Flood vulnerability is an important factor to consider when assessing flood risk and assessing damage. It’s difficult to quantify flood vulnerability because it depends on a variety of factors, including social, economic, environmental, and physical factors. The selection of indicators is the first step in any indicator-based vulnerability assessment. The vulnerability index system has been used to assess flood vulnerability in the Dingapota Haor area. [9] introduced the most advanced and reliable method, the “Flood Vulnerability Index (FVI)” to quantify the vulnerability of floods for an area. The general formula for FVI is calculated by classifying the component into three groups of indicators: exposure (R), susceptibility (S), and resilience (R). The general formula for FVI is calculated by classifying the component into three groups of indicators: exposure (E), susceptibility (S), and resilience (R), Vulnerability   ( V ) =   Exposure   ( E ) / times Susceptibility   ( S ) Resilience   ( R ) ([10]).

With regard to indicators, this equation becomes the following one ([9]).

The total FVI of Dingapota Haor is the sum of these four indicators based on FVI.

This index value indicates the extent of vulnerability. According to [9], the Flood Vulnerability Index value 0.01 indicates very small vulnerability to floods, 0.01 - 0.25: small vulnerability to floods, 0.25 - 0.50: vulnerable to floods, 0.50 - 0.75: high vulnerability to floods, 0.75 - 1.00: very high vulnerability. The general formula for FVI is calculated by classifying the component into three groups of indicators: exposure (E), susceptibility (S), and resilience (R),

With regard to indicators, this equation becomes the following one ([9])

The total FVI of Dingapota Haor is the sum of these four indicators based on FVI (Equations (ii) - (vi)).

Primary data were collected from the Gaglajur, Tetulia, and Suair union households. Similarly, secondary data are collected from the Mohonganj Upzilla Parishad; the Local Government Engineering Department, Mohonganj; the Water Development Board, Netrokona; the Local Public Representative of Gaglajur, Tetulia, and Suair Union; and various web portals, and published research papers. 375 respondents/households were selected through a purposive random sampling method from 12,040 households in the study area. A single person was selected from every household (HH). The head of the family and the oldest person were selected for the questionnaire survey. Primary data were collected from the study area mainly by following two ways: (A) Questionnaire Development and Field testing, and (B) Focused Group Discussions (FGD).

A)QuestionnairesDevelopedand Field Survey

A questionnaire was designed by the authors, presented, and administered at the household level to obtain primary data. The survey was conducted with all classes of people, such as farmers, fishermen, day laborers, and the chairmen of the studied union, through face-to-face interviews to collect information.

B)FocusedGroupDiscussions(FGDs)

A focus group discussion was arranged to collect flood-disaster-related data, and flood vulnerability was assessed at the community level. About ten (10) FGDs were conducted at the selected research sites. The data regarding the frequency of floods, loss and damage due to flood, rainfall, size of the population, Dikes/levees of the haor area, storage capacity, and runoff of the river basin were mainly collected by the FGDs.

Field experiments were conducted at Showair and Tetulia Unions of Mohanganj Upazila of Netrokona district, near the Dingaputa Haor. The experiment was designed with a Randomized Complete Block Design (RCBD), having four (4) treatments, each replicated 3 times. There were twelve (12) plots, each with an area of 48 m². The selected soil amendments, Azolla compost (2 t/ha) and Oystershell powder, were applied in selected field plots one week prior to rice transplanting, and Cyanobacteria (Spirulina) were inoculated (100 ml/plot) in field plots after one week of rice transplanting. The rice cultivar BRRI Dhan-88 is cultivated for the boro season, and BRRI Dhan 39 for the Aman growing season. Composition of azolla compost: Total Carbon 43.6%, T-N 3.10%, C/N 14.06, T-P 0.65, T-K 1.5%. The characteristics of the microalgae are as follows: algal cell density of 8.0 × 106 cells mL⁻¹, chlorophyll content of 4.12 mg L⁻¹, pH value of 7.09, total carbon concentration of 359.70 mg L⁻¹, total nitrogen concentration of 505.55 mg L⁻¹, and total phosphorus concentration of 12.39 mg L⁻¹.

Experimentaltreatments:

28-day-old rice seedlings of cultivar BRRI Dhan-88were transplanted in the field at 25 cm × 25 cm spacing with two seedlings hill⁻¹. Ten days after rice transplanting in the field, ducklings (20-day-old) were released in the plots at the rate of 5 birds/ plot. Ducklings were kept in the plots for 4 hours a day (1st week), then allowed to remain in the plots from morning to evening and removed from the rice fields when the plants reached the flowering stage.

Rice plants, Tiller no/hill, Panicles/hill, grain/hill, grain yield/plot, etc., were randomly measured prior to harvesting. The ducklings were reared in the experimental plot, and growth was measured at weekly intervals using a digital weighing balance. Fish reared in the experimental plots were periodically sampled, and their body weights were measured using a digital weighing balance at two-week intervals during the experimental period. At each sampling event, fish were carefully collected using a hand net to minimize stress and injury. After weighing, the fish were immediately released back into their respective plots to continue normal growth. The collected data were used to evaluate growth performance, including weight gain under different treatment conditions. The specific growth rate of fish and ducklings was calculated to assess growth performance during the experimental period. SGR was determined using the natural logarithm of initial and final body weights over a given time interval.

here, w2 = the final live body weight (g) at time T2 (day), w1 = the initial live body weight (g) at time, t = Time intervals, T2 = time duration at the end of the experiment, T1 = initial time of the experiment (day).

The BCR is a relative measure, used to compare benefit per cost unit. The BCR estimated gross returns and gross costs as a ratio. The formula for measuring BCR is shown below:

Gas samples were collected by the modified closed-chamber method ([35]; [5]) during rice cultivation. Gas samples were collected once a week, starting from 21 DAT until rice harvesting, to get the average CH₄ emissions during the cropping season. During gas sampling, a glass chamber was placed over the rice plants in the middle of the field plot. Gas samples were collected by a 50 ml air-tight syringe at 0 min, 15 min, and 30 min intervals after chamber placement over the rice-planted plot. The samples were analyzed to determine the concentration of CH₄ gas by Gas Chromatograph (Shimadzu/GC 2014, Japan) equipped with a Flame Ionization Detector (FID). The analysis column was a stainless-steel column packed with Porapak NQ (Q 80 - 100 mesh). The temperatures of the column, injector, and detector were set to 100˚C, 200˚C, and 200˚C.

CH₄ emission from an irrigated rice field was calculated from the increase in CH₄ concentrations per unit surface area of the chamber for a specific time interval. A closed chamber equation ([35]; [5]) was used to estimate CH₄ fluxes from each treatment.

where, F= CH₄ flux (mg CH₄ m² hr⁻¹), ρ = gas density (0.714 mg cm⁻³), V = volume of chamber (m³), A = surface area of chamber (m²), H = height of the chamber (m), Δc/Δt = rate of increase of CH₄ gas concentration in the Chamber (mg m⁻³ hr⁻¹), T (absolute temperature) = 273 + mean temperature in chamber (˚C).

The CH₄ emissions data were correlated with and interpreted in relation to plant growth, yield, soil properties, and environmental factors. The seasonal cumulative CH₄ flux for the entire cropping period was computed as reported by [38]: Seasonal CH₄ flux = ∑n_i = (R_i × D_i).

EstimationofGlobalWarmingPotentials(GWPs)

In this study, we used the IPCC factors to calculate the combined GWP for 100 years, GWP = 27 × CH₄ (kg CO₂-equivalents ha⁻¹) + 273 × N₂O (kg CO₂-equivalents ha⁻¹) ([24]). In addition, the greenhouse gas intensity (GHGI) was calculated by dividing GWP by rice grain yield ([30]).

Water Samples were collected from each experimental plot under the four treatments at regular intervals throughout the study period. Water samples were collected from 10 cm below the surface using clean, labeled plastic bottles. Soil redox potential (Eh), flood water pH, EC, TDS, total dissolved Fe (iron) conc. and DO conc. were measured at every week interval during rice cultivation. Nitrate NO 3 − concentration in water samples was determined at 410 nm using a UV spectrophotometer (Brucine-sulfanilic acid method). Ammonium ( NH 4 + ) concentration in water samples was determined by the Indophenol blue method. Dissolved iron concentration was measured by the 1, 10 Phenanthroline method.

After rice harvesting, soil organic carbon (Walkley and Black method; [8]), total N (Micro-Kjeldahl method), available P (Colorimetric method, [33]), and available S (by the calcium chloride (0.15%) extraction method) were determined following standard methods. Exchangeable calcium (Ca), sodium (Na), and potassium (K) were extracted from soil using 1 M CH₃COONH₄ solution, and their concentrations in the extract were directly determined by Flame Photometer (Model: FP 902 PG Instrument).

Statistical analyses were performed with both MS Excel and IBM SPSS. ANOVA and DMRT analyses were computed with SPSS. The figures were generated using SigmaPlot 16.0. ArcGIS 10.8 software was used to prepare the study area map.

Flood vulnerability is an important factor for assessing flood risk and associated damage. Flood vulnerability depends on social, economic, environmental, and physical components.

The FVI (social), FVI (economic), FVI (environmental), and FVI (physical) were calculated through the equation for assessing the flood vulnerability Index. The social value was 0.22, the economic value was 0.04, the Environmental value was 0.68, and the Physical value was 0.01 (Table 1). The sum of the four components revealed an FVI of 0.95, indicating high flood vulnerability due to frequent exposure to flooding, significant susceptibility to flood impacts, and limited adaptive capacity to respond and recover. The flood vulnerability index (FVI) for the Dingapota Haor area was estimated at 0.95 (Table 1), indicating very high vulnerability to floods. According to [9], an FVI value between 0.75 and 1 represents high flood vulnerability because it reflects a combination of frequent exposure to flooding, significant susceptibility to flood impacts, and limited adaptive capacity to respond and recover. It has been reported that the FVI value for the Hatia union was 0.703, indicating that this area is highly vulnerable to flooding ([31]); consequently, the agricultural farming system may be severely hampered.

Table 1. Dingapota haor FVI scale indicators and their value.

According to the respondent’s answers, moderate floods occur in some areas of the Dingapota haor every year, but devastating floods occurred in 1984, 1988, 1996, 2002, 2004, 2008, 2012, 2014, 2017, and 2022. Flood disaster vulnerabilities affected agricultural productivity, such as maximum damage and loss occurred for Boro rice and vegetation, fisheries, house and property, livestock, human, roads, and other infrastructure, respectively, as shown in Figure 1.

Natural CapitalVulnerabilityIndex Assessment

Dingapota Haor was mostly vulnerable because the natural conditions were very fragile. The land vulnerability of agricultural resources focused on land availability, use, and exposure to submergence. The average per-household land area for agricultural activities was limited to 0.5 acres, with a vulnerability index (VI) of 0.25, indicating constraints on land availability. However, 90% of the land was used for rice cultivation, showing optimal use but with a VI of 0.90, indicating dependency on this single crop. Additionally, agricultural land remains submerged for an average of 5 months annually (VI = 0.46), reflecting moderate exposure to waterlogging. Collectively, these factors result in a land vulnerability index of 0.54, indicating moderate challenges in sustainable land use for agriculture (Table 2).

Figure 1. Loss and damage due to floods around Dingapota Haor, Mohanganj.

Table 2. Natural capital vulnerability assessment.

Water vulnerability was focused on the availability, accessibility, and quality of water for agricultural and household use. While 80% of households reported sufficient irrigation water for crops (VI = 0.80), only 60% have access to safe drinking water (VI = 0.60), exposing a significant gap in water quality. A notable 91.2% of households can easily access water sources like tube wells, yielding a VI of 0.91, highlighting good accessibility. However, 18% of HHs reported water conflicts within their community; the vulnerability index was 0.18. Overall, the water vulnerability index was 0.62, reflecting substantial vulnerability due to water quality and availability concerns (Table 2).

The biodiversity component revealed alarming trends in the ecosystem, particularly due to overexploitation and climate change. Fish diversity and fish population were decreasing continuously due to overfishing, with a VI of 0.91, highlighting severe ecological pressure and haor ecosystem degradation. In addition, most households (88%) reported decreasing frog populations in paddy fields, which also accelerated biodiversity degradation, resulting in a high biodiversity vulnerability index of 0.88, representing significant ecological risks that threaten long-term environmental sustainability (Table 2). It was noted that 88% of households were dependent on agriculture as a major income source, and the remaining households, on non-farm activities, were unfortunately affected by floods or other natural disasters. Rice-duck mixed farming provided a source of income for 70% of households. Households reported an average of nine floods that occurred in the past 30 years, with a VI of 0.53, and only 54% receive timely flood warnings (VI = 0.54). The data highlights a gradual increase in climate-related risks over the past decade, including rising floodwater (VI = 0.63), temperature (VI = 0.76), lightning and thunderstorms (VI = 0.84), hailstorms (VI = 0.58), and rainfall (VI = 0.67). Combining the indices for land, water, biodiversity, and climate variability, the overall Natural Capital Vulnerability Index was obtained as 0.67, indicating the natural capital of the Dingapota haor area is in a vulnerable condition.

PhysicalCapitalVulnerabilityIndex Assessment

The housing and assets were seriously affected by floods, as reported by 74% of HH, and the housing and assets vulnerability index was ultimately found to be 0.44 (Table 2). The agricultural machinery vulnerability index was 0.81, while the access to roads/markets and transportation facilities vulnerability index was 0.68 (Table 2). Regarding energy, the energy vulnerability index was 0.66, while the electricity access vulnerability index was 0.64 (Table 2). Overall, the Physical Capital Vulnerability Index was found to be 0.60, indicating a highly vulnerable physical condition in the selected haor area.

CH₄emissionratesandsoilredoxpotential(soilEh)duringthericecultivationperiod

CH₄ emission rates showed significant variation across the four treatments during the rice growth period in the wet Aman season (Figure 2). CH₄ flux measured at 14 to 21 days after rice transplanting was low, which increased significantly with plant growth and the development of soil reductive conditions at both locations of the rice field (Showair and Tetulia Union). In all treatments, CH₄ emissions gradually increased after transplanting, peaked at 42 DAT, and then steadily decreased until harvest. Among the treatments, the rice sole cropping system (T₁) consistently revealed higher CH₄ emissions compared to other mixed farming methods. The maximum CH₄ emission rate, 21 - 28 mg m⁻² h⁻¹, was recorded in rice sole cropping (T₁) at active tillering to early panicle initiation stage, while the CH₄ emission rate dropped sharply at rice maturation to rice harvesting stage at both locations, even though the soil redox value was low enough to produce CH₄. This decline in CH₄ emission could be related to rice plant aging and drainage of water. Similar findings were reported by [20], who found that CH₄ emission rates were low during the early stages of rice growth and gradually increased as soil reductive conditions and plant maturity increased.

Figure 2. Changes in CH₄ emission rates and soil redox potential value under rice, fish, and duck mixed farming during Aman season rice cultivation.

The rice-fish-duck mixed farming (T₄) decreased CH₄ emissions by approximately 22.37%, Rice-Duck (T₂) by 22.90%, and Rice-Fish (T₃) by 13.63% compared to the rice sole cropping (T1) system, respectively. The variation in CH₄ emission may be due to fish species and ducklings’ mobility, activities, and floodwater properties. [26] showed that rice-fish and rice-crab co-cultures reduced CH4 emissions by approximately 23% while improving yields and economic returns. [45] found that the peaks of CH₄ emission fluxes from Rice-Duck (RD) and Rice-Fish (RF) appeared at the full tillering stage and at the heading stage, and the average emission fluxes were significantly (p < 0.05) lower than those from Rice only (CK). In RD and RF, the activities of ducks and fish, such as feeding, disturb the soil, quicken the gas exchange between the soil and the atmosphere, and increase the opportunity of CH₄ emission. In addition, because plankton are consumed by ducks and fish, DO consumption in the water body by weeds and aerobic organisms is reduced, the DO content in the water body is accordingly increased, and thus, CH₄ produced in the soil could be oxidized more quickly. At the same time, the soil Eh value also increased, which inhibited the activities of methanogens and therefore decreased CH₄ production. Frei and Becker’s research showed fish activities boosted the diffusive fluxes of floodwater oxygen. Moreover, fish dropped the floodwater DO and consumed the planktons, reducing the content of floodwater DO and soil Eh value, which might therefore be a cause of higher CH₄ emissions ([18]). The changes in soil redox potential (Eh) under the rice-based farming system during the Aman season are shown in Figure 2. In all treatments, soil Eh values gradually declined after transplanting and reached highly reduced conditions of −201 mV to −217 mV at about 42 DAT. Therefore, soil redox status showed an up and down trend, and finally, before rice harvest, soil Eh value increased due to water draining out. In general, the growing period, T₂ (rice with ducklings), T₃ (rice with fish), and T₄ (rice with ducklings and fish) showed considerably greater Eh values than rice sole cropping (T₁).

CH₄ emission rates during the Boro season exhibited a comparable temporal trend across all treatments (Figure 3). CH₄ emissions increased gradually after transplantation, peaked at 70 - 77 DAT, and then steadily decreased until harvest. The maximum CH₄ emission rate occurred in the rice sole cropping system (T₁), reaching an approximate peak of 36 mg m⁻² h⁻¹ at around 70 DAT, probably due to the most intensive anaerobic soil conditions, more available C from decomposed organic materials, which accelerated methanogenic microbial activity. Comparatively lower CH₄ emissions were observed in all mixed farming systems (rice-duck, rice-fish, rice-duck-fish) than those of the rice sole cropping system (T₁). Among these, the rice-duck mixed farming (T₂) revealed the least CH₄ emissions, trailed by rice-fish-duck (T₄) and rice-fish (T₃). After reaching the peak, CH₄ emissions gradually declined towards rice harvesting stage. [4] stated that the highest CH₄ peak was observed at the flowering to milking/booting stage (77 - 91 days after rice transplanting) of rice plant. This was most probably due to the development of intensely reduced conditions, e.g., Eh value −200 mV to −230 mV in the rice rhizosphere. The maximum CH₄ emissions in rice sole cropping (T₁) may be related to flooded conditions, which create strongly anaerobic soil environments that are favorable for methanogenic microbial activity. The rice-duck (T₂) and rice-fish-duck (T₄) mixed farming system decreased CH₄ emissions to a greater extent compared to rice sole cropping and rice-fish farming. This may be due to ducks’ movement, paddling, and foraging, which disturbed the soil and enhanced oxygen diffusion into the floodwater-soil interface, thereby suppressing methanogenesis. In rice-fish (T₃) mixed farming, fish activity improved water circulation and sediment mixing, which may also increase soil aeration and reduce CH₄ production, although the reduction was lower than in duck-integrated treatments. In rice-fish-duck (T₄) mixed farming, the combined effect of ducks and fish further improved aeration and reduced anaerobic conditions, resulting in lower CH₄ emission compared with rice sole cropping (T₁). On the other side, Ducks and fish may accelerate organic matter decomposition and reduce the accumulation of CH₄-producing substrates in the soil.

[18] stated that the presence of fish in paddy fields boosted net atmospheric CH₄ emissions, perhaps through two mechanisms: 1) dropping the floodwater dissolved oxygen, thus fostering the anaerobic character of the soil environment; 2) releasing a rising portion of CH₄ entrapped in the soil via ebullition. [39] stated that the impact of rice-fish co-culture on greenhouse gas emissions remains controversial.

Figure 3. Trends of CH₄ emission rates and soil redox status (Eh) during Boro season rice, fish and duck mixed farming.

The changes in soil redox potential (Eh) under the rice-based farming system during the Boro season are shown in Figure 3. The soil redox potential (Eh) decreased gradually across all treatments from transplanting to about 70 DAT and reached a highly reduced condition of −213 to −239 mV. The soil redox status showed ups and downs between the treatments, and finally, Eh values increased gradually toward harvest. Among the treatments, sole rice cropping (T₁) generally showed the lowest Eh value (around −228 to −239 mV), while mixed farming systems with fish and ducklings (T₂, T₃, and T₄) maintained comparatively higher Eh values.

Flood water quality parameters significantly affected CH₄ emissions (Table 3). In the rice-duck, rice-fish-duck mixed farming field, water, dissolved oxygen, dissolved iron, phosphate, and nitrate concentrations were significantly higher than in the rice sole farming field plot water (Table 3), which influenced a decrease in CH₄ emissions. This is probably due to negative correlations of CH₄ emissions with the water quality parameters, e.g., DO, nitrate, phosphate, and dissolved iron. Furthermore, nitrate and iron, acting as electron acceptors, decreased methanogenesis and eventually decreased CH₄ emissions.

In both experimental locations, the water quality parameters were quite good for the aquatic living organisms, such as the fish population and ducklings’ growth (Table 3). In this study, a large amount of organic matter was formed from the submerged biomass of rice plants. The decomposition of these organic materials and the dead organisms at the bottom of the haor basin produced CH₄, CO₂, and ammonia ( NH 4 + ) gases, which diffuse from the sediment into the water column and finally to the atmosphere. Fish species must discharge CO₂ to take in fresh oxygen O₂ gas in their bloodstream, which might slow down or be badly affected under higher CO₂ concentration in an anaerobic floodwater paddy ecosystem. Significant amount of NH 4 + -N formed under the intensive reductive conditions of flood water paddy ecosystem, especially in rice sole cropping, rice duck, and rice fish duck mixed farming field plots, which also converted into nitrate-N due to movement of ducklings and fish species, thereby enhanced O₂ penetration from land surface air into flood water column, which increased DO concentration suitable for ducklings and fish population.

Table 3. Water quality parameters in rice-fish-duck mixed farming system.

**indicates significant at the 0.01 level (2-tailed). *indicates significant at the 0.05 level. NS means non-significant.

RiceDuckFishFarmingProductivity,GWP,Netprofit,andBenefitCostratio(BCR)

The effects of different mixed farming treatments on grain yield, economic return, CH₄ emissions, and global warming potential (GWP) during the Aman and Boro rice cultivation seasons are shown in Table 4 and Table 5. During the Aman season, the higher grain yield of 4194 - 4330 kg ha⁻¹ and 3858 - 3980 kg ha⁻¹ were recorded in rice-fish-duck (T₄) and rice-duck (T₂) mixed farming compared to rice monoculture (3550 - 3820 kg ha⁻¹). On average, rice yield was increased by 14.5% and 6.15% over the rice sole cropping. Similarly, in the Boro season, the maximum rice grain yield 6050 - 6230 kg ha⁻¹ was recorded in rice-fish-duck (T₄), followed by 5980 - 6460 kg ha⁻¹ in rice-duck (T2) and 5230 - 5650 kg ha⁻¹ in rice sole cropping (T₁). On average, rice yield was increased by 9.1% and 10.4% over the rice sole cropping system. The increased yield in mixed farming treatments may be due to improved nutrient recycling, soil fertility enhancement, biological pest control, and better nutrient availability resulting from fish and duck activities in the rice field ecosystem. Fish and ducks contributed organic manure through excreta and improved nutrient circulation within the system. [21] reported a higher rice yield (20%) in the rice-duck system compared to the traditional rice sole-cropping system, thereby ensuring about a 50% higher net return and rice-provisioning ability. [37] reported rice yield was increased by 57.8% in rice-duck-fish mixed farming compared to rice sole cropping. [46] also showed that the integrated rice-animal co-culture system, such as fish, frogs, or crayfish raised in paddy fields, increased rice yield by 7.8%, enhanced soil and water nutrients retention, and provided sustainable animal protein while optimizing land and water use.

Table 4. Rice-based mixed farming productivity, net profit, and benefit cost ratio (Two Years Mean Data, Aman Season).

Table 5. Rice-based mixed farming productivity, net profit, and benefit cost ratio (Two Years Mean Data, Boro Season).

In this study, the total cumulative CH₄ flux was recorded as 198.7 - 208 kg ha⁻¹ season⁻¹ during aman growing season, in rice sole cropping (T₁), while 213 - 228 kg ha⁻¹ season⁻¹ during the Boro rice cultivation. The total seasonal CH₄ flux was decreased by 19.5%, 15.4%, and 20.0% under rice-duck, rice-fish, and rice-duck-fish mixed farming, respectively, compared to rice sole cropping in the boro growing season. Similarly, seasonal CH₄ fluxes decreased by 15.7%, 7.4%, and 16.0% under rice-duck, rice-fish, and rice-duck-fish mixed farming, respectively, compared to rice sole cropping in the Aman growing season. Reduced CH₄ emissions in integrated systems might result from the movement of fish and ducks, which disrupted the soil and enhanced oxygen diffusion in flooded areas, consequently inhibiting anaerobic methanogenic activity. [14] reported that rice-duck co-culture decreased CH₄ emissions by 8.8% - 16.7% while maintaining or increasing rice yield. [43] stated that rice-duck systems substantially mitigated greenhouse gas emissions by reducing CH₄ and CO₂, leading to a 19% - 25% decline in global warming potential without compromising rice yield. [4] also reported that rice-duck mixed farming decreased seasonal cumulative CH₄ emissions by 18.0% - 24.0%, while increasing rice yield by 19% - 22.0% compared to rice sole cropping. [28] reported that microalgae biofertilizer combined with a reduced amount of chemical fertilizer significantly decreased CH₄ emissions, the GWP, and GHGI, while increasing the rice yield. Azolla cyanobacterial mixture has been used as biofertilizer to supplement the N demand of the rice crop through partial replacement of the costly chemical N fertilizer under conditions of sustainable agriculture, while its effect on CH₄ and N₂O emissions reduction has been reported by [11], [34], [7], and [25].

The maximum GWP value calculated is 5364 - 5632 kg CO₂ eq. ha⁻¹ for rice sole cropping (T₁) during Aman season, while 5751 - 6156 kg CO₂ eq. ha⁻¹ for the Boro rice growing season. Rice-based mixed farming treatments significantly decreased the GWPs’ value. The Rice-Duck and Rice-Fish-Duck mixed farming systems reduced GWPs by 15% - 19% and 16% - 20%. [39] reported that the rice-crayfish and rice-duck modes significantly decreased GWP by 18.0% and 11.0%, respectively, whereas the rice-fish mode enhanced the GWP by 20.8%. [17] revealed the superiority of the rice-duck co-culture system over the traditional rice monoculture in China by enhancing agricultural sustainability, improving the rural economy and sustainable diets, and, above all, reducing GHG emissions and the overall carbon footprint (9934.0 vs 10875.8 kg CO₂ e/hm2) per hectare of land. In the rice-fish-duck system, the activities of fish and ducks agitated the water, loosened the soil, significantly increased dissolved oxygen content ([41]), greatly reduced soil reductant content, and increased the redox potential. Therefore, the emission of CH₄ was reduced, and the control effect on the peak period of CH₄ emission from the paddy field is the most obvious ([27]).

Economic analysis showed significant variations among rice-based mixed farming practices. During the Boro season, rice-fish-duck (T₄) mixed farming yielded the maximum benefit-cost ratio of 1.88, followed by 1.82, 1.71, and 1.42 for rice-duck (T₂), rice-fish (T₃), and rice sole cropping (T₁), respectively. Comparable patterns were found in the Aman season, where rice-fish-duck (T₄) farming revealed the maximum benefit-cost ratio (BCR) of 1.76, while rice sole cropping

Table 6. Correlation of CH₄ emissions with flood water properties.

*Correlation is significant at the 0.05 level (2-tailed); **Correlation is significant at the 0.01 level (2-tailed).

(T₁) resulted in the least net return with the lowest BCR value of 1.03. These findings suggest that integrated rice-fish-duck farming is more economically feasible than traditional rice monoculture since farmers earn extra income from fish and duck rearing. [43] found that the economic benefits of the rice-fish-duck symbiosis model increased by 17.2% compared with the rice-fish symbiosis model. [29] showed that the rice-fish-duck model also showed higher economic benefits than the rice-fish model and the rice-duck model, with an increase in income by 32.9% and 229.0%, respectively. The integrated rice-fish-duck farming systems improved grain yield and economic profitability while decreasing seasonal cumulative CH₄ emissions. [4] reported an increased net return of Tk. 54,432 - 57,308 with a BCR value of 2.14 from rice duck mixed farming compared to rice sole cropping (net return Tk. 14,686, BCR 1.46) across Dingapota haor. It has also been shown that the net profit is Rs. 3.1 lakh/ha and Rs. 1.56 lakh/ha water area from integrated fish cum duck farming and fish traditional farming, respectively ([36]). In this study, rice-fish-duck mixed farming and rice-duck farming revealed significantly higher productivity and net return compared to rice sole cropping and rice fish mixed farming. Similar results were also reported by [37]. In this study, seasonal cumulative CH₄ emissions were positively correlated with flood water pH, whereas negative correlations were observed with DO, EC, TDS, nitrate, dissolved Fe, Eh, ammonium, and phosphate contents (Table 6), being supported by our previous research studies ([5], 2015). [32] recorded significantly higher rice equivalent yield 7.74 t ha⁻¹, p < 0.001 in Rice-Fish-Duck, followed by 5.48 t ha⁻¹, p < 0.005 in Rice-Duck and 5.34 t ha⁻¹, p < 0.005 in Rice-Fish as compared to rice alone 3.81 t ha⁻¹.