Time Evolution of a PhD Student’s Mood—A Review of My Time as PhD Student of the Research School ESSReS - Towards an Interdisciplinary Approach in Earth System Science - page 18

obvious that reaching the final goals in the three years period was already

improbable. So I started to underline the importance of an alternative “ Plan B ”

already during the first ESSReS PhD committee meeting. After an intense dis-

cussion, all committee members shared the opinion that Plan B has to be set up
while I still try to push the initial .

Luckily, Plan B was established quite quickly. Our partners from the Swedish
Institute of Space Physics gave us willingly access to their set of already measured
CO radiation data. This dataset was obtained with the same observation technique,
but scienti

“ Plan A ”

cally complementary to the one I was going to measure. This allowed
me to breakup the consecutive order of the first two steps of my original plan: I

could gain experience with the retrieval before having finished the work on my own

instrument. In the best case, I could combine both datasets in the analysis, otherwise
I would completely rely on the Plan B data.
But

firstly, I kept on pursuing Plan A for a while. After some more time, the
missing component was finally delivered (2). However, the following tests indi-

cated that it did not work stable and it might take again several months for the
supplier to solve the problems. Of course, this kind of trouble has to be expected if
one tries to go beyond the limitations of state-of-the-art technique. But for me it was
very disappointing and I seriously felt the tension between the uncertainties of
scienti ll formal requirements like the 3-year project

timeframe in this case. So, my optimism was damped again (3) until I really started
to work on Plan B, simultaneously (4).
Some months later, we had one more try with the re-engineered component (5),
but it became clear very quickly that the problem remained and that any solution
will come too late for my project. So, during the second committee meeting, we
decided that I will switch entirely to Plan B and thus skip the technical part
completely. While this was of course disappointing at that moment, I was lucky,
since the work on Plan B was already in progress and allowed for a seamless
transition into the next part of my project.
To keep it short, I started with refreshed motivation into the second phase (6),
but had to face new disappointments after a few months. I made progress in the
numerical data retrieval and got

c work and the need to ful

first results. But these results were not only
inconsistent with previous observations, moreover there were also inconsistencies
within my dataset itself. At this point it became highly questionable if the dataset is
applicable at all. So I spent several months (7) on finding possible explanations in

every direction: de ciencies of the retrieval technique, calibration problems of the

instrument, eventually a new atmospheric feature, or simply my own inability? So, I
had created dozens of pieces of a puzzle, but no idea how to put them together—if — if

they fit together at all. After about two years of my PhD, apparently nothing was

achieved and I felt that it might be better to give up.
Luckily, I had another iron in the

fire. During my research in the months before,
I came across an interesting atmospheric model and I had the idea to replace the
missing technical part of my work with some modeling experiments. With the help
of my PhD committee members, I came into touch with the model developers at the
National Center for Atmospheric Research (NCAR) in Boulder, USA, and we