Data Reconstruction for Remotely Sensed Chlorophyll-a Concentration in the Ross Sea Using Ensemble-Based Machine Learning

Park, Jinku; Kim, Hyun Cheol; Bae, Dukwon; Jo, Young‐Heon

doi:10.3390/rs12111898

Cited by 19 publications

(20 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a result, 17,192,761 pixels were used for model training, and 5,730,921 pixels were used for validation. Furthermore, the critical parameters determining ultimate model performance, number of trees, and number of features were set to 60 and 3, consistent with those in [21]. As such, the R-squared (R 2 ), root mean square error (RMSE), and mean absolute error (MAE) for the training set were 0.99, 0.02 mg m -3 , and 0.01, respectively (Figure 1).…”

Section: Filling Gaps On Chl-a Datamentioning

confidence: 84%

“…We used a machine learning technique to fill the cloud-induced gaps in ocean color, which allows for a more accurate interpretation of phytoplankton biomass in time and space. At high latitudes, ocean color products such as Chl-a contain massive missing data due to various causes, such as the presence of clouds and sea ice [21]. If the spatial and temporal mean is constructed with missing data, the accuracy may be significantly affected [22].…”

Section: Filling Gaps On Chl-a Datamentioning

confidence: 99%

“…The machine learning model used in this study is the random forest (RF) [23]. The overall framework is similar to previous works [21,24]. However, there were several different preprocessing steps for Chl-a images: (i) a 3 × 3 median filter was applied to Chl-a data, (ii) Chl-a images with only 0.1% pixels were excluded, and (iii) the pixels that were not observed 15 days before and after were also removed for continuity.…”

Section: Filling Gaps On Chl-a Datamentioning

confidence: 99%

See 2 more Smart Citations

Phytoplankton Bloom Changes under Extreme Geophysical Conditions in the Northern Bering Sea and the Southern Chukchi Sea

Park

Lee

et al. 2021

Remote Sensing

Self Cite

View full text Add to dashboard Cite

The northern Bering Sea and the southern Chukchi Sea are undergoing rapid regional biophysical changes in connection with the recent extreme climate change in the Arctic. The ice concentration in 2018 was the lowest since observations began in the 1970s, due to the unusually warm southerly wind in winter, which continued in 2019. We analyzed the characteristics of spring phytoplankton biomass distribution under the extreme environmental conditions in 2018 and 2019. Our results show that higher phytoplankton biomass during late spring compared to the 18-year average was observed in the Bering Sea in both years. Their spatial distribution is closely related to the open water extent following winter-onset sea ice retreat in association with dramatic atmospheric conditions. However, despite a similar level of shortwave heat flux, the 2019 springtime biomass in the Chukchi Sea was lower than that in 2018, and was delayed to summer. We confirmed that this difference in bloom timing in the Chukchi Sea was due to changes in seawater properties, determined by a combination of northward oceanic heat flux modulation by the disturbance from more extensive sea ice in winter and higher surface net shortwave heat flux than usual.

show abstract

Section: Filling Gaps On Chl-a Datamentioning

confidence: 84%

Section: Filling Gaps On Chl-a Datamentioning

confidence: 99%

Section: Filling Gaps On Chl-a Datamentioning

confidence: 99%

See 1 more Smart Citation

Phytoplankton Bloom Changes under Extreme Geophysical Conditions in the Northern Bering Sea and the Southern Chukchi Sea

Park

Lee

et al. 2021

Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Note that the bias caused by the different measurement depth was corrected for the satellite-derived SSTs (refer to 2.2). RF has been widely used to examine the regression problems of ocean parameters in recent years [26][27][28][30][31][32]56]. RF is a nonparametric ensemble approach that composes a multitude of bootstrapped regression trees [57][58][59][60][61].…”

Section: Variable Type Variablementioning

confidence: 99%

High-Resolution Seamless Daily Sea Surface Temperature Based on Satellite Data Fusion and Machine Learning over Kuroshio Extension

Jung

Yoo

2022

Remote Sensing

View full text Add to dashboard Cite

Sea SurfaceTemperature (SST) is a critical parameter for monitoring the marine environment and understanding various ocean phenomena. While SST can be regularly retrieved from satellite data, it often suffers from missing data due to various reasons including cloud contamination. In this study, we proposed a novel two-step data fusion framework for generating high-resolution seamless daily SST from multi-satellite data sources. The proposed approach consists of (1) SST reconstruction based on Data Interpolate Convolutional AutoEncoder (DINCAE) using the SSTs derived from two satellite sensors (i.e., Moderate Resolution Imaging Spectroradiometer (MODIS) and Advanced Microwave Scanning Radiometer 2(AMSR2)), and (2) SST improvement through data fusion using random forest for consistency with in situ measurements with two schemes (i.e., scheme 1 using the reconstructed MODIS SST variables and scheme 2 using both MODIS and AMSR2 SST variables). The proposed approach was evaluated over the Kuroshio Extension in the Northwest Pacific, where a highly dynamic SST pattern can be found, from 2015 to 2019. The results showed that the reconstructed MODIS and AMSR2 SSTs through DINCAE yielded very good performance with Root Mean Square Errors (RMSEs) of 0.85 and 0.60 °C and Mean Absolute Errors (MAEs) of 0.59 and 0.45 °C, respectively. The results from the second step showed that scheme 2 and scheme 1 produced RMSEs of 0.75 and 0.98 °C and MAEs of 0.53 and 0.68 °C, respectively, compared to the in situ measurements, which proved the superiority of scheme 2 using multi-satellite data sources. Scheme 2 also showed comparable or even better performance than two operational SST products with similar spatial resolution. In particular, scheme 2 was good at simulating features with fine resolution (~50 km). The proposed approach yielded promising results over the study area, producing seamless daily SST products with high quality and high feature resolution.

show abstract

“…Park et al (2019) also found that the CHL reconstructed for multiple variables through the RF model more accurately simulated spatial patterns and CHL values when compared with other techniques. In addition, Park et al (2020) qualitatively and quantitatively studied the effects of various factors on CHL through RF.…”

mentioning

confidence: 99%

Role of Aerosols in Spring Blooms in the Central Yellow Sea During the COVID-19 Lockdown by China

et al. 2022

Self Cite

View full text Add to dashboard Cite

Reduced amounts of aerosols blowing into the Yellow Sea (YS), owing to the temporary lockdown of factories in China during COVID-19, resulted in a 15% decrease in spring chlorophyll-a concentration (CHL) in March 2020 compared to its mean March values from 2003 to 2021. Particularly, the effect of land-based AOD is insignificant compared with that of atmospheric aerosols flowing into the YS, as indicated by the currents and wind directions. Hence, the main objective of this study was to understand the relationship between atmospheric aerosols and CHL by quantitatively considering relevant environmental changes using a Random Forest (RF) algorithm. Various input physical forcing variables to RF were employed, including aerosol optical depth (AOD), sea surface temperature (SST), mixed layer depth (MLD), wind divergence (WD), and total precipitation (TP). From the RF-based analysis, we estimated the relative contribution of each physical forcing variable to the difference in CHL during and after the COVID-19 lockdown period. The sensitivity of the RF model to changes in aerosol levels indicated positive effects of increased amounts of aerosols during spring blooms. Additionally, we calculated the quantitative contribution of aerosols to CHL changes. When SST was warmer and TP was lower than their climatology in March 2020, CHL increased by 0.22 mg m-3 and 0.02 mg m-3, respectively. Conversely, when MLD became shallower and AOD was lower than their climatology, CHL decreased as much as 0.01 mg m-3 and 0.20 mg m-3. Variations in WD caused no significant change in CHL. Overall, the specific estimations for reduced spring blooms were caused by a reduction in aerosols during the COVID-19 lockdown period. Furthermore, the RF developed in this study can be used to examine CHL changes and the relative role of significant environmental changes in biological blooms in the ocean for any normal year.

show abstract

Data Reconstruction for Remotely Sensed Chlorophyll-a Concentration in the Ross Sea Using Ensemble-Based Machine Learning

Cited by 19 publications

References 51 publications

Phytoplankton Bloom Changes under Extreme Geophysical Conditions in the Northern Bering Sea and the Southern Chukchi Sea

Phytoplankton Bloom Changes under Extreme Geophysical Conditions in the Northern Bering Sea and the Southern Chukchi Sea

High-Resolution Seamless Daily Sea Surface Temperature Based on Satellite Data Fusion and Machine Learning over Kuroshio Extension

Role of Aerosols in Spring Blooms in the Central Yellow Sea During the COVID-19 Lockdown by China

Contact Info

Product

Resources

About