By Miles Evans
Everything failed, but it’s okay. I found out last week that a small error, committed in the very beginning of the summer, has caused my main experiment to not produce any results.
It was a very small distinction that caused all of my problems. My climate model needed to be running in a mode that allows it to compute complex atmospheric chemistry and physical interactions, but it was not. Usually, the default is to have the model in this mode, but the particular setup I used appears to be an exception. This problem caused my test and my control runs to produce the same exact output and so no useable data.
Looking back on the summer, it would have been wise to perform some diagnostics on my model output data early and actively look for problems like this. The skills and knowledge to know what diagnostics to do and how to do them come with experience and I have now compiled a mental list of a number of steps to take to ensure that I don’t make the same mistakes again.
Because of this setback, the focus of my summer research has changed course. Because my models produced valid output, I can still use that data. Fortunately, one of the other undergraduates here ran a model that is almost exactly the same as the control model I ran, but using an older version of the model. Now, my final summer poster will be focused on comparing the older and newer versions and discussing the improvements that were made.
In the end, not all is lost. I am still going to have a poster to present at the symposium on Wednesday with some interesting content. Most importantly, I think, was the amount I learned this summer and the experience I gained in this field. Certainly, future climate modeling endeavors of mine will proceed much more smoothly, now that I know what to look for and can operate a little bit more autonomously.