I have greatly underestimated the complexity of numerical weather prediction (NWP). I took it as a crude guess, and therefore not very difficult to understand. Given grid sizes of tens of kilometers, I did not anticipate the models to give any accurate picture of the weather. However, this is not the case. The amount of detail that goes into the different aspects of these models make them much more accurate than I originally imagined.
One example of this is how different models handle precipitation: some models create precipitation and have it fall instantly. Others allow it to fall over time. This change in mass in certain grids, and how the falling precipitation alters the environment around it (does it evaporate and cool the air?) affects the grid squares around it.
This has also exposed some fundamental flaws in my forecast understanding. I really need to work on fully understanding upper air soundings. I feel like I have “gotten away with” not fully understanding some of the details, and now I am finding that affecting me in NWP.
I’ll let you know how it goes over the next week.
Thank you for reading my post