2
$\begingroup$

I am a student and I have a very simple and yet important question about reanalysis data. Recently, I attended a workshop where various other students discussing the use of ERA5 reanalysis data on their research and at some point, there was the question : "what does the use of reanalysis data means for your research and how this type of data can impact your results?" Nobody had a satisfying answer and now, I am trying to find papers discussing this issue.

For example, can someone use ERA5 parameters on a random forest model as explanatory variables and if so, what does it means for the results?

Apologies if it is a very basic question, I am new in this area.

$\endgroup$
2
$\begingroup$

Welcome to StackExchange SE!

I don't think I fully understand your question, but I'll try to answer it. Reanalysis data is the use of weather models and data assimilation to piece back the weather. It is a 4-dimensional dataset (Latitude, longitude, pressure levels, and time). You can do whatever you wish with that data (provided you follow the legal guidelines with using the ERA dataset), but know that it isn't perfect; it usually is at a coarse resolution, and doesn't always contain everything you may want. Often reanalysis data can be used as input into other meteorological models, such as the WRF model.

As far as a random forest model, it really depends on what your intentions and design are. Admittedly, I don't know much about random forest model or your experimental design. To answer the question of whether the data could be used as explanatory variables, sure if you have some data that isn't within the dataset. But be aware of the limitations. There are other highly used reanalysis datasets, such as the CFSR and MERRA-2 and a variety of other reanalysis datasets, but those have similar limitations too.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.