MoreCast 2.0: Add machine learning to select and bias correct weather models

**As a** weather forecaster

**I need** a machine learning feature to select the best performing weather models and apply a bias correction

**So That** I can quickly produce an accurate weather forecast.

**Acceptance Criteria**
- [ ] We want a biased adjusted precip/windspeed/wind direction for GDPS/HRDPS/RDPS models

**Additional Context**
- WF1 times are on PST solar noon
- Note that parts of eastern BC are on Mountain time. Will need to confirm how forecasts are entered into WildfireOne for these time zones because 12:00 PDT is off by one hour if 12:00 MDT is what is needed.

**Definition of Done**
https://github.com/bcgov/wps/wiki/Definition-of-Done

Start by reviewing the acceptance criteria and the linked Definition of Done, then locate the existing forecast-model selection and weather-data entry points. Define the scope for GDPS, HRDPS, and RDPS bias adjustment, including the WF1 and BC time-zone questions; done should be demonstrated against those criteria.

Written by the indexing model from the issue text.