Top parameters

A model parameter is a value that is included in the model equations. Often, this quantifies a physical, biogeochemical or biological feature of the ecosystem (e.g., a mortality rate of phytoplankton, a feeding preference of zooplankton, a sinking rate of particulate matter). Typically, this value is a constant, but it can be assumed to be variable (e.g. stocastically) in the framework assimilation approaches, to represent uncertainties in the model.
A model parameter is a variable that is internal to the model and whose value can be estimated from data. It is important to determine the top parameters for a given model system so as to:

Reduce the uncertainty on the model output indicators, by improving the estimates of the parameters;
Generate biogeochemical model ensembles, by perturbing the reference values of the parameters;
Improve understanding of biogeochemical processes and fluxes, by assimilating observations in a state-parameter estimation framework.

We ranked the biogeochemical parameters of all the Copernicus Marine Service models investigated in SEAMLESS. This was achieved by using the one-dimensional configurations of the models available to users in the SEAMLESS modelling prototype EAT (https://github.com/BoldingBruggeman/eat). The configurations were chosen to simulate different trophic regimes (oligotrophic deep-ocean station “BATS” in the Sargasso Sea and mesotrophic coastal station “L4” in the English Channel) in contrasting physical conditions (mixed and stratified water column). A Monte-Carlo based sensitivity of the parameters was then run to rank the most important (see the details in the deliverable D3.2 of SEAMLESS: doi: 10.5281/zenodo.6580236). The following tables list the 10 most important parameters for each biogeochemical model.

PISCES (Pelagic Interactions Scheme for Carbon and Ecosystem Studies; doi.org/10.5194/gmd-8-2465-2015)

PISCES
Rank	Notation	Description	Score	Group
1	dom_rem/xremik	DOM remineralization rate	100%	[10]
2	dom_rem/xkdoc	DOC half-saturation constant in limiting bacterial DOM degradation activity (Aumont et al, Eq 34)	99%	[12]
3	Optics/parlux	PAR : SWR ratio	93%	[6]
4	zoo/xprefn	Microzooplankton preference for nanophyto	91%	[17]
5	Dia/mumax0	Diatoms Max Growth	90%	[1]
6	Phy/logbp	Nanophyto temperature sensitivity for growth	88%	[5]
7	zoo/grazrat	MicroZoo maximum grazing rate	85%	[13]
8	phy/mumax0	Nanophyto Max Growth	83%	[1]
9	dia/logbp	Diatoms Temperature sensitivity for growth	81%	[5]
10	phy/padlopers	Nanophyto P-I slope	67%	[1]

ECOSMO II (ECOSystem Model; doi.org/10.1016/j.jmarsys.2013.03.008) 

ECOSMO
Rank	Notation	Description	Score	Group
1	g2	E-folding depth of visible fraction (m)	100%	[6]
2	muPs	Maximum growth rate of small Phytoplankton	83%	[1]
3	mPs	Small Phytoplankton mortality rate	69%	[3]
4	A	non-visible fraction of shortwave radiation	66%	[6]
5	GrZsP	Grazing rate of small Zooplankton on Phytoplankton	64%	[13]
6	gammaZsp	Small Zooplankton assimilation efficiency on Phytoplankton	52%	[13]
7	reminD	Detritus remineralization rate	42%	[10,18]
8	alfaPs	Initial slope of P-I curve for small Phytoplankton	41%	[1]
9	rPO4	PO4 half saturation	38%	[10]
10	Rg	Half saturation rate for Zooplankton	37%	[13,16]

ERGOM (Ecological Regional Ocean Model; doi.org/10.1016/S0924-7963(00)00030-0)

ERGOM
Rank	Notation			Description		Score		Group
1		rp0	Diatoms uptake rate		100%		[17]
2		q10_rec	sediment recycling q10 rule factor		98%		[5]
3		rfr	Redfield ratio P/N		79%		[1]
4		imin_di	minimal optimal light radiation, diatoms		66%		[21]
5		graz	Zooplankton grazing rate		55%		[13]
6		deltao	Phytoplankton mortality rate (pl -> dd)		49%		[10]
7		dn	Detritus mineralization rate (dd -> aa)		45%		[17 or 11]
8		rf0	Flagellates uptake rate		44%		[12]
9		iv	Ivlev constant, quadratic		42%		[11]
10		zcl1	Zooplankton closure parameter		40%		[14]

BFM (Biogeochemical Flux Model; doi.org/10.1016/j.jmarsys.2006.03.006)

BFM
Rank		Notation		Description	Score	Group
1	light/EPS0r		Background shortwave attenuation		100%	[21]
2	light/pEIR_eow		Photosynthetically active fraction of shortwave radiation		57%	[6]
3	B1/p_pu_ra		Activity respiration fraction, bacteria		55%	[9]
4	Z5/p_pu		Assimilation efficiency, microzooplankton		44%	[13]
5	P3/p_q10		Q10 coefficient, picophytoplankton		40%	[5]
6	Z4/p_sds		Exponent of density-dependent mortality, omnivorous mesozooplankton		38%	[14]
7	P3/p_qlcPPY		Reference Chla:C quotum, picophytoplankton		35%	[1]
8	P3/p_qplc		Minimum phosphorus to carbon ratio, picophytoplankton		33%	[4]
9	P3/p_temp		Cut-off threshold for temperature factor, picophytoplankton		32%	[5]
10	P2/p_qlcPPY		Reference Chla:C quotum, nanophytoplankton		32%	[1]

ERSEM (European Regional Seas Ecosystem Model; doi.org/10.5194/gmd-9-1293-2016, 2016)

ERSEM
Rank	Notation	Description	Score	Group
1	light/PEIR_eow,	photosynthetically active fraction of shortwave radiation,	100%	[6]
2	B1/pu	efficiency at high oxygen levels (bacteria)	92%	[9]
3	B1/sR1	maximum turn-over rate of DOM	83%	[12]
4	light/a0w	absorption coefficient of clear water	74%	[21]
5	B1/rR2	fraction of semi-labile DOC available to bacteria	69%	[12]
6	P2/xqcn	threshold for nitrogen limitation (relative to Redfield ratio) in nanophytoplankton	61%	[20]
7	P1/xqn	maximum nitrogen to carbon ratio (relative to Redfield ratio) for diatoms	58%	[20]
8	P1/xqcn	threshold for nitrogen limitation (relative to Redfield ratio) in diatoms	57%	[20]
9	P2/xqn	maximum nitrogen to carbon ratio (relative to Redfield ratio) for nanophytoplankton	57%	[20]
10	P1/sum,	maximum specific productivity at reference temperature for diatoms,	52%	[7]

Assessment of observability/controllability levels of ecosystem indicators in the 5 CMEMS regions.

Indicator	ARC	BAL	NWES	MED	GLO/IBI
Phenology	High	TBD	High	Medium	High
PP	n/a	TBD	High	Medium	Medium
POC flux	n/a	TBD	High	Low	Medium
PFT	Medium	TBD	Medium	Medium	n/a
Trophic efficiency	Medium	TBD	High	Low	Low