Abstract
The study evaluates the performance of the MCC algorithm for near-real-time (NRT) ocean surface current estimation using TIR imagery acquired from the AVHRR sensor onboard EUMETSAT and NOAA satellites. We focused on image pairs with short time-lags (<3 hours), a largely under-explored area that is critical for NRT operational marine applications. To address this, a data enrichment technique was applied to significantly increase the number of usable image pairs with short and long time-lags, thereby reducing data gaps and improving the timeliness of derived currents. A conditional, adaptive framework was employed to dynamically adjust the MCC parameters, eliminating manual tuning, improving computational efficiency, and reducing uncertainties associated with fixed configurations. The method was validated against coastal HFR observations using statistical metrics and visual diagnostics. The results showed robust directional agreement and quantified reliability, with RMSE of 0.11 m/s, MAE of 0.08 m/s, SI of 0.38 and complex correlation magnitudes ranging 0.3-0.9, between HFR observations and MCC-derived currents, particularly for short time-lag image pairs. MCC estimates consistently exceeded HFR depth-averaged measurements, with a speed bias of −0.09 to 0.06 m/s. Overall, the results demonstrate the framework's feasibility for NRT ocean current estimation and its applicability to coastal, marine and maritime applications.