A. Boetius
for Marine Microbiology
Bremen, Germany
F. Janssen
for Marine Microbiology
Bremen, Germany
C. Waldmann
for Marine Environmental Sciences
Bremen, Germany
And the HYPOX Team
Abstract – Hypoxic (low oxygen) conditions are increasing due to eutrophication and climate change. To better understand the dynamics and drivers of oxygen depletion, the EU-funded project HYPOX is starting to build innovative observation systems for continuous oxygen monitoring in aquatic systems. HYPOX further includes experimental and modelling studies on hypoxia drivers and consequences for ecosystems, to gain predictive and decision-making capabilities from the obtained monitoring data. All activities will be embedded into the framework of the Global Earth Observation System of Systems (GEOSS) which aims at integrating all earth related observations into a single source of data.
Aquatic Systems Short Of Breath: How Did It Arise And Why Should We Care?
Most aquatic life depends on oxygen. It is, thus, an alarming finding that the occurrence of hypoxic (low oxygen) conditions is increasing worldwide. This is mainly a consequence of anthropogenic eutrophication (nutrient input) and climate change. In eutrophied waters the excess algal biomass produced is typically not passed on along the food chain. Instead it sinks to the seafloor where it is utilized by micro-organisms consuming oxygen. If bottom water oxygen drops significantly, faunal communities and chemical conditions start to change. Ecosystems undergo successive deterioration, eventually turning into permanently anoxic environments where micro-organisms replace all higher life (see Figure below). This collapse of animal communities leads to a dramatic decline in ecosystem functions and services such as biodiversity, fisheries, aquaculture and tourism. Since the 1960s, the records of such ÛÏdead zonesÛ have doubled every ten years.
Figure 1: Location of oxygen observatory sites in coastal and open seas (blue dots), and land-locked water bodies (green dots) å© 
Oxygen monitoring capacities have to be improved before ecosystems lose functions that may take several decades to restore,. Within the project HYPOX, monitoring of oxygen and related parameters are carried out at several different sites to improve our understanding of hypoxia formation, potential effects of anthropogenic activities, and climate change on future oxygen levels (see Figure 1).
As ecosystem responses depend on frequency, duration, spatial extent and severity of hypoxia events, continuous monitoring of oxygen concentrations is needed. In order to understand the reasons of hypoxia formation and to be able to predict potential effects of anthropogenic activities and global warming on future oxygen levels, monitoring of oxygen and related parameters have to be carried out in a variety of aquatic systems that differ in oxygen status and sensitivity towards change. Experimental and modelling studies of hypoxia drivers as well as consequences of oxygen depletion for ecosystems are needed to gain predictive and decision-making capabilities from the monitoring data obtained.
Location and Characteristics Of Observatories and Target Sites
Oxygen availability at a given site depends on the balance between oxygen supply (mainly physical transport and mixing) and removal (mainly biological oxidation of organic matter) which both vary on different time scales. The response of an ecosystem to hypoxia is determined by even more factors (e.g., adaptation of organisms to hypoxia, availability of nearby fauna for re-colonization, sediment inventory of oxidized and reduced components). To cope with this complexity, monitoring efforts in HYPOX encompass sites that differ both in oxygen availability and sensitivity towards change: Open ocean, oxic areas with high sensitivity to global warming (Arctic Ocean) are included as well as semi-enclosed basins where oxygen is naturally depleted (Black Sea, Baltic Sea), and partly eutrophied landlocked systems that experience hypoxia seasonally or locally (fjords, lagoons, lakes).
Figure 2a: The Scottish Loch Etive as an example for the data acquisition process and modelling approach. Modelling of oceanographic conditions as a first step towards oxygen concentration prediction. Figure: Dmitry Aleynik, 
Simulations of hypoxia dynamics and consequences are an intrinsic part of the HYPOX work plan. Combining physical and biogeochemical modelling of both the water column and the sediments will provide means to extrapolate findings to similar ecosystems and to predict future hypoxia. These capabilities are needed to understand the effect of climate change and eutrophication on ecosystem functions and services and to decide on countermeasures that may be taken.
Continuous measurements of oxygen and associated parameters are an important first step to examine the status of the system that is monitored. In order to extend the gained knowledge, however, modelling efforts need to be made. Modelling is the key tool to turn observations into generalizations that can also be applied to other ecosystems and predictions that extend the current observations into the future. These generalizations and forecasting capabilities are essential to examine the effects of future climate and eutrophication scenarios for oxygen availability and ecosystem functioning. If ecosystems are deteriorating, modelling capabilities will also provide means to decide on adequate countermeasures to be taken. Being aware of these issues, the HYPOX strategy complements oxygen monitoring with modelling efforts. To fully comprehend oxygen dynamics, HYPOX modelling aims to combine physical transport and biogeochemical processes in both the sediment and the water column. The measurements produced by the observatories and during targeted field campaigns will be used to verify the models and, via data assimilation, to improve their predictive capabilities (see Figure 2). Model exploration will be used to extract early warning indicators and tipping points in system behavior. Combining observations and predictions of oxygen availability with existing knowledge about the effects of hypoxia on animal communities and ecosystems will improve our understanding of the potential loss of ecosystem functions and services as a consequence of climate change and eutrophication.
Figure 2b: The Scottish Loch Etive as an example for the data acquisition process and modelling approach. The double mooring configuration that allows for real-time access of the data. Figure: Henrik Stahl, 
One of the major issues of the HYPOX project is to give recommendations on standardized oxygen measurement procedures. This applies to the preparation, calibration and deployment of the according sensor systems. This issue is well known and established for physical sensors. However, as reliable biogeochemical sensors are commonly available for some ten years, standardized procedures for calibration are often lacking. Instead individual laboratories developed their own routines. Within HYPOX oxygen data from different sites have to be made comparable to be able to combine data sets and to identify and quantify global trends. Therefore a comprehensive description of the calibration process has to be provided to assure consistent results that are traceable to standard methods and procedures.
The harmonization process started with a calibration experiment which was carried out to agree on a common calibration routine for optical oxygen sensors (ÛÏoxygen optodesÛ), and to check for the calibration status of individual sensor systems. The formalization of all the individual processing steps will ultimately allow for a definition of a calibration protocol that can be used not just by the HYPOX community but also for any other monitoring purposes. As an example for the formalization process a workflow scheme for calculating the water oxygen concentration from an optode measurement is illustrated in Figure 3. This scheme allows identifying the individual data processing steps and helps to clarify the relevant methods and terms involved.
This scheme only applies to optical oxygen measurements. Other concepts have to be used and formalized for electrochemical Clark type oxygen electrodes. The general approach towards sensor calibration, however, applies to all types of oxygen sensors. However, there may be individual differences that are, for instance, related to the measuring principles that may call for additional processing steps like pressure cycling or for a predefined number of temperature steps to be adjusted.
Figure 3: Processing steps for deriving the oxygen concentration from optode phase measurements. The Phase denotes the phase shift between the modulated excitation light and the light that is emitted by a fluorescent dye (ÛÏfluorophoreÛ) that changes fluorescence properties as a function of the ambient oxygen partial pressure and the temperature.
The obtained polynomial calibration function is then used to calculate the absolute partial pressure as has been illustrated in Figure 3.
Figure 4: Workflow scheme for the proposed oxygen sensor calibrations. Left part: Oxygen calibration steps and temperature cycling. Right part: Workflow describing how the calibration function is obtained.
As a matter of fact quality assurance also applies to the deployment and control of the measurements at the individual sites as well. A comprehensive description is currently under development within the HYPOX project.
Standardization, Integration, Dissemination: Compliance with The Global Earth Observation Initiative
The Global Earth Observation System of Systems (GEOSS) will provide a comprehensive framework to disseminate and retrieve information about earth related processes. HYPOX will particularly contribute by providing recommendations on measurement procedures and best practices with regard to data collection and processing of in-situ oxygen measurements.
Figure 5: Example of a calibration function polynomial of 4th degree for a particular oxygen optode. The green lines are the actual measured values.
From the perspective of a scientific user, GEOSS will offer new opportunities to evaluate and interpret measurement results. For example, meteorological forcing will have an immediate impact on oxygen depleted waters in particular for shallow depths. The reliable availability of meteorological and standard ocean data is important for improved forecasts of oxygen conditions. Within HYPOX different modelling approaches will be used to demonstrate the forecasting capabilities. The HYPOX project will help to develop a vision on how GEOSS will help scientists to improve their knowledge on processes having a direct impact on the society and economy, in particular in regions close to the sea.
For further information on the HYPOX project, please contact:
Dr. Felix Janssen
Max Planck Institute for Marine Microbiology
Celsiusstrasse 1
28359 Bremen, Germany
email: fjanssen@mpi-bremen.de
URL of the project: www.hypox.net
Acknowledgement:
The authors would like to thank the European Commission for their support as part of the Seventh Framework program, Contract number 226213