Honestly because it's easier. Modern scientific computing is actually dominated by C, which powers a lot of the libraries that do the crunching behind the scene

Honestly because it's easier. Modern scientific computing is actually dominated by C, which powers a lot of the libraries that do the crunching behind the scenes. Python has some fantastic connectors that can bind to C libraries and run computationally-intensive code very close to the machine. This makes the programming behind a lot of scientific computing look more like loading a library and running it with different input variables. This way grad students who are doing computational science but who do not have a background in computer science can get running fairly quickly.