Change parameter name passed to read_grid in verify_spatial

[j-fannon]

As suggested by @andrew-MET. Solves problem when using "parameter = AccPcp3h" in verify_spatial.

[adeckmyn]

I'm not sure whether this might break for some file types. Some files might actually contain accumulated precipitation for some fixed accumulation time[s], and passing only $basename strips all accumulation information. But I'm having a closer look. For instance, at RMI, we store output as tar'ed FA files, and a single file contains all lead times, so read_grid can extract the correct accumulation. That is (usually) not the case with GRIB output, so there indeed something needs to fixed.

[j-fannon]

Thanks @adeckmyn, feel free to close if this is not suitable.

[andrew-MET]

There's certainly a problem here - especially given that a large number of institues are using grib files or FA files for single times. I think the cleanest solution would be to extract the precipitation from read_grid() as is and do any decumulation outside of the function - that way all bases are covered. For the edge cases where the precipitation has already been decumulated to the time periods you want you could have a force_param_name argument whereby fullname is passed to read_grid() rather than basename and no decumulation is done.

Ultimately I think harpSpatial should be using read_forecast() and read_analysis() and things like decumulation / accumulation would be handled by those functions. There's a lot of duplication of functionalities of these functions (e.g. file name generation, checking if files exist) in verify_spatial().