A Regional Neural Network Approach to Estimate Water-Column Nutrient Concentrations and Carbonate System Variables in the Mediterranean Sea: CANYON-MED

Abstract

A regional neural network-based method, "CANYON-MED" is developed to estimate nutrients and carbonate system variables specifically in the Mediterranean Sea over the water column from pressure, temperature, salinity, and oxygen together with geolocation and date of sampling. Six neural network ensembles were developed, one for each variable (i.e., three macronutrients: nitrates (NO − 3), phosphates (PO 3− 4) and silicates (SiOH 4), and three carbonate system variables: pH on the total scale (pH T), total alkalinity (A T), and dissolved inorganic carbon or total carbon (C T), trained using a specific quality-controlled dataset of reference "bottle" data in the Mediterranean Sea. This dataset is representative of the peculiar conditions of this semi-enclosed sea, as opposed to the global ocean. For each variable, the neural networks were trained on 80% of the data chosen randomly and validated using the remaining 20%. CANYON-MED retrieved the variables with good accuracies (Root Mean Squared Error): 0.73 µmol.kg −1 for NO − 3 , 0.045 µmol.kg −1 for PO 3− 4 and 0.70 µmol.kg −1 for Si(OH) 4 , 0.016 units for pH T , 11 µmol.kg −1 for A T and 10 µmol.kg −1 for C T. A second validation on the ANTARES independent time series confirmed the method's applicability in the Mediterranean Sea. After comparison to other existing methods to estimate nutrients and carbonate system variables, CANYON-MED stood out as the most robust, using the aforementioned inputs. The application of CANYON-MED on the Mediterranean Sea data from autonomous observing systems (integrated network of Biogeochemical-Argo floats, Eulerian moorings and ocean gliders measuring hydrological properties together with oxygen concentration) could have a wide range of applications. These include data quality control or filling gaps in time series, as well as biogeochemical data assimilation and/or the initialization and validation of regional biogeochemical models still lacking crucial reference data. Matlab and R code are available at https:// github.com/MarineFou/CANYON-MED/.

Multiple Linear Regression Models for Reconstructing and Exploring Processes Controlling the Carbonate System of the Northeast US From Basic Hydrographic Data

“…In response to the sparseness of carbonate system observations, researchers have developed empirical models for estimating carbonate system parameters from basic hydrographic data in other regions (e.g., Alin et al., 2012; Evans et al., 2013; Fourrier et al., 2020; Juranek et al., 2009; Kim et al., 2010; Li et al., 2016; Table 1) as well as globally (Lee et al., 2000, 2006; Millero et al., 1998). The empirical models produced are unique to each region and depend on the balance of global and local processes controlling carbonate chemistry in each system.…”

Deep chlorophyll maximum and nutricline in the Mediterranean Sea: emerging properties from a multi-platform assimilated biogeochemical model experiment

“…In the Mediterranean Sea, a season-dependent sampling frequency that is higher during winter and spring surface blooms and lower during summer slow-dynamic conditions could compensate for the battery-saving needs and the maximization of float impact in joint float-satellite assimilation. In addition, to increase the spatial impact of data assimilation, future operational implementations can be based on pseudo profiles reconstructed by neural network approach such as CANYON-MED that uses the larger coverage of the Argo network and oxygen sensors (Fourrier et al, 2020). The present results of the joint assimilation of multi-stream chlorophyll data showed a mitigation of the increase in the RMSD of chlorophyll with respect to the non-assimilated chlorophyll dataset in the single-stream assimilation simulations (Figs.…”

Efficacy of Feedforward and LSTM Neural Networks at Predicting and Gap Filling Coastal Ocean Timeseries: Oxygen, Nutrients, and Temperature

“…Despite, their widespread use in Oceanography, examples of NNs that predict nutrients or gap-fill temperatures in the water column (i.e., below the surface) are scarce (exceptions include Sauzède et al, 2017, Bittig et al, 2018, and Fourrier et al, 2020 especially after the breakthroughs in training efficiency in 2006 (Emmert-Streib et al, 2020), and the few studies that do accomplish this do not take advantage of modern frameworks, such as TensorFlow. TensorFlow is an open-end ecosystem/framework for neural network modeling.…”