Observing water turbidity in Chinese rivers using Landsat series data over the past 40 years

Abstract

River turbidity is a critical indicator of suspended particulate matter concentration and water quality. Traditional studies on river turbidity have predominantly relied on in-situ measurements, which, while providing high accuracy, are often constrained by their labor-intensive nature and limited spatial coverage. Moreover, these measurements are spatially discontinuous, temporally sporadic, and rarely accessible in remote or inaccessible regions. To address these limitations, this study integrated extensive hourly monitoring data from 1997 stations to develop a Random Forest model capable of quickly estimating river turbidity from Landsat series satellite data and analyzed the spatiotemporal variability of river turbidity across China during 1984–2023. The results demonstrated that the developed RF model effectively inverted river turbidity from Landsat data, with a root mean square error of 19.43 NTU and a mean absolute percentage difference of 38.67% for the validation dataset (N = 367). Relative to the Hu Line, river turbidity across China exhibited a spatial pattern of “low in the west and high in the east”, with western rivers exhibiting a mean value of 22.80 ± 10.27 NTU, compared to 32.50 ± 9.74 NTU for eastern rivers. Additionally, the four largest Chinese rivers displayed a spatial pattern of “clear upstream and turbid downstream”, primarily due to sediment resuspension by high flow in the downstream. Over the past 40 years, approximately 69.51% of the western river areas have experienced increased turbidity due to warming and humidification, whereas about 74.62% of the eastern river areas have become clearer as a result of dam construction and human management. These distinct spatiotemporal changes indicate a reduction in the river turbidity disparity between the eastern and western rivers. In the context of global change, this study provides valuable insights for long-term and large-scale monitoring of river turbidity using Landsat data.