Assessing the impact of climate variability on Australia’s sugarcane yield in 1980–2022

Abstract

Sugarcane is an important crop for global food and energy production. However, its production is greatly affected by inter-annual climate variations in major production regions. While previous studies have assessed climate impacts on sugarcane yield at individual sites, a regional-scale understanding of the climate-yield relationship remains unclear. Here, we collected historical sugarcane yields (1980–2022) and meteorological data from 23 sites across Australia’s eastern coastline sugarcane belt. Three statistical methods, random forest (RF), eXtreme gradient boosting regression (XGBoost), and multiple linear regression (MLR), were used to assess the impacts of climatic factors on sugarcane yield. The results showed that the machine learning methods, particularly RF, outperformed MLR in estimating sugarcane yield. The RF model explained 45–62 % of yield variations in Australia’s sugarcane regions based on climatic means and extreme climate indices. Growing season rainfall was identified as the most important factor influencing sugarcane yield in the Northern region, while CDD (consecutive dry days) was critical in the Central region, and TNn (minimum daily minimum temperature) was the dominant factor in the Southern region. Notably, the dominant factors exhibited a non-linear relationship with yield. In the Southern region, the lowest temperatures above 5 °C produced high yields. By contrast, in the Northern region, yields decreased with rainfall exceeding 1500 mm. Similarly, in the Central region, the increase in CDD substantially reduced yields, with yields reaching a low level after 70 days of CDD. To address these impacts, region-specific adaptation strategies are recommended, including the cultivation of waterlogging-tolerant crop varieties, the development of efficient irrigation systems, and the adoption of low-temperature-tolerant cultivars. This study highlights the critical importance of quantifying the contribution of climate variables to crop yield variability, thereby informing the development of effective, region-specific management practices.