New predictive models promise to revolutionize livestock feeding strategies in one of China’s most important pastoral regions.
A recent study published in the journal Grassland Research examined how near-infrared (NIR) spectroscopy can be used as a quick, efficient, and dependable tool in analyzing the nutritional composition of natural forage (1). This study, which was led by lead authors Binqiang Bai and Lizhuang Hao of Qinghai University, demonstrated how NIR spectroscopy can evaluate forage and help determine livestock feeding strategies. By developing NIR spectroscopy-based predictive models, the research team was able to evaluate the nutritional composition of natural forage in the Qinghai-Tibet Plateau (1).
Located in western China, the Qinghai–Tibet Plateau is known as the “Roof of the World” and is one of the most critical pastoral zones in the country (1,2). The plateau spans over 2.5 million km² and features a harsh highland climate (2). Most Tibetans live above 3500 m, with many exceeding 4500 m (2). Their livelihoods mainly involve farming, herding, and mining. The region also supports millions of yaks (Poephagus grunniens), which are central to the livelihoods of local communities (1,2). Given the heavy dependence on natural pastures for livestock grazing, accurate assessment of forage quality is crucial. However, the problem is that traditional methods are not optimal for this purpose. Traditional methods of forage evaluation are often labor-intensive, time-consuming, and require destructive sampling, making them unsuitable for real-time management decisions (1).
Qinghai Tibet Plateau: Snow-Capped Mountains, Barley Fields, Blue Sky in Stunning 8K Realism. Generated by AI. | Image Credit: © Saba - stock.adobe.com
In their study, Bai, Hao, and their research team sought to improve methods of forage evaluation by building a new predictive model using NIR spectroscopy. As part of their experimental procedure, the research team collected 301 natural forage samples from the Sanjiangyuan Area, which is a key source region for China's three major rivers (1). Then, the samples underwent both conventional nutritional analysis and NIR spectral scanning. Once this process was completed, the team then developed predictive models to estimate crucial nutritional parameters, including crude protein (CP), acid detergent fiber (ADF), neutral detergent fiber (NDF), dry matter digestibility (DMD), gross energy (GE) yield, and methane emissions (1).
Overall, the researchers demonstrated that their model produced encouraging results. The models demonstrated strong predictive accuracy for CP, ADF, and NDF, which are three of the most essential indicators of forage quality (1). CP, ADF, and NDF are often strong indicators that dictate livestock feeding strategies. Crude protein informs energy intake and animal growth, while fiber content affects digestibility and feed intake levels (1).
Despite the success of their NIR spectroscopy predictive model, the researchers acknowledged some limitations with their model. The main one, which the team acknowledged, is that their model needs to improve in predicting traits like methane emissions and energy content (1). The researchers suggested that integrating advanced machine learning (ML) algorithms in future studies could enhance the predictive power of NIR spectroscopy, especially for these more challenging variables (1).
Improving ways to analyze forage quality has huge economic and health implications for both humans and animals. Many societies, including the Qinghai-Tibet region, live an agrarian lifestyle, and they need to be able to rapidly evaluate the nutrient content of their local pastures for their livelihood. The predictive model proposed by the researchers offers a new potential solution that can lead to improved animal health, greater production efficiency, and reduced environmental impact (1).
Moreover, the study underscores the growing relevance of precision agriculture in rangeland management. Combining traditional knowledge with technology like NIR spectroscopy can ensure more sustainable forage utilization (1). This approach aligns with China’s broader efforts to modernize its agricultural sector while preserving ecological balance in vulnerable regions (1).
The team stated in the conclusion of their article that future work they conduct will focus on expanding their sample database and test their models across different seasons and ecological zones of the plateau (1). The researchers hope that by doing so they can create a user-friendly decision-support system for pastoralists that integrates NIR spectroscopy data with satellite imagery and local climate variables (1).
Wearable fNIRS Sensor Tracks Cognitive Fatigue in Real Time
May 7th 2025Researchers have developed a wireless, wearable brain-monitoring device using functional near-infrared spectroscopy (fNIRS) to detect cognitive fatigue in real time. The miniaturized system enables mobile brain activity tracking, with potential applications in driving, military, and high-stress work environments.
New Imaging Method Offers Breakthrough in Forest Soil and Tree Analysis
May 5th 2025A new study published in Geoderma Regional by J. A. Arias-Rios and colleagues at IFAB demonstrates that near-infrared (NIR) spectroscopy is a rapid, cost-effective tool for assessing soil and tree traits critical to forest ecosystem monitoring and management.
How the James Webb Telescope’s MIRI and NIRCam are Advancing Cosmic Exploration
May 2nd 2025Macarena Garcia Marin, an astrophysicist and instrument scientist for the European Space Agency, highlights the role spectroscopy techniques have played in the pivotal research done on the James Webb Telescope since its launch in 2021.
AI and Satellite Spectroscopy Team Up to Monitor Urban River Pollution in China
April 30th 2025A study from Chinese researchers demonstrates how combining satellite imagery, land use data, and machine learning can improve pollution monitoring in fast-changing urban rivers. The study focuses on non-optically active pollutants in the Weihe River Basin and showcases promising results for remote, data-driven water quality assessments.