Diurnal variations in AVHRR SST fields: A strategy for removing warm layer effects from daily images

Different methodologies to estimate the amplitude of the sea surface temperature diurnal variation (DV) and remove it from remotely sensed SST images have been proposed in the last years. Among these, the parameterization proposed by Stuart-Menteth et al. (2004a,b) [Stuart-Menteth, A.C., Robinson, I.S., & Weller, R.A. (2004a). Sensitivity of the diurnal warm layer to meteorological fluctuations Part 1: observations, submitted to Journal of Atmospheric and Ocean Science; Stuart-Menteth, A.C., Robinson, I. S., & Donlon, C.J. (2004b). Sensitivity of the diurnal warm layer to meteorological fluctuations Part 2: a new parameterisation for diurnal warming, submitted to Journal of Atmospheric and Ocean Science] and adopted by the GHRSST-PP (Donlon, 2004) [Donlon, C.J., ad the GHRSST-PP Science Team, 2004: The GHRSST-PP data processing specification v1.0 (GDS v1.0, revision 1.5), GHRSST-PP Report N. 17, Published by the International GHRSST-PP Project Office, pp. 241] appeared as the most promising. In fact, it takes into account wind and insolation variations during the day, that effectively drive the SST diurnal cycle. arameterization has been tested on 6 months of NOAA-16 AVHRR images acquired and processed at CNR with Pathfinder algorithm. The tests evidenced some limits for a correct estimation of the DV in low-wind regimes for any insolation condition, and in high insolation regimes (>600 W/m2) when the wind intensity increases or decreases of more than 2 m/s during the morning. The limits of applicability of the DV correction to NOAA-16 AVHRR data (at least for the Mediterranean area) were thus identified, and data outside these limits were flagged. However, some anomalous heating were not corrected even with these constraints, due to the lack of accuracy in the wind field used for the correction. As a result, a strategy to flag residual outliers in the corrected daily images has been developed, based on the comparison to an optimally interpolated night SST field of the previous day.