Evaluating operational AVHRR sea surface temperature data at the coastline using surfers
de Mora, L
Estuarine, Coastal and Shelf Science
© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Sea surface temperature (SST) is an essential climate variable that can be measured routinely from Earth Observation (EO) with high temporal and spatial coverage. To evaluate its suitability for an application, it is critical to know the accuracy and precision (performance) of the EO SST data. This requires comparisons with co-located and concomitant in situ data. Owing to a relatively large network of in situ platforms there is a good understanding of the performance of EO SST data in the open ocean. However, at the coastline this performance is not well known, impeded by a lack of in situ data. Here, we used in situ SST measurements collected by a group of surfers over a three year period in the coastal waters of the UK and Ireland, to improve our understanding of the performance of EO SST data at the coastline. At two beaches near the city of Plymouth, UK, the in situ SST measurements collected by the surfers were compared with in situ SST collected from two autonomous buoys located ∼7 km and ∼33 km from the coastline, and showed good agreement, with discrepancies consistent with the spatial separation of the sites. The in situ SST measurements collected by the surfers around the coastline, and those collected offshore by the two autonomous buoys, were used to evaluate the performance of operational Advanced Very High Resolution Radiometer (AVHRR) EO SST data. Results indicate: (i) a significant reduction in the performance of AVHRR at retrieving SST at the coastline, with root mean square errors in the range of 1.0 to 2.0 °C depending on the temporal difference between match-ups, significantly higher than those at the two offshore stations (0.4 to 0.6 °C); (ii) a systematic negative bias in the AVHRR retrievals of approximately 1 °C at the coastline, not observed at the two offshore stations; and (iii) an increase in the root mean square error at the coastline when the temporal difference between match-ups exceeded three hours. Harnessing new solutions to improve in situ sampling coverage at the coastline, such as tagging surfers with sensors, can improve our understanding of the performance of EO SST data in coastal regions, helping inform users interested in EO SST products for coastal applications. Yet, validating EO SST products using in situ SST data at the coastline is challenged by difficulties reconciling the two measurements, which are provided at different spatial scales in a dynamic and complex environment.
We thank Gavin Tilstone and Vassilis Kitidis for help comparing the temperature sensor (Tidbit v2 temperature logger) used in the study with the VWR1620-200 traceable digital thermometer. We thank Giorgio Dall’Olmo and Dionysios Raitsos for encouragement, motivation and useful discussions. We thank Stefano Ciavatta for help with statistical interpretation. The work was initially presented at the 2016 European Space Agency Living Planet Symposium in Prague before being updated and written up. The work was supported by and NERC's UK National Centre for Earth Observation. JS time was funded by the NERC project Candyfloss NE/K002058/1. We thank the NERC Earth Observation Data Acquisition and Analysis Service (NEODAAS) for the AVHRR imagery. Finally, we thank the editor and two anonymous reviewers for useful comments on our manuscript.
This is the author accepted manuscript. The final version is freely available from Elsevier via the DOI in this record.
Vol. 196, 5 September 2017, pp. 276–289