| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5.54 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Este estudo foca-se na análise dos registos de Temperatura de Superfície do
Mar (TSM) oriundos de boias derivantes (GDP), de um modelo numérico (MOD) e
de satélites (SAT), na região da Macaronésia, no período de 30 de junho de 2019 a 30
de junho de 2020. A principal finalidade é identificar as limitações e potencialidades
do uso de cada um destes métodos, nesta área.
Para tal, foram utilizados o Root Mean Square Error (RMSE), Mean Bias
Error (MBE), Mean Absolute Error (MAE) e correlação de Kendall para avaliar a
diferença entre os três conjuntos de dados. Comparou-se estes erros por subárea de
forma a estudar a variabilidade espaciotemporal. Dadas as características de cada
metodologia, considerou-se as TSMGDP como a ground truth.
As boias medem a TSM através de termístores e apenas têm uma cobertura
de 77%. Estas medem os dados in-situ (mais confiável) e em tempo real. Quanto ao
MOD, este assimila dados de boias fundeadas, derivantes, navios e imagens satélite,
e nunca poderá ser independente destes. No entanto, fornece dados instantâneos
a todas as horas do dia, efetua previsões e análises aos seus registos e tem uma
abrangência global espacial e temporal. Por último, o SAT combina vários resultados
de imagens microondas e infravermelhos. Carecem de calibração, validação por
dados in-situ e tem uma cobertura global. O resultado das diferenças é a medição
de diferentes profundidades da camada de superfície.
Neste estudo, as diferenças entre os conjuntos de dados verificam-se maiores
perante fenómenos de mesoescala, nomeadamente eddies, e estes encontram-se mais
frequentes nas regiões do Arquipélago dos Açores e Cabo Verde, nos meses mais
frios.
This study focuses on the analysis of Sea Surface Temperature (SST) records from drifting buoys, a numerical model (MOD) and satellites (SAT), in the Macaronesia region, from June 30, 2019 to June 30, 2020. The main purpose is to identify limitations and the capabilities of the usage of each of these in this particular area. The Root Mean Square Error (RMSE), Mean Bias Error (MBE), Mean Absolute Error (MAE) and Kendall correlation were used to evaluate the differences between these three datasets. These errors were compared by sub-area in order to study the spatiotemporal variability. Regarding the characteristics of each methods, the GDP data was considered the study’s ground truth. Buoys use thermistor to take measurements of SST but only have a 77% coverage. They measure data in-situ (most reliable) and in real time. As for the MOD, it assimilates data from anchored buoys, drifts, ships and satellite images, and can never be independent from them. Moreover, it offers instantaneous data at each hour of day, it provides forecast and analysis of its records, and has a global spatial and temporal scope. SAT combines various microwave and infrared imaging results. Requires calibration and validation by data in-situ.The result of the differences is the difference in surface layer depths. In this study, the differences between the data sets are greater in the face of mesoscale phenomena (meanders and eddies), and these are more frequent in the regions of the Archipelago of the Azores and Cape Verde in the colder months.
This study focuses on the analysis of Sea Surface Temperature (SST) records from drifting buoys, a numerical model (MOD) and satellites (SAT), in the Macaronesia region, from June 30, 2019 to June 30, 2020. The main purpose is to identify limitations and the capabilities of the usage of each of these in this particular area. The Root Mean Square Error (RMSE), Mean Bias Error (MBE), Mean Absolute Error (MAE) and Kendall correlation were used to evaluate the differences between these three datasets. These errors were compared by sub-area in order to study the spatiotemporal variability. Regarding the characteristics of each methods, the GDP data was considered the study’s ground truth. Buoys use thermistor to take measurements of SST but only have a 77% coverage. They measure data in-situ (most reliable) and in real time. As for the MOD, it assimilates data from anchored buoys, drifts, ships and satellite images, and can never be independent from them. Moreover, it offers instantaneous data at each hour of day, it provides forecast and analysis of its records, and has a global spatial and temporal scope. SAT combines various microwave and infrared imaging results. Requires calibration and validation by data in-situ.The result of the differences is the difference in surface layer depths. In this study, the differences between the data sets are greater in the face of mesoscale phenomena (meanders and eddies), and these are more frequent in the regions of the Archipelago of the Azores and Cape Verde in the colder months.
Description
Keywords
Comparação TSM GDP satélites modelo numérico Macaronésia