most of the weeds emerged. Every year, weeds were identified and counted species wise from four places in each plot/scenario by using a quadrate of 1.0 m 2 . Formal identification of the weed species was done by the first author in consultation with Dr. Virender Kumar, weed scientist. Total weed density (TWD) was calculated by summing the individual density of all weeds and expressed as (no. m −2 ). Similarly, the density of grassy and broadleaf weeds (no. m −2 ) was counted by summing the individual densities of grassy and broadleaf weeds, respectively. Weed biomass (g m −2 ) was determined by drying the collected weed samples at 65 °C for 72 h. Twelve weeds (representing grassy and broadleaf weeds) viz. Phalaris minor Retz, Polypogon monspeliensis (L.) Desf, Rumex dentatus L., Solanum nigrum L., Anagallis arvensis L., Coronopus didymus (L.) Sm, Melilotus indicus (L.), Medicago denticulata Willd, Chenopodium album L., Convolvulus arvensis L. and Cirsium arvense (L.) Scop were observed over the years, while Oxalis corniculata found negligible in wheat crop (Table S1). The voucher specimen of the weed species has not been deposited in a publicly herbarium, as there is no available a publicly herbarium.
Weed diversity indices. Relative weeds density (RD, %) of species in the whole weed community was calculated as the ratio between the density of a given weed specie to total weed density in each scenario. Weed's biodiversity was measured, including (1) the species richness (S), i.e., the number of species existed in a quadrat; (2) the species diversity, which was calculated using Shannon-Wiener index as H Statistical analysis. The data of weed parameters were analyzed using analysis of variance (ANOVA) according to Gomez and Gomez 65 , for randomized block design using SAS 9.1 software (SAS Institute, 2001). The treatment means were separated using Tukey's honestly significant difference (HSD) at 5% level of significance. The mean effects of tillage, cropping systems and irrigation methods were determined using linear con-