Assessing Inter-Provincial Climate Connectivity in Indonesia Using Detrended Cross-Correlation and Network Analysis
This study employs Detrended Fluctuation Analysis, Detrended Cross-Correlation Analysis, and Complex Networks to characterize the multi-scale, long-range persistent inter-provincial climate connectivity across Indonesia's 38 provinces, identifying distinct regional hubs and connectivity patterns driven by monsoon circulation and organized convection to inform adaptation strategies and monitoring network optimization.
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
Climate is not a collection of isolated weather events happening in one place; it is a vast, breathing system where a shift in the air over one ocean can ripple across continents, altering rain and heat thousands of miles away. In the tropics, this system is driven by massive, slow-moving forces like the monsoon winds that sweep across the Indian and Pacific Oceans, and by the warming and cooling of the sea surface known as El Niño. Understanding how these forces connect different regions is vital for predicting floods, droughts, and heatwaves, especially in places where the landscape is broken into thousands of islands. However, traditional ways of measuring these connections often fail because they assume the weather behaves in a simple, predictable pattern, ignoring the fact that climate data is messy, full of trends, and changes its character over time. To see the true links between distant places, scientists need tools that can strip away the noise and find the underlying rhythm that ties the weather of one province to another, even when that connection only appears over long periods.
A team of researchers set out to map these invisible threads across Indonesia, the world's largest archipelago, which stretches over 5,000 kilometers and contains more than 17,000 islands. They analyzed fifteen years of daily weather records from 38 provincial stations, covering five key elements: air temperature, humidity, rainfall, sunlight, and wind speed. Instead of looking at these variables in isolation, they used a method that treats the weather data as a complex web, allowing them to see how strongly the climate in one province is tied to another at different time scales, from a few days to several years. Their goal was to find out which provinces act as the central hubs that hold the regional climate together and which areas are so isolated by their unique geography that they operate on their own.
The researchers first confirmed that the climate across all of Indonesia has a deep, long-term memory. This means that weather patterns do not just reset every day; instead, the conditions of today are statistically linked to conditions from weeks, months, or even years ago. This persistence grows stronger the longer the time scale, suggesting that the region's climate is a unified system that remembers its past. However, when they looked at how different provinces talk to each other, the story changed depending on the time scale and the specific weather variable. For short periods, like a few days, most provinces showed almost no connection to their neighbors. The weather in one place was largely independent of the weather in the next, as if the islands were speaking different languages. But as the researchers looked at longer periods, spanning seasons and years, the connections began to emerge, revealing a clear hierarchy of influence.
The study found that the geography of the islands plays a decisive role in how these connections form. In the western part of the country, the islands of Sumatra and Java showed moderate links, but these were not uniform. Some provinces, like those in the interior lowlands, moved together, while coastal areas or those with complex terrain, such as the volcanic highlands of Yogyakarta or the urban heat of Jakarta, remained isolated. These isolated spots are so influenced by their immediate surroundings—like the heat of a city or the blocking effect of a mountain—that they do not follow the broader regional signals. In stark contrast, the eastern islands of Bali, Nusa Tenggara, and Maluku showed almost no connection to each other at any time scale. Despite being neighbors, their climates are so distinct, shaped by different ocean currents and dry seasons, that they do not synchronize. The researchers concluded that treating these eastern island groups as a single climatic unit would be a mistake, as their weather patterns are fundamentally fragmented.
The most surprising finding was the behavior of the province of Papua in the far east. While it appeared on the edge of the network for temperature and humidity, it became the central hub for wind and rainfall. At the longest time scales, the southern part of Papua emerged as the most connected node in the entire network, linking six different island groups simultaneously. This happens because Papua sits at the intersection of two massive wind systems, acting as a structural nexus where the winds from Australia and the Pacific converge. This makes it a critical point for understanding how wind and rain travel across the entire Indo-Pacific region. Conversely, the province of East Kalimantan, despite being on the same island as its neighbors, was found to be completely cut off from them in terms of temperature and humidity, likely due to its unique exposure to the sea and local land changes, though it remained connected through wind patterns.
The type of weather variable being measured also dictated how the connections looked. Rainfall and wind speed showed the most extensive cross-island links, connecting distant provinces almost immediately, driven by large, organized storm systems that sweep across the archipelago. Sunlight, however, told a different story, splitting the network into two distinct camps: one group in the west and another in the east, separated by a persistent divide in cloud cover that has held steady for fifteen years. Temperature and humidity, which are more sensitive to local land features, took longer to show connections, only forming large, unified clusters when the researchers looked at seasonal and yearly patterns driven by the monsoons.
These results offer a clear guide for how to monitor the climate of Indonesia. The study suggests that a single network of weather stations is not enough; instead, monitoring efforts must be tailored to the specific variable and the region. Provinces that act as hubs, such as South Sumatra for temperature or South Papua for wind, can serve as strategic observation points to understand the climate of many surrounding areas. However, the isolated provinces, including the highlands of Papua and the eastern islands, require their own dedicated monitoring infrastructure because no nearby station can accurately represent their unique weather. By understanding these specific patterns of connection and isolation, scientists and policymakers can build better early warning systems and adaptation strategies that respect the complex, multi-layered reality of the Indonesian climate.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.