Began working on a new problem this week. What would be nice is if there was one algorithm that we could apply to a handful of filters that would be able to identify everything equally well. As you know, biology doesn't make things very similar except for a few handful of interesting mimics. Therefore, one thing that may work for the detection of one species may not work for another. In fact, life stages may throw off detection within species. This week I was able to show how different adjustments to filters could isolate and ID the organisms we were missing and prove that one size filters do not fit all. The beginning of the week we tried Hough-based tracking to get the other half of scotoplanes (his feet have been far more interesting to the computer than the rest of his body)
Before
After
The next obstacle is making the adolescent stage of echinocrepis stand out. the young are white and the old are black, both work well with contrast. The adolescent stage is a horrible sand colored monstrosity that blends into the background.
I have finished my Intro, Methods and Materials, and lesson plan. Im moving into the results and discussion soon. Going to be done in no time.

This week has been moderately slow. I've been working on RStudio Tutorials and getting acquainted with MBARI's banzai pipeline. Unfortunately there was a hiccup with the DNA sequences & primers sent to Stanford. The forward, index and reverse primers were all correct and yet there was some problem where they could not be sequenced, Kris believes its a mistake on Stanford's part in which case they will pay for another kit and hopefully it will not take too long. Kayaking on Monday was quite fun, I did not suck as much as last time. I'm helping out with the MBON bbq next week even though I won't be there for the actual event because NASA!!! The seminar yesterday was super cool and it was nice to see a young woman being super successful, it gives me hope (also I think her shrimp shirt was FIRE). The end of the program is coming up quick so I'm also playing with paper ideas, although having the proposal for the REU program is a big help to getting started. Still though, the end of the program means the end of seeing all MBARI humans and thats saddening. BUT STILL 4.5 weeks left.

It has been a moderately frustrating week! I received feedback from my mentor and unfortunately, we have different approaches when it comes to our way of communicating scientific ideas. The feedback word document I received looked the first essay you wrote to your English teach in middle school - it was 90% red marks. It was definitely disheartening to see.

Going through the revisions that he wanted me to make further illustrated why it is important for me to create this story map. While there were many great suggestions that he made regarding phrasing, there were parts that he wanted me to elaborate on that honestly no one would care to read in the real world. Acronym soup and discussion about who's funding who and why does not cater towards a general audience as they are likely to read my story for 1 minute or less.

Now comes the part where I will negotiate with him regarding what is important and what isn't. Unfortunately, he is on vacation this week and I do not want to bother him while he is on vacation so hopefully the story will be done next week so that I can start on the next one.

Last Thursday, after a nearly 12-hour process, I am happy to say I successfully regenerated radiolarian photoproteins (for both shallow- and deep-water species). This is exciting not only because it hasn't been done before, but also because I now have a significantly greater amount of material to work with (if you take care of your photoproteins/apo-photoproteins, i.e. keep them frozen, you can often regenerate photoproteins several times). This week, with my newly generated material, I have: determined that the optimal pH for the biolum reaction in a Collozoum sp. (shallow-water rad) is 8.5, found that Collozoum (and even Tuscaretta - deep-sea rad) photoproteins are stable even after being left at room temperature for 2 hours, and that filtering out excess, unwanted calcium ions from a solution of apo-photoprotein is a pain in the butt (which is a very important step in regeneration). While I have had small victories, it is apparent that this project has a long, long way to go, so let's see how far I can get by the end of this internship!

In the meantime, I get to go out on the Paragon today for a mini-cruise on the bay! Some of my lab mates are diving for samples, so I'm coming along for the ride (and if I'm lucky snag some samples for myself). If you're outside any time between 10:30-11am, you'll be able to spot us by our "Life Aquatic"-inspired headwear: red beanies!

I've spent the majority of this week learning about the engineering side of my project, the parts I've just kind of taken for granted up to now but have decided I need to fully understand to take proper advantage of. Not that I have any regrets, but I would say that an ecology degree has not quite set me up for an easy time understanding all of this. I'm up for the challenge, though.

At this point I think I have a pretty good grasp on Fourier transforms, a moderately good grasp on Welch's Power Spectral Density Estimates, and I can see an understanding of Daubechies wavelets in my future. I am a very visual learner, I've come to find, which makes the textbooks and webpages and most resources really hard for me to understand without graphs and other visual representations. Engineers don't have enough pictures in their books, I'm telling you! Jokes aside, though, I'm feeling pretty good about the progress I've made towards understanding these concepts since I had virtually no useful background in any of these things. Representations of amplitude, frequency, time, scale, translation- they're all very abstract, and half the battle is wrapping my head around how to think about them in terms of actual physical sound waves.

I went to the Moss Landing Marine Lab Library this morning to get a book from the joint MBARI library, and they were very nice! Moss Landing Marine Lab is also very beautiful. The librarian, Katie, helped me find a book that she thought would be helpful to me, so now I have a very intimidating-looking book called "An Introduction to Random Vibrations, Spectral, & Wavelet Analysis." I'm going to understand this stuff if it's the last thing I do!

I also have a PAMGuard classifier I'm feeling pretty good about! After looking at the Welch's Power Spectral Density Estimates and tweaking my frequency parameters appropriately (shoutout again to Nick for being my Matlab Signal Processing/Engineering consult! He's very good at breaking things down so I can understand them without a master's degree in engineering, and I really appreciate the help) I'm detecting what I believe is about half the beaked whale clicks in my file of interest, which might not sound like a big deal, but considering when I started it didn't detect almost any, it's an accomplishment that I'm feeling good about. Now I'm going to test it on some other files to make sure I've not specified it too much to this file, but I got my parameters from looking at the spectral density estimates from a variety of files and all of them seemed pretty consistent. I also now believe that I understand what every parameter PAMGuard uses is actually looking at and doing, so now I know why expanding the frequency bands for some of the parameters ended up hurting me more than helping. It's funny how much easier it is to fix things when you understand how they work in the first place.

In a super fun throwback, yesterday my computer broke again. It was being very laggy and the mouse wasn't showing up, so I made the lethal mistake of trying to reboot it- it's been just long enough that I forgot I wasn't supposed to do that, because when I booted it back up there was a Windows error (I have a Mac, so it's particularly awkward). Sheldon, however, fixed it in about 20 minutes because he is the best, and now I'm back in business. I was almost nostalgic for my first two weeks at MBARI, while sitting under my desk trying to figure out which phone jack to plug my ethernet cable into to make the really old computer in my office boot up so I could keep working while Sheldon was fixing my laptop.

I've also hammered out a structure for my final paper, which I think I'm going to try to start next week. My mentors are both on board and I will (hopefully) have some useful data to present, so hold on to your hats and prepare for some really exciting spreadsheets! (I'm going to slip some whale sounds in there too, so hopefully the audience won't fall completely asleep!)

Another fun week! Its always a blast to get out on the slough, I hope that everyone enjoyed that. I also really like the speaker this week. I'm really excited about the explorer program, it is so refreshing to see projects like that in our current funding climate.

Its definitely starting to feel like the home stretch. My analysis is going very well, and its time to start making pretty visuals for the talk. I have a general idea where I'm going with it, and I would like to have a head start on what I know I always run out of time on. As far as nuts and bolts go, I need to take the syslog from the rover(which is like a logbook of everything the rover does) and pull out the CM data raw so that I can plug it into matlab a lot easier.

I finally finished with my VARS measurements last week! This week I have been doing data analysis and running stats on the data to look for patterns. I am currently creating knotched box plots in order to compare pulses, as well as size class and color. This will allow me to assess whether size distribution of 1 pulse or 1 color is significantly different from the others. I am also working on creating a cumulative sizing curve for the data.

After I have finished doing that, I will be working on spatial statistics in order to see whether individuals are clumped, random, etc. In order to do this I will be creating 50 meter bins for each transect so that I can compare the bins within each transect. I plan on mapping the transects on GIS, so that I can visualize where each of the pulses/transects are located in reference to one another. This will be useful when running the spatial statistics because if there are patterns within the bins of each transect, I can reference the bin locations on the GIS map in order to figure out what is occurring in that particular area.

Lastly, I am taking 10% of the images (best quality) that I used for measuring individuals in VARS and running them through a separate software program called Solidworks that has an auto-trace tool that automatically creates a 2D sketch from an image that you can analyze. It also calculates the area of the image, which will allow me to compare my VARS measurements with the Solidworks measurements, so I can account for error in my methods.

Week 7. Reflecting on the progress I've made with coding skills. I had never touched MatLab before week 1 and spent most of my time pestering my mentors for help with every little thing and and now I spend most of my time at my computer and they come to ME to ask for updates on what I'm doing . Still working up all the datasets I have amassed to perform all of the mathematical analyses I want to do. Most of my time is spent figuring out the best way to create "perfect" datasets (no gaps or "NaNs") out of NaN-riddled datasets that reflect sporadic sampling, instrument errors, inconsistent sampling depths, etc, and chopping up the results every which way--monthly averages! No, bi-monthly averages! No, daily! In addition to performing PCA on all of the different parameters, I am working on correlating the resulting time series to the Multivariate ENSO Index, the El Niño Modoki Index, the North Pacific Gyre Oscillation, and the Pacific Decadal Oscillation to see which of these indices can best explain the variability we see over time in our data. Once I have done that and gone back to "clean up" the other figures I have produced through the preceding weeks, it will be time to really interpret the results, make some inferences as to what processes are driving the changes and what I think will happen as time goes on (ie. start writing the paper and preparing for the presentation). Work, work, work, work, work, work. He say me have to...

This week has been actually super productive. I have hammered out the direction I want to go in with my project which is dominance between diatoms, autotrophic dinos and heterotrophic dinos (proxies) in space and time. So basically, I will be looking at when and where these species are dominating in comparison to each other, and then seeing what the physical conditions of the Bay are at the time (i.e. stratification, nitrate and oxygen levels, thermocline depth, bathymetry, upwelling, and SST).

I have learned a lot from my mentor in regards to MatLab, code writing, and the scientific process. Part of what I have to do as well is justifying my parameters, like showing why I decided to set my depths to find the mean proxy values over (as to not skew my data) to 40 meters rather than 30 meters for example. Which is something I do not think about because I just do it, but someone could ask during my presentation and I would not be able to answer the question fully!

Spent part of this week figuring out what went wrong on the flyer cruise last week, and it turns out we didn't actually break both spares (the second time at least) on the flyer, but just one of them. We did, however, discover that the connection between the laser and the fiber cable is incredibly finicky which is why we thought it had broken last week. All a moot point now though, because we ended up breaking the spare shortly after discovering just how sensitive the port alignment is. We have plans to talk with the manufacturer and see if maybe it's the port that's messed up and not the fiber connector, and also to investigate whether changing over to a different port system makes sense.

We have also essentially given up on doing a full fit of the five-peak stretching regime in seawater, since we can get fits with wildly different parameters that all look just as correct. We briefly discussed investigating an isobestic point in the regime instead, but decided that was silly once we realized there's no way for the math to work out between five peaks. Instead, I might look at the bending regime (which only has two peaks, and therefore an isobestic point may exist) and see if there is a well defined temperature dependence there.

Almost on time this week! I've been pretty under the weather and have found it hard to get work done. After my update last week, I finished implementing a first pass at the translator. It takes the abstract syntax tree (a data structure that represents all the commands the user is trying to communicate) and produces XML code. Currently, this XML code does not comply with the schemas to which actual mission scripts have to comply. This is okay, as even a working prototype on a subset of the schema would be really cool. Other good news: our meeting with the LRAUV operator went very well! He was impressed with what I had so far which honestly was surprising. I had misunderstood some key aspects of the mission scripts, which he corrected me on (for example, "Behavior" constructs are written in a lower-level language and are not defined in the mission scripts, but rather called and run. This wasn't clear from reading the code). He also reminded me that the online interface they use to run missions also has a "virtual" mission option, so I could run produced code that way on a virtual LRAUV just to see what it does. While this isn't necessarily helpful for the project (I verify everything at the code level), it would make a cool visual aid for the seminar, as I could show an actual mission and the same mission produced from my new language running side-by-side (ideally they would, in fact, do the same thing).

I'm slowly adding new features and thinking about how I want to handle the misconceptions I had with how the scripts work. Some new features that I wouldn't have thought would be tricky are turning out to be challenging, which is always frustrating -- especially when it requires restructuring large sections of the code. It will get done!

Kayaking was fun! Despite living in Monterey, I had never kayaked up in the slough. Hope everybody has a good weekend!

After weeks of battling with C#, I finally have the sensor reliably reading data at 100 Hz and saving all the acceleration values to the SD card. Once the data is imported into Matlab, I remove the DC offset from the signal and then apply a moving average filter to smooth out the noise. Using Welch's method it is possible to look at the spectral power density of the filtered and unfiltered signal to see a reduction in noise and identify that all the signals we are interested in are below 2 Hz. Finally, I apply a numerical integrations scheme using the trapezoid rule to approximate the velocity profile measured while the sensor was moving on the linear translation stage. In the image below, the linear translation stage was programmed to have a constant acceleration of 1 mm/sec^2 and reaches a terminal velocity of 25 mm/sec before decelerating. The black line is the velocity profile obtained from the raw data, and the blue is the estimation of the velocity once the data has been smoothed (it is looking good!).

Next I will program longer motion paths so that we can collect up to 1000 seconds of data and start to access the error in the sensor over longer sampling times.

I think I'm the last one writing this but that means that I had been super busy that I didn't even had enough time to write it!

During this week I have been super busy but it had been less hard than the last one. I have been collecting the data that we get during the 24h experiment in different files, making some graphs and trying to understand what's going on with micromonas. I'm also preparing a little powerpoint about what I have been doing during this month using the data that I have been organizing because every Monday I have a meeting with the people form my lab, and next Monday we need to explain how are our projects going. I'm a little bit nervous about it but I know this would help me to get more confident for the symposium. Also, everyone in my lab knows something about algae and they're super smart so I'm sure they will help me a lot making questions or giving me advice about how to make a great presentation for all of you in a few weeks.

After all this work, we can't confirm that micromonas is eating because the percent of micromonas + beads is very low compare to the total of micromonas, but I think that we have some interesting stuff that I will show you soon and the only thing that we would need to confirm if they're eating or not it's more time (something that, unfortunately, I don't have).

It's been a long week of analyzing samples from my experiment. Everything started out fine but, turns out, the solution we used to preserve my abalone larvae started to degrade their fragile bodies when they were stored overnight...Therefore, I have to run my experiment again. At the end of week 8... Tiffany and I went up to Davenport this morning to pick up more abalone for next week's experiment, so it's going to be a week of just preparing. It's going to super stressful but I'm ready!

On another note, I've started gathering information I am planning to include in my presentation at the end of the internship and I've realized I have a lot of "basic" information about my project I need to cover. Problem is, I don't think I'll have enough time to present it all. So I'm trying to figure out some ways to summarize it without cutting out information critical to understanding what my experiment is about.

Another busy week. I was having some issues verifying the parameters of sperm motility analysis with the ImageJ CASA plugin. While I got data, the process was fully automated so I wasn't completely sure which cells were being tracked. Another issue was that sperm were going in out of focus from swimming up and down the water column which could have led to error in the data.

With all of these issues, my mentor and I decided that we should try a different tactic. Some digging and talking to other scientists led us to a new plugin called ManualTracking. This plugin does exactly what the name sounds like. It allows you to manually track sperm frame to frame. The data it outputs is distance traveled and velocity (frame to frame). From this we can determine sperm motility. So now the game plan is to track 10 sperm per each sample and analyze the data from this to compare sperm motility between high and low pH exposures.

The symposium is rapidly approaching so I'm on a tight schedule to process all of these videos. My life for the next week will be endlessly tracking sperm. In between that I will be helping Aileen and Daryll with their experiments. All of us are taking on a different lens of effects of climate change stressors on different life stages in the red abalone so I have loved working together and learning about their projects.

Better late than never, right? I completely spaced on this week's Confluence, in part to my Thursday being tied up (more on that in a bit)...

Meetings, meeting, meetings - this week was all about meetings!

Monday's kayaking powwow was fabulous. The conditions were awesome and there were plenty of critters to ooh and ahh at! Kayaking the slough had been on my bucket list for some time, so I was stoked to finally cross it off. I'm eager to do it again!

Later that day, I filled in on an after work panel with aquarium staff along with a group of college students (mainly freshmen/sophmores...?) who are assisting with the aquarium's teen programs this summer. We had an open discussion about career-related tips and suggestions. Though the panel was designed to support the students, I think we all came away with learning a few new things! It was also an awesome opportunity to network with aquarium staff.

Tuesday, Desmond and I met with the social media intern from NOAA's MBNMS office. She was interested about the work we're doing and vice versa. We chatted strategies, ideas, and insights while discovering similarities and differences between our respective institutions.

After delivering her seminar on Wednesday, I met with Amanda Netburn to learn more about her work so far with NOAA's OER. Having also spent time on the E/V Nautilus, she described the differences between the programs and how they are played to their individual strengths. Amanda also provided some great recommendations regarding my interest in marine science communications, including a fabulous reference for my project.

Thursday, I finally got sit in on the aquarium's weekly social media meeting and it was AWESOME! Observing their individual teams brainstorm and collaborate reminded me of previous experiences I've had working with content/creative teams. Although I was there as a fly on the wall, I was impressed with the thought processes and execution of social media tactics, many of which MBARI could just as well implement and benefit from. They were equally enthusiastic about an MBARI presence at the meeting and left an open invitation for my return.

Last, but not least, I joined Susan von Thun and Kelly Lance in attending a Friday morning seminar at the aquarium. Shark biologist and conservationist David Schiffman discussed the roles and effects of social media in ocean conversation and shared observations from his personal uses. He had some interesting perspectives that gave me food for though on researchers' perspectives of social media and public engagement. Some of the aquarium's social media folks were also in attendance, so there was a bonus of getting to see them again and carry conversations over from the previous day.