Author Archives: Admin

Earlier this week many of the researchers and scientists involved in OSNAP presented their work, based on the first two years of continuous monitoring in the North Atlantic, at the Ocean Sciences Meeting in Portland, OR. While there are still a lot of implications and details forthcoming, Nature – News reported on these findings in a short summary linked below.

 Read full Nature article

by Feili Li

It seems to be a quiet year for OSNAP – just one OSNAP cruise took place last summer and so most of the moorings have stayed in the water for a year and a half.  But OSNAP scientists have been working intensively on analyzing the first two years of data and we have started to obtain some very interesting results based on measurements from individual arrays.  Some of those results were presented and discussed at a workshop held in Southampton, UK in early November (blog post).  Much more is coming and will be presented at the Ocean Sciences Meeting 2018 in Portland, OR.  Please stay tuned as we will soon publish a dedicated blog post with a list of all OSNAP-related presentations at OSM 2018.

In addition to analyses by individual groups, all OSNAP scientists have been working closely on the first data products from the full array.  We are now in the process analyzing preliminary results and finalizing the flux estimates.  Final products are expected to be delivered in spring of 2018.  These final products include the overturning volume and associated heat and freshwater transport time series along with the cross-sections of velocity, temperature, and salinity (Figure 1 below shows the mean velocity and property fields at OSNAP).  It is always worth mentioning that OSNAP is not an isolated program as our results are based on many existing observational efforts in the region (e.g., Argo, AVISO) and the results will be analyzed in coordination with ongoing programs (e.g., OVIDE, RAPID).

During the first two years of the OSNAP deployment, the subpolar North Atlantic experienced a widespread cooling with two successive intense winters (2014/15 and 2015/16).  Strong air-sea heat fluxes during those two winters led to intensified deep convection with an enormously large production of Labrador Sea Water.  All the changes make us wonder about any concurrent changes in circulations (overturning and gyre) at the subpolar latitudes.  Once we have firmed up the flux estimates at OSNAP, we will soon begin the task of investigating those observed changes in the region and linking them to local and/or remote forcing mechanisms. 

Next year sure will be a productive year for OSNAP – be prepared for exciting news!

Happy Holidays!

by Laura Castro de la Guardia

When I am ask by friends: What is it that I study? I generally give them the quick answer: I study biological-oceanography at the University of Alberta. But when they look-up “University of Alberta” on Google map for example (Figure 1), they  always point out: there are no coastlines near the University of Alberta! In fact, the province of Alberta in Canada, has NO coastlines at all. So, how is it then, that I can study the oceans?  Although one way will be to spend a lot of time travelling to either western, eastern or northern Canada to do my field work, I can also study the oceans from my own desktop at university!

I use a mathematical model on the computer to create a virtual ocean with some biology and chemicals; it is sort of like a video game, but the model attempts to be as realistic as possible. The core of the model is based on the most current understanding of physical and mathematical relationships that exist between the ocean, the atmosphere, the sea ice and the biology.

There are many models available. The ocean model I used is called NEMO (http://www.nemo-ocean.eu/) that comes together with a sea ice model LIM. The biological and chemical model I used is embedded within NEMO and it is call BLING (https://sites.google.com/site/blingmodel/). Cool names acronyms, right?! Both models are free to use by any user, but it requires some understanding of computing science, programing, and a very powerful computer. We have to run our model on super-computers that are shared across Canada  (Compute Canada/Calcul Canada).

Unlike what you may have imagined from my video game analogy, the output of the model is not a movie, but lots of numbers (a.k.a simulated data). The “simulated data” is what I use to do statistical analysis of many different things, for example, I can see the current state of the ocean, or the sea ice, or the marine algae (phytoplankton). We can also make movies with the simulated data  (e.g. http://knossos.eas.ualberta.ca/vitals/outcomes.html)

Although models are still not able to reproduce an identical ocean to our real ocean, one of many advantages of an ocean model is that I can study how one single event/phenomena/or property in the atmosphere affects my simulated ocean  or biology. This type of studies are called sensitivity studies, and they are like experiments in a lab. This is important because in our current climate, many things are changing at once (for example in the Arctic Ocean, sea ice is decreasing, temperature is increasing, the rate of river flow into the ocean is larger, there is more rain, there are more storms during the autumn), but we only observe the response of the oceans to all changes. While with the model I can have the response of the model to all changes, but also the response of the phytoplankton to only one change (e.g. more storms during fall (Figure 2)). Depending on what I am studying, I can then answer which of all these changes is the most important, which one is the one I should be most concern with? These are the kind of questions I would like to focus on for my sensitivity experiments, because these questions can help us prepare for the changing future: e.g. they could help shape or guide the adaptive tactics and conservation programs.

Figure1. Google map showing the locations of the University of Alberta, Canada.

Figure 2. A simulation with storms (a) compared to a simulations without storms (b). The differences between each panel shows the regions where the storms have a greater impact on phytoplankton.

We want to bring everyone’s attention to the launch of the OSNAP Challenge located on this site under News and Events. Anyone is welcome to submit a prediction (or technically a hindcast because the data have already been collected) for the first two years of AMOC data from the OSNAP line. This contest is similar to the one organized by the RAPID program last year but in this contest there are no past observations of the AMOC from the OSNAP line so the level of difficulty is higher!

We are going to open up this blog to write-ups of the methods from each prediction and will be announcing the winners here in the Spring. The deadline for the submissions is April 1^st. For more information including instructions on how to submit a prediction and how the submissions will be judged, see the site here.

We wish everyone luck and may the best model win!

-Nick

By Elizabeth Comer

As part of the OSNAP array, the Extended Ellett Line (EEL) is a repeat hydrographic section that crosses between Iceland and Scotland. This line measures part of the Atlantic Meridional Overturning Circulation in particular capturing the majority of warm water flowing northwards from the Atlantic into the Nordic Seas and around half of the returning cold deep water (Figure 1). The heat that is transported northwards is released along its journey transferring heat and moisture to the atmosphere. The amount of heat being carried determines how much is released, therefore making it an important factor in climate predictions. By making measurements along the EEL we can investigate the currents’ structure and long-term changes in heat and freshwater transport. The EEL provides the perfect platform for investigating the heat and freshwater changes over time through its yearly measurements over 40 years.

Figure 1. A schematic of the Atlantic Meridional Overturning Circulation (Curry and Mauritzen, 2005).

The EEL has measured velocity from the ocean’s full depth using an instrument called the Lowered-Acoustic Doppler Current Profiler (LADCP). This instrument is lowered through the water column and relies on the changes in return frequency of acoustic pulses to determine the water’s speed (Figure 2). The LADCP data is an exciting chance for us to see the in-situ velocity of the entire water buy valtrex australia column. Combining this velocity and hydrographic salinity and temperature measurements from each survey will provide the heat and freshwater transports across the EEL.

This is the research that I am currently carrying out and alongside this I will be taking part in the 2016 EEL research cruise, which requires being at sea for a month. So far, I have only been on weekly length research cruises so this will be a first and exciting experience for me. I am not only looking forward to collecting and processing my own data, but joining in with other scientists and learning new methods of data collection. Another first for me will be attending the Ocean Sciences Conference in New Orleans this February. This will be a great opportunity for me to meet researchers in my field, share experiences with other early career scientists and gain feedback on my research. These experiences will both not only enhance my learning, but build my confidence when explaining my research to different audiences.

Figure 2. This diagram shows what happens to the acoustic pulses when they reflect off of moving particles of water (https://www.whoi.edu/instruments/viewInstrument.do?id=819, Credit: Sontek)

by Nick Foukal, graduate student at Duke University

As the RAPID team prepares to release the next 18 months of AMOC measurements from the mooring array at 26°N, I have been busy building a statistical model to predict those observations. Statistical models extrapolate into the future using data on past states of the system and differ from physical models in that there is no dynamical constraint placed on the predictions. Whereas physical models might demonstrate how the AMOC responds to wind and air/sea buoyancy fluxes and build predictions based on that information, statistical models only need to know what the system has done in the past to predict the future. So in many ways, statistical models are not as useful as physical models; they cannot tell you why a system behaves the way it does, or how future changes to the environment may affect the system, but oftentimes statistical models can tell you the minimum amount of information you need to make accurate predictions.

Another useful trait of statistical models is that they provide a baseline metric from which to judge the performance of physical models. Weather forecasting is an example of this: until advances in computational capability and the advent of continuous satellite measurements improved the numerical weather forecasting models, the best-performing weather forecast models were statistical models. My goal in this project is to evaluate where oceanography is on the journey toward predictive skill: can physical models outperform a relatively simple statistical model in predicting the next 18 months of the AMOC?

State-space analysis is one of many ways to build a statistical model. The basic tenet of the state-space model that I use here is that the future state is a function of the current state. This type of state-space analysis also requires stationarity in the system, thus trends or oscillations with periods longer than the period of measurement must be removed. In addition, autocorrelation and known oscillations at periods shorter than the period of measurement should also be removed (if the oscillations are assumed to be stationary into the future) so that the state-space model can focus on the ‘unpredicted’ aspect of the data.

Given these requirements, I downloaded ten years of RAPID data (April 2004 – March 2014) at 12-hourly resolution, averaged the data to 10-day resolution due to the 10-day time scales of flow compensation between the upper and lower limbs of the AMOC as reported in Kanzow et al. [2007], calculated the integral auto-decorrelation time scale (36 days) and then averaged the data at 40-day resolution to produce a time series of independent observations. To remove the seasonal cycle, I calculated a continuous seasonal climatology (Fig. 1) by taking a 30-day running mean of the data padded with the December data at the beginning and the January data at the end. This padding ensured that the climatology was not biased by when the year began and ended and the running-mean ensured that the climatology was a continuous function rather than based on monthly means.

Figure 1. The climatological seasonal cycle of the RAPID AMOC data (2004-2014). The seasonal cycle has an amplitude of 4.68 Sv., RMSE of 2.98 Sv. and explains 24% of the variance in the data. The minimum occurs in March and there is a broad maximum from July through November.

To analyze trends or oscillations beyond the study period, I fit the data with five models: a linear trend line, a step-function with the mean from April 2004 to April 2008 and the mean from May 2008 to March 2014 (based on results from Smeed et al. [2013]), two linear trend lines for the same time periods as the step function, a quadratic fit, and a sine curve. The fit with the lowest RMSE is the sine curve (Fig. 2).

Figure 2. The sine curve fit to the AMOC observations without the seasonal climatology. The sine curve has an amplitude of 2 Sv., period of 10.41 years and phase shift of 6.16 years. This sine function has the lowest RMSE (2.6 Sv.) when compared to a linear fit, a step function fit (2004-2008 and 2008-2014), a quadratic, and two linear fits (2004-2008 and 2008-2014). The maximum of the sine curve occurs at the end of October 2005 and the minimum occurs in early January 2011. The next maximum predicted by just this component is in the Spring of 2016 while the most recent inflection point occurred in mid-2013.

To predict the AMOC signal that remained after the seasonal and long-term oscillations were removed, I fit the parameters of a state-space model to the ten years of anomalies (Fig. 3). The two parameters that require optimization are the number of dimensions and the number of nearest neighbors. Dimensions refers to the number of previous observations in time to use in the prediction, and the number of nearest neighbors refers to the number of time periods with similar AMOC variability (each consisting of the number of dimensions) to use. I tested models with zero to 25 dimensions and zero to 25 nearest neighbors by calculating each of the models’ RMSE when compared to the observations for the MOC observations from 2004-2014. The model with the lowest RMSE (2.46 Sv) has 10 dimensions (each prediction uses information from the past 400 days), and 14 nearest neighbors. The fact that the model needs just over one year of previous data implies that there may be residual seasonality that the seasonal climatology did not remove.

Figure 3. The state-space model fit to RAPID AMOC observations without the climatological seasonal and sinusoid cycles. The model uses 10 dimensions (400 days) and 14 nearest neighbors. State-space models with many nearest neighbors typically under-predict the amount of variance in the original data because the number of values that are averaged to create a prediction is too large.

When the three components (seasonal cycle, long-term oscillation and state-space model) are combined (Fig. 4), they recreate 48.5% of the variability in the observations from 2004-2014 and have a cumulative RMSE of 2.46 Sv. In comparison, models with just the mean MOC (RMSE = 3.42 Sv. and 0% of variance), the climatological seasonal cycle (RMSE = 2.98 Sv. and 23% of variance) and the climatological seasonal cycle plus the long-term sinusoid (RMSE = 2.60 Sv. and 42.1% of variance), do not fit the data as well. The combined model also produces a prediction for the next 18 months of the AMOC (Fig. 4, blue). Of the 6.11 Sv. amplitude in the predicted values, over 75% is due to the seasonal cycle, with the increasing sine component (Fig. 2, blue) slightly compensated by the negative state-space component (Fig. 3, blue). The two peaks in the combined model’s prediction (Fig. 4, blue) of 20.28 Sv. and 20.14 Sv. occur in October 2014 and August 2015, respectively, and the trough of 16.06 Sv. occurs in February 2015.

Figure 4. A comparison of statistical models with predictions for the next two years of RAPID AMOC based on the model that combines the seasonal, long-term and state-space models. The average standard deviation for the next two years (blue shading) is +/- 2.4 Sv. The error does not diverge because it depicts the amount of spread in each individual prediction of the next time step provided that the previous prediction was accurate.

References

Kanzow, T. et al. (2007) Observed Flow Compensation Associated with the MOC at 26.5°N in the Atlantic. Science, vol. 307, pp. 938-941.

Smeed, D. et al. (2013) Observed decline of the Atlantic Meridional Overturning Circulation 2004 to 2012. Ocean Science Discussions, vol. 10, pp. 1619-1645.