NSI & Yamaguchi Univ. Develop AI Model to Predict Cropland Vulnerability

A joint research paper by Yamaguchi University, New Space Intelligence Inc. (NSI), and Assam Down Town University was published in the international peer-reviewed journal Land (MDPI) on January 16, 2026.

Research Overview 　This study developed a high-precision methodology to predict cropland vulnerability in Bangladesh, a nation highly susceptible to climate change, by analyzing 22 years of long-term Earth observation data using machine learning (Random Forest). The findings demonstrated that the model can successfully visualize drought risks and cropland stress with high accuracy 2 to 3 months prior to harvest, identifying key climate drivers such as March precipitation.

Practical Significance 　The methodology enables a critical shift from "reactive crisis management" to data-driven "proactive prevention" before agricultural damage occurs. Highly versatile as it utilizes existing satellite data without requiring specialized ground equipment, this approach is expected to significantly contribute to sustainable climate adaptation, including optimized water resource allocation, accelerated crop insurance processing, and strategic seed distribution.

Publication Information

• Title: A Comparative Analysis of Machine Learning Approaches for Predicting Cropland Vulnerability in Bangladesh Using Long‑Term Earth Observation Data

• Authors: Arnob Bormudoi, Masahiko Nagai

• Journal: Land 2026, 15(2), 174

• DOI / URL: https://www.mdpi.com/2073-445X/15/1/174