Total phosphorus (TP) concentration significantly impacts river ecosystems, yet its dynamics remain poorly understood in the Contiguous United States (CONUS). This study leverages remote sensing and machine learning to developed the CONUS’s Landsat-based Estimation and Assessment of Riverine TP (CLEAR-TP) database. Using this extensive database, we explore the spatial and temporal change of TP across CONUS rivers.The GitHub repository includes the source code used in the analysis of the developed database.
Total phosphorus (TP) concentration impacts river ecosystems, yet its dynamics remain poorly understood in the Contiguous United States (CONUS). We leveraged remote sensing and machine learning to study riverine TP dynamics. We developed a remotely estimated riverine TP database (CONUS’s Landsat-based Estimation and Assessment of Riverine TP, CLEAR-TP), covering 33,497 river reaches (107,000 km in length) from 1984 to 2018. Our results suggest significant long-term TP trends in 3,166 reaches (9.45%), with 2,636 (7.87%) showing declining (-0.995%/year on median) and the remaining showing increasing (1.216%/year on median). CLEAR-TP effectively captured longitudinal variations in TP concentrations along river profiles. Our results suggest that, at the entire river-course scales, climate, hydrological, and landcover factors were the primary drivers of longitudinal TP variation across most CONUS rivers.
-
CLEAR-TP Database: The database will be publicly available after the publication (Currently on review).
-
Coverage: The database spans 33,497 river reaches (a total of 107,000 km in length) from 1984 to 2018.
-
Approach: Combines remote sensing technologies with machine learning algorithms for TP estimation.
-
Remote Sensing Data: https://doi.org/10.1029/2020GL088946
-
ML Model: https://doi.org/10.1029/2024JG008121
Corresponding Author: ramtelpp@mail.uc.edu