| 23 Jul 2026
Sea Surface Temperature Upwelling Index for West Iberia (1982–2021)
Abstract. Cloud cover introduces extensive gaps in sea-surface temperature (SST) fields derived from infrared satellite measurements, and gap frequency rises further when differencing data to compute an SST-based upwelling index. Averaging can eliminate gaps but degrades spatio-temporal resolution, whereas interpolation preserves resolution but requires robust methodology. The CoRTAD database — a 21-year global ocean satellite SST dataset, gap-free, at 4 km resolution and weekly averaged — in which missing observations were mitigated by merging night- and day-passes from the Pathfinder V5.3 satellite record and computing weekly means. Residual SST gaps were filled first with a 3 × 3-pixel spatial median and, where necessary, with a piecewise cubic Hermite interpolating polynomial. The resulting spatial and temporal resolution supports observation of sub-mesoscale to mesoscale coastal-ocean processes. From CoRTAD, SSTs at near- and offshore locations for West Iberia (37°–44° N) were extracted and a coastal upwelling index (UIsst) calculated for the period January 1982 to December 2021. The nearshore (midshelf) climatology derived from this dataset successfully identifies established recurrent upwelling centres along the West Iberian coast, which were previously detectable only in cloud-free synoptic SST imagery, thereby supporting the validity of the dataset. Monthly climatology reveals no north-south differences in upwelling intensity or season end-date, but shows that the central region between Cabo Carvoeiro (~39° N) and the River Douro (~41° N) has a notably later upwelling season onset (July/August versus June/July). With continuous satellite scatterometer winds available since 1999, the stage is set to compare an Ekman-derived upwelling index with the SST-derived index. This study demonstrates that a high-resolution, gap-free satellite SST dataset combined with targeted interpolation enables robust characterisation of coastal upwelling dynamics over multi-decadal timescales.
The paper presents a new sea surface temperature (SST)-derived coastal upwelling index (UI^SST) for the western Iberian margin (37°–44° N), covering the period from January 1982 to December 2021. The dataset is based on the gap-free, 4-km, weekly CoRTAD v6 SST product, with UI^SST defined as the temperature difference between midshelf and offshore waters along the coast. The accompanying analysis characterizes the climatological variability of coastal upwelling at weekly, monthly, and seasonal (3-month) timescales, as well as its spatial structure, in support of the dataset's description and validation. The results successfully reproduce the well-known recurrent upwelling centres associated with major capes and identify a delayed onset of the upwelling season in the central sector between Cabo Carvoeiro and the River Douro. The dataset is publicly available through Zenodo via the CoastNet infrastructure, and future developments include extending it to near-real-time operation and evaluating it against a wind-based (Ekman) upwelling index.
General Comment:
This is a solid and valuable dataset paper, well suited to ESSD, and the dataset is clearly described, well documented, and appropriately validated. I recognize that the scope of a data descriptor is to establish the quality, provenance, and usability of the dataset rather than to deliver an in-depth scientific analysis, and the climatological characterization in Section 4 fulfills that role well as a demonstration of the dataset's utility.
That said, I have one suggestion, offered to strengthen the demonstration value of the dataset rather than as a requirement for publication. Several of the climatological features highlighted as illustrations of what the dataset can reveal (the timing of the onset and termination of the upwelling season, the apparent lack of a north–south gradient in upwelling intensity, and the delayed onset in the Cabo Carvoeiro–River Douro region) are currently identified through visual inspection of the weekly, monthly, and seasonal (3-month) climatologies in Figures 5–8. These are interesting and, in my view, well-supported interpretations. Since these features will likely serve as reference points for future users of the dataset, it could add value to support them with simple, non-parametric statistics alongside the visual presentation, purely as an optional enhancement to the dataset's documentation. For example:
None of this bears on the acceptability of the dataset itself, which I consider ready for publication. These are suggestions for making the accompanying climatological description a more robust and reusable reference for future users of the data, should the authors wish to incorporate them.
As a final note, a few typographical/grammatical issues that could be corrected: l.39 draw up -> drawn up; l. 203 and 485 costal -> coastal; l. 479 instigate -> instigated; l. 492 week ->weak; L. 680 References Kilpatrick et al. (2001) and Kravtsov et al. (2008) merge.