Error cleanup

[SnowHydrology]

Issue #26 notes that the unit test values hard-coded into the PET routines don't match the values computed by the model. I found two reasons for this:

1. The pet_zulu_time var was set to be an int but it's a double
2. The assumed latitude and longitude were incorrect

Additionally, issue #27 found an uninitialized parameter.

Both issues are now fixed.

Additions

• NONE

Removals

• NONE

Changes

• src/pet.c Updated print statements with correct values for lat & lon
• include/pet_tools.h Changed pet_zulu_time from int to double
• src/bmi_pet.c Updated to include heat_transfer_roughness_length_m initialization

Testing

1. Ran ./make_and_run_read_forcings.sh

Closes #26 and #27

[madMatchstick]

@SnowHydrology - although completely unrelated to your changes here, I believe we need to incorporate the -DNGEN_IS_MAIN_PROJECT=ON flag to NOM lib build in failed Ngen Integration Tests workflow. Not completely sure if this is yaml-speak and/or cmake-speak. This change will apply to a couple other formulation repo workflows as well (topmodel, NOM, etc.) Looking into...

Also - tagging ngen issue #761 for ref