Comparison between remotely-sensed sea-surface temperature (AVHRR) and in situ records in San Matías Gulf (Patagonia, Argentina)
In situ records of sea surface temperature collected between 2005 and 2009 were used to compare, for the first time, the temperature estimated by the Multichannel algorithms (MCSST) of the Advanced Very High Resolution Radiometer (AVHRR) sensors in San Matías Gulf, in the north of the Argentinean Pa...
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Lenguaje: | English |
| Publicado: |
Pontificia Universidad Católica de Valparaíso. Facultad de Recursos Naturales. Escuela de Ciencias del Mar
2014
|
| Materias: | |
| Acceso en línea: | http://www.scielo.cl/scielo.php?script=sci_arttext&pid=S0718-560X2014000100016 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | In situ records of sea surface temperature collected between 2005 and 2009 were used to compare, for the first time, the temperature estimated by the Multichannel algorithms (MCSST) of the Advanced Very High Resolution Radiometer (AVHRR) sensors in San Matías Gulf, in the north of the Argentinean Patagonian Continental Shelf (between 40°47'-42°13'S). Match-ups between in situ records and satellite sea surface temperature (SST) were analyzed. In situ records came from fixed stations and oceanographic cruises, while satellite data came from different NOAA satellites. The fitting of temperature data to a Standard Major Axis (SMA) type II regression model indicated that a high proportion of the total variance (0.53< r² <0.99) was explained by this model showing a high correlation between in situ data and satellite estimations. The mean differences between satellite and in situ data for the full data set were 1.64 ± 1.49°C. Looking separately into in situ data from different sources and day and night estimates from different NOAA satellites, the differences were between 0.30 ± 0.60°C and 2.60 ± 1.50°C. In this paper we discuss possible reasons for the above-mentioned performance of the MCSST algorithms in the study area. |
|---|