Hello meteorologists and atmospheric scientists around the globe,
The day is finally here. After months, and honestly years, of development, I am officially finished building the full suite of Python charting scripts I set out to create for NSF NCAR’s Weather Research and Forecasting model, WRF.
Over the years I have been developing Python tools that highlight WRF’s capabilities for operational forecasting and real-world analysis. These scripts were built not only to showcase WRF, but also to be shared with the community for free, especially for meteorologists and atmospheric scientists who may not have much Python experience but still want access to high-resolution graphics, high automation and batch processing, and a wide range of meteorological diagnostics and products. Every script has been standardized across the collection, so the workflow, structure, and commenting style are consistent, making it easier to understand what each script is doing, adapt it to your region, and modify it for your own needs.
In the Google Drive link and on my GitHub page, you will find a folder called “Python Charts that work World.” Inside that folder you will find 96 Python scripts and 5 Bash shell scripts that run the full suite of graphics, including multicore execution options. The products span synoptic and mesoscale dynamics, severe weather diagnostics, and simulated remote sensing, all based directly on WRF output. The goal is to make a complete, automated, operational-ready post-processing package that takes raw WRF output and turns it into briefing-quality graphics with minimal user intervention.
GitHub profile: https://github.com/HathewayWill/ WRF Python Scripts repo: https://github.com/HathewayWill/WRF_Python_Scripts
These scripts are designed to work with the WRF-MOSIT Conda environment using wrf-python, but they can also be used outside of Conda if you install the required Python modules:
numpy, matplotlib, metpy, wrf-python, netCDF4, pillow, scipy, cartopy, geopandas
Each Python script can work with any WRF supported map projection, including Polar, Lat-Lon, Mercator, and Lambert. Each script is also designed to handle multiple WRF output patterns, including a single wrfout file, a single file containing multiple times, or multiple files containing multiple times. The intent is that you should not have to reorganize your output files into a special format to make the scripts work, the scripts are meant to adapt to how WRF is commonly run in real workflows.
A major focus of this project was making the maps operationally useful the moment they are generated, without requiring manual GIS setup, shapefile hunting, or post-editing. Across the gridded map scripts you will see a consistent approach to adding geographic context through Cartopy features, plus automated city labeling for quick situational awareness, so the output is immediately interpretable in a forecasting or briefing environment.
Many scripts include common Cartopy features so the meteorological fields remain anchored to real geography. Depending on the domain and product, this can include:
- Coastlines, with resolution appropriate for the map scale
- International borders and state or province boundaries, when relevant
- Lakes, rivers, and other water bodies for orientation
- Land and ocean shading or masking for clean contrast between continents and water
- Gridlines with labeled latitude and longitude ticks, formatted consistently for readability
- Optional terrain or topographic context where it matters, such as streamline and terrain plots
Geographic elements were standardized across the collection so line weights, label sizes, and background features remain consistent from plot to plot. The goal is readability, geographic context supports the meteorological fields without overpowering them, and the overall look and feel stays familiar whether you are viewing upper-air diagnostics, surface fields, cloud products, or precipitation maps.
Many scripts include city labels to make the output immediately useful for operations and briefings. Labels are intended to support quick orientation and readability in typical WRF domains, often including major cities, airports, and other commonly referenced points, with placement chosen to reduce overlap with filled contours and line contours whenever possible.
Because the scripts support Polar, Lat-Lon, Mercator, and Lambert projections, the geographic overlays and labeling logic are built to work anywhere in the world. The intent is that you can run the same workflow whether your domain is over North America, Europe, the Middle East, the tropics, or any other region, and still get a map that is geographically meaningful and presentation-ready.
You will notice that some scripts have similar names. That is intentional. I developed both Metric and Imperial versions, so the same diagnostic can be produced in the most common unit sets used around the world. In many cases this means paired scripts like millimeters versus inches, kilometers versus miles, meters versus feet, Celsius versus Fahrenheit, knots versus miles per hour. The goal was for a forecaster or researcher to run the exact same suite regardless of location, without having to rewrite unit conversions or maintain separate code bases.
Description: Pressure-level maps designed for diagnosing jet structure, trough and ridge placement, forcing for ascent, moisture transport, and overall synoptic and mesoscale evolution. These include classic operational analysis fields adapted for WRF output, produced automatically. Many of these plots combine multiple fields in a single graphic, for example height with isotachs and wind, or humidity with thickness and sea-level pressure, so the output supports real forecasting decisions rather than showing a single field in isolation.
- 1000hpa_equiv_temp_k_pressure_wind_speed_dir.py
- 250hpa_wind_height_isotachs.py
- 300hpa_wind_height_isotachs.py
- 500hpa_vorticity_wind_pressure.py
- 500hpa_wind_height_isotachs.py
- 700hpa_relative_humidity_slp_thickness.py
- 700hpa_wind_height_isotachs.py
- 850hpa_frontogenesis.py
- 850hpa_qvector_divergence_wind_pressure.py
- 850hpa_temp_advection_height_wind_speed_dir.py
- 850hpa_temp_degc_height_wind_speed_dir.py
- 850hpa_wind_height_isotachs.py
- 925hpa_temp_degc_height_wind_speed_dir.py
- 925hpa_wind_height_isotachs.py
Description: Products focused on moisture availability and cloud field structure for aviation forecasting, convective initiation context, and general cloud interpretation. These scripts are designed to quickly identify deep moisture plumes, saturation patterns, cloud field coverage by layer, and cloud-top thermal structure.
- cloud_frac_high_feet.py
- cloud_frac_high_meters.py
- cloud_frac_low_feet.py
- cloud_frac_low_meters.py
- cloud_frac_mid_feet.py
- cloud_frac_mid_meters.py
- cloud_top_temperature.py
- cloud_top_temperature_rainbow.py
- precipitable_water_cm.py
- precipitable_water_inch.py
Description: Stability and thermodynamic diagnostics intended to summarize the convective environment and changes in instability, inhibition, and parcel buoyancy. These are the types of fields forecasters commonly use to assess convective initiation potential, severe weather environment quality, and evolving mesoscale instability gradients.
- convective_cape_cin.py
Description: Surface and near-surface products designed for operational forecasting, situational awareness, and briefing graphics. Many include sea-level pressure and wind overlays while emphasizing a target surface parameter. These scripts cover classic surface analysis variables plus human-impact indices commonly used in operations, including heat and cold stress metrics.
- surface_dewpoint_degc_slp_wind_speed_dir.py
- surface_dewpoint_degf_slp_wind_speed_dir.py
- surface_heatindex_degc_slp_wind_speed_dir.py
- surface_heatindex_degf_slp_wind_speed_dir.py
- surface_humidex_degc_slp_wind_speed_dir.py
- surface_relative_humidity_slp_wind_speed_dir.py
- surface_slp_wind_gust_speed_knots_direction.py
- surface_slp_wind_gust_speed_mph_direction.py
- surface_streamlines_terrain_ft.py
- surface_streamlines_terrain_m.py
- surface_temp_degc_slp_wind_speed_dir.py
- surface_temp_degf_slp_wind_speed_dir.py
- surface_terrain_ft_slp_wind_speed_dir.py
- surface_terrain_m_slp_wind_speed_dir.py
- surface_thi_degc_slp_wind_speed_dir.py
- surface_visibility_km.py
- surface_visibility_miles.py
- surface_windchill_degc_slp_wind_speed_dir.py
- surface_windchill_degf_slp_wind_speed_dir.py
- road_icing_index.py
Description: Time-accumulated precipitation and winter-weather products at operationally useful intervals. These cover liquid precipitation, snowfall, and snow water equivalent across hourly, multi-hour, daily, and total accumulation windows. These products are meant to support forecasting of rainfall rates, storm totals, winter storm impacts, and hydrologic context, with Metric and Imperial variants built in.
- surface_1hr_precip_inch_slp_isotherm.py
- surface_1hr_precip_mm_slp_isotherm.py
- surface_1hr_snow_inch_slp_isotherm.py
- surface_1hr_snow_mm_slp_isotherm.py
- surface_1hr_water_equivalent_snow_inch_slp_isotherm.py
- surface_1hr_water_equivalent_snow_mm_slp_isotherm.py
- surface_24hr_precip_inch.py
- surface_24hr_precip_mm.py
- surface_24hr_snow_inch.py
- surface_24hr_snow_mm.py
- surface_24hr_water_equivalent_snow_inch.py
- surface_24hr_water_equivalent_snow_mm.py
- surface_3hr_precip_inch.py
- surface_3hr_precip_mm.py
- surface_3hr_snow_inch.py
- surface_3hr_snow_mm.py
- surface_3hr_water_equivalent_snow_inch.py
- surface_3hr_water_equivalent_snow_mm.py
- surface_daily_precip_inch.py
- surface_daily_precip_mm.py
- surface_daily_snow_inch.py
- surface_daily_snow_mm.py
- surface_daily_water_equivalent_snow_inch.py
- surface_daily_water_equivalent_snow_mm.py
- surface_total_precip_inch.py
- surface_total_precip_mm.py
- surface_total_snow_inch.py
- surface_total_snow_mm.py
- surface_total_water_equivalent_snow_depth_inch.py
- surface_total_water_equivalent_snow_depth_mm.py
Description: Simulated observational-style products intended to provide a radar-like view of convective structure and precipitation intensity derived from WRF fields. This is meant to help bridge the gap between model output and what forecasters are used to interpreting in real time.
- surface_simulated_radar_reflectivity.py
Description: Fire weather focused diagnostics designed to highlight atmospheric stability and dryness patterns commonly used in wildfire forecasting and fire behavior discussions. These provide quick environmental context for plume-dominated fire potential and stability regimes.
- haines_index.py
- c_haines_index.py
Description: Site-specific and vertical diagnostic plots for stations or user-defined latitude and longitude points. These complement the gridded maps with time series, thermodynamic profiles, and vertical structure views. They are meant for forecasting at specific airports, cities, incident locations, research sites, and decision support points.
- meteogram_degc.py
- meteogram_degf.py
- skewt_diagram.py
- enhanced_skewt_diagram.py
- vertical_wind_profile.py
- vertical_wind_profile_4km.py
- vertical_wind_relative_humidity_cross_section_height_pressure.py
- vertical_wind_speed_cross_section_height_pressure.py
Description: Tropical-focused surface products emphasizing sea-level pressure, winds, gusts, and sea surface temperature, with Metric and Imperial options. These are designed to support tropical cyclone monitoring, trade wind regime diagnosis, and tropical marine forecasting workflows.
- tropical_surface_slp_wind_gust_speed_knots_direction.py
- tropical_surface_slp_wind_gust_speed_mph_direction.py
- tropical_surface_slp_wind_speed_knots_direction.py
- tropical_surface_slp_wind_speed_mph_direction.py
- tropical_surface_sst_degc_slp_wind_speed_dir.py
- tropical_surface_sst_degf_slp_wind_speed_dir.py
- tropical_hurricane_track.py
Description: WRF-native air parcel trajectory tools designed to calculate back trajectories, forward trajectories, and paired back/forward trajectories directly from WRF output. These scripts use WRF model fields directly rather than requiring an external trajectory package. They are intended for source-region analysis, downstream transport analysis, boundary-layer pathway review, and research workflows where the trajectory calculation should remain tied to the same WRF simulation being analyzed.
These scripts read raw WRF U, V, and W fields, destagger and rotate winds into earth-relative flow, use WRF vertical velocity for passive 3D parcel motion, and automatically support nested-domain fallback. For example, if a parcel starts in d03 and exits that nest, the script continues the trajectory using d02, then d01 if needed. This allows inner-domain trajectory calculations to continue across the full model pathway without stopping at the nest boundary.
The default trajectory timestep is 60 minutes, which keeps the output cleaner and works well for standard hourly WRF history output. Users can still choose a finer timestep with --dt-min, such as --dt-min 15, when higher-frequency WRF output is available or when a sensitivity test is needed.
- WRF_Back_Trajectory.py
- WRF_Forward_Trajectory.py
- WRF_Back_Forward_Trajectory.py
The default academic workflow uses:
- Raw WRF U and V wind fields, destaggered and rotated to earth-relative flow
- WRF W for passive 3D vertical parcel motion
- Model-level lower-boundary clipping and continuation when a parcel reaches the lowest usable WRF mass level
- File-aware timing, where the scripts automatically detect available WRF output times
- Nested-domain fallback, such as d03 → d02 → d01
- Cartopy mapping with a vertical parcel-path panel below the map
- CSV output for each trajectory and combined CSV output when multiple heights or locations are used
python3 WRF_Back_Trajectory.py \
d03 SanMarcos \
29.8899 -97.9961 \
--wrf-dir /home/workhorse/WRF_Intel/WRF-4.7.1/run \
--height-levels-m 100,500,1000This script starts from the latest available WRF output time by default and traces parcels backward through the available model period.
python3 WRF_Forward_Trajectory.py \
d03 SanMarcos \
29.8899 -97.9961 \
--wrf-dir /home/workhorse/WRF_Intel/WRF-4.7.1/run \
--height-levels-m 100,500,1000This script starts from the first available WRF output time by default and traces parcels forward through the available model period.
python3 WRF_Back_Forward_Trajectory.py \
d03 SanMarcos \
29.8899 -97.9961 \
--wrf-dir /home/workhorse/WRF_Intel/WRF-4.7.1/run \
--height-levels-m 100,500,1000This script uses the middle of the available WRF period by default, traces parcels backward to the beginning of the model period, and traces parcels forward to the end of the model period. In the combined CSV, negative age_hours values are backward in time, 0 is the launch time, and positive age_hours values are forward in time.
Each trajectory script supports additional launch locations and multiple starting heights:
python3 WRF_Back_Trajectory.py \
d03 SanMarcos \
29.8899 -97.9961 \
--wrf-dir /home/workhorse/WRF_Intel/WRF-4.7.1/run \
--height-levels-m 50,100,250,500,750,1000,1500,2000,2500,3000,4000,5000 \
--extra-location Austin,30.2672,-97.7431 \
--extra-location CorpusChristi,27.8006,-97.3964
You can also use a CSV file with city, lat, and lon columns through the --locations-file option.
Typical trajectory outputs include:
- One CSV file per launch height and location
- One combined CSV file when multiple heights or locations are used
- One PNG image with the map and vertical parcel-path panel
- Domain-tracking columns such as
wrf_domain_used - Parcel-height columns such as
height_agl_m,height_agl_used_m, andheight_msl_m - Wind columns such as
u_east_mps,v_north_mps, andw_mps
The five Bash shell scripts are designed to automate execution. They can activate the appropriate environment, locate or accept a WRF output directory, create date-stamped output folders, and run the chart suite in a structured way across domains, including multicore parallel processing where configured. The trajectory scripts can also be run directly from the command line because they include their own file-aware WRF output discovery and timing logic. The intent is that you can generate an entire forecast graphics package in one run, organized by domain and location, without manually running dozens of individual commands.
The purpose of this project is simple, make high-quality WRF post-processing graphics easier for the broader community to use, including those who have never written a Python script. If you are already comfortable in Python, you can also treat this as a base framework, the scripts are standardized so you can add additional diagnostics in a consistent way, and keep your own additions aligned with the rest of the suite.