Docs: output verification in examples + document fixed_season (#128)

[Sammyjoseph999]

Addresses the website feedback in #128.

Changes

1–3. Examples now show their output. The quickstart and API "Fetching data" examples ran without any confirmation of what was fetched. Added df.shape / df.columns / df.head() so users see the returned DataFrame, and switched them to the credential-free nasa_power path so they run as-is.

2. Seasonal output as a table. docs/getting_started.md "First analysis" now tabulates analyze_climate_statistics results (year / rain / NDWS / WRSI) with pandas instead of showing a mixed dict, and drops the Earth-Engine evaluate_hazards call so the whole section is credential-free.

4. fixed_season documented — and the warning fixed. The reporter noted the "season counts differ" warning tells you to use --fixed-season but there’s "no documented way to pass this parameter." In fact fixed_season already exists in analyze_climate_statistics (it takes a "MM-DD:MM-DD" string, not True). So:

• The Getting-started page now shows fixed_season="03-01:05-31" with a note on why/when to use it.
• The warning message itself (climate_statistics/statistics.py) now points to fixed_season="MM-DD:MM-DD" (Python API) in addition to --fixed-season (CLI), so it stops sending Python users to a CLI-only flag.

Verification

• Ran the exact Getting-started example end-to-end: fetch returns a 366-row DataFrame; the seasonal table renders (2016–2020, rain/NDWS/WRSI).
• mkdocs build --strict passes; ruff clean; full suite 598 passed (the preserved warning-prefix keeps existing assertions green).