User Guide¶
This guide covers PyFIA's core concepts and workflows for FIA data analysis.
Overview¶
PyFIA implements design-based estimation following the methodology described in Bechtold & Patterson (2005). This ensures statistically valid estimates that match official USFS results.
Key Concepts¶
EVALID (Evaluation Identifier)¶
FIA data is organized by evaluation groups. Each EVALID represents a complete inventory cycle for a state or region. PyFIA automatically selects the most recent evaluation when you filter by state.
from pyfia import download, FIA
db_path = download("GA") # Download Georgia data
db = FIA(db_path)
db.clip_by_state("GA") # Auto-selects most recent EVALID
Domain Filtering¶
Domains define the population of interest:
- Land Type: Forest land, timberland, or all sampled land
- Tree Type: Growing stock, all live, or sawtimber
- Custom Domains: SQL-like conditions for specialized analysis
Expansion Factors¶
FIA uses a stratified sampling design. Each plot/tree has an expansion factor indicating how many similar plots/trees it represents. PyFIA handles these automatically.
Guides¶
- Downloading Data: Download FIA data directly from the DataMart
- Domain Filtering: Control which plots, conditions, and trees are included
- Grouping Results: Group by forest type, ownership, species with auto-enhanced names
- Spatial Filtering: Filter by polygon boundaries and group by polygon attributes
- Variance Estimation: Understanding uncertainty in estimates
- Lazy Evaluation: Memory-efficient workflows for large datasets