Measuring human impacts is vital to protect biodiversity and respond to climate change. Located in the third pole of the world, the Sanjiangyuan area (SA) is known as the source of the Yellow River, Yangtze River, and Mekong River, with an area of 363 thousand km². It is rich in biodiversity and highly sensitive to climate change and human activities. Protected areas, e.g., the Sanjiangyuan National Park and the Coco Xili Nature Reserve, have been established in the SA to restrict human activities for biodiversity conservation. To measure the human impacts of this area, we developed a human footprint (HF) dataset cumulating population density, land use activity, grazing intensity, nighttime lights, roads and railway factors. This dataset is at a 1 km resolution covering the whole area and the years 1995, 2000, 2005, 2010, and 2015. Our dataset is more accurate for the SA than the global HF dataset. This dataset will not only help to understand the human-nature relationship in the area, even in the third pole of the world, but can also be used to assess the effectiveness of nature reserves.

Spatial human footprint (HF) dataset contains 5 raster files, namely HF in the Sanjiangyuan area (SA) in 1995, HF in the SA in 2000, HF in the SA in 2005, HF in the SA in 2010, and HF in the SA in 2015, respectively, which can be read by ArcGIS software.--------------------------------1. The name of each tif filesHF in the SA in 1995: Human footprint in the Sanjiangyuan area in 1995;HF in the SA in 2000: Human footprint in the Sanjiangyuan area in 2000;HF in the SA in 2005: Human footprint in the Sanjiangyuan area in 2005;HF in the SA in 2010: Human footprint in the Sanjiangyuan area in 2010;HF in the SA in 2015: Human footprint in the Sanjiangyuan area in 2015.--------------------------------2. The uncertainty of the datasetThe population density data in the SA for 1995 were replaced with those of 2000 because the data for 1995 was lacking. As such, roads data for 1995 was replaced with those for 1996, and grazing intensity data for 1995 was replaced with those for 2000. Thus the HF for 1995 actually showed that in the early 2000s.