Although only 16% of households practiced AWD, this technique proved to be effective for paddy producers who reached a higher efficiency score of 1.144 in the AW season.This implies that the positive differences in the efficiency of the two farmer groups were mainly due to input reduction strategies.A similar circumstance amongst other groups was also evident; however, these figures were not statistically significant in the AW season.difficulties for farmers in 1M5R and AWD application still exist, including weather conditions , the surface of rice field not adequately leveling, and inefficient operation of irrigation systems and water management by small households that are some distance from canals.In addition to SBM scores, it is important that the information on input slacks is fully understood because it demonstrates the excessive usage of inputs and how well the farmers respond in managing their resources during production in order to achieve the greatest efficiency.The slacks of the main input components by rice group are shown in Table 5.First, in the SA season, the largest and smallest slacks of seed density belonged to HR at 45.46 kg/ha and AR group at 29.70 kg/ha, respectively.In the AW season, farmers decided to reduce the amount and not waste seeds.The seed slack of the AR group appeared unchanged at approximately 29 kg/ha.The MR group generated equal slack of seeds at 30.21 kg/ha to AR; meanwhile,vertical grow rack of the figure for the HR group was up to 38.24 kg/ha although it was reduced by 7 kg/ha compared to the SA season.

The largest slacks of N, P2O5, and K2O fertilizers were observed in the HR group in the SA season, followed by the MR and AR groups.Second, in the AW crop, farmers used N at 35 kg/ha with AR varieties.Due to the high risk of pests and disease in the SA season, farmers wanted to spray more pesticides and herbicides, which led to larger slacks in every rice variety group.Irrigation slacks between AR, HR, and MR in the AW season were much higher than those in the SA season.Producers have to pay for water pumped in and out of the field to maintain the continuous flooding status of rice plants.Between the rice groups categories, MR cultivation required an irrigation cost approximately 1.78 times higher than that of AR.In AW season, it is feasible to apply multiple drainage or AWD techniques to reduce water use , while simultaneously reducing the GWP of CH4 emissions.Water-saving irrigation and field drying will reduce CH4 emissions, while reducing the use of chemical fertilizers which will reduce N2O emissions from rice cultivation.The same conclusions were drawn from studies by Linquist et al., Snyder et al., Yang et al.and Zhang et al.; i.e., reducing the flooding status in paddy fields by applying AWD and N fertilizer management will help households mitigate GHG emissions.Therefore, it is important to understand information on input slacks combined with CSA practices as presented in Table 6.Regarding N fertilizer use, the 1M5R adopters had significantly smaller slacks than non-adopters in the two seasons of 13.27 kg/ha and 12.84 kg/ha, respectively.In addition, CSA practice was also effective in reducing the wastage of water, as evidenced by smaller slacks in irrigation costs for 1M5R and AWD adopters.In the SA season, the difference in excessive irrigation cost between 1M5R and conventional farmers was approximately 177,000 VND/ha, while the difference between AWD adopters and non-adopters was 148,000 VND/ha.In the AW season, although the slacks were all larger than in the SA season, 1M5R and AWD were still efficient when the participants had significantly smaller slacks than non-participants.Managing nitrogen and irrigation slacks more efficiently by practicing 1M5R and AWD, farmers can better adapt to climate change, sequester more carbon in the soil, and reduce GHG emissions, such as N2O emission intensity.Thus, paddy producers can significantly reduce N2O emissions from their fields by more precisely estimating N fertilizer needs and reducing N slack in each production step.Based on the manual of GHG emissions estimation in agriculture , we attempted to connect GHG mitigation with the reduction strategy of N.GHG emissions from synthetic N fertilizer consist of direct and indirect N2O emissions.

Direct N2O emissions occur at the source by microbial processes of nitrification and de-nitrification and indirect N2O emissions are produced from atmospheric after re-deposition and leaching processes from managed soils.Since there has been no regional N2O emission factor for paddy cultivation in the MKD, China’s direct N2O EF  for rice production, and the FAO N2O EF are available for reference.However, China’s EF does not provide an indirect EF for rice production.Thus, we used the FAO N2O EF for synthetic fertilizer use for both direct and indirect emissions.Table 7 presents the possible reduction of N2O emissions by rice groups when the slacks of N fertilizer were reduced by 100% and 50%.In the SA season, it was estimated that if the slacks of N could possibly be reduced by 100%, AR, HR, and MR households can mitigate N2O emissions by 62, 74.2 and 70.7 kg/km2, respectively.If farmers can only reduce 50% of slacks, those figures will be 31, 37 and 35.35 kg/km2 of N2O, respectively, for the three rice groups.In the AW season, although the N application rate was similar to the SA season, the AR group had a larger slack of N.Thus, there was no need to apply approximately 100 kg of N per ha for AR production in the AW season.If it is possible to minimize 100% of N slacks, AR, HR and MR paddy producers should be able to decrease the amount of N2O at 73, 64.7 and 66.3 kg/km2, respectively.Regarding mitigation strategies, it is important to apply nitrogen fertilizer following these four principal management factors: right source at the right rate, right time, and right place.These 4Rs should be used together in a comprehensive plan appropriate for the cropping system and also account for all sources of nitrogen input to crop fields.If the 4Rs are practiced and monitored well, they will increase crop yield and profitability, while also greatly reducing GHG emissions.Information on the 20 most super-efficient paddy farms is presented in Table 8.The AW season with advantageous weather conditions helped households reach much higher super-efficiency scores.Households with super-SBM scores > 4 were super-efficient DMUs and could be considered as outliers in the sample.There were households that appeared to be super-efficient in the two seasons with their unchanged AR and HR varieties, including BL032, BL034, and BL036 with Nang Hoa 9 variety; AG078 and DT032 with OM5451 variety; and BL028 and DT100 with Dai Thom 8 variety.BL032, BL034, and BL036 were cooperative members, which is why they were confident of the support from cooperatives when producing AR in the wet and rainy seasons.Although both AG078 and DT032 produced the same OM5451 variety, their products were sold at very dif-ferent prices during the two seasons.AG078 sold their paddy to traders at a price of 4950 VND/kg and 5200 VND/kg in the SA and AW seasons, respectively.DT032, with a pre-arranged farming contract, could obtain much higher prices at 6100 and 6200 VND/kg in the two seasons.Moreover, AG078, DT032, BL028, and DT100 all belonged to local paddy cooperatives.In addition, some households preferred to change their rice varieties when the AW season began.They were: AG062 and CT027 changed from OM5451 to Dai Thom 8; BL081 changed from Nang Hoa 9 to Dai Thom 8; BL047 changed from Dai Thom 8 to Nang Hoa 9; and CT019 changed from sticky rice CK92 to Dai Thom 8.

Among those households that practiced variety switching and were still super-efficient, household CT019 produced sticky rice CK92 in the SA season and was even more successful when changing to AR Dai Thom 8 in the AW season, with super-SBM scores of 3.77 and 6.48, respectively.This was good evidence of efficient farming owing to their cooperative membership and 1M5R package application in the field.Household BL047 changed from Dai Thom 8 in the SA season to Nang Hoa 9 in the AW season because of the contract between the enterprise and farmer.Thus, the selling price of paddy rice was also very high at 6300 VND/kg and 7200 VND/kg in the SA and AW seasons, respectively.Households AG062 and CT027 changed from OM5451 to Dai Thom 8 because they wanted to obtain higher output prices from individual traders at harvest.The appearance of Dai Thom 8 and OM5451 varieties in the list of super-efficient households confirmed their super-efficiency when grown in different ecological zones in the MKD.In particular,vertical grow tables aromatic Dai Thom 8 and Nang Hoa 9 varieties cultivated by households in Bac Lieu province appeared to be strongly adaptive to the environment and continued to be efficient overall.Thus, producing AR and HR using CSA techniques can not only bring economic benefits to farmers but can also protect the environment and mitigate GWP by reducing GHG emissions.The farming area and quantity of MR should strictly follow the government’s recommendations.Agricultural businesses face a vast number of simultaneous challenges.Shrinking marginals, complicated pan-European regulations and external, as well as internal, demands to mitigate their environmental footprint are all examples of requirements to be met.As a response, several different techniques are proposed to meet the needs of farmers.Even though farming has been developing technologically for centuries, the 21st century offers a wide range of technological possibilities that could deeply affect the future of farming.One of them is Artificial Intelligence.AI is to a large extent responsible for the smartness in smart farming.The term ‘smart farming’ constitutes a wide scope and demanding expectations.In this paper, smart farming is defined as the system of data driven tools for decision support in one or several parts of a farm’s production, not restricted to nor limited by the agricultural sector they belong to.Smart farming could enable increased yield volumes, mitigate the workload for farmers, contribute to climate change adaptation and future-proof farming for the coming centuries.With this in mind, smart farming is expected to affect several areas within the agricultural sector.To mention a few, some trained AI models are implemented to predict the optimal time for planting and harvesting crops, prevent nutrient deficiencies and the spread of diseases, and guarantee food safety.

Contrary to most earlier research, this study investigates technical aspects as well as non-technical aspects of smart farming.Several earlier studies have scrutinized technical aspects such as optimal remote sensing picture resolution and important cyber security aspects to sensor systems.Here, those aspects are considered but other essential, practical aspects such as data ownership and data sharing are also analyzed.Furthermore, non-technical aspects to smart farming, for instance trust and profitability, are discussed.By this interdisciplinary approach, new insights into the possible application of AI in agriculture are provided.Additionally, the wide scope allows for a comparison between three different agricultural sectors: arable farming, milk production and beef production.The study took place in Sweden, but a large part of the findings may be relevant in an international context as well.Thereafter, qualitative data was collected through a semi-structured interview study examining how different agricultural stakeholders regard smart farming technology.All interviews took place in the first half of 2021.In total, 21 respondents were interviewed from different parts of the agricultural sector in Sweden.Table 1 shows an overview of the respondents.The respondents were grouped by their occupation, where ten are individual farmers , seven are people working at commercial enterprises and cooperatives in the agricultural sector , two are researchers at Swedish research institutes and the remaining two work at a governmental agency.Out of the seven respondents from commercial enterprises and cooperatives, four come from organisations that are deemed to have some economic interest in the agricultural sector, although they also have a cooperative function.Two other respondents in the same categorization are strictly cooperatives, one with ties to the public sector.The final is a certification organ.The interviewed researchers are hired by a Swedish university with agricultural focus, and the respondents from the governmental agency are also tied to different functions within a Swedish authority focused on agriculture.The interview questions are included in the Supplementary Materials.The lists of questions were created prior to the interviews and have not been altered.different lists of questions were used for farmers and for organizations, companies, governmental authorities and researchers.For farmers, a different set of questions was used for those farmers familiar with the concept of smart farming , vs.those farmers not familiar with the concept of smart farming.For organizations, companies, governmental authorities and researchers, only one list of questions was used, since all were selected based on their experience with smart farming.