title={{Local Mixing Determines Spatial Structure of Diahaline Exchange Flow in a Mesotidal Estuary: A Study of Extreme Runoff Conditions}},
volume={54},
year={2024}
}
@article{Blauw2009,
abstract={The set-up, application and validation of a generic ecological model (GEM) for estuaries and coastal waters is presented. This model is a comprehensive ecological model of the bottom of the foodweb, consisting of a set of modules, representing specific water quality processes and primary production that can be combined with any transport model to create a dedicated model for a specific ecosystem. GEM links different physical, chemical and ecological model components into one generic and flexible modelling tool that allows for variable sized, curvilinear grids to accomodate both the requirements for local accuracy while maintaining a relatively short model run-time. The GEM model describes the behaviour of nutrients, organic matter and primary producers in estuaries and coastal waters, incorporating dynamic process modules for dissolved oxygen, nutrients and phytoplankton. GEM integrates the best aspects of existing Dutch estuarine models that were mostly dedicated to only one type of ecosystem, geographic area or subset of processes. Particular strengths of GEM include its generic applicability and the integration and interaction of biological, chemical and physical processes into one predictive tool. The model offers flexibility in choosing which processes to include, and the ability to integrate results from different processes modelled simultaneously with different temporal resolutions. The generic applicability of the model is illustrated using a number of representative examples from case studies in which the GEM model was successfully applied. Validation of these examples was carried out using the 'cost function' to compare model results with field observations. The validation results demonstrated consistent accuracy of the GEM model for various key parameters in both spatial dimensions (horizontally and vertically) as well as temporal dimensions (seasonally and across years) for a variety of water systems without the need for major reparameterisation. {\textcopyright} 2008 Springer Science+Business Media B.V.},
author={Blauw, Anouk N. and Los, Hans F.J. and Bokhorst, Marinus and Erftemeijer, Paul L.A.},
abstract={A new empirical equation is introduced that describes the photosynthesis by phytoplankton as a single, continuous function of available light from the initial linear response through the photoinhibited range at the highest levels liable to be encountered under any natural conditions. The properties of the curve are derived, and a procedure is given for fitting it to the results of light-saturation experiments for phytoplankton. The versatility of the equation is illustrated by data collected on natural phytoplankton assemblages from the eastern Canadian arctic and from the continental shelves of Nova Scotia and Peru.},
author={Platt, T. and Gallegos, C. L. and Harrison, W. G.},
file={:home/og/references/general-references/Photoinhibition of photosynthesis in natural assemblages of marin.pdf:pdf},
issn={00222402},
journal={J. Mar. Res.},
number={4},
pages={687--701},
title={{PHOTOINHIBITION OF PHOTOSYNTHESIS IN NATURAL ASSEMBLAGES OF MARINE PHYTOPLANKTON}},
title={{Photoinhibition of photosynthesis in natural assemblages of marine phytoplankton}},
OxyPOM (Oxygen and Particulate Organic Matter) and DiaMO (Diagnostic Model for Oxygen) are aquatic biogeochemical models that consider key processes for dissolved oxygen (DO) dynamics, such as re-aeration, mineralization, and primary production.
Both models are implemented in the Fortran-based Framework for Aquatic Biogeochemical Models [FABM, @Bruggeman2014], which enables their deployment in different physical drivers in realistic and idealized applications.
Additional routines for calculating the attenuation of photosynthetically active radiation are included.
The processes represented in OxyPOM and DiaMO enable studying DO in fresh, marine, and transitional waters.
With this model, we include a testcase for simulating DO in Cuxhaven Station in the Elbe estuary from 2005 to 2024.
This testcase uses the General Ocean Turbulence Model (GOTM) [@Burchard2002] to simulate vertical 1D hydrodynamics with realistic parameterization for tidal dynamics [@Reese2024] and scripts for downloading real meteorological forcing from [kuestendaten.de](https://www.kuestendaten.de).
OxyPOM (Oxygen and Particulate Organic Matter) and DiaMO (Diagnostic Model for Oxygen) are aquatic biogeochemical models that consider key processes for dissolved oxygen (DO), such as re-aeration, mineralization, and primary production, in fresh, transitional and marine waters.
Both are implemented in the `Fortran`-based Framework for Aquatic Biogeochemical Models [FABM, @Bruggeman2014] for interoperability in a variety of hydrodynamic models, in realistic and idealized applications, and for coupleability to other aquatic process models.
With these models, we include an updated light profile implementation and testcases for simulating DO at Cuxhaven Station in the Elbe estuary 2005--2024; for this, we use the hydrodynamic General Ocean Turbulence Model (GOTM) [@Burchard2002] including tides, and `R` and `bash` scripts for including weather and river data from [kuestendaten.de](https://www.kuestendaten.de).
# Statement of need
Dissolved oxygen (DO) is a key variable for water quality assessment of the ecological state of running and standing aquatic ecosystems [@EC2006], and an intermediate step between atmospheric extremes and biological impacts [@Tassone2022].
Models for DO in waters are thus necessary; most existing models, however, describe DO dynamic as a side product of more or less complex biotic and abiotic dynamics. These new models OxyPOM and DiaMO focus on the key processes that produce or consume oxygen while removing the complexity of adjacent processes.
OxyPOM was implemented by @Holzwarth2018 using DELWAQ [@Blauw2009] and coupled with UnTRIM [@Sehili2014] as the physical driver in an idealization of the Elbe estuary.
This implementation, however, was limited to this specific application and thus lacked portability.
Implementing this model in the Fortran-based Framework for Aquatic Biogeochemical Models (FABM) [@Bruggeman2014], OxyPOM can be used with many physical drivers; in different geographical domains; and coupled with other biogeochemical models.
OxyPOM uses vertically explicit formulations for re-aeration, primary production, and light attenuation, which are lacking in the @Holzwarth2018 implementation.
DiaMO is a simplified model for quick assessments for DO dynamics in applications where modelling complete bio-geochemical dynamics is not needed.
Dissolved oxygen (DO) is a key variable for assessing water quality and ecological state of aquatic ecosystems [@EC2006].
Most existing models describe DO dynamics as a side product of more or less complex (a)biotic dynamics.
OxyPOM and DiaMO remove much of this complexity and focus on the key processes that produce or consume oxygen.
A predecessor 1D long-channel version of OxyPOM was initially implemented by @Holzwarth2018 specifically for the closed-source UnTRIM-DELWAQ hydrodynamic and water quality system.
This implementation lacked portability, and was neither findable, nor accessible, interoperable or reusable (FAIR). The reimplementation with the FABM API ensures FAIR principles, foremost the interoperability with (1) many hydrodynamic models, (2) other aquatic process models, and (3) reusability in different topological domains.
For vertically resolved applications, we added formulations for re-aeration, primary production, and light attenuation.
Where a complete representation of bio-geochemical dynamics is not needed, DiaMO is an even more simplified model for quick assessments of DO dynamics.
## OxyPOM: Oxygen and Particulate Organic Matter
The model OxyPOM resolves the dynamics of
dissolved oxygen (DO),
particulate organic matter (POM)--fresh and refractory--,
DO dynamics is based on a mass balance equation accounting for re-aeration and photosynthesis as oxygen sources, and respiration, mineralization and nitrification as sinks:
This equation is applied to each water volume considered by the physical driver.
Re-aeration occurs in the surface-most layer as a function of temperature, salinity and wind speed.
Photosynthesis is limited by nutrient concentration and light intensity in each layer.
Respiration includes oxygen consumption for both micro-algae classes.
Mineralization is the oxygen consumed to transform matter from organic to inorganic forms.
Nitrification is the oxygen consumed to oxidize ammonia into nitrate.
Temperature-dependent rates limit these processes.
In OxyPOM, POM and DOM have an explicit elemental composition (carbon, nitrogen and phosphorus).
POM is present in two qualities, fresh and refractory, which transition in the sequence fresh $\rightarrow$ refractory $\rightarrow$ dissolved.
POM and DOM mineralize to inorganic dissolved inorganic nutrients: nitrogen and ortho-phosphate.
Dissolved inorganic nitrogen is further subdivided into ammonium and nitrate.
Ammonia transitions to nitrate--nitrification-- as a function of DO.
Silicon is present in dissolved--bio-available-- and particulate mineral forms.
The two micro-algae classes (ALG1 and ALG2) have growth rates that depend on temperature, light intensity and inorganic nutrient concentrations.
Both ALG1 and ALG2 growth depend on dissolved nitrogen and ortho-phosphate concentrations, and only ALG1 depends on free silicate, thus representing diatoms.
Microalgae uptake dissolved inorganic nutrients and release dissolved nutrients when they die with a temperature-dependent mortality rate.
While the full description of the processes in OxyPOM is available in @Holzwarth2018, some changes were required for a vertically explicit implementation within FABM:
Re-aeration occurs in the surface layer as a function of temperature, salinity and wind speed [@Weiss1970].
Photosynthesis is limited by nutrient concentration and light intensity with an exponential saturation [@Platt1980].
Respiration includes oxygen consumption for micro-algae.
Mineralization and Nitrification is the oxygen consumed to transform matter from organic to inorganic forms and to oxidize ammonia into nitrate, respectively.
1. Re-aeration is calculated as surface oxygen transference using saturation concentration as a function of temperature, salinity and wind speed [@Weiss1970].
This approach is commonly used in other models, such as ERSEM [@Butenschon2016].
POM and DOM have an explicit elemental composition (carbon, nitrogen and phosphorus);
POM is present in two qualities, which transition in the sequence labile $\rightarrow$ semi-labile $\rightarrow$ dissolved.
POM and DOM mineralize to inorganic dissolved nitrogen and ortho-phosphate.
Dissolved inorganic nitrogen is the sum of ammonium and nitrate; ammonium transitions to nitrate as a function of DO.
Silicate is present in dissolved --bio-available-- and particulate mineral forms.
The two micro-algae classes (one with dependence on dissolved silicate, thus representing diatoms) have growth rates that depend on photosynthesis; their growth rates depend on dissolved nitrogen and ortho-phosphate concentrations.
Micro-algae take up and release dissolved nutrients when they die with a temperature-dependent mortality rate.
2. Light limitation for photosynthesis is calculated with an exponential saturation [@Platt1980].
OxyPOM shows high skill by reproducing surface DO at the Cuxhaven station in the Elbe estuary (\autoref{fig:validation}).
3. Settling velocities are set constant, and vertical redistribution of matter is carried out by the physical driver.
We validate both models in the Cuxhaven station in the Elbe estuary, where OxyPOM shows high skill by reproducing surface DO.
{ width=99% }\
<div>
\label{fig:validation}](figure1.png){ width=99% }
</div>
## DiaMO: Diagnostic Model for Oxygen
DiaMO resolves the dynamics of DO, living and non-living organic particulate carbon forms (phytoplankton and detritus, respectively) under the assumption that light, not nutrients, is the limiting factor for primary production.
DiaMO is thus a carbon-based implementation.
DiaMO resolves the dynamics of DO, living and non-living organic particulate carbon forms (Phytoplankton (Phy) and Detritus (Det), respectively) under the assumption that light, not nutrients, is the limiting factor for photosynthesis; DiaMO is a carbon-only implementation.
DO is solved with the mass balance equation of OxyPOM (\autoref{eq:do}), setting nitrification to zero.
In DiaMO, aggregation rate is a mortality term for phytoplankton [@Maerz2009].
As in OxyPOM, all rates in DiaMO are temperature-dependent.
Aggregation rate is a mortality term for phytoplankton [@Maerz2009].
As in OxyPOM, all rates in DiaMO are temperature-dependent. DiaMO was equally validated and shows high skill reproducing surface DO at the Cuxhaven station.
## Light in OxyPOM and DiaMO
<!-- this paragraph needs rewriting as light is its own model -->
Together with OxyPOM and DiaMO, this repository includes the model oxypom/light as an alternative model to the FABM implementation of the light model used by GOTM [@Burchard2002].
While the default light model assumes that the photosynthetically active radiation (PAR) in a vertical layer $z$ of thickness $\Delta z$ is in the centre of the layer, oxypom/light calculates PAR in the representative depth $\bar{z}$, which satisfies the mean value theorem.
PAR evaluated at $\bar{z}$ is thus the mean PAR intensity on the layer.
Since this calculation can be computationally expensive to evaluate, first- and second-order approximations are possible.
Together with OxyPOM and DiaMO, this repository includes the model `oxypom/light` as an alternative model to the FABM implementation of the light model used by GOTM.
While the default light model assumes that the photosynthetically active radiation (PAR) in a vertical layer $z$ of thickness $\Delta z$ is in the centre of the layer, `oxypom/light` calculates PAR in the representative depth $\bar{z}$, which satisfies the mean value theorem, such that
PAR evaluated at $\bar{z}$ is the mean PAR intensity in the layer.
Since this calculation can be computationally expensive to evaluate, first- and second-order approximations are implemented.
The first-order solution is equal to the centre of the layer $z + \frac{1}{2} \Delta z$, and the second-order approximation is
\begin{equation}
...
...
@@ -125,8 +109,8 @@ where $\alpha$ is the light extinction coefficient for the layer, accounting for
# Model documentation and license
The models are documented in short form in the `ReadMe.md` section of the repository and a complete description of the science behind OxyPOM is in @Holzwarth2018.
Open access data from third parties are not included with the model, scripts for their download are however included.
Our own models, scripts and documentations are are released under open source licenses, mostly Apache 2.0, CC0-1.0, and CC-BY-SA-4.0; a comprehensive documentation of all licenses is provided via REUSE Software.
Open access data from third parties are not included with the model, and scripts for their download are included.
Our own models, scripts and documentations are are released under open source licenses, foremost Apache 2.0, CC0-1.0, and CC-BY-SA-4.0; a comprehensive documentation of all licenses is provided via REUSE Software.
