All algorithms
Reviewed analysis
Historical calibration
Compare an algorithm with multiple references and understand where it succeeds or fails.
What we need
Area, period, algorithm or outputs to evaluate, and available references. Training and evaluation periods are defined before the work.
What you receive
Reference comparisons, metrics by date and elevation, uncertainty and calibration recommendations. Code and licensing are agreed in the quote.
What the analysis can establish
More years do not guarantee more snow examples. Without valid snow references, only false alarms can be assessed. Regional results do not establish performance elsewhere.