modverif

Model verification for multidimensional gridded and station data, following MET/METplus standards.

Overview

modverif evaluates model output by comparing gridded datasets or comparing model grids against station observations. It computes standard verification metrics, produces publication-quality plots, and handles cyclone tracking and storm composite visualization. All data I/O uses the cfdb format with CF conventions.

Evaluation Types

Type	Class	Description
Grid-to-Grid	Evaluator	Compare two gridded datasets cell-by-cell or domain-aggregated. Supports FSS, wind vectors, and diurnal analysis.
Grid-to-Point	StationEvaluator	Compare gridded model output to weather station observations with automatic interpolation and height matching.
Cyclone	track_cyclone + compare_cyclone_tracks	Track cyclones independently in two datasets and compare depth, timing and track separation. evaluate_cyclones additionally computes per-variable metrics inside the cyclone region.

Key Features

• Cell-level and domain-aggregated metrics -- NE, ANE, RMSE, Bias, MAE, Pearson correlation, POD, FAR, CSI, GSS, Frequency Bias
• Fractions Skill Score (FSS) -- multi-scale spatial verification for precipitation and threshold-based fields
• Vector wind evaluation -- Vector RMSE, speed bias, direction bias from U/V components
• Diurnal cycle analysis -- metrics grouped by hour-of-day with UTC offset support
• Station evaluation -- automatic grid-to-point interpolation, per-station and aggregate metrics, lagged correlation
• Cyclone tracking -- SLP-based tracking with radius estimation, an optional time window, and support for projected (y/x + CRS) model grids
• Cyclone comparison -- pressure bias, timing offset and track separation between two independently tracked storms, plus per-variable evaluation within the cyclone region
• Storm composite plots -- PWAT/MSLP/VIMF overlays with side-by-side model comparison and animated WebP output
• Verification plots -- scatter, station map, time series, performance diagram, Taylor diagram, FSS, wind rose
• Spatial and temporal filtering -- bounding box, boolean mask, start/end time
• All I/O via cfdb -- consistent CF-conventions database format for inputs and outputs