An Advanced-Analytics Assessment of Drought Risk in Common Bean (Phaseolus vulgaris L.) Production in Rwanda: A 23-Year District-Level Bayesian, Machine-Learning, and Copula-Based Framework (2003–2025)

In the highlands of East Africa, where the air is thin and the soil is rich, a single crop feeds a nation. The common bean is more than just a vegetable; it is the primary source of protein for millions of people in Rwanda, a small, densely populated country where most families rely on the rain to grow their food. For decades, farmers have watched the sky, hoping for the right amount of water at the right time. But the climate is changing. The rains are becoming less predictable, and the heat is intensifying. When the soil dries out while the sun beats down, the bean plants suffer, and the harvest shrinks. Understanding exactly how this happens, and where it will hurt the most, requires looking beyond simple averages. It demands a way to see the invisible patterns in the weather, to measure the risk of a bad harvest before it happens, and to pinpoint the specific places where a small change in strategy could save a community from hunger.

A team of researchers from the University of Rwanda has taken on this challenge by building a new kind of map, not of land, but of risk. They did not just count the beans harvested over the last twenty-three years; they built a complex digital model that combines weather data, computer learning, and advanced statistics to tell a detailed story of what is happening to the country's most important crop. Their work covers every single district in Rwanda, from the high volcanic peaks in the north to the dry lowlands in the east, tracking the journey of the bean from 2003 to 2025. The goal was to move past general descriptions of drought and instead create a precise, district-by-district diagnosis of why yields are falling and where the danger is greatest.

The story the data tells is one of a sudden and sharp turn. For the first part of the study period, bean production in Rwanda was slowly climbing, reaching a peak around 2013. But then, something changed. The researchers found a clear break in the trend in 2019, a moment when the steady progress stopped and a steep decline began. Since that year, the average amount of beans harvested per hectare has dropped significantly. In the years following 2020, the decline accelerated, with yields falling to levels not seen in two decades. The analysis shows that this is not just a series of bad years; it is a structural shift in the climate system that supports the crop. The country has moved from a pattern of occasional dry spells to a new reality where drought is becoming a frequent, systemic threat.

To understand why this is happening, the researchers looked at the specific conditions that kill the bean plant. They found that the crop is most vulnerable when it is flowering and filling its pods. During these critical weeks, the combination of too little rain and too much heat is devastating. The study revealed that these two dangers rarely happen alone; they tend to arrive together. When the air is hot, the soil dries out faster, and the plant cannot recover. The researchers used a sophisticated statistical tool to measure how often these hot and dry conditions occur at the same time, finding that they are tightly linked. This means that trying to solve the problem by addressing only the lack of rain or only the high temperature would be like trying to put out a fire with a single bucket of water when the whole house is burning. The risk comes from the combination.

The power of this study lies in its ability to show exactly where the danger is concentrated. By analyzing data from all thirty districts, the researchers identified a specific cluster of seven districts in the eastern and south-eastern parts of the country that are facing the highest risk. These areas, including Bugesera, Kayonza, and Kirehe, form a continuous zone of vulnerability. In these lowland districts, the risk of drought is so high that the local climate has become hostile to the traditional way of growing beans. In contrast, the high-altitude districts in the north and west, such as Musanze and Burera, remain much safer, with cooler temperatures and more reliable rainfall. This sharp divide means that a single national strategy will not work; the solution must be targeted specifically at the seven districts that are suffering the most.

The economic cost of this shift is staggering. The researchers calculated the total financial loss caused by these droughts over the last decade, adjusting for the changing value of money. They estimate that between 2015 and 2025, the country has lost more than one billion dollars in bean production. The vast majority of this loss, about three-quarters of the total, has occurred in that single eastern cluster of districts. This is a massive blow to the local economy, where many families depend on selling beans to buy food and pay for school. The study suggests that without intervention, this trend will continue. Using a computer model that learns from past patterns, the researchers projected that by 2030, yields could fall even further, potentially dropping by another nine percent if nothing changes.

The researchers did not stop at identifying the problem; they also proposed a way to fix it. They designed a blueprint for a new kind of decision-support system that uses the internet of things and artificial intelligence to help farmers. Imagine a network of simple, low-cost sensors placed in the fields of the seven high-risk districts. These sensors would measure the moisture in the soil and the temperature of the air in real time. This data would be sent to a central computer that runs the same models used in this study. The system would then send simple, clear messages to farmers' phones, telling them exactly when to plant, when to water, and which types of beans are best suited for the current weather. It would also help insurance companies create fairer policies that pay out when the weather turns bad, protecting farmers from total ruin.

The study confirms that the era of relying on old patterns of rain is over for Rwanda's bean farmers. The climate has shifted, and the risk is now concentrated in specific places where the heat and dryness combine to crush the harvest. The path forward requires acknowledging this new reality. By focusing resources on the seven districts that are most at risk and using modern technology to guide farmers through the uncertainty, it is possible to build a system that is resilient enough to feed the nation. The data is clear, the location of the danger is known, and the tools to respond are ready. The question now is whether the country can move quickly enough to protect its food security before the next dry season arrives.