In the heart of America’s Corn Belt, a technological revolution is brewing, one that could redefine how farmers manage risk and insurers underwrite policies. At the forefront of this innovation is Sachini Wijesena, a researcher from the Centre for Advanced Modelling and Geospatial Information Systems (CAMGIS) at the University of Technology Sydney. Her work, published in Environmental Research Communications, is set to enhance weather index insurance (WII), a tool that could become increasingly vital in the face of climate change.
WII is not new, but it has traditionally relied on a single weather index, a method that often falls short in capturing the complexity of weather events and their impact on crops. Wijesena’s research, however, proposes a novel approach that leverages machine learning and remote sensing data to create a multi-weather index that is highly predictive of crop yield.
The key challenge, Wijesena explains, is to maintain transparency while harnessing the power of complex machine learning models. “Neural networks can model the multifaceted nature of factors influencing crop growth, but they are often perceived as ‘black box models,'” she says. “Our solution is to use a surrogate model, a generalised linear model (GLM) in this case, to approximate the neural network’s predictions.”
The results are promising. The GLM achieved a mean absolute error (MAE) of 8.2%, comparable to the neural network model’s MAE of 7.6%. But more importantly, the weather index derived from the surrogate model incorporates multiple remote sensing indexes and weather variables, providing a more comprehensive picture of the factors affecting crop yield.
This multi-weather index includes Potential Evapotranspiration (PET), Evapotranspiration (ET), Land Surface Temperature (LST), Vegetation Condition Index (VCI), and minimum temperature. By considering these variables, the proposed WII product demonstrated a substantial hedging efficiency, reducing downside risk by 21%.
The implications for the agricultural industry, and indeed the energy sector, are significant. As climate change continues to exacerbate extreme weather events, the demand for innovative risk management tools like WII is set to grow. This research could shape the future of crop insurance, making it more accurate, transparent, and financially viable.
Moreover, the energy sector, which is increasingly reliant on biofuels and other agricultural products, could benefit from more stable crop yields. By reducing the risk of significant yield losses, WII could contribute to a more stable supply of agricultural products, thereby supporting the energy sector’s sustainability goals.
Wijesena’s work, published in Environmental Research Communications, which translates to Environmental Communication Research, is a testament to the power of interdisciplinary research. By combining machine learning, remote sensing, and agricultural science, she has developed a tool that could revolutionise the way we manage risk in the face of climate change.
As we look to the future, it is clear that technology will play a pivotal role in shaping the agricultural industry. This research is a step in that direction, a beacon of innovation in the face of uncertainty. It is a reminder that even in the most challenging times, there is always room for progress, always room for growth. And in the heart of America’s Corn Belt, that progress is already underway.