River surface water temperature (RSWT) is a major variable controlling freshwater biogeochemistry, ecological function, and downstream heat transport, yet long-term observational records remain limited and biased toward instrumented basins. Even if long-term satellite thermal observations offer a means to overcome this gap, their use in characterizing non-linear changes in large regulated rivers remains underexplored, reducing our understanding of RSWT dynamics in systems where both climate forcing and anthropogenic regulation act simultaneously.
We exploit the full Landsat 7, 8 and 9 Collection 2 Level-2 Surface Temperature archive (2000-2025) over a representative reach of the lower Yangtze River (China) downstream of the Three Gorges Dam. Cloud-, snow-, and water-confidence masking is applied at the pixel level, followed by per-scene reach averaging using the JRC Global Surface Water mask.
The Landsat Collection 2 archive yielded a near-continuous per-scene RSWT record for the Yangtze reach across the 2000-2025 period, with consistent overlap between the L7, L8 and L9 sensors. The series preserves a coherent annual seasonal cycle and shows a visually apparent upward drift in summer thermal maxima during the second half of the record. Quantitative trend testing is currently in progress.
The constructed multi-decadal Landsat thermal record for the Yangtze provides a robust observational basis for quantifying decadal RSWT change in a large regulated river. The processing framework is fully transferable, and an ongoing extension to additional rivers across climatic and continental gradients aims at a globally consistent assessment of decadal river thermal change.