Archive for the ‘Cooks Crossing’ Category

Disappearing Act on Magical Line

Wednesday, July 14th, 2010

As of 7/7/2010 at 14:37 Cook has been missing in action. It has been 170 hours since Cook last called. We have talked to Webb and they don't see any scenario where it will come back. We currently have no indication why we lost Cook.

Cook was out for 108 days, and was about 350 km from where Drake went down after 104 days last year. We are examining possible causes for Cook's loss and seeing if there is any correlation with Drake.

Cook was a brave droid who died in battle. He collected excellent data which we will continue to analyze.

We'd also like to take this moment to thank our partners in Spain. Your forecasting and analysis have been incredibly critical while piloting the glider. We'd like to thank you for your time, effort, blogs and Skype sessions. Working with you on this mission has be an honor and a joy. We look forward to continuing our relationship in the future.

A U.S. July 4 Holiday Report on July 5

Monday, July 5th, 2010

Figure 1: Here we take a look at the track of Cook on the basin scale.  At the end of last week we set the next waypoint for Cook as the top of the Mid-Atlantic Ridge.  This will be the section of the flight that the summer students concentrate on.  Lets see if we can make the European side of the Ridge before they leave.   Once on the European side of the Atlantic, we set our sights on the fall transit to the Canaries.   The present distance to the Canaries is about 3,500 km.  If we look at our progress along 26.5 N, we made the turn back on May 11.  Since making the turn, we have traveled 1,150 km.  In 6 days we'll take another check to see how far we have made it in 2 months.  Still we know the approximate numbers.  We took about 2 months to go 1/4 of the distance to the Canaries.  Maintaining this average speed gets us to the Canaries in another 6 months - again - not a great time to be out on the North Atlantic in small boats.  We also have to worry about the thermal engine, and how much temperature change we need from top to bottom.  We like to have a 15 C temperature difference, but can we live with less?  A good question. That is about the time that the surface to 1200 m temperature difference reaches a minimum, giving us another envelope to push.  So there is a lot of talk right now between our students in the U.S. and our friends in the Canaries.  What temperature range do we expect this winter, and how will Cook handle it?  Should we give up trying to maintain the 26.5 N line, and instead use the currents to our advantage to speed across like we did with RU27?

Figure 2: Lets look at the HyCOM forecast, this time for sea surface height (color) and surface currents (black arrows).  The Cook currents have been switched to red so they show up.  We are presently going against the current that is circulating clockwise around that sea surface high to our north in red.  HyCOM says small currents are on the way as we cross into the blues.

Figure 3: Overlaying the HyCOM currents on the satellite sea surface temperature map gives us a lot of confidence in HyCOM.  We see lots of features in the forecast currents lining up with the fronts and features in these satellite observations.

Figure 4: But now we overlay the HyCOM currents on the satellite altimeter sea surface heights and geostrophic currents, and we loose a little confidence in the model in some regions.  Here we see that some features are lining up with HyCOM & the satellite SST, giving us strong confidence in these features.  But some are not lining up.  These we have less confidence.  if we had time, and our mission was data assimilation, we would head Cook straight into these regions of disagreement.  If we are racing against time, we do the opposite, and head Cook into regions with known favorable currents.  But in our case, we are still on a heat flux science mission, so we'll maintain this line for as long as we can and still make the Canaries.

Will We Make It?

Friday, July 2nd, 2010

We took a look at Cook's current position and how long it has been flying in attempts to predict when it will be arriving in the Canary Islands. Since Cook is a Thermal Glider, powered by a change in temperature, we needed to look at the modeling data from the past year and make sure we would have a significant enough change in temperature between the surface and 1200 meters so that we can fly Cook efficiently. We think that in order to maximize Cook's flight efficiency we must see a change in temperature of at least 15 degrees Celsius.

Potential Path for Cook

Looking at a potential path and doing some calculations based on how we have been flying, we were able to predict Cook's arrival in the Canary Islands around December 12, 2010 at the earliest. We are thinking that Cook will be arriving later, probably mid-January.

We looked at temperature plots for the Canaries for the winters of 2009  and 2010.

During December 2009 there was a temperature difference of about 16 degrees Celsius. This should be a large enough difference to properly fly the glider.

These are temperature plots from January 2009 and 2010 respectively. In 2009 flying would have been difficult, 2010 there was about a 15 degree difference.


These are temperature plots from February 2009 and 2010 respectively and both show a temperature difference less than 15 degrees Celsius. We really don't want to be flying in the Canaries during that time.

The winter of 2009-2010 seemed to be warmer than 2008-2009, allowing the temperature difference to be greater. If we are going to be around the Canaries after December, we will be hoping for a warmer winter in that area.

Once we looked at the temperature plots for the Canary Islands, we also needed to look at the plots for near Cape Verde, our destination after the Canaries.

December 2009 HyCOM Temperature

This is a plot of the HyCOM model temperatures for December 2009 in Cape Verde. There appears to be a difference in top and bottom temperature of about 20 degrees Celsius.

January 2010 HyCOM Temperature

This is a plot of the HyCOM model temperatures for January 2010. The end of January will probably be the earliest we will get close to Cape Verde. The temperature difference appears to be about 19 degrees Celsius.

February 2010 HyCOM Temperature

This is a plot of the HyCOM model temperatures for February 2010. This will be when we will be flying near Cape Verde. The difference in top and bottom temperatures appears to be about 18 degrees Celsius.

We are able to look at the past year's HyCOM temperature profiles as a guide to what this year's temperature profiles should look like. Using this information, we are able to see that there is enough difference in top and bottom temperatures to exceed a difference of at least 15 degrees Celsius. We will still be able to fly efficiently near the Canary Islands in either December or January and we will be able to fly near Cape Verde in January or February, so we will make it!

July 1 Update

Thursday, July 1st, 2010

Cook continues to make progress against a strong head current. Most recently reported currents from Cook (white lines along the track) are 22 cm/sec to the northwest.  We have the waypoint pulled in tight to stay close to the 26.5 N line, making Cook's speed 34 cm/sec to the southeast.  So we are still making 12 cm/sec progress in the right direction. Last year 22 cm/sec would have pretty much stopped us in our tracks. Amazing what a difference a year makes.  The HyCOM model forecast sea surface height (color fill) and surface currents (black arrows) are in perfect agreement with Cook.  So some of these features extend deep.  Perhaps we are out of the "water lily" feature.   Today we will work on the next waypoint for Cook.  We want to have it set up for the long weekend.  Do we move north up to 27 N to the weaker currents?  Or do we maintain the line?  Cook is doing such a good job maintaining this line, we are all reluctant to give it up.  We are collecting a new dataset along this critical line for monitoring the maximum in the north-south heat flux.  Since Cook is healthy, it looks like we can take the time we need to get across.  That will be the focus of today's discussions.

Cook’s Flight Characteristics

Thursday, June 24th, 2010

While analyzing Cook's flight characteristics, we have noticed that the glider is experiencing a drastic heading error at the start of each dive and climb.  In the figures below, the blue lines represent downcasts, and the red lines represent upcasts.  At the start of each dive, Cook veers 100 degrees or more to the left, and at the start of each upcast veers about 50 to 60 degrees to the right.  The glider seems to experience a larger range of heading error on downcasts than on upcasts as well.  The first plot shows typical heading error, and the second plot shows a more drastic error in which the glider actually rotated a full 360 degrees.

Taking a closer look at the previous plot (shown below),   the yellow points show that Cook's heading error went from about -2 radians (-120 degrees) to +3 radians (+170 degrees) in less than 10 meters.  From this data, we conclude that the glider rotated a full 360 degrees before heading in the correct direction.  Typically, the fin should be able to steer the glider in the desired direction.

The fin responded to Cook's heading error, but was not able to correct the deviation.  Once the glider was facing more than 180 in the wrong direction, the fin quickly changed position so that Cook would continue spinning in the same direction back to its correct heading.  The figure below shows the position of the fin in radians.

The pitch and roll do not seem to be affected by Cook's heading error and fin position.  The roll, however, does see a slight difference during upcasts and downcasts.  The roll never reaches more than 5 degrees in either direction, but is always slightly to the right and larger during upcasts.

Cook and HYCOM match up

Tuesday, June 22nd, 2010

The glider's data and the HYCOM surface model both show the currents going in the same direction. Despite a head current Cook is still managing about .24 m/s.

The Navy’s Predictions are Correct

Friday, June 18th, 2010

So today's task was to look at Cook's data and compare it with the Navy's HYCOM model predictions. Now knowing how to do both in MATLAB we set off to produce some profile plots to make the comparison easier.

Cook Temperature Profile

HYCOM Temperature Profile

This is a comparison of Cook and HYCOM Temperature profiles. Both were constructed using data from Cook and HYCOM on June 18, 2010. Seeing that both show the same descending trend and are close in top and bottom temperatures Cook has confirmed what the HYCOM model predicted.

Cook Salinity Profile

HYCOM Salinity Profile

This is a comparison of Cook and HYCOM salinity profiles. Both were constructed using data from Cook and HYCOM on June 18, 2010. Seeing that both show a similar descending trend and have similar top and bottom salinities, the HYCOM prediction was proven to have been correct using Cook's data. The small difference seen at the surface between the two profiles is due to the heading error of Cook, which we are currently still working to fix.

Learning MATLAB

Thursday, June 17th, 2010

After a crash course in MATLAB from John yesterday, we started the morning by pulling the newest data file from Cook and plotting it in MATLAB. The blue and yellow show the down casts and the red and cyan show the up casts.

June 17, 2010 Depth v. Temperature

This plot is a Depth versus Temperature from Cook's last up and down casts. It shows the surface temperature being around 30 degrees Celsius and declining with depth to a minimum of 5 degrees Celsius. The thermocline can be seen from about 200m to 600m.

June 17, 2010 Depth v. Salinity

This is a plot of Depth versus Salinity from Cook's last up and down casts. The surface salinity is 36.35 psu and the salinity at 1200m is 34.6 psu. The halocline appears around 200m and extends down to about 900m.

June 17, 2010 Surface Salinity

While analyzing our plots we came across something that doesn't look quite right. We had some questions as to why the upper 40m of the salinity plot.  We think this could be due to Cook's current flight pattern which would not allow the sensors to have the right flow through them to provide accurate data.

June 17, 2010 Depth v. Density

This is the plot for Depth v. Density for Cook's last up and down casts. The density at the top is around 1023.5 Kg/m^3. The density decreases with depth to a density of 1032.5 Kg/m^3 at a depth of 1200m.

We will continue to plot CTD data for Cook and will be having another lesson from John today in order to learn to plot the HyCom data.

Cook May See an Unmarked Eddy

Thursday, June 17th, 2010

From Cook's data it appears to be in a clockwise eddy, however the Altimetry says that it should be in a counterclockwise eddy.

The HyCOM data says that the currents are southward and nearly perpendicular to eastward currents reported by Cook. The differences in the information provided by Cook, the Altimetry, and the HyCOM make this an interesting area to examine with the glider.

Cook CTD sensor working great

Wednesday, June 16th, 2010

So we plotted some CTD data from yesterday's casts and it appears this sensor is functioning well. There does not appear to be any lags, meaning the CTD does not show any delay when reading the temperature data on the down and up casts at the thermocline.  When this occurs, we do not get an accurate reading of the position and strength of the thermocline.  The plots below show the results of yesterday's casts and we determined some good news.

The first plot shows one entire down and up cast from June 15, 2010.  The other two plots are zoomed-in figures of the upper 250m of the water column and also the water column between 250m - 850m.  Both of the zoomed-in figures show a randomness in the data, meaning that the results are an indication of the oceans natural currents.  The CTD doesn't show any delays in the readings between the thermocline where you have a change from warm to cold waters.  If that were the case, we would have seen a difference in the up and down casts.

We also saw a similar result with the salinity profile.  Below is a plot displaying the salinity data.

We are going to continue to monitor the CTD over the weeks ahead and we will keep everyone updated with how it is functioning.