Archive for the ‘Espresso & Biospace’ Category

Plans going forward into tomorrow and the weekend

Thursday, November 5th, 2009

Today was an exciting day during the OSSE.  We had friendly debate and lots of great ideas.  This is what the future of oceanography can be, and I feel very lucky to be part of it. We also had give the strategy to the JPL path planner to provide the gliders the next set of waypoints.

The crux of the debate was where do we send the gliders? The AUV? Where should we send the U. Deleware glider? What should be the glider configuration for the saturday Hyperion satellite pass? What science do we want to address?

Several in the group wanted to get gliders to the north to resolve some interesting transport issues.  See Yi's thoughts in the earlier blog.  Others wanted to head the gliders in nearshore to confirm the presence of the winter bloom?  Others wanted to fan out in the mid-shelf in parallel lines to validate models?  Discussion of the way forward continued in discussion in many of the groups late into the afternoon.  Given uncertainities that need to be resolved about shipping lanes, to figure out where are the safe northern zones, the inital decision to give the JPL scientists time to run the path planner, a decision was to fan out the gliders as they head in back to the mid shelf.  Tomorrow we will figure out the distribution to cover uncertainity in the north and anchor the shelf wide modeling efforts. The gliders are making good progress. They are now heading onshore, this path planning is cool!


In the afternoon, discussion continued for over an hour with Julia Leven, Scott Glenn, Wendell Brown, Oscar Schofield, and Javier Zavaka-Garay. Currently the path planner uses the ensemble model  to guide the path planner.  One issue is that we know all the models have different uncertainities and errors.  We focused our discussion on the variability in the currents as this is what influences the ability of the gliders to fly. Below is ensemble currents, below that are variance in the u and v components between the models.  Our discussion focused on using the gliders to minimize the variance (decrease the red areas),  and how that might be a path forward for how we prioritize missions in the next two weeks.  The cross shore lines tonight should help minimize the uncertianity in the mid shelf where the winter bloom is occuring, and then we plan how to resolve the northern area that is of interest to a bunch of the OSSE science team.  What a great day to be an oceanographer!




Thought from JPL after today’s excellent telecon!

Wednesday, November 4th, 2009

Thoughts from Yi:

As we discussed at the telecon, there is a really interesting outflow feature from the NY harbor,  see attached SST images for yesterday from GOES (11+12 at 6-km), TMI+AMSRE (25-km),  and MODIS (1-km).  All these data are blended using 2DVAR into a 1-km image, also attached.

We just post the SST comparisons with the four models on the CI portal, see attached difference maps (model-observation) for HOPS, NYHOPS, COAWST, and Rutgers ROMS.  As you can see, all four models are warmer in that outflow region, suggesting a weak or no outflow from the NY harbor.  Presumably  there will be a big difference for the salinity signature, that we don't have any observations now.
It would be great if we can send one glider to this area to first map this feature, and second assimilate it into the models to improve their forecasts.








JPL Executive summary for 11/4/2009

Wednesday, November 4th, 2009

JPL Executive Summary of 11/04/2009

Winds have switched to Northwesterly. NAM forecasts indicate weak winds Thursday in from the southeast direction. Northwesterly winds will resume Friday but increasing amplitude (up to 30 knots by the end of Friday). We continue to receive four ocean model forecasts daily. These model forecasts are compared with the daily mean observed surface current. It should be pointed out that the server that has HF Radar failed overnight. At least the Rutgers’ group is rerunning their model that will be available later today. We continue to generate the multi-model ensemble forecast based on the equal weighting method. There is a good agreement on the current flowing in the southwest direction, although the ensemble model forecasted current is significantly weaker than observed. There are discrepancies for the in/out flow at those major fresh-water input locations. All three gliders are moving along the predicted path, but at a much slower speed

Avijit’s late night thoughts

Wednesday, November 4th, 2009

Late night thoughts from Avijit Gangopadhyay:

Hi Oscar:

I was reading through your blog of this evening and was intrigued by the Glider current map...

Then I was looking at the model out at our site
and was trying to see what the model is showing at that location for today and next couple of days. You can go frame by frame or run
a quick animation...

What struck me is that the model is showing a variable pattern of shelf eddies interacting with each other making the currents go in different directions depending on which side of an eddy (of which kind) and the front you are on... It is quite variable near the Glider location around
73.5W, 39.2W. To begin with there were two shelf eddies opposing each other to the east of the Glider location, and then new eddy-like
formations are happening..

I have not checked the other models yet (its late now), maybe they show similar things even better resolved. Ours is pretty coarse resolution.

I am copying Wendell who can also take a look and let me know what you think... Its interesting nonetheless...

You can go to our site and then choose Mid-Atlantic shelf zoom and browse through the recent forecast...

Good night!


Evening OSSE update 11/3/2009

Tuesday, November 3rd, 2009

Evening sets on the second day of the OOI CI  OSSE. There was a productive 2PM call, and we are filling the write board in the cool room with suggestions provided by the IOOS and OOI CI team.

For the ocean conditions:
The  changes in the Mid-Atlantic Shelf tonight are a big current reversal in 24 hours.  Last night I was bloggin' about the northward flowing current, and tonight I lament the south flowing surface currents.  These strengthening southern flowing current has been building over the day according to the CODAR data.  You would think the shelf could make up its mind by now.  The forecasts are provided in the earlier JPL summary blog.


So do is the reversal supported by model outputs?  Sure enought the ocean models show the south flowing currents nicely.  We show below the output from the JPL models.  Clearly in the CODAR and models the Mid-Atlantic Bight the currents are flowing to the south.  Given this how are the gliders doing?


The reversals, in 6 depth averaged currents show a muddled and nasty looking glider fleet.  In general the gliders showed southward flowing currents in agreement with the CODAR and numerical models.  A few random vectors showed northward flowing currents, and we will need to dig through the data to see if it is local variability ?  So given that we are testing the CI planning and prosecution software, how is the ASPER/CASPEN software faring tonight?  The gliders over the last data transfer show progress to the north.  For the experienced glider watchers, watch the next update to see how we feel the southern flowing current.  Will we keep the the steady nothern progress? The way points for the gliders for two southern gliders are two the south, which is surprising given the southern current.  This to me looks like it shows the earlier influence of the prior period of northern flowing current.  We will see, next call in tonight should be interesting. At the 2PM call we decided as a group to keep the glider on the path planning waypoints from yesterday (blue line in the 2nd figure below).  Lets see how we progress.




What are the gliders seeing?  A quick recap of the vertical structure seen by the gliders. Salinity shows the same cross-shore variability with near-shore waters showing slightly lower saline waters.  Temperature shows an interesting slug of warm water in the mid shelf, looks like two water masses based on the salinity structure, as there is vertical structure in the salinity and the temperature profiles are near isothermal. The outer half warmer waters and cooler waters in the mid-shelf show enhanced chlorophyll.  Values are close to ~3 micro-g/L, close to the mean winter bloom values.  I have a graduate student who has been analyzing a decade of satellite data, her analysis suggests the winter bloom is found from around water depths of ~20m - 60m. We seen this enahnced slug of plant material growing at depths 25-40 meters.  As a betting by next week, my bet is the blooms extends to around 55m. Is this the start to the winter bloom?  Man if it is, as a geek scientist, I am stoked! Interestingly,the nearshore shows higher backscatter.  This is not true offshore in my hopeful winter bloom.  This would be consistent with low phytoplankton waters but the shallow water depths where significant waves lead to sediment resuspension.  The resuspended sediment often shows enhanced optical backscatter values, and these waters showed low chlorophyll fluorescence indicating low phytoplankton biomass.  Colored Dissolved organic matter showed slight variability cross-shelf, but nothing dramatic with slightly higher values nearshore.  these enhanced values correlate with the low salinity water which is consistent with freshwater inputs. For the ocean geeks, we have several different water masses indicated by the salinity, temperature, particle, chlorophyll glider data.  We have a valuable data set to study the transition period in the Mid-Atlantic from stratified to well mixed conditions!






The JPL daily update

Tuesday, November 3rd, 2009

JPL Executive Summary of 11/03/2009

NAM forecasts indicate winds in the range of 15~20 knots for the next two days. Northeasterly winds Monday and Tuesday will switch to Easterly by Wednesday and Northeasterly again by Thursday but with much weaker amplitude (less than 10 knots). Even though with more clouds on Sunday, there is still reasonable satellite coverage for SST. We receive three ocean model forecasts Monday. Models still show warmer SST on the shelf break. The variance of the ensemble forecast shows major differences in salinity at the fresh-water input locations. On Monday, ensemble predictions were used by mission planning software to compute optimal current-sensitive submersible paths and retask three Rutgers gliders: ru05, ru21, and ru23. The planned paths from each of the three available models showed good agreement over the area of interest. However, all three gliders didn’t go very far during the last 24 hours, in contrast to the eastward direction as predicted by the ensemble model forecast. Examination of the ensemble forecast variance indicates major differences in subsurface current among the three model forecasts around the current glider locations.

November 3, Morning Update

Tuesday, November 3rd, 2009

The gliders have been dancing a nice tango last night.  Evening and morning conversations with Chip & Dave, has us thinking this largely reflects the strong conditions on the shelf.



The currents measured by the CODAR support this interpretation.  Below are the currents from last night and remember the nominal speed of a glider is 20-30 cm/s.  Hopefully currents will lay down and we can start making headway.


What do the forecasts show for currents? Two forecasts for 11/4 are shown below, and they  suggest currents will begin to lay down.  That would be most excellent.



How well are the current forecasts doing?  To look at this we compared model outputs against measured Codar currents.  While there are small differences, on the whole the models appear to be doing very well.  This gives me faith that we will start to make progress today.  We shall see.


November 2 Evening update

Monday, November 2nd, 2009

Why are the gliders heading in a circle?

The first big victory of the OOI OSSE occurred on Sunday night.  The planning and prosecution software used model forecasts to provide a series of the waypoints that were loaded into the glider flight plans.  The waypoints guided the gliders for the day.  At 2:00 EDT we in the team conference call agreed to a series of offshore waypoints fanning the three gliders out across the shelf.  Upon some of the afternoon call backs we notice that the gliders we heading back to shore!  That reflects, our guess, the glider heading back to a waypoint inshore by the path planner, and the glider had went past the assigned waypoint during its flight.  The glider therefore backed up and headed back to the missed point.  The gliders now however have been given a series of new waypoints and have begun to now swing around and head back offshore.  It will be an interesting experiment as the currents are strong and perhaps the gliders will be advected offcourse during the next few hours.  We will have some fun exploring ways to optimize the software as part of this OOI construction phase.



The conditions of the shelf have not changed dramatically over the last 12 hours. There are strong northward flowing currents.  The particle loads are enhanced in the nearshore bottom waters.  Not shown, but the no stratification in temperature and mild cross shore gradients in salinity.



Where are we going? Below the yellow line shows the actual glider path.  The blue line shows the CI path planning course (here using the path planning based on the Rutgers ROMs model).  We will follow to see how well the glider does.  There will be several things to follow in the coming days, we have several different ocean models, you can check to see how different the predictions are between the models.  We can see how well the glider does.  Remember if the glider does not follow the path planning it can reflect unrealistic models, it can reflect changing ocean conditions, it can reflect ocean conditions stronger than glider flight capabilities.  Updates to follow in the AM!  Real ocean tests are always the best!


State of the Ocean and System, Nov 2, 2009

Monday, November 2nd, 2009

Join the OSSE team!  We really encourage people to check out the sites at:

It is an open site that will allow anybody to join the workgroups.  The goal again is to explore the Mid-Atlantic Bight and to provide as much feedback as possible to the OOI CI team.  The site is outfitted with a series tools and resources. Also check out the web site of the data, in simple check out our ocean!

OK onto business, what is the status of the Mid-Atlantic Bight today?

From our JPL buddies we have the daily update. Weather forecast for 11/01/2009 - Southerly winds are observed during the weekend with a speed around 10. NAM forecasts indicate that the wind will switch to northerly winds by Monday and last for about two days with about the same magnitude. Because of the improved weather conditions, there is therefore excellent satellite coverage for SST. We continue to receive four ocean model forecasts on the daily basis. A multi-model ensemble forecast is constructed based on the equal weighting method. The variance of the ensemble forecast is also estimated and will be used to guide the glider deployment in the coming days. Four model forecasts are also compared with observed SST and surface current. A Google Earth (GE) based web interface is also developed to track the four gliders being deployed.

For the gliders, the three nearshore glider are flying extremely well and are being passed by the offshore glider moving in. We are making good progress.   We had the first CI assisted way-points sent to the gliders last night.  A big step, that now we iterate on to make it more useful. So over the last 12 hours we have had our formation of four gliders.


What are the gliders seeing?  They this morning were seeing strong northward flowing currents. no temperture stratification with a mild cross shore gradient.  As before the particle loads as indicated by optical backscatter.  The regions of enhanced backscatter are also associated with high chlorophyll.






AUV team is arriving robots in Hand.

Monday, November 2nd, 2009

The Cal-poly team has arrived in Tuckerton and is busy getting the systems ready for deployments.  The deployment schedule for the week is post below and was provided by the AUV team leader Arjuna Balasuryia.  The MIT and NURC teams arrive tomorrow.  The plans & team are provided below.


Preliminary Test Plan for AUV operations


1. MIT - On-board autonomy (Back-seat driver) and Ocean Model

2. NUWC - IVER AUV (Hammerhead)

3. Cal.Poly - IVER2 AUV

4. JPL - Mission planning, EO-1 and Ocean Model

5. Rutgers - REMUS AUV, SLOCUM gliders, MODIS, AVHRR and Ocean Model, R/V Arabella

6. UMass - Ocean Model

7. Stevens - Ocean Model

8. UNC - Ocean Model


This trial is scheduled for the period of Nov. 4 - 6, 2009 in the mid-Atlantic bite region to demonstrate the planning and prosecution sub-system under the OOI-CI.


1. Deployment of mobile assets to capture an Ocean feature. Identification of the Ocean feature based on Ocean models and plan missions to deploy a fleet of mobile assets.

2. Feed data from the field nodes back to the models

3. Demonstrate the on-board autonomy of the mobile assets for adaptive ocean sampling

4. Demonstrate the cluster autonomy capabilities of the mobile assets in an acoustic communication network for adaptive ocean sampling.

Field Operations

Datum: Latitude 39o 25/ 06.82//N, Longitude 74o 11/ 55.08//W

Gateway Buoy: Latitude 39o 27/ 46.17//N, Longitude 74o 08/ 23.20//W


Fig. 1: AUV operation area 10km x 10km box

Communication Infrastructure

All the AUVs are equipped with a WHOI micro-modem. R/V Arabella will have a modem on a side connected to the top-side command and control station. Top-side command and control station consist of a display tracking all the mobile assets in operations and this view can be shared through the Google Earth. It also has a control panel to re-direct all the assets including commands for adaptive ocean sampling. In addition to the modem on the side, Arabella will have a RF link to the gateway buoy for wider acoustic communication coverage.

Day by day operations




Operations near the field station. Setup the lab and command and control station on Arabella. Check the vehicle status and run couple of front seat missions around the field station to check the low-level controllers and sensors. Check on acoustic communication links. Run back-seat missions. On recovery upload data to MIT server.


Go out to sea deploy the gateway buoy and run missions in the box in Fig. 1. Deploy the assets to the loiter points determined by ASPEN/CASPER. Check track-trail behaviors on IVERs for collaborative missions with REMUS. Run a collaborative sampling mission with REMUS running a grid box and IVERS trailing it. REMUS will do a race-track out of the box and back and IVERS will re-acquire REMUS. Run thermo-cline missions using IVERS to demonstrate the adaptive ocean sampling capabilities. There will be few virtual vehicles simulated on board the ship collaborating with the real AUVs to test cluster autonomy behaviors. On recovery transfer data for analysis.


Half day ops at sea running similar missions as on 11/05/09. After lunch recover the gateway buoy and get back to shore for unloading.