Quantifying CDOM and DOC in major Arctic rivers during ice-free conditions using Landsat TM and ETM+ data

Document Type



As high-latitudes warm, permafrost thaws, and the hydrological cycle accelerates, ground-based monitoring of riverine organic matter may be supplemented by satellite remote sensing during ice-free conditions. Recent programs, namely the Arctic Great Rivers Observatory, have established methodologically consistent sampling across the hydrograph, and shared the resulting data publicly. However, these efforts are limited by frequency, funding, and length of record. Satellite remote sensing can be used to estimate chromophoric dissolved organic matter (CDOM) as a riverine constituent that influences optical properties in surface waters. In this study, daily CDOM absorption was first estimated using discharge-constituent regression-based models for 2000–2013. We then regressed these discharge-based CDOM estimates against Landsat TM and ETM+ surface reflectance data from Google Earth Engine for the six largest rivers draining the pan-Arctic watershed (the Kolyma, Lena, Mackenzie, Ob’ Yenisey, and Yukon rivers). These CDOM results were converted to dissolved organic carbon (DOC), using the strong relationship (R2 = 0.88) between direct measurements of the two constituents. Using river-specific remote sensing models, R2 could be as high as 0.84. Grouping all rivers into a single “universal” regression reduced R2 and increased root mean square errors, such as in the Yenisey River where R2 dropped by 0.63, and RMSE rose by 1.1 m−1. Seasonally varying discharge drove much of the variation in satellite-derived CDOM and DOC, corroborating recent studies. Satellite imagery can increase the frequency of monitoring observations, particularly during summer and fall when riverine CDOM absorption may be most sensitive to thawing permafrost.

Publication Title

Remote Sensing of Environment

Publication Date




First Page


Last Page







Arctic, chromophoric dissolved organic matter, dissolved organic carbon, Google Earth Engine, Landsat, remote sensing, rivers