updates to use time fron NGEN

[mkarim-rtx]

https://jira.nextgenwaterprediction.com/browse/NGWPC-10207

Centralizing time control via NGen across UEB, TopoFlow-Glacier, and Noah-OWP

This task standardizes how simulation time (start, end, timestep) is handled across multiple models—UEB, TopoFlow-Glacier, and Noah-OWP-Modular—by shifting control from model-specific configuration files to the NGen Framework.

Objective

Ensure all models:

Ignore hardcoded config time (start/end/dt)
Use NGen-provided realization time dynamically via BMI
Maintain compatibility with existing workflows with minimal code changes

What was done

1. NGen (Core Change)
   Added logic to pass realization time through BMI inputs:
   start_utime
   end_utime
   dt_seconds
   Ensures a single source of truth for time control
   Applies uniformly across all BMI models

PR: NGWPC/ngen#172

2. UEB
   Replaced reliance on config timing with runtime override
   Implemented:
   Temporary rewrite of UEB control file before Initialize()
   Injected NGen time into rewritten file
   Result:
   UEB runs using NGen time without modifying original config
   Zero changes required in UEB internal physics

PR: NGWPC/ueb_bmi#14

3. TopoFlow-Glacier
   Added full BMI-level time handling:
   adapter_set_realization_times()
   _apply_realization_time_from_strings()
   Logic:
   Prefer NGen time if provided
   Fallback to config only if missing
   Result:
   Fully decoupled from YAML time configuration
   Works seamlessly with NGen forcing window

4. Noah-OWP
   Removed dependency on internal RunModule time handling
   Implemented:
   BMI-side apply_realization_time_config logic
   Updated NamelistRead.f90:
   Made startdate and enddate optional
   Added placeholder defaults instead of fatal error

Result:
Model initializes without config time
Real time applied later via BMI inputs from NGen
Why this change was needed

Before
Each model:
Required its own config-based time
Could conflict with NGen realization
Caused errors or inconsistencies

After
NGen controls time globally
Models become:
Consistent
Easier to integrate
Less error-prone

Final Outcome
All three models now:
Use NGen-driven time
Do not require config start/end dates
Maintain backward compatibility via fallbacks
Eliminates duplicated logic and synchronization issues across models

[jswade-rtx]

With Mohammed's updates to noah-owp-modular and ngen, I was able to run a noah-owp formulation after removing the start and end time from the bmi config file. I also successfully ran a calibration + validation + forecast workflow to demonstrate the time handing updates to the forcing configs.