What a week to come back to after my retreat! With all the data finally in the database and properly examined for quality before I left, this week we started analysis of the historic data Bridget and I spent so long working with. Before we could do any analysis though, there was a lot of reading to do. Over Monday and Tuesday we learned 3 species diversity measures, a new statistical analysis and started to use a new program with its own coding language neither of us had ever seen before. It was a lot of work but it has been making me feel much more like a serious researcher, as well as regret not retaining more from the stats class I took last year.
Two of the diversity measures we learned were the Shannon-Weiner Diversity Index and the Simpson's Diversity Index. Both of these indexes are calculated with abundance data including how many species are present in a sample, how many individuals of each species there are and these species being studied need to occupy the same or similar habitats. The Shannon index, represented by H' when its calculated, ranges from 0-5. It measures uncertainty in correctly predicting the species of the next hypothetical individual collected. If H'=0 all the individuals in the sample would belong to one species. H' has never been calculated higher than a 5, meaning that every individual in the sample would be unique and there would be a near equal number of individuals and species. If every individual is unique, there's a high uncertainty that one could correctly predict species of the next individual. The Simpson's index, represented by D when calculated, measures the likelihood that 2 individuals chosen at random will be part of the same species and ranges from 0-1. If D=0, then all individuals are of the same species and if D=1 all the individuals are unique.
We also learned about rarefaction and how it can be used to balance species richness across multiple sample groupings and put more weight in less represented species and take weight away from dominant species in a community.
The final statistic we learned about was MRPP (Multi-response Permutation Procedures). It's a non-parametric multi-variable statistical tool for testing the hypothesis that a group of a priori selected samples is identical to a group chosen by random chance. Before conducting an MRPP, a similarity matrix must be created from abundance and species data using a program like R, the statistics program utilizing a coding language mentioned earlier. A similarity matrix takes all the species in a sampled community and plots them in an abstract space where more taxonomically related species are closer together than distance relatives. The matrix measures their difference from each other in special units, in our case Bray-Kurtis units and plots records the differences in a matrix. Thankfully we had R to do that for us so all we had to do was type up a script importing and transforming the data until we were able to type a couple lines and the MRPP ran itself 1000 times.
Needless to say, learning complicated new statistics tests and coding was mentally exhausting, so Wednesday's trip to OIMB was a refreshing change of pace. We were able to go out in Cape Arago to the tide pools and out on the rocks near the kelp beds to see where Peter and Leyia collected their plankton samples. It was a lot of fun. We also got a nice tour of some of the labs and classrooms, including one full of stuffed seabirds, some baleen and a sea otter fur! We were also treated to an amazing cafeteria lunch of chowder, salad, cheesy garlic bread and marionberry cobbler. It was a fun day with a beautiful long drive there and back through amazing forests, dunes and over many bays.
It was a long and eventful week, but I wouldn't have it any other way. Until next week...
Thanks for reading,