Cloud-free Sea Surface Temperature Research Articles

Cloud-Free Sea-Surface-Temperature Image Reconstruction From Anomaly Inpainting Network

Sea-surface temperature (SST) images obtained by satellites contain noise and missing SSTs due to cloud covers. We propose a method for reconstructing denoised cloud-free SST images via deep-learning-based image inpainting. For denoizing, we use data-assimilation images to train a reconstruction network by considering the physical correctness of SSTs. For reconstruction stability, we introduce <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">anomaly inpainting network</i> , which does not directly complete missing SSTs but estimates the difference between the unobserved SSTs and the average SSTs. SSTs do not fluctuate much over a few days; thus, we can use recent average SSTs as a rough estimation of SSTs and can assume that the SST difference will be within a specific range. We conducted experiments to evaluate our method with satellite SST images and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in situ</i> SST data. The results indicate that our method with anomaly inpainting network qualitatively and quantitatively outperformed conventional SST image inpainting methods.

Oceanic Response to Tropical Cyclone Gonu (2007) in the Gulf of Oman and the Northern Arabian Sea: Estimating Depth of the Mixed Layer Using Satellite SST and Climatological Data

The category 5-equivalent tropical Cyclone Gonu (2007) was the strongest cyclone to enter the northern Arabian Sea and Gulf of Oman. The impact of this cyclone on the sea surface temperature (SST) cooling and deepening of the mixed layer was investigated herein using an optimally interpolated (OI) cloud-free sea surface temperature (SST) dataset, climatological profiles of water temperature, and data from Argo profilers. SST data showed a maximum cooling of 1.7–6.5 °C during 1–7 June 2007 over the study area, which is similar to that of slow- to medium-moving cyclones in previous studies. The oceanic heat budget equation with the assumptions of the dominant turbulent mixing effect was used to establish relationships between SST and mixed layer depth (MLD) for regions that were directly affected by cyclone-induced turbulent mixing. The relationships were applied to the SST maps from satellite to obtain maps of MLD for 1–7 June, when Gonu was over the study area. Comparing with the measured MLD from Argo data showed that this approach estimated the MLDs with an average error of 15%, which is an acceptable amount considering the convenience of this approach in estimating MLD and the simplifications applied in the heat budget equation. Some inconsistencies in calculating MLD were attributed to use of climatological temperature profiles that may not have appropriately represented the pre-cyclone conditions due to pre-existing cold/warm core eddies. Estimation of the diapycnal diffusion that quantified the turbulent mixing across the water column showed consistent temporal and spatial variations with the calculated MLDs.

Two-Decade Variability of Sea Surface Temperature and Chlorophyll-a in the Northern South China Sea as Revealed by Reconstructed Cloud-Free Satellite Data

Monthly sea surface temperature (SST) and chlorophyll-a (Chl-a) concentration during 1997–2018 in the Northern South China Sea (NSCS) were gap-filled using the Data INterpolating Empirical Orthogonal Functions (DINEOFs) approach. Good accuracy of the reconstructed data was verified based on an independent cross-validation data set. Significant spatial heterogeneities and seasonal variations were observed in the cloud-free SST and Chl-a data. An apparent decreasing trend in SST was found along the coast of Guangdong, east of Hainan Island, and in the Taiwan Strait with an average decreasing rate amounting to 0.01 °C per year. Increase in SST was observed in Beibu Gulf and the deep offshore region in the NSCS at an average rate as high as 0.009 °C per year. An increasing trend in Chl-a was detected in the northern Beibu Gulf, near the mouth of the Pearl River estuary, and around Shantou with an average rate of 0.005 mg <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\cdot \text{m}^{-3}$ </tex-math></inline-formula> per year. EOF analysis of SST and Chl-a time series showed that the first three modes of both SST and Chl-a can explain most of their variances with accumulative contributions of 99.1% and 82.3%, respectively. The effects of surface net heat flux (NHF), Oceanic Niño Index (ONI), wind speed, and river flux on the first three modes of SST and Chl-a were quantitatively elucidated. Factors that affected the temporal and spatial variations of Chl-a and SST in the NSCS were qualitatively discussed, including El Niño-Southern Oscillation (ENSO), monsoon, upwelling, typhoon, internal waves, and Kuroshio intrusion.

Machine learning techniques for regional scale estimation of high-resolution cloud-free daily sea surface temperatures from MODIS data

High-resolution sea surface temperature (SST) estimates are dependent on satellite-based infrared radiometers, which are proven to be highly accurate in the past decades. However, the presence of clouds is a big stumbling block when physical approaches are used to derive SST. This problem is more prominent across tropical regions such as Arabian Sea(AS) and Bay of Bengal(BoB), restricting the availability of high-resolution SST data for ocean applications. The previous studies for developing daily high-resolution cloud-free SST products mainly focus on fusion of multiple satellites and in-situ data products that are computationally expensive and often time consuming. At the same time, it was observed that the capabilities of data-driven approaches are not yet fully explored in the estimation of cloud-free high-resolution SST data. Hence, in this study an attempt has been made for the first time to estimate daily cloud free SST from a single sensor (MODIS Aqua) dataset using advanced machine learning techniques. Here, three distinct machine learning techniques such as Artificial Neural Networks (ANN), Support Vector Regression (SVR) and Random Forest (RF)-based algorithms were developed and evaluated over two different study areas within the AS and BoB using 10 years of MODIS data and in-situ reference data. Among the developed algorithms, the SVR-based algorithm performs consistently better. In AS region, while testing, the SVR-based SST estimates was able to achieve an adjusted coefficient of determination (Radj2) of 0.82 and root mean square error (RMSE) of 0.71 °C with respect to the in situ data. Similarly, in BoB too, the SVR algorithm outperforms the other algorithms with Radj2 of 0.78 with RMSE of 0.88 °C. Further, a spatio-temporal and visual analysis of the results as well as an inter-comparision with NOAA AVHRR daily optimally interpolated global SST (a standard SST product available in practice) the suggest that the proposed SVR-based algorithm has huge potential to produce operational high-resolution cloud-free SST estimates, even if there is cloud cover in the image.

Application of DINCAE to Reconstruct the Gaps in Chlorophyll-a Satellite Observations in the South China Sea and West Philippine Sea

The Data Interpolating Empirical Orthogonal Functions (DINEOF) method has demonstrated usability and accuracy for filling spatial gaps in remote sensing datasets. In this study, we conducted the reconstruction of the chlorophyll-a concentration (Chl-a) data using a convolutional neural networks model called Data-Interpolating Convolutional Auto-Encoder (DINCAE), and we compared its performance with that of DINEOF. Furthermore, the cloud-free sea surface temperature (SST) was used as a phytoplankton dynamics predictor for the Chl-a reconstruction. Finally, four reconstruction schemes were implemented: DINCAE (Chl-a only), DINCAE (Chl-a and SST), DINEOF (Chl-a only), and DINEOF (Chl-a and SST), denoted rec1, rec2, rec3, and rec4 respectively. To quantitatively evaluate the accuracy of these reconstruction schemes, both the cross-validation and in situ data were used. The study domain was chosen to be the Northern South China Sea (SCS) and West Philippine Sea (WPS), bounded by 115–125°E and 16–24°N to test the model performance for the reconstruction of Chl-a under different Chl-a controlling mechanisms. The in situ validation showed that rec1 performs best among the four reconstruction schemes, and that adding SST into the Chl-a reconstruction cannot improve the reconstruction results. However, for cross validation, adding SST can slightly improve spatial distributions of the root mean square error (RMSE) between the reconstructed data and the original data, especially over the SCS continental shelf. Furthermore, the potential of DINCAE prediction is confirmed in this paper; thus, the trained DINCAE model can be re-applied to reconstruct other missing data, and more importantly, it can also be re-trained using the reconstructed data, thereby further improving reconstruction results. Another consideration is efficiency; with similar reconstruction conditions, DINCAE is 5–10 times faster than DINEOF.

Adding value to INSAT-3D sea surface temperature fields using MODIS data over the tropical Indian Ocean

ABSTRACTAccurate estimates of sea surface temperature (SST) are crucial for climate studies, numerical weather prediction and air–sea interactions. Following the launch of India’s advanced geostationary satellite – INSAT-3D with two thermal infrared split window channels in 2013, it is now possible to monitor land and ocean surfaces more reliably at higher spatiotemporal scale. In this article, an attempt has been made to develop a more accurate infrared-based cloud-free SST estimates over the tropical Indian Ocean by the synergistic use of geostationary INSAT-3D and polar-orbiting Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) satellite measurements through a successive correction method. This method is applied primarily for the month of May 2015 at coarser spatial resolution and at a daily scale. Results are compared independently with multi-satellite SST estimates and also against in situ observations such as Argo floats and buoys. The merged SST product shows noticeable improvement over INSAT-3D-based estimates alone. Comparison of the merged SST product with Argo observations shows that the root mean square difference (RMSD) has been improved from 1.23 to 0.79 K, and bias and correlation are also significantly improved. Overall results indicate that the synergistic use of INSAT-3D and MODIS satellite observations has potential for more accurate SST estimation over the tropical Indian Ocean at finer temporal resolution and larger spatial coverage for several near-real time meteorological and oceanographic applications.

Global space-time scales for day-to-day variations of daily-minimum and diurnal sea surface temperatures: their distinct spatial distribution and seasonal cycles

Spatial and temporal scales of day-to-day sea surface temperature (SST) variation were estimated globally based on the microwave cloud-free SST data. By employing the diurnal correction method of SST using sea surface wind (SSW) and solar shortwave radiation data, the scales of daily-minimum SST and diurnal SST amplitude were independently investigated along with seasonal and regional dependency. The daily-minimum SST showed moderate spatial scales of O(150–400 km) and relatively long temporal scales of O(3–4 days). In eddy-rich areas, the small spatial and long temporal scales were found to be associated, which suggests that the sub-mesoscale SST structures could be reproduced by spatial–temporal interpolations. Diurnal SST amplitude, which is strongly affected by air–sea interaction and high activity in the warming seasons, showed extremely long/short scales in the zonal/meridional directions of up to 1000 km/O(100–200 km), and its temporal scales were less than 1 day.

DAILY BLENDED MICROWAVE AND INFRARED SEA SURFACE TEMPERATURE COMPOSITION

A simple system for daily cloud free sea surface temperature (SST) composition based on thermal AVHRR and microwave TMI data is presented in this paper. Barnes’ objective analysis (Barnes, 1964) is applied as an interpolator to merge these two data sources, which have different spatial and temporal resolutions in a daily SST composition and in a regular grid product. Three comparisons were carried out as follows. First, in situ SST (daily average) measurements from eleven PIRATA´s buoys were compared. The correlation coefficients results varied from 0.89 to 0.99, and RMSE, MAE and MBE values have not exceeded 0.57 for period from 2002 to 2010. Second, comparisons between daily SST composition and average daily in situ SST collected from twenty three drifting buoys for the period from May 2008 to October 2010. The statistics results are 0.94, 0.25, 0.19 and -0.002 for correlation, RMSE, MAE and MBE, respectively. Third, SST (daily average) time series generated by OSTIA project was compared. The temporal and spatial RMSE (considering the study area) values ranged from approximately 0.21oC to 1.50oC and its average was 0.47oC for the period from January 1st to May 31st, 2006. Besides, an investigation about the influence of the data homogenization in the SST interpolation is discussed. Validation results are quite consistent (with SST composition accuracy less than 1.0oC). Thus, aiming to fulfill the numerical oceanographic model assimilation purposes and other oceanographic features studies, the developed SST product may be recommended as a candidate.

Spatial and temporal variability of SST and ocean color in the Gulf of Maine based on cloud-free SST and chlorophyll reconstructions in 2003–2012

The spatial and temporal variability of sea surface temperature (SST) and Chlorophyll-a (Chl-a) in the Gulf of Maine (GOM) is examined using daily, cloud-free Data INterpolating Empirical Orthogonal Function (DINEOF) reconstructions during 2003–2012. The utility of the DINEOF SST and Chl-a is demonstrated through direct comparisons with buoy- and ship-based observations. EOF analyses of cloud-free products are further used to quantify the SST and Chl-a variability on seasonal to inter-annual timescales. The first mode of SST is dominated by an annual cycle in response to net surface heat flux, with SST lagging surface flux by ~57days. The second mode of SST underscores interactions between GOM, the Scotian Shelf, and the slope sea in response to the basin scale atmospheric forcing represented by the North Atlantic Oscillation. The third mode correlates well with the evolution of Scotian Shelf-slope frontal displacement. The first EOF mode of Chl-a is dominated by a winter–spring bloom and a fall bloom, with a spatial distribution modified by the tidal mixing that facilitates nutrient delivery from the deep ocean. The second EOF mode is likely associated with a winter bloom in the warm slope sea, where the low-frequency variations of second modes of SST and Chl-a are in phase, suggesting a possible coupling between physical and biological responses to atmospheric forcing. The third mode of the Chl-a is likely associated with freshening events associated with advection of the Scotian Shelf Water, which enhance stratifications in the eastern GOM.

Cloud-free sea surface temperature and colour reconstruction for the Gulf of Mexico: 2003–2009

The data interpolating empirical orthogonal function (DINEOF) method is applied to concurrent MODIS sea surface temperature (SST) and chlorophyll-a (chl-a) data to produce daily, 4 km, cloud-free SST and chl-a analyses for the Gulf of Mexico (GOM) from 2003 to 2009. Comparisons between SST analysis and in situ buoy temperature measurements indicate that the DINEOF method can accurately resolve temperature variability and solve the cloud-cover problem, which is a typical issue of remote-sensing observations. Based on significant correlations between cloud-free chl-a, SST and sea surface height (SSH) data in the GOM, a simple chl-a statistical prediction model is further developed. Favourable comparisons between model solutions and independent satellite chl-a observations indicate that the statistical model provides a feasible means to predict GOM chl-a field based on existing SST and SSH information, allowing for observational gap fillings when concurrent chl-a data were not available.

In-situ diurnal sea surface temperature variations and near-surface thermal structure in the tropical hot event of the Indo-Pacific warm pool

Diurnal Sea Surface Temperature (SST) variations and the near-surface thermal structure of the tropical hot event (HE) have been investigated using advanced in-situ equatorial observations with hourly temporal resolution. The information on the HE area defined by the satellite cloud-free SSTs is used to sample the in-situ observations. The in-situ SSTs sampled for the HE conditions show that a maximum (minimum) SST has a histogram mode at 30.8°C (29.0°C), and frequently appears at 15:00 (07:00) local time. The amplitude of the diurnal SST variation (DSST) is defined by the difference between the maximum and minimum SSTs. The mean DSST during HEs is greater than 0.5°C, and has a maximum of about 0.75°C at the HE peak. The time series of mean DSST gradually increases (rapidly decreases) before (after) the peak. The satellite SST has a systematic positive bias against the corresponding daytime SST measured by the Triangle Trans-Ocean buoy Network. This bias is enhanced under conditions of large in-situ DSST. One-dimensional numerical model simulation suggests that the systematic bias is caused by the sharp vertical temperature gradient in the surface layer of HE. The near-surface thermal structure is generated by conditions of high insolation and low wind speed, which is the typical HE condition.