Image clipped from Gledhill et al 2015 A t the boundaries Wind Sea surface temperature Precipitation and salinity Carbonate chemistry Dissolved oxygen Nutrients Should we treat ID: 310849
Download Presentation The PPT/PDF document "Does consideration of global or regional..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Does consideration of global or regional change alter our prediction of ecosystem response to nutrient abatement?
Image clipped from Gledhill et al 2015
A
t the boundaries
Wind
Sea surface temperature
Precipitation and salinity
Carbonate chemistry
Dissolved oxygen?
Nutrients?Slide2
Should we treat regional trends simply as boundary changes?
Working backwards from field data, can formal inverse modeling sort out changes in the biology of the system
How do we address the inevitable need to compare models or evaluate the benefit of a new state variable (e.g., grazers)?Slide3
Gledhill et al 2015
ARAGONITE SATURATION STATESlide4
Gledhill et al 2015 and NOAA NFSC
SALINITYSlide5
Wallace et al. 2014
Narragansett Bay, 2013
Metabolic CO
2
(and Reduced Buffering ?)Slide6
Salinity (
psu
)
---- Lower
---- Upper
NuShuttleSlide7
Does any of this matter for predicting ecosystem response?
Consider carbon to chlorophyll ratios Key parameter in water quality models Current best estimate is based on algal monocultures
And is therefore sensitive to…
Things that alter community structure
e.g., salinity, temperature, grazing, nutrients, light, carbonate chemistry
C:Chl is determined by….
phytoplankton community structure (not predictable from monocultures)
e.g., cell size-abundance, species abundance
Slide8
Grear, Rynearson, Montalbano,
Govenar
and Menden-Deuer,
submitted
S
ize-abundance spectra of incubated whole plankton communities from Narragansett Bay
224
uatm
p
CO
2
361
uatm
p
CO
2
724
uatm
p
CO
2
Shift toward smaller cells
Shift toward smaller cells
No
change
RI STACSlide9
Grear, Rynearson, Montalbano,
Govenar
and Menden-Deuer,
submitted
Pre
224
361
724
Pre
224
361
724
Pre
224
361
724
Small Cells
Medium Cells
Large Cells
Each size class affected differently by
p
CO
2Slide10Slide11
Fixed water quality model parameters that may change annually or
decadally: CtoChl Carbon to chlorophyll ratio K0 light attenuation due to phytoplankton
BR Benthic remineralization Rphyt River phytoplankton load
Others?
WHAT ABOUT THE MODELS?Slide12
What value of
θ maximizes the likelihood, given the 2006 nushuttle data,
the 2006 physics, and the rest of the WQ model ?
Fixed water quality model parameters that may change annually or
decadally
:
θ
= {
CtoChl
, K0, BR,
Rphyt
}
Does this model:
θ
= {
CtoChl
A
,CtoChl
B
, K0, BR,
Rphyt}, improve fit (likelihood) without adding variance?Slide13
Corrected Chlorophyll (ug
L
-1)
---- Lower
---- UpperSlide14
Dissolved Oxygen (mg L
-1
)
---- Lower
---- UpperSlide15
θ
= {CtoChl, K0, BR, Rphyt}Use of likelihood methods requires joint probability model.Bayes methods provide a workaround, by sampling the joint posterior probability.
Markov Chain Monte Carlo methods “randomly walk” through the joint space and “sniff out” the peak in the likelihood.
y
x
z
Parameter hyperspace
i.e., state variable
i
(DO,
Chl
a) in segment
j
on day
t
Observations:Slide16Slide17
Each stop (> 10
4) in the random walk requires a full annual simulationCoded ODE in C8 days on a Linux with parallel processors (very tricky with MCMC)
FUTURE DIRECTIONS:I would like to use correct initial conditions, boundary fluxes, and both GEM and OBM-based exchangesNeed a geospatial model of observations for better linkage to WQ model (parse out observation and process error).Slide18
EPA ATLANTIC ECOLOGY DIVISION
Bay Ecosystem Time Series (BETS)
MonthlyTemperatureSalinity
Dissolved oxygen
Nutrients
Chlorophyll
Total suspended solids
Carbonate chemistry
Stable
isotopes
Contact: Autumn OczkowskiSlide19
NuShuttle
Salinity
1998
2004
2010
J F M A M J
J
A S O N DSlide20Slide21
Ullman, D. S., and D. L.
Codiga
http://www.crmc.ri.gov/samp