At the start of this week I organized the radiometer data and paired the readings with the appropriate subject. At first glance the majority of the spectra looked like one green spike. However, after taken the median of a reading and dividing it by the median of the the closest plate reading, the spectra look a bit more familiar. Under ambient light the Didemnum spectra look somewhat similar to what we got with the DiveSpec. Below is a graph of the seven different ambient Didemnum spectra taken on the second dive. All the spectra have been divided by the closest ambient plate reading.

The radiometer was also used with a dive light. This produced more variability in the reflected light. This is most likely due to the distance and angle the light was from the subject. Below are the five different spectra of Didemnum using the dive light. Again, all these spectra have been divided by the closest dive light illuminated plate reading.

I was surprised that the ambient light readings looked similar to the DiveSpec readings. When looking at the raw data and seeing a single spike around 550, I figured there was to much organic material between the radiometer and the subject. However, to my surprise I discovered by using the plate readings, this unwanted light can be removed. I expect that if the radiometer were to be deployed on an AUV, it would be in less turbid waters then these.

In other news I should be finishing the visibility report soon. I will be happy to be done with it. Having never used matlab before I second guessed myself on a lot of the operations I did. However, this has helped me build confidence in my matlab skills. That being said, I find it quite frustrating that I can not find any strong correlations. I compared diver reports of currents and MVCO reports and they do not match at all. I assume the currents are variable and can change within a short time span. I have looked at a comparison between what the diver reports the current as and what the diver reports the visibility as and there is a relationship. However, the diver reports of current do not match the MVCO reports so I do not see the same relationship between MVCO current and diver visibility. So this gets me no where. I have been experimenting with three dimensional graphs in matlab. Even with the three dimensional graphs, swell height is still the only condition I can find any correlations with. Here are two three dimensional graphs,

Again, I don't see the correlation with period I have been expecting. I have looked at wave and swell period, and its just not there. I have no doubt that swell period does affect particle suspension, but it is not apparent in this data. I have recently learned that someone else at WHOI is doing the same comparison as I am. I look forward to comparing results.

This week I also got a quick introduction to the LISST. Next week I will be using it to determine the attenuation coefficient of Bentonite clay.

Early this week, I started compiling graphs and data from the MVCO visibility study. I am organizing this information so I can write a paper on what i found, or more realistically what i did not find. On Monday I ran a test to see if my data was matched correctly by comparing temperature recorded by the divers and by the 12 meter node. If the data is matched correctly they should be close to the same. The graph below shows the relationship.

As you can see they match up well. This gives me confidence that the matlab scripts i wrote were operating correctly. I will include this information in the paper.

The rest of the week has been spent either collecting or sorting radiometer data. All the data taken on Tuesday has been saved to my computer and organized using matlab.

Lets start with the Satlantic data taken on the second dive. With the Satlantic radiometer we logged data for 36 minutes, during this time the radiometer took 1294 light frames and 259 dark frames. 383 of the light frames had the lens cover on. That leaves us with 912 frames of usable data. I am still in the process of finding the best way to separate the spectralon plaque readings from readings of Didemnum and other objects. The plaque readings seem to fall into two categories, high reflectance and low reflectance. I'm assuming the high reflectance was when the dive light was in use. I can separate the brighter plaque readings easily but separating the rest of them is still in the works. Next week I will have some spectra from the second dive. Below is a picture of Didemnum taken on one of the dives.

The DiveSpec gave us surprisingly good results. On the first dive two usable spectra were taken with the DiveSpec. Afterwords spectra were taken of various organisms in the flowing seawater lab. The graph below shows some of the results.

When using the LED illumination of the DiveSpec, Didemnum produced a rather consistent reflectance spectra. The black lines are spectra taken while diving and the red lines are spectra taken in the flowing seawater lab.

Another characteristic of Didemnum we will have to look into is its ability to fluoresce. In the flowing seawater lab we used the DiveSpec to take one fluorescence reading of Didemnum. here are the results,

The spike at the beginning is blue light emitted by the blue LED's of the DiveSpec, however the small spike at 675 is not being emitted by the LED's. Its possible that Didemnum could fluoresce under certain conditions. In the future we will have to repeat this and see if we get similar results.

Also this week, Alina and I have been discussing the best way to get a picture of back scattered light off of suspended particles in water. We have come up with a few good ideas, and should be able to image only the back scattered light without the surface reflected light. In the lab I was able to eliminate all visible surface reflected light with Wayne's polarized lens. However, its my understanding that light reflecting off the particles will be polarized as well. So by using the polarized lens we may be eliminating a far amount of back scattered light. We will work this out and hopefully take the pictures next week.

I apologize, some of the graphs have very thick lines, i will correct that in future posts. Just as a side note i wanted to add this last graph of swell period plotted against swell height. This data was taken by the 12 meter node.

This graph shows some correlation between swell height and visibility. However it does not show a relationship between swell period and visibility. All three visibility categories cover a wide range of swell periods. The three categories begin to separate from one another when we look at swell height. This is the strongest relationship I've seen yet.

Hello and welcome to my new blog. As Emmanuel has requested I will post an entry every Friday about the weeks events/findings.

Early this week I finished, for the most part, comparing data from the WHOI dive logs to the data taken from the MVCO nodes. This was a good exercise in learning matlab, the amount of data taken by the MVCO nodes is enormous. Matlab was terribly confusing at first, but after working throught this data I have a good understanding of matlab.

That being said, my results are not what I expected. I expected taht the correlation between oceanic conditions and diver visbility would be obvious. When graphing teh data i see only weak trends. this leads me to believe that i may have made a mistake in matching the data up, or chaning the time in the dive logs from EST to GMT (I also took into account daylight savings time). However, I don't Believe this to be the case. One sign that my calculations are correct is that when divers were in the water at the MVCO site, the node was not logging data. this makes sense because the divers were making adjustments to the node. When matching the data I accepted the closest node reading within two hours of the dive start time. In some cases over 200 dives had no node reading whthin two hours, this leads me to believe my calculations are correct. Below is a graph of how the 12 meter nod readins deviate from the dive times.

In the dive logs visibility was rated A-E: A is 0-2, B is 3-10', C is 11-20', D is 21-40', and E is >50'. At the MVCO sight visibility was very rarely above C, and for the dives that visibility was higher then C there was no node data within two hours. Below is a graph of the distribution of visibility for all 929 dives done at the MVCO site in the last ten years.

The following two graphs are wave period and wave height over time, the color and shape of the data points describe the visibility.

I expected that visibility would be layered, with the best visibility occurring at low wave height and low wave period, and vise versa for the worst visibility. However, this is not the case. There is a large amount of data including numerous more graphs that show similar findings, that I am not going to post here. I will continue to see if I can make any conclusions based on the data I have.

In other news, we gathered spectra using both the DiveSpec and the Satlantic radiometer. The DiveSpec is functional, and I compared the Spectralon 99 and 95 plaques above and below water. I used the 99% as a reference for the DiveSpec, the spectra above and below water looked like this,

I then read the 95% plaque,

The reflectance of the 95% plaque was between 80-95% for both above and below water readings. These readings were at a 90 degree angle to the plaque. I also used the signature mode the DiveSpec to examine so other reflectance patterns. This graph shows reflectance of fluorescent light looking down into water and a second reflectance pattern from below the water surface, the difference between these two spectra give you a rough estimate of reflectance off only the waters surface.

The next set of graphs show reflectance spectra taken by the Satlantic radiometer. The spectra were taken by pointing both the light and the radiometer at the Spectralon 95 plaque. The first is of a white LED light and its very smooth and consistent. The graph shows 35 different readings from the radiometer. The second graph shows the reflectance of a xenon strobe. This shows 35 different readings as well. This gives you an idea of what the reflectance spectra looks like, however the radiometer and strobe were not synchronized, thus producing the variation you see.

Hopefully next week we will have reflectance spectra from various objects in the water including Didemnum. I have written a matlab program for importing and graphing data from both the Dive Spec and the Satlantic radiometer. I look forward to seeing what Didemnum spectra look like next week.