10020 Biological, physical, and chemical oceanographic research to enhance and support resilient and healthy marine ecosystems – Part 1
Convenors:
Heather Andres, Fisheries and Oceans Canada
Gwénaëlle Chaillou, Institut des Sciences de la MER (ISMER), Université du Québec à Rimouski
Martine Lizotte, Fisheries and Oceans Canada
Gary Maillet, Fisheries and Oceans Canada
Nancy Soontiens, Fisheries and Oceans Canada
The interplay of physical, chemical and biological oceanographic conditions forms the foundation of marine ecosystems and partly determines their productivity and health. The health of marine ecosystems can exhibit natural variations and can also be directly and indirectly affected by human activities, such as anthropogenic emissions of carbon dioxide, aquaculture, exploratory drilling, shipping and fishing activity, tourism, etc.. This session aims to explore the linkages between physical, chemical and biological oceanographic processes and their impacts on ecosystems, as well as how those linkages vary and change via natural and human causes. Further, studies aiming to characterize ecosystem status and health by incorporating oceanographic data are encouraged. Specific topics may include:
Marine carbon dioxide removal
Environmental controls on and changes to primary and secondary productivity, either through observational or modelling studies
Oceanographic impacts of changing sea ice and surface freshwater (from land, glacial melt, icebergs and hydrological cycle changes), as well as other climate changes
Ocean acidification, deoxygenation and other changes to habitat suitability
Seasonal predictability of oceanographic conditions
Extreme events, and
Environmental conditions and indices applied to the study of marine ecosystems.
10:30 AM
Timing and Controls of Spring Restratification on the Newfoundland and Labrador Shelf Based on GLORYS12v1 Data
Heather Andres
10:45 AM
Modelling Analysis of the Spring Phytoplankton Bloom on the Newfoundland Shelf Under Climatological Forcing Conditions
Jared Penney (Virtual)
11:00 AM
On Indexing Habitat Suitability in the Gulf of St. Lawrence
Rick Danielson (Virtual)
11:15 AM
Examining Hydrodynamic Connectivity Among Marine Protected Areas in Atlantic Canadian Waters Based on Trajectories of Bioparticles With Various Movement Capabilities
Kyoko Ohashi
11:30 AM
Applying Doppler Sonar Techniques in Northern Newfoundland for Fish Tracking Near Mussel Farms
Shane Anderson
noon um we have a bit of time so what we’ll do we’ll uh still strict to um stick to the 12m minute presentation with three three minute for um questions and then if we have time at the end um uh we’ll uh we can have a little bit of discussion um so actually the first talk is by my co-chair here Heather Andres and um she’ll be talking about the timing and controls of spring reratification on the Newfoundland and Labrador shelf based on Gloris 12 version one data and I also want to mention that we have three posters in this session uh so tomorrow morning don’t miss those and I’ll invite Heather to start her talk and oh yes I just wanted to say that I’ll I’ll kind of flag at 3 minutes and then at 1 minute and um at zero minutes and then we can start oh yes and for people online welcome as well and we have um Q&A and chat please um potentially um put your Q&A your questions in the Q&A so we can monitor those so Oh yes and I forgot to mention there is a 30 second lag between what’s happening here and what’s happening online so we just keep that in mind hello everyone uh it’s a pleasure to speak to you this morning oh apparently my first screen won’t come up all right i’ll be speaking to you today um colleagues of mine at DFO Newfoundland and Labrador as well as Dr frederick Hixier and Peter Galbra are pleased to talk to you about some work we’ve been doing about looking at stratification in the Newfoundland and Labrador region um so for background uh the spring phytolanton bloom forms the basis of the marine food web in the Newfoundland and Labrador region and some work has been done previously to try and assess how you can relate the timing of this phytolanton bloom to physical oceanographic conditions and some initial work by Wuadell in 2007 linked the spring phyto timing of the spring phytolankton bloom at station 27 which is let’s see if I’ve got I kind of have it which is the star red star near uh on the coast of Newfoundland uh to sea ice retreat and more recently Dr frederick Seir has linked it instead to timing of the stratification minimum or the initiation of restratification in the spring now observational data sources in the region are predominated by the Atlantic zone monitoring program which was designed to monitor interanual variations in oceanographic conditions uh it’s primarily includes three surveys a year at set transexs which are marked on the plot in the red dots and the red uh and the red labels but unfortunately this is insufficient temporal resolution to be able to capture changes in something like the timing of a spring reratification at any locations other than at station 27 so the goals of this work is to use model data to characterize springatification timing and controls on the Newfoundland and Labrador shelf and then that’s going to include some evaluation relative to the AENMP observations where we have them uh I’ll present the stratification seasonal cycle on at these transexs and analyze the contributions from surface freshening and heating via an along shore transct and finally do some initial exploration of the relationship between the timing of the restratification and sea ice retreat in the model so the model I’m using is the glories 12 global ocean reanalysis which spans years 1993 to 2020 i’m not going to give you a lot of the details here but I will I will point out that the runoff is based on a monthly climatology for based on years 1948 to 2004 there is some attempt to try and capture the impact of freshwater hydraological changes that are ongoing through the application of an open ocean freshwater flux has that’s u prescribed at a fixed rate that’s supposed to try and compensate for or capture some of the impact from icebergs ice sheet melt and other land water storage changes the data that’s assimilated is predominantly from satellite fields sea surface temperature sea level anomalies and sea ice concentration but it also includes the insitu temperature and salinity profiles which includes the A7MP data itself so I’ll show you some anal in initial evaluation that we performed so we calculated stratification in this case as the buoyancy frequency in the model and I’ve plotted here on the right the on the vertical axis the model buoyancy frequency versus the observed for three seasons spring summer and fall when we have uh ASMP data and the different colors for the dots represent the different transsects and this work was published just recently in a paper last year and what we showed for each transsect we create we create a single stratification value it’s based on the vertical maximum stratification at each station then averaged across the all stations in the transsect and what we found was that there were significant correlations in the interannual variations between AZ and P and glories 12 we’re getting values between 7 and 0.9 however as you can tell by looking at these plots the glories 12 significantly underestimates peak stratification by a factor of up to 3/4 now I’m showing here the climatological season pattern of for stratification again buoyancy frequency in the in the y- axis on the x it’s it’s the week of year and the diff the lines here are based on glories 12 and the different colors correspond to the station the transexs where we have the most northern transsect from the top moving down southward and the shading around those lines corresponds to the interannual variation the dots are colored and indicate the trans the p data value from a given year so you get a sense of both the observational data and the and the model data on this field and what you can see is that stratification really starts increasing first in these northern Newfoundland transexs but the those changes are very very weak for quite a while whereas what we see in the model is that the stratification peaks earliest at the most upstream of trans transexs on Labrador so this is beachy island in Makovic Bank if we look at the timing for the stratification decrease in the late summer then what we see is that proceeds initially at most northern transexs and then on down southward so to then analyze the contributions from temperature and freshwater to this stratification field I’ve created an alongshore transct and that’s marked here by these purple dots where this transct intersects the ANP transexs i’ve used the ASP station as my reference and I’m decomposed the contributions or estimated them from temperature and salinity via the equations here and those are plotted here so on the leftmost panel this is the maximum stratification at each station so vertical max in the middle is the contribution from salinity and in the right hand side is the contribution from temperature and the top at the top of the zero essentially for distance is northern coast the northern part of Labrador shelf and it proceeds down here and the different lines correspond to the transexs so these three are all from the Labrador shelf and the rest of this down here is all the Newfoundland shelf and what we see is particularly in northern Labrador that reratification is preceding any noticeable heating the instead what we see is there is a significant contribution from fresh freshening at that time period if we look at the contributions for temperature we see that they are occurring at pretty much exactly pretty much the same time of year for all of these transexs regardless of where they are north to south and you can see the signature of that in our total stratification however this is much later than it’s in the midsummer and it’s much later than the timing that typical timing we would consider for phytolanton uh bloom so here I’ve taken the in these yellow boxes the the estimated times or typical times for phytolankton bloom start and peak and this is published from sieral 2024 and what you can see is for except for this station here on the Labrador shelf the rest of the time the time period we’re actually interested in is precedes any really strong stratification so we’re looking at very a period of very weak signals and if we look at our contributions then really the only contribution we see is from freshing which raises the question of the role of sea ice during this stratification initiation so to examine this further I’ve want to define a time for stratification restratification and I want to define a time for sea ice retreat so I’ve defined string restratification as the first day after we reach the minimum and the vertical maximum buoyancy frequency where we exceed 10% of the range between the minimum and the maximum for the a given year so that that that threshold would be different each year and the minimum is marked then in these black lines and this the restratification date is then marked in the red lines and I’ve calculated this timing both from glory’s 12 and from the asmp data at station 27 which as I say is the only station we have observational data of sufficient temporal resolution to be able to compare and what we see is compa if we look at the the general trend of stratification change and growth growth and maximal value throughout the year what we’re looking at is something that’s actually quite earlier than most of that so like I said this is a period of very weak change and very short-lived and you can see in both the model and in the observations this is these are short-lived events that we’re talking about that are coinciding with our typical phytolanin bloom timing and if the there’s a challenge okay if we look at the oh sorry sometimes our observational data is sparse and it looks like that got cut off by the thing so I’ll just move on so what happens here is you can see that sometimes we measure this this point but then it turns out that actually there were no data points prior to that so we really can’t judge where in that period it would have happened so when I compare those and I plot the modeled stratification timing versus the observationally based ones we see we actually have fairly good correspondence for a subset of the data and a subset of the data shows glories underestimating the observed one the observed timings timing of sea ice retreat is given by when we reach 10% less than the maximum sea ice uh area or volume or thickness for that year and that is a very noisy signal if you look at smaller areas and it can sometimes be very difficult to say this is clearly se ice retreat so if I plot then my reratification date a assessed on a grid cell basis on the left here averaged over all years and I show the sea ice retreat dates averaged over all years on the right we see a few features one is the earliest dates are here at the outflow of Lake Melville and the timing there greatly precedes that from sea ice retreat that is a general trend for all of the Labrador shelf we’re seeing that the timing of reratification is much earlier than the timing of sea ice retreat that’s whereas over the new northern Newfoundland shelf the timings are pretty similar and based on the White Bay transct which is given by these dots we see a weak but significant correlation with sea ice retreat over just north of that upstream of that in the southern Labrador shelf so essentially what we’re seeing is that we have the in the model there are seasonal patterns of stratification that depend where you are the timing of the phytolanton bloom only surface freshening appears to be contributing to this weak stratification there’s based on the analyses I’ve performed there’s little evidence that sea ice is playing an important role there we see the the earliest stratification at the where we have freshwater export in the mouth of Lake Melville and also Hudson Strait we see the Labrador shelf the sea sea ice retreat is occurring much later than the restification which raises the question about potential for under ice blooms um for northern Newfoundland there is some there’s a potential based on timing that there could be a relationship but I want to note and this is for future work that sea ice melt can occur without a sea ice volume change due to the advective nature of the conditions in this area so a future work is going to look at actual sea ice try to quantify the sea ice melt that’s occurring in the model as opposed to just the volume changes and then see if we get a better relationship there so thank you [Applause] [Music] okay to ask for questions oh yeah okay so if you have any questions we have a microphone for those in the room if if you’re just too eager I can repeat the question as well i think it’s okay thank you very much for your talk and I just have a question because I didn’t understand um the stratification so you showed the stratification um as a histogram over time and you showed the salinity contribution and the temperature contribution but as far as I could tell the salinity plus the temperature didn’t equal the density so what am I missing i I I would I that is an open question right now we’re trying to work out so we were calculating them using two different methods and I think the difference there is we’re using like I said the equations I showed for the temperature and salinity components and for the total I was using using the GSW built-in N squared calculator whereas these other ones I was calculating separately so there is a question mark there as to why they don’t add up and it definitely doesn’t thanks very much uh I have a question but also I just going to come on the last one flores uses data assimilation maybe there’s an assimilation term taking place in those calculations uh but I was going to ask when you showed the restification timing compared to the melt timing i saw very nicely there were two pathways yeah red especially along the inshore then the shelf break Labrador current which to me is suggesting then that fresh water is actually a signal maybe of the previous year’s melt and runoff in Hudson Bay and or farther off that’s then being exported is that likely what you’re seeing yep and I think if you I mean this is without me having done any look when I look at this what I’m interpreting is exactly what you’re saying you see the Hudson straight outflow right here and you can see it kind of following along the coastline here there’s little connectivity originally here but from this point along you can see all the wa water that would be coming off the land right there but what’s interesting I think is what’s happening around Lake Melville where we have a source term essentially and there you can kind of see some pathways where this water is mixing out off to the outer shelf as well and uh but as well continuing along in through Straighter Bell Isle etc so there’s definitely uh yes I would say what we’re seeing is very much an advective regime happening great thank you okay thank you so thank you very much Heather um if you have questions you can u uh put them in the chat and maybe we can address them later on in the discussion part uh we’ll try now to move on to under climatological forcing conditions uh Jared is joining us um virtually online and we had we did have a few issues with the sound so we’ll we’ll try to see if that works now jared uh yeah can you hear me now oh yes that’s good all right i’ll I’ll explain later did you try the music i’ll explain later and I think you’re going to need I think you’re going to need to share your presentation from there Jared is my understanding oh I thought it was being shared through the other end if you want I’ll open it up i just Yeah if you can open it up from that end that’s Yeah okay I’ll do that okay so share where’s the share button on this one again oh here it is okay thank you okay go and Uh yeah I’m seeing presenter mode hello uh sorry can y’all hear me now okay uh just go ahead sorry I’m I’m not getting any confirmation i can’t hear right now if someone could just drop in the chat whether or not to start uh room i’m not sure if you heard me but I I can’t hear you right now so if you could just drop in chat when I’m good to start I’ll just go ahead oh okay wonderful um sorry is I can’t see the slides now it’s okay no I can’t see it yeah yeah you can see it already okay just maybe I’ll go back screen okay uh I see slides i think people can hear me are we good uh yeah uh I’m just going to go ahead uh okay hi everyone uh today I’m going to talk to you about some biohysical modeling work uh that we’ve been doing to simulate the spring phytolanton bloom near the coast of Newfoundland so this project was funded a couple years ago under DFO’s competitive science research fund uh the proposal was spearheaded by Nancy Sunines which allowed us to hire uh Chongqen for a couple of years as a physical scientist uh and it was his work on the development of the numerical model that I’m going to talk about that really drove a lot of the research that we’re going to look at today uh next slide please so in the northwest Atlantic uh the spring phytolanin bloom has a direct impact on the health of the marine food web uh the rapid increase in the amount of available phytolankton feeds the zup plankton which in turn feeds larger and larger marine animals uh off the coast of Newfoundland the bloom is observed to take place between February and April in the past and past work has shown a correlation between timing of the bloom and timing of sea ice retreat the more recent work has suggested a timing with minimum stratification and the common theme there is that the increase in fresh water uh leads to a change in stratification which is going to influence bloom development next slide please so we initially proposed this project uh two of the planned objectives were given as written here uh we wanted to develop a model relating physical and biogeochemical processes to simulate conditions in the water column leading up to and during spring bloom and considering that we are under a changing climate uh we want to look at how the bloom responds to a v variety of hypothetical forcing conditions um this latter one is important when traditional linkages between physical variables in the bloom weaken for example due to an ongoing loss of sea ice so the question I’m uh putting forward today is how does our model represent the spring bloom when forced with typical climatological conditions uh and what happens when we modify those conditions in some way for example making the air temperature warmer next slide uh to outline the rest of this talk I’m going to briefly describe the model and the data that we use to force it uh then I’m going to detail a reference simulation forced using climatological conditions from uh January through to July uh and then following that we’re going to modify our forcing in three ways to see how the bloom responds when we are using this model so we’re going to increase our winter winds use overall warmer air temperatures and shift the freshwater forcing conditions a little bit earlier next slide uh we call the model that we have been working with SOG NL which was branched from SOG which was a 1D biohysical model developed by Susan Allen’s group at UBC to simulate the spring phytolankton phytolankton bloom in the straight of Georgia uh we made some changes to simulate conditions at station 27 which is about 4 km offshore and 10 km from St john’s uh and it’s a long-term monitoring station uh going back decades that is regularly surveyed as part of DFO monitoring missions particularly the Atlantic zone monitoring program and the site is strongly affected uh affected by the invection of water from the Labrador current the governing equations for the physical model are given here uh where this cursive V indicates the mixing of the current velocities temperature and salinity by the standard KP formulation and surface forcing on the physical model is done by wind air temperature with some contributions from relative humidity and cloud fraction now the non-turbulent heat flux is scaled down in the column to prevent uh the water from overheating uh and additionally we nudge the salinity towards the profiles uh extracted from the glorious 12 reanalysis product over a fairly short relaxation time scale next slide please uh the BGC model uh is a nutrient phytolankton zup plankton detrous model uh we’re using a single class of phytolankton with biological parameters based on fellasiosyra northern skill d um nitrogen and silicon are the uh are the nutrients uh with the former being split between ammonium and nitrate uh ammonium being consumed first the equations governing the evolution of the phytolankton and nutrient profiles are given here uh these are mixed as passive tracers through KP formulation while phytolanin also has a sinking term but I would like to focus on the uh growth rate so next slide please which is divided into three parts uh so we have the cellular division of the phytolanin which is going to be either light limited or nutrient limited we have a natural mortality of the phytolankton uh which is going to be temperature dependent only and then we have grazing by zuoplankton which in simulations I’m going to show is negligible uh here uh so in each of the nutrient equations uh the cellular division term that shows up as consumption of either nutrient by the phytolankton then each of these terms also has a temperature dependence that basically says the reaction rate is going to go up exponentially as the temperature increases while phytolankton’s growth rate peaks at 10° before decreasing down to zero at 18° uh next slide please uh the forcing boundary and initial conditions are all listed here uh the weather conditions used to construct uh climatologies were taken from observational data from the St john’s airport and the Canadian weather energy and engineering data sets the temperature and salinity data used for initial and bottom boundary conditions come from the CAST data set the Canadian Atlantic shelf temperature salinity data set which assembles profiles from a variety of different sources then nitrate silicate and chlorophyll for the boundary conditions and the initial conditions uh come from bottle data that’s been collected as part of the ASMP and then finally the salinity profiles like I said before they’re nudged to climatology of the glory’s daily mean salinities uh and we know that glories does assimilate data from the AZMP and we seem that it does a pretty good job of representing uh conditions at station 27 so in a typical year as shown by these uh TNS profiles from Glories we uh start in the winter with near freezing salty waters uh which gradually warm at the surface and uh slowly freshen before rapidly freshening a little bit later and then as seen in the uh surface mean chlorophyll uh we get our highest chlorophyll concentrations in March April and fairly low for the rest of the season next slide please so uh on this slide we’ve got the uh reference simulation that I’m uh going to uh refer to going forward which is forced using the climatology I described uh previously from January to the end of July in the upper panel we have the wind and air temperature that force the model in blue and black and the surface temperature and salinity from the model are given by green and red the bottom panel is a little complicated but I’ll do my best to explain uh shows the time evolution of the phytolanton concentration profile is given by the color bar where cyan represents the surface concentration of chlorophyll to better uh depict when the bloom is actually happening the magenta lines indicate the depth of active mixing while the blue contours denote where the phytolanton growth rate is equal to zero so below that contour for most of the year uh down around minus30 there is phytolanin loss additionally to the top right uh above that contour there’s also phytolanin loss then finally the hatching indicates if there is nutrient limitation on the growth rate while no hatching means light limitation so this simulation we have relatively low phytolanton concentrations before the bloom initiates around midFebruary which gradually increases to a peak at the beginning of May after this peak we have a subsurface maximum for the rest of the simulation because the phytolanin has consumed all the silicon that’s been available in the system as depicted by these hashes there in general the phytolankton concentration increases as the mixing layer depth decreases and the phytolanin is primarily mixed at shallower depths for which growth conditions are more favorable i just want to note that the blue contours that I’m plotting here they’re only plotted at noon every day so while it looks like the phytolankton uh the the mixing layer depth is always within a region of positive growth there’s a dal cycle there that I’m going to talk about uh in the next couple of slides we can also see a relationship between the mixing layer depth and some of the physical variables depicted above uh such as the stronger winds leading to the deep deepening of the mixing layer depth in about midFebruary uh which is then followed by weaker winds and an increased stratification due to surface freshening in March and April that leads to an overall shallower mixing layer depth so this is the first modification that we’re going to make to the simulations uh we’re going to scale up the winter winds by uh wind speeds by 10 25 and 50% as shown in the upper panel here and then the mixing layer depths for each of these simulations respond as you would expect uh the stronger winds cause deeper mixing layer depths before they all converge back to the same mixing layer depth once the winds have been scaled back their climatological values the increase in the wind uh delays the bloom initiation and peak also leads to a higher peak concentration uh next slide please now the reason we’re seeing these lower concentrations when the winds are stronger uh which we can see from these plots of growth rate here on the right uh where red indicates growth and blue indicates loss is that the deeper mixing layer depth keeps mixing phytolanton into loss favorable depths which lasts longer during the short winter days um as the days become longer and the mixing layer depth schos the phytolankton is primarily mixed and resides in these growth favorable depths for more uh for a greater proportion of a given day next slide please the second set of forcing modifications we’re going to consider is warmer air temperatures uh which we implement by uh forcing with the 50th through 95th percentile air temperatures over 93 to 2022 uh for the reference case through to the 75th percentile uh the difference in surface temperatures for those simulations is at most three while in our warmest 95th percentile air temperature case that’s a few degrees warmer than even the previous most warm case uh these small temperature differences generally only correspond to minor mixing layer differences uh with which leads to warmer temperatures giving slightly earlier plume initiations and peaks and you can see this in these profiles of temperature salinity and density on the right uh we have these small uh you know couple degree changes in temperature and because the stratification is primarily dominated by salinity the overall change in stratification is not too big and it’s the same for you know each of the simul it’s similar for each of the simulations however uh in the warmest case uh the bloom initiates and peaks much earlier than the other cases as indicated by these blue lines and part of the reason for this appears to be an effect of the temperature dependence on the growth rate which we can see from the bottom panel which is much higher uh for this warmest case uh when compared to the other ones uh next slide please uh final set of modifications we’re going to shift the salinity profiles that we nudged to a bit earlier which generally means shifting fresher conditions earlier and earlier so for the smallest temporal shifts where the surface salinities are still relatively salty uh this leads to deeper mixing layer depths uh as the stratification is weaker uh which means lower phytolanton concentrations and later bloom initiations and peaks for the largest temporal shifts uh we see the freshest water earlier thus really early strong stratification and shallower mixing layer depths uh which mean higher winter concentrations and earlier peaks and blooms uh next slide and finally just to sum up uh so the bloom initiation uh as we’ve seen in these simulations happens when the phytolankton is mixed within depths favorable to growth for a sufficient period of time and uh this mixing layer depth is going to be influenced by wind strength and stratification with the latter being mostly controlled by salinity on the uh due to freshwater uh influx when we’re talking about the waters around Newfoundland uh here are some additional notes but I I’ve gone a bit long so I’ll gladly take any questions now do you want to come up to the microphone so Jar can hear you thank you hi it looks like a a great process study i’m a little bit worried about taking St john Met data directly to station 27 um we know that at least some of the characters cha characteristics of at least the wind change very quickly as you move offshore and I’m just wondering whether you’ve made any allowances for that or or have any way of checking that uh not that’s a really good question and it’s been a bit since we’ve looked at different forcing data so when we had initially started this we were also forcing with uh we had also tried forcing with the ER5 reanalysis product um and there you know there are obvious differences there like the change the you know the dial I guess temperature difference is going to be smaller and that there was a bit of difference in the wind but the patterns look the same at the time uh when we were trying to compare to observation the St john’s weather station was giving better results if we went back now with the modifications that we’ve made to the model since I would probably say that five year five might end up doing a bit better than it did before okay and then this was hourly data you were using yes yeah hourly okay thank you thank you so much Jared we’ll move on to our next uh presenter um also yes Rick to share your screen stage [Music] do you have your microphone on Rick can you hear us it’s like looking He’s Rick we’re having issues with the with the um sound so we’ll move on and we’ll come back to you at the end so um Kiyoko Uashi will be presenting next on examining hydrodnamic connectivity among marine protected areas in Atlantic Canadian waters based on trajectories of bioparticles with various movement capabilities thank you um my name is Kyoko i’m now at DFO but this is work that I did when I worked at Dalhauszy University with Gin Sheng as my supervisor bruce Hatcher at Cape Breton University was the other PI sarah Smith works for Bruce and they were in charge of coming up with the biological behaviors this work was funded by DFO the development of the circulation model was funded by the Ocean Frontier Institute and we would like to thank the digital research alliance of Canada on whose clusters we ran our experiments so as I think everyone here has heard of Canada has pledged to to protect 30% of its marine space by 2030 we are about halfway there now and as Canada as the government deliberates on which areas to protect next it would be nice to know the potential connectivities among existing and potential NPAs so that for example if the ecosystem in one NPA is disturbed then another MPA can for example act as a source of larve for that ecosystem whose um the ecosystem that was disturbed and the use of numerical models can provide some information on how MPAs can be connected so that’s what we are doing and we are look um our main area of interest is the schol and in this work work we release particles in experiments from three existing MPAs one is called bankis American it’s of the Gaspay Peninsula in Quebec so that’s there the second MPA is St ants Bank which is just off Cabba Street and the third one is Basin Head which is off the east coast of Prince Edward Island and most of this presentation is about passively drifting particles but in the end I will talk a little bit about work that we just started on adding swimming behavior to the particles so there are two parts to our methodology one is the circulation model and this is something that was developed for another project with two other research groups the circulation component is ROMs the CI component is sites the horizontal grid size in our area of interest is about six and a half kilometers on each side we have 40 terrain following layers in the vertical for lateral open boundary and surface inputs we are using reanalysis data gloris for the ocean and ice rafi for the atmosphere we are also specifying tides from the Oregon State University title model at the lateral boundaries and we also use freshwater input from rivers and also from melting glaciers and tundra that’s mainly off Greenland and Baffin Island and one thing that is different from our published paper is that um for the the simulation used here we are adding methods called the semi-prognostic method and spectral nudging to counteract model drift and bias and I won’t get into that here but there is a forthcoming manuscript on that work the model was initialized in September of 2013 and we are using model results from 2015 through 2018 so the other component of our methodology is the particle tracking model we are using something called ROMs path this is an open-source model that was developed for years with ROM’s output and we are using seasonal mean simulated currents from 2015 through 2018 so there are 16 experiments for each MPA and we add something called random walk so that’s small random movements just in the horizontal and this is to account for movement due to circulation features that the model is forward um that the model cannot simulate because the horizontal grid is too coarse and we set the horizontal diffusivity to 10 m square per second and if I have time I will get into the effect of this parameter and we do not add this small vertical small random moments in the vertical and particles are released near the surface from 3 MPs they are released in a regular grid so that means the the number of particles released depends on the area of each MPA and in the following slides I will be showing snapshots after 60 days or after 90 days that are composited over four years and they’ll the number of uh and out the distribution of particles will be normalized by the total number released over four years so starting with the St ansbank MPA of Cape Breton so this is after 60 days and as I mentioned for each season I am compositing the snapshots after 60 days for the four years over which the experiments were conducted and the shading is the percent of particles released in four years per square kilometer and we can see that there is a lot of seasonal variability and for example in the summer after um in the summer after 60 days there is a relatively high concentration of particles of Sable Island and which is quite intriguing for us because there is another MPA in this area and also we can see that particles tend to bypass the Atlantic coast of Nova Scotia which we think might be due to this little gap between the Cape Breton coast and the western boundary of the of the MPA so here as I said I am just summing everything over the four years but another way to plot these results is to add the outline of these patches um for each year so that’s what I’m doing here and the the outlines are colorcoded by the ears and we can see that there is a lot of inter angle variability too in the particles movements and one thing to note is that if we look at spring and if you look at the 2017 outline which is this olive green outline the that’s when the particles travel the furthest equator along the shelf break and that is going to be a recurring theme theme for the other MPA too so moving on the next MPA is um Bankis America which is off the Gaspay Peninsula in Quebec and so once again there is a lot of seasonal variability in how much the particles move and this time I’m showing the snapshots after 90 days because after 60 days they’re still in the Gulf and so and then we can also see that this bank is this American MPA can also act as a source area for the St ansbank MPA especially in winter so once again we can add the outlines for each year and then once again we can see that spring of 2017 that’s the olive green contour so 2017 is when the art the particles travel furthest equatorward along the shelf break and then the last MPA we’re going to look at is basin head which is of the east eastern end of PEI and once again there is a lot of um there is a lot of seasonal variability um once again this is um sorry this is uh 90 days after release and oops and sorry so um so after 90 days in the summer the um none of the particles have made it out of cab straight onto the Scotian shelf whereas in winter in in winter and fall the particles tend to travel along the coast of Nova Scotia and then in spring the particles tend to spread across the shelf and once again we tend we see that spring of 2017 is when we see the particles traveling the furthest equator along the shelf break so then we know that um the the outflow from the St louis River I forgot to mention the um the circulation pattern but the outflow from the St louis River is an important factor in the circulation in this region so I plotted the monthly mean discharge of the St lawrence River for the four years that we are looking at this is discharge estimated by a linear regression model from the water level at Quebec City and we can see that the discharge in 20 in 2017 is a lot higher than in the other years so that’s the year when the particles traveled the furthest equator along the shelf break so um this is outside the scope of this study but it does seem that the St louis River discharge does have a large effect on the bifurcation of the outflow from Cabba straight and hence on the movement of larve and juvenile animals in this area so this was all passive particles and we have just started working on adding swimming behavior and right now we’re working on dial vertical migration or DVM this is behavior that is thought to be undertaken as a predator avoidance strategy so in standard DVM um animals are near the surface at night to feed and then in deep waters at night to avoid being seen and being eaten and in for programming DVM in ROMsp we used code that was included in draft form in by the developers so we just cleaned it up and got it to work so that’s the schematic I am showing on the left and then for the parameters we try to use values that are relevant to the Scosian shelf and for surface radius uh we are using a temporal mean for our study period from era 5 so that means it is a temporal mean but it does take into account both the seasonal and intannual variation in the solar radiation so here I’m just showing results from the St an’s Bank uh 60 days after release and so once again there is a lot of seasonal variability there is a lot of retention within the within or near the released area in the summer and then to show the difference between the passive case and the DVM case I’m showing in this slide just the spring and summer cases so spring in the top row summer in the bottom row u with the DVM on the left um just passive particles on the right so we can see that although DVM is supposed to be a or thought to be a predator avoidance scheme um it does make uh it does affect the movement of particles so um to conclude the I in particle tracking experiments can be useful in um determining the locations of potential MPAs and how the existing MPAs and track and we plan to add more uh swimming behaviors release areas and we hope to repeat these experiments using project uh future projections thank you any questions in the audience dave just a quick thought here um uh you’re looking at 60 and 90 days and in future work you’re thinking of uh adding swimming behavior but um not being a biologist um I wonder um uh that seems like a long time for anything to go unmodified by natural chemical or biological processes and I’m thinking are are you thinking of putting any kind I’m not too sure you you made it very general as to what you’re uh what you’re tracking through there but I’m just wondering any thoughts on adding more to the model that’s going to say well if this is a plankton it’s going to die go to heaven or something in this time so um yes so I think once we add behaviors um the passive cases will be only a reference because I think in nature larvi are passive for only a few days that’s my understanding so um they’re they’re not going I don’t think larve of any species are actually passive for several month so the passive cases I think will be just for comparison once we add the biological behaviors and we would like to add settling after a month or so i think I included that in the future work so that is planned so did that answer your question yes in other words you are thinking of it good yes yes thank you any just any other questions online or yes it thanks Nancy for submitting a question uh thanks for the great talk Kiyoko I’m wondering besides the deal vert vertical migration are there other factors that influence the vertical motion of the particles Um well there is a lot of motion due to adection um if you could can Oh actually I [Music] um so just the passive particles do undergo a lot of movement so this is um I think this was yes so this was the case from St an’s Bank so the left hand side is passive particles and then the right hand side is DVM particles so it’s depth versus time so just the just vertical invection does uh does have a lot of effect effect on the vertical positions if that answers your question and thank you Nancy thank you perfect so hi everyone thanks uh I’m Shane i am currently a master student at uh Memorial And my project uh I’m actually working with the DFO so what we’ve actually done um is uh they had these Cooper do uh acoustic Doppler current profilers uh deployed off of the coast of northern Newfandland in Notre Dame Bay um and they’ve remotely or uh and continuously mount uh collected data for the deployment of a whole year between August 2021 and 2022 and uh the whole question is is well they deployed them near a muscle farm so they put one in by the muscle farm inside the bay the other one was outside of the bay more or less just so we can compare the two and the big picture question is is do these farms affect the wild population but uh we are no well I’m just a physicist i’m no biologist by any means but uh so I’m more less interested in the data that we’ve gotten here and then once I hand it back over to DFL they can try and come to some conclusions there and help help me with that so our deployment site that you see up at the the top figure here is all the way up in northern Newfandland it’s called Notre Dame Bay it’s a little bit west of Lewisport i don’t know if anyone’s familiar with that uh so both ADCPS were operating at a 300 kHz frequency and it placed 35 meters below the surface uh NDB05 was just a tag put on and NDB04 was the tag put on the two ADCPS for station IDs uh 05 is at the mouth of the bay so outside I’ll be referring to this one as the outside uh transducer and DB04 is the inside transducer so when I say out inside that’s what I mean by that um and then the three green stations you see down here so the bottom picture is just a zoomed in portion of the top one uh the muscle farm or the sorry the uh three green stations are just different stations that the DFO went out every month and they did uh monthly data sampling of dissolved oxygen and chlorophyll so with my acoustic uh data I did some uh just comparisons with how it compares with dissolved oxygen and chlorophyll so just a brief introduction here nothing too serious about Doppler sonar um so initially you have the sonar that’s in the water uh you can have them oriented up or down to place depending on what you want to do ours actually were placed 35 meters below so they were upward oriented so they emitted sound pulses upward but this one uh you just emit a sound pulse and it’s a pulse pair which is important so two pulses travel down through the water column they bounce off scatters in the ocean it could be anything and then that sound is reflected back and uh uh initially is used to measure water velocity measurements which is correct but we use it in a different way so uh and Doppler and sonar you might think obviously is frequency it measures the change in the frequency but we don’t do that either we measured we measure the shift in the phase between the pulse pair and then that shift phase is then compared to goes through a mathematical process called autocorrelation and then if these things are correlated then they get coined as say a target detection if there’s not no correlation then it’s just no target detection it’s just disregarded as volume back scatter so and one more thing about the instruments they need to be calibrated of course and down here at the bottom of your screen here um you see this big long equation that is the volume back scatter equation um and when we calibrate our instruments the highlighted in red down here in the uh in in the equation is KC factor and once we calibrate the instruments this is the constant we are looking for to plug into our equation because initially the instrument measures the data in counts but we want to convert from counts to dB so that’s what that uh constant here that we through the calibration process we calculate and it goes into our equation so how do we actually detect a target so this data these two plots here this is not mine this is what my supervisor graciously gave me permission to use just as an example so when I talk about the autocorrelation process um you see up at top this is the intensity reading and uh you see these bright spots in the thing in the in the plot sorry u these are high intensity points within the data set and down bottom is a corresponding correlation uh image and then you see that the portions where’s the laser thingy here perfect so you have like a high intensity reading here and then down here you have high corresponding correlation so these two pieces the correlation corresponds with the the intensity so these two are correlated so I guess regarded as a fish detection or some sort of target detection and that is in a nutshell of how we coin if what we actually uh determine target with so looking at our volume back scatter data so so the big spikes here correspond to um the volume back scatter and as we can see inside the bay here there’s like this little secondary peak here and but these big peaks mainly they’re just is most of it is just big volume back scatter or whatnot uh and the top figure here is just uh detected targets or detected uh yeah detected targets inside the bay and then the bottom piece here is just the calibrated intensity and I just put them side by side just for to show that uh there’s so many within the calibrated intensity targets here there’s just so much data that you notice this little tail here toward the end between I don’t know minus 65 to -40 dB this is what we’re actually interested in because when we look at the detected target data this stuff here this is where our uh predominantly more uh single fish detections or target detections are coming into play and whereas here is can be a little noisier so we’re actually more interested in this region here so I chose a threshold of about -65 dB so anything um greater than – 65 dB that gets coined as a uh target detection anything less than that it just gets thrown away as volume back scatter so uh so I took this little red area of both regions and I just just kind of plotted that for instance so uh inside uh sorry outside the bay you’re seeing more of a of a secondary peak here so it shows more of a distribution of more I guess uh target detections within the bay as compared to inside the bay here um and it it what it’s showing essentially is just the size distribution of our targets within the fish column and here uh this is just uh some sonogram or echoggram sorry of our uh target detections so left is inside the bay right is outside the bay the the more uh brighter colors you’re seeing that’s more or less what we’re getting as fist detections and we’re seeing not a whole lot they look pretty similar which is good I guess um we’re seeing sort of I guess this is sort of like a plankton bloom of sort I guess you can see um but the most interesting thing we’re seeing here is that it seems like there are bigger targets inside the bay as opposed to outside the bay so we’re kind of in the process of investigating that now and just to get some more information on this I took that circled region i filtered our data and I just plotted histograms of this so top image is inside bottom image is outside uh we can see our main peak here at around minus 80ish it is shifted a little bit as you can see as compared from inside to outside don’t really know what that is quite yet uh we do have some suspicions um however when we do look at this thank you you can see these little secondary humps here which indicate again larger targets and that’s just the distribution of the size of our targets here and uh it does as it does seem like there are some bigger ones in inside the bay but mainly there is an overlap between the two regions and uh just some more visualization stuff here uh we actually took the number of we call them number of fish but it’s just number of detections um so this is just an example for inside the bay and this is all I’m here the only reason I really put this here is just to show that we do have similarities here of the top and bottom so we have high intensity here and we have high number of fish camps which we would expect but we’re also seeing points of not much intensity as opposed to highly have a lots of fish detection so this is just to show that just because we have points of lower intensity it doesn’t mean we’re not detecting fish and uh just some I’m running low on time so I’m just going to have to zoom here pretty quick uh this is just the the comparisons with the chlorophyll we can see some uh correlations here but we’re having some not so correlated stuff here so it is correlated in a sense but is not entirely correlated because we’re getting high chlorophyll counts here but we’re not getting much back scatter here and then another comparison with dissolved oxygen we’re seeing a correlation here but it’s a negative correlation so here where we’re getting high intensity readings up around here we’re not getting so much dissolved oxygen and here we’re getting lots of dissolved oxygen but not so much intensity so the next steps of course would be just to investigate these questions okay and some conclusions so overall seems like there’s larger targets inside our bay in around the muscle farms than there are outside the bay uh comparison of the chlorophyll data shows a little bit of correlation but not a whole lot uh and so like I said I’m not a biologist so I’ve just done a little bit of reading but uh high chlorophyll usually suggests high phytolankton numbers which should be an indication of high zop plankton so you think high chlorophyll would mean high like fish detections but we’re not always seeing that it’s not always the case and uh where’s my dissolved oxygen too and oh yes and uh our second conclusion of course we’re seeing a negative correlation with our dissolved oxygen and uh lastly but not least our crucical backer data is oh that’s right there uh but what I meant to say was that the two data sets the intensity they do seem to be fairly similar fairly the same which is always good it’s just we see apparent bigger fish or detections in um inside the bay rather than outside so the next steps of course the biggest pressing issue would be why are we getting larger targets inside the bay rather than outside one possible uh thing we’ve talked about it could be a calibration issue within the instrument and if it is we have to fix that um and then also compare with plankton survey so uh we’ll have to connect with the DFO again and get some more data on plankton surveys within the area and uh yeah so thanks everyone [Music] [Applause] thank you Shane we have time for one question from the audience or online i was wondering about the size discrimination sure so I was kind of I was kind of curious about this the questions you were talking about it seemed like something you were making a connection between the size of an object and the frequency and yet you also were you were rejecting a certain frequency band as being just back okay volume back scatter so I wonder if you can dig into that a bit and kind of talk about size please okay so size we’re talking about intensity is the intensity reading right as opposed to the frequency so um I’m not sure if it’s here anywhere so okay okay just for an example here so for our intensity so I chose a threshold of – 65 so typically anything around minus40ish minus 30ish you’re getting into let’s say more discrete single targets or say bigger targets such as something maybe like a codfish or something like that we don’t exactly know anything beyond anything less than say minus 40 or minus 50 you’re getting into zop plankton territory for example right so smaller targets but it’s just you know and you can get clusters and you know you can get clusters of zop plankton so so essentially in a nutshell you’re getting with a lower DB is generally smaller targets higher DB is generally bigger targets more like discrete single detections okay and what about movement if you had like say a COD for example and it was moving would that affect this the frequency you would get uh typically uh no but because they generally move cod for example generally move with the current and current is not really that fast especially inside here it’s millimeters per second right or centimeters per second i can’t remember the magnitude but uh any for more discrete targets if you’re getting movement like that it’s not going to affect it a whole lot but if you had say like a cluster of fish maybe and because one detection one pulse comes down if it moves certain distance or orients around like that and by time the other pulse gets and it goes back to the receiver then it’s not going to show much correlation so then it will throw away so that’s an issue with the instrument as well right okay we need to move up all right thank you thanks [Music] okay we’re going to try again