Integrated farming as an adaptation strategy to climate change and variability in frequently drought prone regions of Southern India

1 Introduction

Agriculture is one of the climate-exposed sectors where economic damages from climate change have been detected in IPCC (2023) and agricultural growth in rainfed ecosystems is even more impacted by climate-related risks (Jain et al., 2025; Bora et al., 2025). Increasing weather and climate extreme events have exposed millions of people to severe food insecurity and reduced water security, with the largest adverse impacts in Africa, Asia, Central and South America, Small Islands and the Arctic, and globally for small-scale food producers and low-income households (IPCC, 2023).

Extreme temperatures, decrease in rainy days, and frequent droughts are the predominant climate change induced challenges in India (Rao et al., 2016; Singh and Bhargavi, 2024). As a result, small holders have suffered from crop failures, increased vulnerability, and ultimately the loss of livelihoods (Singh and Sinha, 2019). Climate risk is usually manifested in drylands as an incidence of droughts and high intra-season variability in rainfall. India, one of the most drought prone countries of the world with about 53% of its geographical area is under arid and semi-arid regions that frequently experience drought (Samuel et al., 2022). In India, vulnerability to climate change and variability was assessed district wise and the results indicated that the districts with higher levels of vulnerability are located in the largely arid and semi-arid districts in western and peninsular India (Rama Rao et al., 2019). In the semi-arid region of India, climate change-induced extended dry spells, altered crop growing seasons, reduced species variety, and changes in the onset and end of the crop growing season (Bal et al., 2022) have the potential to harm natural resources (Todmal, 2019). Rainfed agriculture in India accounts for about 51% (72.98 m ha) of the net cultivated area and supports 40% of food grain production and 60% of the livestock population, which mainly depends on monsoon rainfall ranging from 750 to 2,000 mm (Rao et al., 2015). This is also mainly monsoon-reliant and risk-prone, and it often encounters extreme variations in rainfall, resulting in a wide range of fluctuations. The heavy reliance on erratic monsoon rains has left nearly 50% of rainfed farmers with low climate resilience, significantly more pronounced than that of their irrigated counterparts (Murthy, 2019). A dryland farm household must manage different types of risks which affect the productivity of crops and livestock and incomes, leading to food insecurity. Research shows that among production risks farmers perceive that drought is the primary risk they often encounter (Rama Rao et al., 2020; Kurup et al., 2021). In India, the estimated loss in agriculture by 2030 due to droughts would be more than US$7 billion and these losses can be mitigated by 80% through the adoption of climate resilient technologies and adaptation strategies (Sutton et al., 2013; Samuel et al., 2022).

Small and marginal landholders in India account for 86% of the total landholders, and per capita availability of land is 0.1 ha in 2020 and projected to reduce further to less than 0.1 ha by 2030 (Agriculture Census, 2016). Small-scale farms are considered the backbone of the agricultural sector, contributing significantly to food security, rural livelihoods, and employment generation in India (Lowder et al., 2021). In addition to supporting rural livelihoods, small farms play a role in food security by generating a significant portion of India's staples, including wheat, rice, and pulses. Their production accounts for over 50% of these items (Prusty et al., 2025). Despite such importance, small farming systems encounter bottlenecks and challenges that impact their productivity, economic viability, and sustainability (Srinidhi et al., 2025). Small farming systems have (1) limited access to irrigation resources limiting their resilience and productivity, especially in drylands (Kumar et al., 2014); (2) restricted access to institutional credit often compels small landholders to depend upon high-interest informal lenders, thereby constraining their investment in agricultural inputs and technology; and (3) constrained market access due to inadequate infrastructure and price fluctuations that often result in the sale of their produce at minimal prices (Sharma and Agarwal, 2018).

In this context, the adoption and effective integration of multiple climate adaptation strategies at the farm level play a pivotal role in improving the resilience, especially in rainfed production systems characterized by frequent and high climatic risks. Climate-resilient interventions (CRIs) ensures enhanced agricultural productivity in a sustainable and inclusive manner while directing the agricultural sector to adapt to current and future climate changes, and strengthen the resilience (Teklewold et al., 2019; Prusty et al., 2022). Strengthening of the existing farming systems especially with diversification of crops, livestock and associated management practices contributed to better sustainable livelihood security plus improved ecological security, economic efficiency and social equity of households (Behera et al., 2018; Ravisankar et al., 2022). The intrinsic variability of farming systems frequently results in a reduced uptake of high-potential technologies, as these do not align well with the heterogeneous nature of smallholder systems, which necessitate tailored technological solutions. It is suggested in this scenario that studies be conducted to characterize the current farming systems, pinpoint production constraints, and improve production and profitability with interventions that are climate resilient, socially acceptable and cost-effective.

Numerous studies were conducted to define and characterize existing farming systems (Haileslassie et al., 2016; Alvarez et al., 2018), constraint analysis of existing farming systems (Innazent et al., 2022) and assessment of the impact of targeted interventions in farming systems approach in irrigated ecosystems (Kaur et al., 2021). Only a few studies (Kumar, 2019; Samuel et al., 2024; Prasad et al., 2024) have assessed the impact of climate-resilient interventions (CRIs) in rainfed ecosystems, and these evaluations were largely based on secondary data derived from household surveys. Moreover, most studies focused on isolated interventions, such as the adoption of improved crop varieties over traditional ones or the use of improved livestock feed in place of conventional feed. However, comprehensive impact assessments of CRIs from a whole-farm perspective under on-farm research conditions remain scarce. In this context, the present study is distinct in its focus on evaluating integrated CRIs specifically tailored for marginal and smallholder farmers in rainfed ecosystems under real farm conditions. Interventions such as improved, drought tolerant, high yielding cultivars, intercropping systems, in-situ moisture conservation furrow, rainwater harvesting and its effective utilization through drip and sprinkler systems, nutrient sprays, integrated-crop livestock systems, fodder production and conservation, etc. can potentially minimize the impact of drought and enhance farm productivity and profitability (Srinivasarao et al., 2016; Srinivasarao and Gopinath, 2016; Chary et al., 2019; Hansen et al., 2019; Diallo et al., 2020; Bhargavi et al., 2025).

Resilience in agricultural systems is multidimensional and extends beyond yield and income stability to include absorptive, adaptive, and transformative capacities of farming households in responding to climatic stresses (FAO, 2016; Meuwissen et al., 2019). In the present study, resilience was assessed not only through improvements in productivity and economic returns, but also through the capacity of farming systems to withstand drought impacts, recover from climatic shocks, and sustain production under variable rainfall conditions through diversification and climate-resilient interventions. The selected indicators such as system productivity, sustainable yield index, economic returns, resilience score and resilience gain were used as practical and measurable proxies to evaluate the adaptive performance of integrated farming systems under real farm conditions. These indicators collectively reflect the ability of farming households to minimize production losses, improve livelihood security, and enhance long-term sustainability in drought-prone rainfed ecosystems (Folke et al., 2010; Meuwissen et al., 2019; Prasad et al., 2023, 2024). Rainfed agriculture is characterized by frequent droughts, uncertain rainfall intensity and distribution which in turn hampering the resilience and sustainability of small farming systems in India. Strengthening the farming systems with interventions in whole farm perspective have significant potential to improve productivity, profitability, livelihood security, and resilience under climate stress conditions. Previous studies mainly focused on farming systems characterization, and evaluation of individual interventions using secondary household survey data. However, comprehensive on-farm assessment of multiple interventions in rainfed integrated farming systems with respect to resilience and sustainability remain limited. Our study demonstrates a methodology for prioritizing, assessing the impact of potential interventions and comparing the climate resilience of farming systems' case studies from various contexts. We believe that this study will be useful to researchers and policymakers engaged in agriculture, food security, sustainable livelihoods and water management in semi-arid areas, particularly in the context of climate change.

2 Methodology

2.1 Location

The study was undertaken to evaluate the impact of CRIs in Karnataka, a predominantly semi-arid state in South India (KSNDMC, 2017) through multi stage sampling procedure. The study district Tumkur was purposively selected based on the vulnerability to drought and it is one of the most vulnerable semi-arid districts in India (Samuel et al., 2024). In Tumkur district, two rainfed blocks, i.e., Tumkur and koratagere, were randomly chosen based on the agroecological representation and predominance of rainfed crops (Rangegowda, 2014). From each block, three villages were randomly selected by using random number table. In total six villages (three villages from each block) were randomly selected for the study. The study villages were drought prone areas predominantly occupied by small and marginal farming communities; Dominated with rainfed crops viz., finger millet, groundnut and pigeonpea. The undulating topography subjected to quick runoff, soil erosion, less adoption of soil and water conservation measures, less intensified cropping systems, poor water harvesting structures were the major constraints observed in the selected area. The geographical location details of the project site were depicted in Figure 1.

A comprehensive list of all farm households in the selected villages was prepared, categorized by farm size: marginal farmers: up to 1 ha, small farmers: 1–2 ha, medium farmers: 2–4 ha and large farmers: morethan 4 ha. Households were further classified based on irrigation status: completely rainfed, less than 30% area under irrigation, more than 30% area under irrigation. From the prepared list of farm households, a total of 40 farmers were randomly selected in each village. This included 20 farmers practicing exclusively rainfed farming and 20 farmers with less than 30% of their net cropped area under irrigation (partially irrigated). Within rainfed and partially irrigated groups, farmers were selected proportionally from the marginal, small, and medium farm categories based on their relative representation in the village. For example, if marginal, small, and medium rainfed-only farmers in a village are in the ratio 40:40:20, then the selected sample of 20 farmers would include eight marginal, eight small, and four medium-sized farms. Thus, a total of 240 farm households were taken under consideration from six selected villages for the study.

2.2 Rainfall trend in the study area

The rainfall pattern and distribution of the study villages during 2019–2023 are presented in Table 1 and Figures 2A, B. The normal rainfall of the study area during the crop-growing period (June–December) was approximately 832 mm, with the major share received during the southwest monsoon season (June–September). The rainfall distribution during the study period was highly variable. The years 2019 and 2020 received near-normal rainfall, whereas 2021 and 2022 recorded excess rainfall. In contrast, 2023 experienced drought conditions with prolonged dry spells during critical crop growth stages, particularly in August and September, adversely affecting crop productivity. The June–November season of 2023 was the fifth driest season during the 35-year period (1990–2025), with an SPI6 value of −1.03 (Figure 2B), indicating moderate drought conditions. Among the study years (2019–2023), 2023 recorded the lowest SPI6 value and was the driest year; therefore, it was considered the drought year for the present study. These rainfall variability across the years indicate that although total annual rainfall varied moderately across years, temporal variability and uneven rainfall distribution posed major challenges to agricultural productivity and livelihood stability in the study region. Long-term rainfall data of Tumkur indicate that the district experience 3–4 drought years in every 10-year period. Of these, two to three are in moderate and one or two may be of severe intensity (KSNDMC, 2020; Kumar et al., 2022).

Table 1. Rainfall distribution in the study villages.

2.3 Description of the participatory farming systems approach and CRIs

2.3.1 Benchmarking the farming systems

The study followed a participatory farming systems research approach involving four major steps: (1) benchmarking of existing farming systems; (2) classification of farm households into relatively homogeneous farm types; (3) identification of production constraints and suitable climate resilient interventions (CRIs); and (4) on-farm evaluation of CRIs.

A benchmark household survey was conducted during 2018 across six selected villages of Tumkur district to characterize the existing farming systems and generate baseline information. A total of 240 farm households were surveyed using a semi-structured questionnaire. The sampling procedure followed a multi-stage sampling approach to ensure adequate representation of different farm-size categories and irrigation situations prevailing in the study area. Information was collected on farm resources, crop and livestock enterprises, labor availability, irrigation access, input use, farm income, and major production constraints. The benchmark survey database, supported by secondary information obtained from village records and local agricultural departments, was used to identify dominant farming systems and classify households into relatively homogeneous groups based on farm size and access to irrigation facilities.

Out of the 240 households surveyed, 199 households (83%) belonged to small and marginal farmer categories. From this pool, 24 farmers were selected purposively for detailed participatory on-farm research considering their representativeness of the farming systems, willingness to participate, and continuity for long-term monitoring during the study period. The selected farmers were grouped under two farming situations: rainfed farming systems, which were entirely dependent on monsoon rainfall, partially irrigated (PI) farming systems, where lessthan 30% of the cultivated area was irrigated through borewells. Under each farming situation, farmers were further categorized into marginal and small farm households. Each category consisted of treatment and control groups for comparative evaluation of interventions. The treatment group received climate resilient interventions (CRIs), whereas the control group continued with existing farmers' practices without interventions. Under rainfed conditions, 12 farmers were selected comprising six marginal farmers (three treatment and three control) and six small farmers (three treatment and three control). Similarly, under partially irrigated conditions, 12 farmers were selected following the same distribution. The average farm size was 0.9 ha for marginal farmers and 2.0 ha for small farmers. The assignment of treatment and control groups was carried out purposively to ensure comparability between groups with respect to farm size, resource endowment, irrigation access, and farming system characteristics. Subsequently, farmers in the same village with similar agroclimatic, socioeconomic, and farming characteristics, but without exposure to the interventions, were selected as the control farmers to ensure comparability between treatment and control farmers. The baseline/benchmark values used for comparison were derived from the mean performance of farmers belonging to four categories during the initial 2018 survey, namely: rainfed marginal benchmark (Rainfed_M_B), rainfed small benchmark (Rainfed_S_B), partially irrigated marginal benchmark (PI_M_B), and partially irrigated small benchmark (PI_S_B). The abbreviated forms used in the entire manuscript are: rainfed_M_I - Rainfed marginal farmer with interventions; Rainfed_M_NI - Rainfed marginal farmer with no interventions; Rainfed_S_I - Rainfed small farmer with interventions; Rainfed_S_NI - Rainfed small farmer with no interventions; PI_M_I – Partially irrigated marginal farmer with interventions; PI_M_NI - Partially irrigated marginal farmer with no interventions; PI _S_I - Partially irrigated small farmer with interventions; PI_S_NI - Partially irrigated small farmer with no interventions. The sample selection is systematically presented in Figure 3 and Table 2.

Table 2. Sample size (farmers) of on-farm research (2019–2023).

Values in the parentheses indicate sample size (farmers) considered for the study period-2019–2023.

To enhance the productivity and profitability of these farming systems, farm type specific interventions were implemented with various IFS components/modules (Table 3). Additionally, capacity-building programs were organized to provide need-based training to farmers on improved crop cultivation techniques, fodder production, animal nutrition and healthcare, and resource management practices. These climate resilient interventions/technologies in integrated farming system approach aimed at strengthening and diversifying rainfed and partially irrigated farming systems to make them more resilient, productive, and economically viable for marginal and smallholder farmers. The CRIs adopted (Table 3) were evaluated for a period of 5 years (2019–2023) to assess their impact on system productivity, profitability, resilience, employment generation, sustainability and market dependency for inputs and outputs.

Table 3. Interventions adopted for on-farm study.

2.4 System productivity and profitability

Direct comparison of farming systems is possible only when the productivity of all IFS components is expressed in terms of a single reference crop using crop equivalent yield based on economic value. Finger millet was used as the reference crop because it is the predominant staple crop cultivated across the study area and forms the principal component of the existing farming systems. The yields of different IFS components, including crops, livestock, and horticultural enterprises, were converted into finger millet equivalent yield/system productivity using the market value of both main products and by-products. The FMEY was calculated using the following formula:

FMEY (kg)=crop/milk yield (kg )×crop/milk price (INR kg−1)finger millet price (INR kg−1)

The system profitability calculation used the partial budget analysis method that includes total variable cost, gross field benefits and net field benefits under each system. The variable costs included human labor costs for plowing, sowing, weeding, and harvesting and cost of inputs such as seed, fertilizers, pesticides, and farmyard manure (FYM). The unit of human labor was based on labor days and was calculated by recording the time required for each agricultural operation and converting them to labor days (8 h being equivalent to 1 labor day). The cost of labor was calculated using the minimum wage rate for the study years as per the Labor Law of the Government of India. Gross returns were calculated by taking the price that farmers receive for the crop/milk/produce when sold and net returns were calculated as the difference between gross returns and total variable costs. The market prices of inputs and outputs prevailing during the respective production periods of the study were considered.

2.5 Diversified farming systems' resilience

The ability of a system to adapt to change, maintaining the same structural and functional controls and producing the same outputs despite stress or turbulence is known as resilience (Prasad et al., 2024).

The year 2023 was considered as the drought year for resilience assessment because the study area experienced prolonged dry spells during critical crop growth stages, reduced number of rainy days, and deficit rainfall distribution compared to normal years. The resilience analysis was carried out by comparing the performance of farming systems with and without interventions under drought conditions relative to benchmark normal-year performance.

Two quantitative indicators, namely, resilience score and resilience gain, were used to evaluate the effectiveness of CRIs in sustaining productivity and income under climatic stress conditions following the approach suggested by Rama Rao et al. (2019) and Prasad et al. (2024). These indices provide a simple and interpretable framework for measuring the extent to which adaptation interventions minimize adverse impacts of drought.

The resilience score indicates the relative capacity of a farming system to maintain yield or income under drought conditions compared to normal conditions without adaptation. A higher resilience score reflects greater system stability and adaptive capacity under climatic stress. The resilience gain approach is particularly useful in climate adaptation studies because it estimates the avoided yield or income loss attributable to interventions during stress years. The methodology has been widely used in assessing adaptation benefits under rainfed farming systems and drought-prone regions (Birthal et al., 2015).

Resilience score(%)=(Ya1/Yn)×100

Where, Ya1 = Yield or income with adaptation under drought conditions.

Yn = Yield under normal conditions without adaptation.

The higher the proportion more is the resilience. This measure of resilience is computed for with and without adaptation measures and the difference in the resilience measure can be attributed to the adaptation of technologies.

Resilience gain(%)=[(Ya1/Yn)-(Ya0/Yn)]×100

Where, Ya0 = Yield or income without adaptation under drought conditions.

To compare the performance of the various technologies in comparison to traditional practices during drought situation, t-test assuming equal variance was carried out to test whether difference in performance with and without technologies is statistically significant or not. The t-test adopted to test the significance of the impact of technologies is as follows.

t=x1̄-x2̄S2(1n1+1n2)S2=∑i=1n1(xi-x1̄)2+∑j=1n2(xj-x2̄)2n1+n2-2

Where, x₁ and x₂ are the mean yield with adoption of intervention and with no intervention, respectively; s² is the pooled sample variance; n₁ and n₂ are the sample sizes; and t is a Student t with n₁ + n₂ – 2 degree of freedom. Regression analysis was done to assess the impact of technologies adoption over non adoption and also the interaction between the technologies and the drought stress. The regression model adopted to assess the impact of technology is as follows:

With the interaction effect, the regression model becomes:

Yi=ß0+ß1D2i+ß2D3i+ß3(D2i×D3i)+ui

Where Y_i = outcome variable (yield, income, resilience score, etc.); D₂ = farm size dummy, 1 = marginal, 0 = small; D₃ = climate resilient intervention dummy, 0 = non-beneficiary/control, 1 = beneficiary/intervention; D₂ x D₃ = interaction team; u_i = error term.

2.6 Market dependency ratio (MDR)

The market dependency ratio (MDR) can be considered as an important indicator that reflects the dependency on markets for obtaining stable income. It evaluates the material flows of inputs and outputs at market prices in order to assess the performance of the farming system. The MDR for inputs is the ratio of value of inputs generated internally to value of the total inputs used in the system and MDR for outputs is the ratio of value of outputs generated internally to value of the total outputs generated from the system (Rao et al., 2017).

2.7 Farming systems' sustainability indices

Sustainability indices show how successful will be the diversified farming over a period of time. Sustainable yield index (SYI) and sustainable value index (SVI) denotes the ability of a technology/intervention in generating minimum assured yield/returns as a per cent to the highest observed yield/returns with high probability. Under rainfed situations, maximization of minimum sustainable yield over years is very important. The sustainability of the agricultural practices is guaranteed by the higher sustainability index ratings. Two indices viz., sustainable yield index (SYI) recommended by Ram et al. (2016) and sustainable value index (SVI) suggested by Bohra and Kumar (2015) were computed in the present investigation by using the following formulae.

Sustainable Yield Index (SYI)=y -σymax

Where, y is the average yield of a system over the years, σ is standard deviation (SD) and y_max is maximum yield of a system over the years.

Sustainable Value Index (SVI)=v -σvmax

Where, v is average net profit over the years, σ is standard deviation (SD) and v_max is maximum net profit obtained in any of the year.

3 Results

3.1 Impact of CRIs on system productivity across farming systems

The CRIs adopted across farming systems contributed significantly in improving the productivity under both rainfed and partially irrigated (PI) conditions when compared to the control farmers (Figure 4 and Tables 3, 4). Among rainfed marginal households (HH), the productivity significantly increased from 4,800 to 6,892 kg HH⁻¹ yr⁻¹ (43%) with adoption of interventions, compared to benchmark level (3,716 kg HH⁻¹ yr⁻¹), the productivity enhancement was 85%. Among rainfed small farmers, with the adoption of interventions productivity increased from 6,064 kg to 10,781 kg HH⁻¹ yr⁻¹ (77%), but when compared to benchmark level (5,284 kg HH⁻¹ yr⁻¹), the productivity enhancement was 104%. A similar but more pronounced trend was evident in PI marginal farmers, where productivity rose significantly from 7,793 kg to 9,399 kg HH⁻¹ yr⁻¹ (21%) with interventions but when compared to benchmark level (5,733 kg HH⁻¹ yr⁻¹), the productivity enhancement was 64%. Among PI small farmers, with the adoption of interventions productivity increased from by 29%, but when compared to benchmark level, the productivity enhancement was 110%. Farmers under PI condition diversified with multiple enterprises particularly high income generating horticultural crops + livestock, in a small area with the limited availability of water realized higher productivity.

Table 4. Impact of CRIs adoption on system productivity across farming systems.

^***indicate statistical significance at the 1% level and ^**at 5%.

The system productivity improvement with the adoption of interventions against control farmers was more in case of small farmers in both rainfed (77%) and PI conditions (29%) compared to marginal farmers under rainfed (43%) and PI (21%).

Under rainfed conditions, the difference in the system productivity due to adoption of interventions was statistically significant across all farming systems. The difference in the systems productivity with the adoption of CRIs under rainfed conditions for small farmers was 4,716 kg HH⁻¹ yr⁻¹ and for marginal farmers was 2,092 kg HH⁻¹ yr⁻¹ (Table 4). Under PI conditions the impact of interventions resulted in 3,712 kg HH⁻¹ yr⁻¹ higher system productivity in case of small farmers and for marginal farmers it was 1,606 kg HH⁻¹ yr⁻¹. The impact of interventions was more in rainfed conditions compared to PI conditions and with respect to land holding the impact was more in small HHs compared to marginal HHs. The success of any resilient intervention/technology depends on the major resource base, i.e., land. The land size of the farmer influences the adaptive capacity of farmers.

The higher productivity of improved cultivars of crops coupled with in-situ moisture conservation furrows resulted in higher system productivity compared to that of no interventions. Improved fodder varieties viz., fodder sorghum varieties-CO-FS-29, CO-FS-31, super napier (CO-5), fodder maize (African tall) and fodder from perennial trees Melia dubia, Sesbania grandiflora during off season provided higher and year round biomass production. In addition, livestock feed interventions, multi-nutrient mixture (40 g day⁻¹ animal⁻¹), and a concentrate mix in 50:25:25 ratio (energy source, protein source, bran) at 4 kg day⁻¹ animal⁻¹ enhanced the milk yield as well as milking period. These livestock interventions enhanced the livestock productivity by 53 and 45% in case of rainfed marginal and small HHs, respectively, compared to the counter control farmers. Whereas under PI situations compared to the control, the livestock productivity of marginal and small HHs got improved by 24 and 31%, respectively (Table 5).

Table 5. Impact of livestock interventions on system productivity across farming systems.

^***indicate statistical significance at the 1% level, ^**at 5% and ^* at 10%.

Regression analysis (Table 6) showed that under rainfed conditions, without adoption of interventions the system productivity was 6,720 kg HH⁻¹ yr⁻¹. The adoption of interventions enhanced the productivity by 3,404 kg HH⁻¹ yr⁻¹ compared to control. The system productivity of marginal HHs was less by 2,576 kg HH⁻¹ yr⁻¹ compared to that of small HHs. In PI situation, without adoption of interventions the system productivity was 13,526 kg HH⁻¹ yr⁻¹ and the interventions enhanced the productivity by 2,659 kg HH⁻¹ yr⁻¹ compared to control. The system productivity for marginal HHs was less by 6,259 kg HH⁻¹ yr⁻¹ compared to that of small HHs. The diversified farming systems' productivity improvement was 51% in rainfed situation and 20% in PI condition.

Table 6. Impact of CRIs vs. land classes on system productivity across farming systems.

^***Significance at 1%.

Variables - Interventions: With interventions = 1; No interventions = 0; Land holding: Marginal = 1; Small = 0.

Rainfed Y = 6,720 + 3,404 X₁ – 2,576 X₂.

PI – Y = 13,526 + 2,659 X₁ – 6,259 X₂.

3.1.1 Contribution of different enterprises to total system productivity

The results from Figure 5 indicated that under rainfed situation, the major share of system productivity for a marginal HH and small HH from livestock component was 74 and 62%, respectively, followed by crops (23 and 35%) and horticulture (3% each). Similarly, the marginal and small HHs under PI conditions recorded more productivity from livestock component (68 and 44%), followed by crops (23 and 26%) and horticulture crops (9 and 31%). In rainfed agriculture, livestock plays a significant role as 62% of small and marginal farmers are directly involved in the livestock sector. The salient finding was that even though, the system productivity enhanced through the adoption of interventions across all the farming systems, the share of crops and horticulture components to the total productivity significantly increased indicating the impact of interventions was pronounced. The contribution of horticulture to the total system productivity was same for both marginal and small HHs under rainfed situations. Though the small farmers' land holding size was more, the proportionate productivity from horticulture was similar to that of marginal farmer. Whereas in PI condition, the system productivity share from horticulture component was higher among small farmers than marginal farmers. Due to better management of rainwater and its efficient use, helps in shifting staple/common crops to high value horticultural crops. Moreover, better harvest of rainwater and its efficient use enables replacement of the rainfed land by irrigable land which inturn helps the rural people to diversify their income sources.

3.2 Impact of interventions on system profitability across farming systems

System profitability is an important and overarching objective of any farming system. The farm size, multi-enterprise possession, adoption of CRIs were found to significantly determine the income of farmers. The diversified farming systems with the adoption of interventions, significantly improved the profitability (Figure 6). Among rainfed marginal HHs, with the adoption of interventions the mean returns increased from Indian Rupee (INR) 101,908 HH⁻¹ yr⁻¹ to INR 137,688 HH⁻¹ yr⁻¹ (35%) and for small HHs from INR 143,528 HH⁻¹ yr⁻¹ to INR 214,931 HH⁻¹ yr⁻¹ (50%), but when compared to benchmark level, the mean profitability enhancement was 82 and 85%, respectively. A similar but more pronounced trend was evident for PI small farmers, where mean profitability increased from INR 365,541 HH⁻¹ yr⁻¹ o INR 379,505 HH⁻¹ yr⁻¹ (3%), but when compared to benchmark level, the average profitability enhancement was 203%. The system profitability with the adoption of interventions against control farmers was more in case of small farmers in both rainfed (50%) and PI conditions (3%) compared to marginal farmers under rainfed (35%) and PI situation (−3%). The returns of marginal farm HH with adoption of interventions were less than the returns of the small farm HHs indicating that increase in the farm size increased the income. The better performance of improved cultivars of finger millet, pigeon pea, groundnut and field bean along with in-situ moisture conservation furrows resulted in higher system profitability. Improved fodder varieties provided higher and year round biomass production. In addition to this livestock feed interventions, multi-nutrient mixture (40 g day⁻¹ animal⁻¹), and a concentrate mix in a 50:25:25 ratio (energy source, protein source, bran) at 4 kg day⁻¹ animal⁻¹ enhanced the milk yield as well as milking period resulting in significant improvement in net returns.

Many of the researchers evaluated the economic impact of set of interventions like crop, cropping systems, livestock, and natural resource management but not in whole farm perspective. This study has important implications for the practical value of improved technologies when adopted on farmers' fields.

3.2.1 Economic indicators

The inputs and outputs were valued at the market price and the results revealed that input- output ratio was highest for all the farming systems with adoption of interventions when compared with no interventions. The input- output ratio of rainfed_M_I farming system is 2.83 indicating that the farmers were able to get INR. 2.83 for every one rupee spent compared to 2.52 of rainfed_M_NI farming system (Table 7). The lowest input-output ratio (2.08) was observed in the case of the Rainfed_S_NI. The market dependency ratio (MDR) can be considered as an important indicator that reflects the dependency on markets for obtaining stable income. The material flows were valued at the market price and market dependency ratio was calculated for different farm types. A lower value of MDR for inputs was better as it would mean that farmers incurred less expenditure on purchase of inputs. Farmers who adopted the interventions have lesser MDR for inputs compared to control farmers due to better resource recycling among the components of IFS. The farming systems that can minimize the need for external inputs have a key role in sustaining the agricultural systems in rainfed agriculture. The lower MDR for inputs (49%) was noticed with the marginal HHs with interventions compared with control farmers (57%) under rainfed conditions and under PI situation MDR for inputs (38%) was noticed with the marginal HHs with interventions compared with control farmers (69%). The similar trend was observed in case of small HHs under both rainfed and PI situation. On the other hand, a lower MDR for outputs indicates that farmers are more dependent on markets for the sale of their surplus produce. The MDR for outputs was found to be higher in the case of farmers who adopted interventions compared to control farmers across all the farming systems. The higher MDR for outputs in case of small HH compared to marginal HH is due to higher surplus produce for market sale and diversified enterprises in small farm HHs.

Table 7. Economic indicators of different farm types (1 INR = 0.011 USD).

3.3 Resilience of diversified farming systems to drought stress

In this study, two indices, resilience score and resilience gain were used for assessing the performance of climate resilient interventions during the stress year (2023) vis-‘a-vis that of the normal year (mean values of four years, i.e., 2019, 2020, 2021, and 2022 were considered as the normal year). Resilience score gives an idea about the performance of the interventions during a stress year vis-a-vis that of the farmers' practice during the normal year, which indicates the ability of the technology to withstand stress. Resilience score is deemed higher when interventions performance during stress year closely aligns with that of normal rainfall year. The Figure 7 illustrates that resilience-building interventions during drought year significantly enhanced the system yield of all the farm types. The interventions resulted in highest resilience scores under rainfed small (158%) and marginal (132%) farm HHs. This suggests that during the drought/stress year, interventions like drought-tolerant crops, intercropping systems, in situ moisture conservation furrows and livestock feed additives resulted in 158 and 132% higher yields than the yields of normal year without adoption of interventions. Practically, small farmers in these regions often face challenges related to land degradation, water scarcity, and low profitability. Thus, we emphasize that millets and pulses/oilseeds are companionable intercropping blends, which would augment the productivity of the whole system. But under PI conditions, the resilience score of small (119%) and marginal (112%) HHs was less compared to the small and marginal HHs of rainfed situation. The water availability, though limited, reduced the yield differences of horticulture crops like tomato, chili and arecanut during normal year without interventions and drought year. This would have contributed to less resilience scores under PI conditions compared to rainfed conditions.

The resilience gain indicates the extent of yield advantage due to the adoption of interventions over that of the farmers' practice without interventions during the drought year, considering the fact that the famers' practice can withstand stress to some extent. The resilience gain due to adoption of interventions for rainfed small and marginal farmers was 65 and 40%, respectively, indicating that the extent of yield improvement was 65 and 40% due to adoption of interventions during the drought year over that of no interventions in drought year. Comparatively, under partially irrigated condition the resilience gain due to adoption of interventions in marginal and small farmers fields was 26 and 21%, respectively, indicating that interventions avoided yield loss 26 and 21% as a proportion of what would have occurred in the absence of interventions in drought year. Higher performance of small farmers over marginal farmers in resilience score was likely due to lower baseline incomes and higher risk exposure, making the benefits of interventions more pronounced.

3.4 Employment opportunities through strengthening of farming systems

The data pertaining to the employment generation (Figure 8) revealed that both in rainfed and PI situations, man days for 5 years (mean of 2019–2023) were 361 for marginal HH with adoption of CRIs and 317 for marginal HH without interventions. The interventions generated employment of 13% increase over no interventions in case of marginal HH. In case of small HH, human labor requirement for 5 years (mean of 2019–2023) were 416 with interventions and 344 for control farmers generating 20% higher employment. In PI farming situations, man days for 5 years (2019–2023) was 407 for marginal HH with interventions and 326 for marginal HH without interventions, enhancing 24% man power requirement and 476 for small HH with interventions and 353 for small HH without interventions enhancing 35%. The various interventions across different components of IFS, especially livestock and horticulture required more labor which makes the farm family engaged round the year.

3.5 Sustainability indicators of the diversified farming systems

The sustainability analysis of the integrated farming system (IFS) interventions, based on sustainable yield index (SYI) and sustainable value index (SVI) revealed a marked improvement in both yield stability and economic viability across farming systems (Figure 9). SYI and SVI denotes the minimum guaranteed yield/value as a per cent to the maximum observed yield/value with high probability. In rainfed farming situations, the marginal farmer category with interventions showed a significantly higher SYI of 0.73 compared to the SYI of 0.44 without interventions. This reflects more stable and consistent yields resulting from adoption of CRIs. The corresponding SVI of the marginal farmer category with interventions was 0.69, which was also notably higher than control (0.33), indicating stronger economic sustainability too.

Similarly, for small farmers under rainfed conditions, SYI with interventions was 0.71 compared to the SYI of 0.25 without interventions. The SVI was 0.71 in the intervention group, indicating substantial improvement over the control (0.29) conditions.

In partially irrigated conditions, marginal farmers with interventions recorded SYI and SVI of 0.74 and 0.7, respectively, significantly higher than their counterparts without interventions (0.41 and 0.12, respectively). Among small farmers' category, SYI was 0.72 with interventions compared to 0.28 in case of control. The SVI of small farm HHs (0.71) also surpassed the control category (0.15) values. The enhanced indices across both categories highlight the role of integrating horticulture and livestock with cropping along with adoption of climate resilient technologies in improving farming systems' sustainability and economic returns.

4 Discussion

4.1 System productivity and profitability

The present study demonstrated that integration of climate resilient interventions (CRIs) significantly improved system productivity and profitability across both rainfed and partially irrigated farming systems. The positive response was more pronounced among small farm households compared to marginal households, indicating that farm size and resource availability influenced the effectiveness of interventions.

The increase in productivity was mainly attributed to the combined effect of improved crop cultivars, in-situ moisture conservation practices, diversified cropping systems, improved fodder production, and livestock nutritional interventions. Improved cultivars of finger millet, pigeon pea, and groundnut coupled with moisture conservation furrows enhanced crop growth and reduced moisture stress during dry spells. Similarly, improved fodder varieties such as fodder sorghum (CO-FS-29 and CO-FS-31), Super Napier (CO-5), fodder maize (African Tall), and perennial fodder trees such as Melia dubia and Sesbania grandiflora ensured year-round biomass availability, thereby improving livestock productivity.

Livestock-based interventions including supplementation with multi-nutrient mixture and concentrate feed improved milk yield and extended lactation period, which substantially contributed to overall farm income. In addition, recycling of crop residues and livestock manure reduced dependency on external inputs, thereby lowering production costs and improving profitability.

The findings of the present study are in agreement with earlier reports indicating that improved crop varieties and integrated farming approaches enhance productivity and farm income under rainfed conditions. Ogada et al. (2020) reported that improved crop cultivars contributed to surplus production and increased household income. Similarly, Lisson et al. (2010) and Prasad et al. (2023) observed that improved fodder and mineral supplementation significantly increased milk yield and economic returns. Chand et al. (2011) and Bijla (2018) also emphasized the important role of livestock in improving livelihood security and reducing rural poverty.

The higher profitability observed under treatment farming systems further confirms that integration of crop, livestock, and natural resource management interventions enhances resource-use efficiency and economic sustainability. Earlier studies mainly evaluated individual interventions such as crops, livestock, or natural resource management separately (Pradhan et al., 2018; Anantha et al., 2021; Layek et al., 2023), whereas the present study evaluated the cumulative effect of interventions from a whole-farm perspective.

4.2 Resilience of the diversified farming systems

The study clearly demonstrated that adoption of climate resilient interventions improved the resilience of diversified farming systems during drought conditions. The resilience gain and resilience score were substantially higher under treatment plots compared to control plots, particularly among rainfed small farm households. The higher resilience observed under rainfed conditions can be attributed to the combined effect of drought-tolerant crop varieties, intercropping systems, moisture conservation furrows, supplemental irrigation, and livestock feed interventions. In-situ moisture conservation practices improved soil moisture retention and reduced the adverse effects of prolonged dry spells on crop growth. Intercropping systems such as finger millet + groundnut and finger millet + pigeon pea enhanced resource-use efficiency, stabilized yields, and minimized production risk under moisture stress conditions. The improved performance of treatment plots during drought years indicates that integrated crop and livestock interventions enhanced the adaptive capacity of farming systems. Green fodder availability and livestock nutritional management also contributed to income stability during drought periods by sustaining milk production.

The findings are consistent with previous studies reporting the benefits of intercropping and moisture conservation practices in dryland agriculture. Honnatti (2015) and Sukanya et al. (2024) reported that finger millet-based intercropping systems improved productivity and resource-use efficiency. Similarly, Reddy et al. (2016) observed that finger millet + pigeon pea intercropping with moisture conservation furrows improved productivity and reduced runoff losses. Baig et al. (2013) also reported that in-situ moisture conservation practices enhanced rainwater retention and drought resilience. The present study therefore highlights the importance of integrating climate-resilient technologies into farming systems for improving yield stability and livelihood security under increasing climatic variability. Future research should extend the analysis beyond production-oriented indicators to include direct measurement of absorptive, adaptive, and transformative capacities, thereby enabling a more comprehensive assessment of resilience in integrated farming systems under climate variability and drought conditions.

4.3 Resource recycling and market dependency

The lower market dependency ratio for inputs under treatment farming systems indicated improved internal resource recycling and reduced reliance on external inputs. Integration of crops and livestock facilitated efficient recycling of farm resources through utilization of livestock manure for crop production and crop residues for livestock feeding. The reduced dependence on purchased inputs contributed to lower production costs and improved sustainability of farming systems. Moisture conservation practices also reduced the need for supplemental irrigation, thereby improving resource-use efficiency.

The findings are in agreement with Descheemaeker et al. (2019), who reported that integrated farming systems improve nutrient recycling and reduce external input dependency. Similarly, Duncan et al. (2013) and Bhargavi and Behera (2023) observed that livestock integration substantially reduces fertilizer requirements through manure recycling. The comparatively lower market dependency among marginal farmers may be attributed to limited purchasing power and lower risk-bearing ability, which often compels farmers to rely more on internal farm resources.

4.4 Employment opportunities

The integration of diversified enterprises under climate resilient farming systems significantly enhanced employment generation across both rainfed and partially irrigated situations. The increase in labor requirement was mainly associated with livestock management, horticultural activities, fodder cultivation, and value-added farm operations.

Enhanced employment opportunities within the farm may help reduce seasonal migration from rural areas by ensuring year-round livelihood opportunities. Diversified farming systems also improve household engagement in agriculture and strengthen livelihood security.

The findings support earlier reports indicating that migration from rural areas is often driven by declining profitability and climatic risks in agriculture (Deshingkar and Start, 2003; Singh, 2019; Bhargavi and Behera, 2020). The present results suggest that integrated farming systems can serve as an effective strategy for improving rural employment and reducing distress migration.

4.5 Sustainability of the diversified farming systems

The higher sustainable yield index (SYI) and sustainable value index (SVI) observed under treatment farming systems indicated improved yield stability and economic sustainability due to adoption of climate resilient interventions.

The improved sustainability indices can be attributed to diversified farming systems, integration of livestock and horticulture components, adoption of improved crop varieties, and efficient natural resource management practices. These interventions reduced production risks and improved stability of yields and income across years despite rainfall variability.

The findings corroborate earlier reports that integrated farming systems and resilient technologies improve sustainability under rainfed agriculture. Barnes et al. (2019) reported higher sustainability indices under diversified farming systems integrated with resilient interventions. Similarly, Yadav et al. (2024) observed that integrated farming systems recorded higher SYI values, indicating improved yield stability.

5 Conclusion

The present study demonstrated that integration of climate-resilient interventions within existing farming systems improved productivity, profitability, sustainability, and resilience under drought-prone rainfed conditions of southern Karnataka. The combined adoption of improved drought-tolerant crop varieties, intercropping systems, moisture conservation practices, fodder interventions, and livestock integration enhanced system productivity by 77% in small farmers and 43% in marginal farmers compared to farmers' practices. Similarly, resilience gains during the drought year were higher in intervention farms, with rainfed small and marginal farmers recording 65 and 40% improvement, respectively, over control. Evaluating the impact of individual interventions on overall system productivity is difficult and beyond the scope of the present study. However, assessing the impacts and economic viability of the predominant interventions provides valuable insights for policymakers to promote and scale up suitable interventions on a larger scale. The study further indicated that diversified integrated farming systems contributed to better resource recycling, reduced market dependency for inputs, and improved employment generation at the household level. The findings highlight that farm-type-specific and location-specific climate-resilient interventions can strengthen the adaptive capacity and livelihood stability of small and marginal farmers in vulnerable rainfed ecosystems.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Vijay Singh Meena, ICAR - Mahatma Gandhi Integrated Farming Research Institute, India

Reviewed by: Prashant Sharma, ICAR - Indian Institute of Soil and Water Conservation, Research Center, India

Tamilarasu Arivelrasan, Kaveri University, India