Bittig, Henry C. (2014) Towards a Quantum Leap in Oceanic Oxygen Observation - From Oxygen Optode Characterization to Autonomous Observation of Gas Exchange and Net Community Production. (PhD/ Doctoral thesis), Christian-Albrechts-Universität Kiel, Kiel, Germany, 215 pp.

Preview

Text Bittig_Diss.pdf - Published Version Available under License Creative Commons: Attribution 3.0. Download (40MB) | Preview

Abstract

This thesis presents ways to use autonomous oxygen observations to quantitatively estimate biological productivity and gas exchange. For this purpose, high quality oxygen sensor data are essential and the metrological prerequisites are established through dedicated experiments and procedures. Oxygen optodes promise to be long term stable due to their optical detection principle. Thus they seem predestined for autonomous deployments. However, their response to oxygen is non-linear and they require a most accurate calibration. A new calibration setup is described which is based on electrochemical oxygen generation. It can be used to multipoint calibrate oxygen optodes with an accuracy of 1 μmol kg−1. This setup also helped in the identification of optode sensor drift which is stronger at high oxygen levels than at low oxygen. Such a drift severely compromises the field deployability of optodes as well as the current routine of optode factory or laboratory calibration, sensor shipment, and subsequent deployment with an associated delay of a week to a year. Drift rates of up to −3.5 % in 2 months strongly impair the potential to reach a targeted field accuracy of 1 μmol/kg (Gruber et al., 2010). An in-situ referencing method is thus crucial to detect and quantify an optode drift. Based on near-surface and in-air measurements, a method to accurately reference optodes on Argo floats is described. This referencing method is available throughout the entire instrument’s lifetime and thus also capable to correct not only pre-deployment drift but also any further drift during deployment. Profiling applications rely on a fast time response to adequately resolve gradients. Optodes, however, do have a slow response compared to electrochemical oxygen sensors. The time response of oxygen optodes is extensively characterized in laboratory and field experiments and the impact of the response time on sensor data is quantified. Response times can be drastically reduced when optodes are pumped and a framework is presented to predict the response time based on temperature and the flow conditions of the deployment platform. This metrological work lays the foundation for a new quality of autonomous oxygen observations. Concurrent underway measurements of oxygen together with nitrogen and carbon dioxide in Southern Ocean surface waters are presented, revealing the different level of physical or biological control. A mixed layer gas exchange model is applied and adequately reproduces the spatial and temporal distribution of physical super- and undersaturation. It is used to improve oxygen-based net community production (NCP) estimates by removing physical contributions, e.g., due to entrainment. Finally, biogeochemical and biooptical data from five floats in the North and South Atlantic subtropical gyres are discussed and analyzed. In both gyres, there are three depth layers with a distinct biogeochemical characteristic. First, a net productive layer (layer I) is found between the surface and the node of apparent oxygen utilization (AOU = 0) around 110 m where nitrate is completely depleted and there still is a net oxygen production. This layer encompasses a shallow oxygen maximum at ca. 60 – 70 m that forms below the mixed layer in spring, summer, and autumn. Second, there is a net heterotrophic layer (layer II) below until approx. 200 m characterized by a stoichiometric nitrate deficit compared to the oxygen consumption. This layer is split into two subdomains, one where nitrate is in deficit and no measurable nitrate is present above the nitracline (layer IIa), and one where there is measurable nitrate present but it is still in deficit (layer IIb). Third, waters below are net heterotrophic and oxygen and nitrate are stoichiometrically balanced (layer III). The boundary between layer IIb and III is determined by the advective supply of surface nutrients through subtropical mode water. From the 2 years of float data, quantitative estimates of NCP are derived by using a 1D abiotic model with a dedicated gas exchange module that successfully separates physical and biological effects on oxygen. Production and respiration as well as the stoichiometric imbalance (oxygen excess in the surface and nitrate deficit below) follow the above depth horizons. A mean annual NCP of 1.2 ± 1.1 mol C m−2 yr−1 and 1.6 ± 0.8 mol C m−2 yr−1 is obtained for the North and South Atlantic gyre. Optodes drifted during the deployment and a conservative in-situ correction was necessary. The NCP estimates therefore likely represent a lower bound. At the same time, this echoes the need to implement a proper in-situ reference approach as described in this work.

Actions (login required)

View Item