Researchers have developed a new method using machine learning and multiple data sets to estimate irrigation water use (IWU) across China, and also revealed how water use might change in the future based on different climate and economic scenarios.

Irrigated agriculture accounts for about 20 percent of global croplands and contributes more than 40 percent of food production. China is home to the world's largest irrigated farmland, accounting for over half of the country's arable lands.

Therefore, water for irrigation is critical to the crop growth and yields. Accurately characterizing changes in IWU is conducive to formulating optimal water resource allocation policies, particularly in the context of climate change.

However, due to limited data availability and model constraints, it is hard to use existing IWU estimation methods on a national scale or under future climate change scenarios.

The researchers from the Aerospace Information Research Institute under the Chinese Academy of Sciences developed a new machine-learning model to estimate nationwide IWU under a data-driven framework.

The new model integrated a range of high-precision hydrological satellite remote sensing products, meteorological drivers, economic statistics and numerical model simulations. It showed high precision in IWU projections with over 90 percent accuracy rate.

Taking a series of climate and socioeconomic scenarios into consideration, the researchers combined the established machine-learning framework with four advanced Earth system models to forecast the future trends and additional costs of IWU across China over the next 70 years.

Under different greenhouse gas emission scenarios, China's IWU is projected to increase in about 60 percent of provinces, especially in the northwest and northern regions.

IWU is also predicted to increase by 8.5 percent to 17.1 percent by 2050 compared to the historical period (1981-2010), with higher emissions leading to more significant increases.

This rise comes with an estimated additional cost of 1.65 to 3.91 billion U.S. dollars per year, highlighting the urgency for sustainable water management.

The research findings could provide crucial information for policymaking regarding agricultural water, contributing greatly to the sustainable use and management of water resources in the context of climate change and socioeconomic dynamics, the research team stated, adding that the results are also conducive for other countries confronted with increasing irrigation demands.

An article about this discovery was published in the journal Earth's Future.