Monthly Archives: June 2015


What an exciting time to be alive! The 29th congress of the PSSA (Phycological Society of South Africa) is next week. Team Kelp and Co. will be flying up to Durban Monday morning and returning on Friday evening. There are heaps of talks, posters and two field trips. Mostly I am excited to be meeting all of the other phycology people in the country. I will also be giving a talk about the SACTN. Spreading the gospel as it were. I have worked out a first draft of the talk and will be touching it up over the next couple of days. I also volunteered to help other Team Kelpers with their figures for the conference.

I plan on making a lot of contacts and getting people interested in using the SACTN to help push the whole process along as waiting for SAEON to get off their buns is a fools game.

Circuit breaker shenanigans

After a long weekend in which I wrote a bit of code to create some boxplots for someone in my lab as well as run an experiment on the time series to see how many extremely hot and cold days there were, it was good to get back to working full time… And then the circuit breaker for all the plugs in my flat popped. I did however read a good article before then about how climate change may or may not be affecting EBUSs. Interestingly, as the land heats up, this increased warming was assumed to increase offshore advection, allowing for more upwelling however(!), as the surface waters warm even more, this increase in the stratification of the water column may actually prevent colder deep water from rising into the euphotic zone. At least to a degree in which an effect will be seen on the respective ecosystems. This could then lead to these deep waters becoming more and more potentially productive so that when they did finally break through one would be dealing with some potentially record breakingly bad algal blooms.

After finishing up on the paper I went on a four hour quest to sort out the circuit breaker issue. And when I returned home apparently the circuit breaker had taken a moment to consider it’s desire to be a team player and switched back on… so that’s cool. Anyway, next week is the PSSA and I need to get my presentation in order for that. So that will occupy the next couple of days. I am mostly going to butcher up my PhD Prezi I made for UWC and cut that down to the right length while staying action packed and exciting!!

Another Week

The inevitable forward march of time. With the citation work done, the chapter now requires a bit more attention to the relationships between missing data, time series length and the size of the trends detected by the linear models. This needs to be thoroughly expounded upon to show that the large amount of high changes in decadal temperatures detected in this study are understood to be potential artefacts of the quality of the time series’ themselves. Beyond this, the nagging issue of rounding decimal places still requires support from published literature before the article can be submitted for proper review and publication.

R Shirt

Mendeley has been a wonderful addition to my work flow. I am very disappointed with myself for not adopting this technology when I was first introduced to it during my MSc.¬† In combination with LaTex/ BibTex, all through RStudio, referencing articles gets munged into one seamless time saving extravaganza. I can’t recommend it enough! All of my citations¬† for the first chapter are now sorted and should I inevitably need to make any changes it will be done auto-magically.

In other worthwhile news, AJ gave a talk today to the R user group of Cape Town about R work flows and a bit of what we are up to and I was cited as one of the co-authors on the presentation. And throughout the presentation I helped field questions etc. It is a very rewarding feeling to be so knowledgeable about something. If only I were so clever about what I am actually meant to be studying… Also, I won an R shirt with a monkey on it! WIN

Back in the saddle

Last week I completed a level three emergency first aid course, which is one of the prerequisites for the Class IV commercial diver supervisor ticket I am working towards. Now I just need 100 hours of supervised supervising and resit the class IV exams. Then Team Kelp will always have a dive supervisor on hand, giving us more freedom in our research endeavours.

I also had my first draft edited and returned to me, thanks be to AJ. I’ve started cruising on that as well as introducing yet another figure. This time I show what the decadal trends of warming or cooling are along the coast, using colour, and the power of the results is shown in the size of the dots (coast_DT). I am also toying with the idea of interpolating the DT values along the coast and have the alpha value (transparency) of the ribbon change depending on the power, which would also be interpolated. I think that would look pretty cool, but would be much different from the other figures and so wouldn’t be very feng shui.

Before jumping on the revisions I also read two short papers about recent findings that discredit the IPCCs (Intergovernmental Panel on Climate Change) findings that there has been a warming “hiatus” since 1998. By reanalysing the possible errors in the multiple types of data used for their calculations, a secondary revision found that with the appropriate corrections, the “hiatus” disappears, and the decadal trend of warming seen since 2000 has now become the same as that seen since 1950, when these things started to be monitored. This rate is 0.116C+-0.067 increase per decade for the land and sea of the entire planet since 1950. This is not including the warmer that occurs in the Arctic as coverage here is limited. However, when adding non-linearly interpolated data, based on what does exist, this would shift the current decadal pattern up by 0.02 – 0.03C per decade as the Arctic is warming much faster than the rest of the planet. Interestingly, the average increase in South Africa’s coastal temperature as seen from all of the in situ data available to us is very similar at 0.119C+-1.198, which yes, is ridiculous. This wild standard deviation comes from the very strong negative (-7.234) and positive (3.6) values detected by the shorter time series. Much thought and care is going into how to appropriately manage this issue.