📄 earth_science

Developing a Composite Indicator and Coupling-Coordination Framework for Quantifying Climate Vulnerability-Adaptive Capacity Interactions: Validation Across Agroecological Typologies in Indonesian Rice Farming

This study develops and validates a novel, objective framework integrating the Livelihood Vulnerability Index, CRITIC-Entropy weighting, and the Coupling Coordination Degree model to quantify climate vulnerability-adaptive capacity interactions among Indonesian rice-farming households, revealing distinct coordination patterns across agroecological typologies and gender groups.

Original authors: Adi Firmansyah, Rafnel Azhari, Sumardjo Sumardjo, Siti Syamsiah, Leonard Dharmawan

Published 2026-08-11

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Adi Firmansyah, Rafnel Azhari, Sumardjo Sumardjo, Siti Syamsiah, Leonard Dharmawan

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Weather-Proofing Puzzle

Imagine you are trying to figure out which houses in a neighborhood are most likely to survive a massive storm. You wouldn't just look at the roof; you'd check the foundation, the windows, the emergency kit, and even how strong the neighbors are to help each other. In the world of climate science, this is called vulnerability. It's not just about how bad the weather gets (that's "exposure"), but also how sensitive a farm or family is to the damage, and most importantly, how well they can bounce back or adjust (that's "adaptive capacity").

For a long time, scientists have tried to measure this using a "Livelihood Vulnerability Index," which is like a report card for a community's ability to handle climate change. But there's a catch: usually, the people making the report card decide which grades matter most, which can be a bit subjective. Another tool, called the "Coupling-Coordination Degree" model, is like a dance partner score. It checks if two things—like the storm's intensity and a family's ability to dance—are moving in sync or tripping over each other. This paper asks a big question: Can we combine these tools, use math to remove the human bias in grading, and see if they work for individual rice farmers in Indonesia, not just for whole cities?

The Paper's Story: A Dance Between Storms and Survival

This research is a deep dive into the lives of 316 rice-farming families in Indonesia, specifically in two very different places: the island of Lombok and the province of South Sulawesi. The authors, Adi Firmansyah and his team, wanted to see how the type of land a farmer works on changes their ability to survive climate change. They didn't just look at the weather; they looked at the "dance" between the climate threats and the farmers' resources.

To do this, they built a super-charged measuring system. Instead of guessing which factors are most important, they used a mathematical method called CRITIC-Entropy weighting. Think of this as a robot referee that looks at all the data and automatically decides which clues are the most useful based on how much they vary, removing any human bias. They then used the Coupling-Coordination Degree (CCD) model to see if the farmers' "survival skills" (adaptive capacity) were keeping up with the "storm threats" (vulnerability).

The Three Types of Farms

The study grouped the farmers into three distinct "personalities" based on their land:

Irrigated Lowlands (IL): These are the "VIPs" of the rice world. They have permanent water systems, flat land, and can grow rice twice a year.
Rainfed Uplands (RU): These are the "hikers." They live on steep slopes, rely entirely on rain, and can usually only grow rice once a year.
Coastal Wetlands (CW): These are the "surfers." They live near the sea, dealing with tides and saltwater, and sometimes mix fish farming with rice.

The Big Findings: Who is Dancing in Sync?

The results showed that the type of land a farmer owns matters more than just their personal education or age. The "dance" between vulnerability and adaptation looked very different for each group:

The Lowland Dancers (Irrigated): These farmers are doing the best. Their "Coupling-Coordination Degree" score was 0.547, which the authors call "slightly coordinated." It's like a couple who is mostly in step. They have a bit of extra energy (adaptive capacity) to spare, meaning they are slightly better prepared than the current weather threats require. They have good roads, access to credit, and strong irrigation.
The Coastal Surfers (Wetlands): These farmers are in the middle, with a score of 0.391, labeled "slightly uncoordinated." They are struggling a bit more. The tides and storms (like El Niño) are hitting them hard, and while they have some ways to diversify their income, it's not quite enough to fully balance the threat.
The Upland Hikers (Rainfed): This group is in serious trouble. Their score was 0.284, classified as "moderately uncoordinated." The authors describe this as an "adaptive capacity lag." Imagine a runner who is being chased by a tiger (the climate threat) but is running in mud boots with no shoes (low adaptive capacity). The threat is way ahead of their ability to cope. These farmers face the highest vulnerability because they are isolated, have poor roads, and lack access to money or government help.

The Gender Gap

The study also looked at who was leading the farm. In every single type of land, families led by women had lower scores for "survival skills" than those led by men. The gap was widest in the struggling Rainfed Uplands. It turns out that women often have less access to land rights, credit, and training, making it harder for them to dance in sync with the challenges.

What the Math Says

Using a statistical test called binary logistic regression, the researchers found that the biggest things helping farmers adapt were:

Having access to agricultural extension services (experts who teach new methods).
Being able to get credit (loans).
Having at least 9 years of formal education.
Being part of a farming group.
Living in an irrigated area.

Being a female-headed household was a significant hurdle, lowering the chances of having high adaptive capacity.

What This Means for the Future

The paper suggests that we can't just use a "one-size-fits-all" plan for climate change. The "Rainfed Upland" farmers don't just need better seeds; they need roads, banks, and institutions to show up. The "Irrigated Lowland" farmers need to keep their infrastructure strong so they don't lose their advantage. And for everyone, programs need to specifically help women farmers, who are currently lagging behind.

The authors are careful to note that this is a snapshot in time (a cross-sectional survey), so they can't say for sure how these scores will change over the years. They also admit that while their math method is great at reducing bias, it still needs to be tested against other methods to be absolutely certain. However, they believe this new way of measuring—combining the vulnerability index with the coordination dance model—works well for individual farms, not just big cities, offering a clear, objective way to see where help is needed most.

Technical Summary: Developing a Composite Indicator and Coupling-Coordination Framework for Quantifying Climate Vulnerability-Adaptive Capacity Interactions

Problem Statement
Smallholder rice farmers in Indonesia face escalating climate threats, yet national adaptation policies often rely on a "one-size-fits-all" approach that fails to account for the distinct biophysical and socioeconomic conditions of different agroecological typologies. While the concept of adaptive capacity is well-established, its measurement remains plagued by methodological subjectivity, particularly in the weighting of composite indicators. Furthermore, existing analytical tools, such as the Coupling-Coordination Degree (CCD) model, have been primarily validated for macro-spatial urban or regional data but have rarely been rescaled to household-level agrarian systems. This study addresses the gap in understanding how agroecological typology mediates the relationship between climate vulnerability and adaptive capacity, while simultaneously proposing a more objective, transparent methodological framework for quantifying these interactions.

Methodology
The study employs a unified analytical framework integrating three components, validated across 316 rice-farming households in three distinct agroecological typologies in Indonesia: Irrigated Lowland (IL), Rainfed Upland (RU), and Coastal Wetland (CW).

Data Collection: Primary data were gathered via structured surveys in Lombok and South Sulawesi between February and May 2026. The sample was stratified by typology (123 IL, 113 RU, 80 CW).
Composite Indicator Construction:
- Livelihood Vulnerability Index (LVI): Calculated based on Exposure, Sensitivity, and Adaptive Capacity components.
- Adaptive Capacity Index (ACI): Derived from five livelihood capital dimensions (Human, Social, Natural, Physical, Financial) comprising 25 indicators.
- Weighting: To reduce subjectivity inherent in expert-elicited or equal weighting, the study utilizes an objective CRITIC-Entropy weighting method. This approach derives weights based on the variance (contrast intensity) and correlation (conflict) of the observed data, ensuring indicators with higher information content receive proportionally greater weight.
Coupling-Coordination Degree (CCD) Model: The study adapts the CCD model, traditionally used for urban sustainability, to the household level. It calculates the degree of balanced interaction between the Vulnerability subsystem ( $V$ ) and the Adaptive Capacity subsystem ($AC$) using the formula $D = \sqrt{C \times T}$ , where $C$ is the coupling degree and $T$ is the coordination index.
Validation: The framework's validity is assessed through discriminant analysis across typologies and gender-disaggregated comparisons. The study acknowledges that the CCD thresholds are adapted from macro-level literature and treats the household-level application as a methodological proposition to be tested empirically.

Key Results
The application of the framework revealed distinct patterns of vulnerability and coordination across the three typologies:

Vulnerability Profiles: Rainfed Upland (RU) systems exhibited the highest overall vulnerability (LVI = 0.623), driven by high climate exposure (temperature anomalies, drought frequency) and low adaptive capacity. Coastal Wetland (CW) systems faced high exposure to rainfall variability and ENSO events, while Irrigated Lowland (IL) systems showed the lowest vulnerability (LVI = 0.387).
Adaptive Capacity: IL households demonstrated the highest adaptive capacity (ACI = 0.541), particularly in Physical and Financial capital, due to better infrastructure and market access. RU households had the lowest capacity (ACI = 0.318), with significant deficits in physical capital (irrigation, roads) and financial access.
Coupling-Coordination Patterns:
- IL (Slightly Coordinated, D=0.547): Characterized by an "adaptive capacity surplus," where resources modestly exceed current vulnerability levels.
- RU (Moderately Uncoordinated, D=0.284): Characterized by a severe "adaptive capacity lag," where vulnerability significantly outpaces available resources (V-AC = +0.305).
- CW (Slightly Uncoordinated, D=0.391): An intermediate state where high exposure is partially offset by livelihood diversification.
Gender Disparities: Female-headed households consistently scored lower across all capital dimensions and typologies. The adaptive capacity gap was most pronounced in RU systems, compounding existing structural disadvantages.
Determinants: Logistic regression identified extension access, credit access, formal education, and typology status as significant positive predictors of high adaptive capacity, while female-headed household status was a significant negative predictor.

Key Contributions

Methodological Innovation: The study successfully rescales the CCD model from macro-spatial urban applications to household-level agrarian data. It demonstrates that the logic of coupling-coordination (assessing the alignment between subsystems) remains diagnostically valid at the micro-level, provided both subsystems are measured on comparably constructed composite indices.
Objective Weighting: By employing the CRITIC-Entropy method, the study offers a transparent, replicable alternative to subjective weighting schemes, reducing analyst bias in the construction of vulnerability and adaptive capacity indices.
Typological Differentiation: The research provides empirical evidence that agroecological typology is a structural driver of adaptive capacity, explaining a significant portion of variance (23.4%) independent of individual socioeconomic characteristics. This challenges the efficacy of uniform policy interventions.
Diagnostic Utility: The framework effectively identifies specific "lag" conditions (e.g., the adaptive capacity lag in RU systems), offering a nuanced tool for policymakers to distinguish between areas needing structural intervention versus those requiring maintenance or enhancement.

Significance and Claims
The authors claim that this integrated framework provides a transparent, objective, and transferable approach for measuring climate vulnerability-adaptive capacity interactions in smallholder farming systems across the Global South.

The study posits that its primary significance lies in demonstrating that place-based structural conditions (captured by agroecological typology) exert an independent and powerful influence on adaptation outcomes, often outweighing individual household characteristics. Consequently, the authors argue for a reorientation of adaptation policy toward typology-sensitive targeting. Specifically, they suggest that rainfed upland systems require structural enablement (infrastructure, institutional development) rather than just technology transfer, while coastal and irrigated systems require targeted strategies addressing specific exposure risks and sustainability.

The paper maintains a modest stance regarding its methodological contributions, acknowledging that the household-level CCD thresholds are adapted rather than independently re-derived for agrarian data. It explicitly calls for future sensitivity analyses comparing weighting schemes and longitudinal studies to validate the dynamic nature of the "lag" classifications, framing the current work as a robust, replicable template for comparative vulnerability assessment rather than a definitive, final solution.