In the vast, high-altitude expanse of the Tibetan Plateau, a silent transformation has been taking place over the past few decades. Cultivated pastures have been rapidly expanding, a trend that has raised concerns about grassland degradation and the need for effective ecological protection. However, until now, the high-resolution spatial distribution of these cultivated pastures has remained poorly understood, hindering efficient agricultural and livestock management.
A recent study, published in the journal Earth System Science Data (which translates to “地球系统科学数据” in Chinese), has shed new light on this issue. Led by B. Han from the State Key Laboratory of Herbage Improvement and Grassland Agro-Ecosystems at Lanzhou University, the research team has mapped the cultivated pastures on the Tibetan Plateau at an unprecedented 30-meter resolution for the years 1988 to 2021.
The team utilized Landsat data from the Google Earth Engine cloud computing platform to build a random forest binary classification model. This model was trained using carefully selected training samples and validated against 2000 independent random reference points in two pilot study regions with different climates and landscapes. The model achieved an impressive overall accuracy of 97.05%, with an F1 spatial consistency score of 82.51%, indicating high confidence in its ability to map the distribution of cultivated pastures.
“This study provides a crucial baseline for understanding the extent and dynamics of cultivated pastures on the Tibetan Plateau,” said Han. “It offers valuable insights for scientists, policymakers, and ecological conservation practitioners.”
The dataset produced by this study covers 77% of the Tibetan Plateau, including Qinghai Province and the Tibet Autonomous Region. It reveals a significant expansion of cultivated pasture areas from 0.46 million hectares in 1988 to 1.57 million hectares in 2021, with an average annual growth of 33.5 ± 2.5 thousand hectares.
The implications of this research extend beyond the Tibetan Plateau. The random forest binary classification approach developed by Han and his team holds promise for identifying cultivated pastures in other regions of the world. This could prove invaluable for supporting ecological protection and restoration efforts, as well as for informing agricultural and livestock husbandry management practices.
As the world grapples with the challenges of climate change and the need for sustainable land use, this research offers a timely and important contribution. By providing a detailed and accurate map of cultivated pastures on the Tibetan Plateau, it paves the way for more informed decision-making and effective ecological management.
The dataset is available on Zenodo, ensuring that the findings are accessible to researchers and practitioners worldwide. As Han noted, “This is just the beginning. We hope that our work will inspire further research and collaboration in this critical area.”