Development and Validation of an Empirical Ocean Color Algorithm with Uncertainties: A Case Study with the Particulate Backscattering Coefficient. Issue 5 (3rd May 2021)

• Record Type: Journal Article
• Authors: McKinna, Lachlan I. W.; Cetinić, Ivona; Werdell, P. Jeremy
• Is Part Of: Journal of geophysical research. Volume 126:Issue 5(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-05-03
• Subjects: model validation -- ocean color -- optical backscattering -- remote sensing -- uncertainties; Oceanography -- Periodicals; 551.4605
• DOI: 10.1029/2021JC017231
• Languages: English
• ISSNs: 2169-9275

Abstract

Abstract: We explored how algorithm (model) and in situ measurement (observation) uncertainties can effectively be incorporated into empirical ocean color model development and assessment. In this study we focused on methods for deriving the particulate backscattering coefficient at 555 nm, b bp (555) (m −1 ). We developed a simple empirical algorithm for deriving b bp (555) as a function of a remote sensing reflectance line height (LH) metric. Model training was performed using a high‐quality bio‐optical dataset that contains coincident in situ measurements of the spectral remote sensing reflectances, R rs (λ) (sr −1 ), and the spectral particulate backscattering coefficients, b bp (λ). The LH metric used is defined as the magnitude of R rs (555) relative to a linear baseline drawn between R rs (490) and R rs (670). Using an independent validation dataset, we compared the skill of the LH‐based model with two other models. We used contemporary validation metrics, including bias and mean absolute error (MAE), that were corrected for model and observation uncertainties. The results demonstrated that measurement uncertainties do indeed impact contemporary validation metrics such as mean bias and MAE. Zeta‐scores and z ‐tests for overlapping confidence intervals were also explored as potential methods for assessing model skill. Plain Language Summary: If we repeat a scientific measurement multiple times, we expect to record slightly different values each time. The average of these data is reported as the measurement and the spread of data either side of the average as the measurement uncertainty. With the knowledge of measurement uncertainty sources, such as a measurement sensor's internal instability, we can transfer the uncertainty through mathematical models. Satellite sensors measure light reflected from the ocean. The "ocean color" reflectance signal contains information about seawater optical properties. In this research, we explored how measurement uncertainties can be treated when ocean color models, that decipher the reflectance signal, are constructed and evaluated. In a case study, we developed an ocean color model to predict the optical particulate backscattering coefficient; a quantity that describes how particles scatter light in a backwards direction. In a process called validation we assessed the skill of our model by comparing model‐estimated values with directly observed, or "sea‐truth, " values. We found that when measurement uncertainties were considered, the validation results changed. The line height model was also compared with two other existing methods and found to perform with similar skill in the open ocean and potentially better skill in murky coastal waters. Key Points: A reflectance line height metric was used as a predictor of the particulate backscattering coefficient at 555 nm The degree of overlap metric was used to correct validation skill metrics for measurement uncertainties in modeled and observed data Bland–Altman and zeta‐score plots were explored as alternatives to one‐to‐one scatter plots routinely used for algorithm validation