Choose the forecast
Select a station from the searchable map, then choose a model and run. Use the forecast-hour slider or the left and right arrow keys to move through valid times.
Penn State Meteorology
Sounding Viewer User Guide
How to explore forecast soundings, compare guidance, and understand the calculations shown in the viewer.
Font Size
Back to viewer
Using the Sounding App
Explore the plots
Use the Skew-T toggles for moisture, the dendritic growth zone, wet bulb, frost point, and parcel traces. Axis controls adjust the displayed ranges. Hover over profiles, meteograms, and inset plots for exact values.
Compare and export
Add model or run lines in comparison mode and select a line from either chart to make it active everywhere. Export buttons download the visible plot as PNG; Export JSON downloads the selected dataset.
Viewer modes
Standard
View current deterministic model guidance for one station, model, and run.
Archive
Load historical deterministic guidance by selecting a UTC date and cycle hour.
Model Comparison
Overlay models or runs at a shared valid time and compare their meteogram traces.
Ensemble
Explore REPS members plus mean and percentile summaries for forecast uncertainty.
Calculation Reference
Parcel definitions and diagnostic conventions are aligned with SHARPpy, using an independent implementation and virtual-temperature-corrected buoyancy. See SHARPpy parcelx and the recommended citation: Blumberg et al. (2017), doi:10.1175/BAMS-D-15-00309.1.
| Parameter | Formula or algorithm | Use | Applied in | Source |
|---|---|---|---|---|
| SB parcel | Initial p, T, Td are taken from the lowest valid thermodynamic level. | Represents immediately surface-based convection. | Skew-T trace; SB diagnostics | App-computed; SHARPpy-aligned |
| 100 hPa ML parcel | Pressure-weighted mean potential temperature and mixing ratio through the lowest 100 hPa, restored to surface pressure. | Reduces sensitivity to a shallow surface anomaly. | Skew-T trace; ML diagnostics | App-computed; SHARPpy-aligned |
| 300 hPa MU parcel | The maximum Bolton equivalent potential temperature on a 1 hPa grid in the lowest 300 hPa selects the source level. | Finds elevated or surface instability. | Skew-T trace; MU diagnostics | App-computed; SHARPpy-aligned |
| LCL | TLCL = [1/(1/(Td−56) + ln(T/Td)/800)] + 56; pressure follows the dry Poisson relation. | Approximate cloud base of a lifted parcel. | Parcel markers and derived panel | App-computed; Bolton (1980) |
| Parcel trace | Dry Poisson ascent below the LCL; pseudoadiabatic saturated ascent follows environmental pressure levels above it, with exact source and LCL points retained. | Compares lifted parcel temperature with the environment. | Skew-T SB/ML/MU overlays | App-computed; SHARPpy-aligned level sampling |
| CAPE / CIN | B = g(Tv,p−Tv,e)/Tv,e; trapezoidal height integration. Positive energy is CAPE; negative energy above 500 hPa is excluded from CIN. | Estimates updraft energy and inhibition. | Derived Parameters | App-computed; SHARPpy convention |
| LFC / EL | Pressure-interpolated zero crossings of virtual-temperature buoyancy after the LCL. LFC begins positive buoyancy; EL ends the first positive layer. | Bounds the primary freely convecting layer. | Parcel markers and derived panel | App-computed |
| Potential temperature θ | θ = (T + 273.15)(1000/p)Rd/cp. | Compares temperature across pressure levels and diagnoses stability. | Time-height; ML parcel selection | App-computed for ECMWF; BUFKIT may supply profile values |
| 850 mb temperature | Read directly at 850 mb when present; otherwise linearly interpolate temperature between the surrounding BUFKIT sounding levels in log-pressure coordinates. | Tracks lower-tropospheric thermal changes even when native model levels do not fall exactly at 850 mb. | MSLP/RH/850 mb temperature meteogram | App-parsed/interpolated from the BUFKIT profile as surfaceTimeSeries T850; external-model profiles provide their 850 mb level |
| Equivalent potential temperature θe | Bolton (1980) approximation using pressure, temperature, dew point, mixing ratio, and LCL temperature. | Combines heat and moisture; MU selection and θe inset. | MU parcel; side-panel inset | App-computed for parcels/ECMWF; BUFKIT THTE otherwise |
| Saturation vapor pressure | Liquid: 6.112 exp(17.62t/(243.12+t)); ice: 6.112 exp(22.46t/(272.62+t)) hPa. | Converts temperature/dew point/frost point to vapor pressure. | RH bars, time-height RH, DGZ moisture | App-computed; WMO/CIMO Magnus |
| Frost point | Compute e from liquid-water dew point, then invert the ice Magnus equation: Tf = 272.62L/(22.46−L), L = ln(e/6.112). | Canonical moisture coordinate for subfreezing RH products. | Frost-point overlay; RH calculations | App-computed |
| RHi / RHw | 100esi(Tf)/esi(T), or /esw(T) for RHw. Values are not capped. | Diagnoses saturation relative to ice or liquid. | RH bars; time-height; surface meteogram; DGZ stats | App-computed |
| Wet-bulb temperature | BUFKIT TMWC when supplied; ECMWF uses the Stull (2011) empirical T/RH approximation and prevents Tw > T. | Supports precipitation-type and evaporative-cooling assessment. | Optional Skew-T overlay | Mixed: upstream/app-computed |
| Precipitation type | The modified Bourgouin/Birk et al. (2021) method diagnoses cloud-ice probability and wet-bulb melting/refreezing energies. Only complete profile levels with pressure ≥100 mb are considered. REPS runs the method independently on every perturbed member and control. Its mean averages available member diagnostics; each P10/P25/P50/P75/P90 product uses the complete p-type result from the real perturbed member with the lowest pressure-thickness-weighted temperature RMSE from that percentile temperature profile. | Estimates rain, snow, freezing-rain, and ice-pellet probabilities. The representative-member percentile method preserves a physically coherent member profile instead of taking independent percentiles of phase probabilities. | Precipitation-type panel and deterministic/ensemble comparisons | App-computed from model temperature, dew point, wet-bulb temperature, height, and pressure. REPS representative selection excludes the control, requires at least two common above-ground temperature levels, and may choose a different member each forecast hour. Stored payloads require reingestion. |
| DGZ layer | Contiguous −18 to −12 °C segments between 0.75 and 12 km AGL, minimum 0.5 km deep. Primary layer favors depth, then RH, then ascent; statistics are depth-weighted. | Highlights temperatures favorable for dendritic snow growth. | Skew-T DGZ overlay and tooltip | App-computed |
| Wind components/speed | u = −V sin(direction), v = −V cos(direction); V = √(u²+v²). m s⁻¹ converts to knots with 1.94384449. | Supports hodographs, barbs, and surface wind. | Skew-T, hodograph, meteogram | App conversion from model/BUFKIT winds |
| Hodograph kinematics | Profile winds are sorted by AGL height, surface is held from the lowest valid wind when needed, and layer endpoints are linearly interpolated by height. Bunkers motions use 0-6 km mean wind, 0-500 m and 5.5-6 km mean-wind shear direction, and a 7.5 m s⁻¹ right/left deviation. Bulk shear uses surface-to-top vectors for 0-1, 0-3, 0-6, and 0-8 km. SRH and layer-mean streamwise vorticity follow the supplied Karl reference code for 0-1 and 0-3 km SRH and 0-1 km streamwise diagnostics. | Diagnoses storm motion, storm-relative curvature, low-level rotation, and deep-layer organization. | Full-size hodograph modal | App-computed; Bunkers et al. convention; Karl reference implementation; SHARPpy-style presentation |
| Precipitation | Millimetres convert to inches by division by 25.4. Meteogram bars display each available P01M, P03M, or P06M forecast-interval accumulation, and the total line is their running sum. | Displays forecast-interval and storm-total precipitation. Most-probable precipitation type is plotted only for intervals with at least 0.01 inches. | Surface snapshot; meteograms and model comparisons | App conversion/aggregation of upstream data |
| Cloud cover | BUFKIT total cloud is min(100, LCLD + MCLD + HCLD), after bounding each layer to 0–100%. ECMWF and RDPS total cloud cover is supplied directly by the model and normalized to percent. | Shows the fraction of sky covered by cloud; BUFKIT layer overlap is intentionally approximated by a summed-and-capped total. | Cloud-cover meteogram using TCLD, HCLD, MCLD, and LCLD | BUFKIT total is app-computed; layer values and external-model totals are upstream. Existing stored runs require raw-file backfill or reingestion. |
| PWAT | BUFKIT PWAT is converted from mm to inches; its upstream vertical integration is not recomputed here. | Measures total-column water vapor. | Derived Parameters | Upstream BUFKIT value |
| REPS mean and percentiles | At each available above-ground isobaric level and valid time, the app applies a missing-value-aware arithmetic mean or NumPy linear P10/P25/P50/P75/P90 independently to temperature, specific humidity, geopotential height, and U/V wind components. A separate synthetic surface level aggregates member surface pressure and 2 m thermodynamics. Surface parameters use the same fieldwise method. Contributing-member counts are retained for every field. | Summarizes the 20 perturbed members without averaging wind direction across the 0/360° discontinuity. | REPS ensemble profiles and meteograms | App-computed from upstream MSC member fields; units remain those of the deterministic payload fields. |
| REPS terrain mask | MSC HGT_SFC supplies model terrain. The surface profile level is fixed at 0 m AGL and uses surface pressure with 2 m temperature and RH-derived dew point. An isobaric level is excluded for a member when its pressure is at or above member surface pressure or its geopotential height is at or below model terrain. Summary levels aggregate the remaining above-ground contributors. | Prevents underground temperature, moisture, height, and wind values while keeping the displayed surface tied to 2 m thermodynamics. | All REPS member, control, mean, and percentile profiles | Terrain and atmospheric fields are upstream MSC values; masking and AGL conversion are app-computed. Existing REPS payloads require reingestion. |
| REPS synthetic-profile diagnostics | Dew point is derived from aggregated specific humidity and pressure; wind direction/speed are reconstructed from aggregated U/V. Parcel and hodograph diagnostics are then recomputed from each synthetic mean or percentile profile using the same algorithms listed above. | Keeps diagnostics internally consistent with the displayed synthetic profile; a percentile diagnostic is not the percentile of member diagnostics. | Active ensemble summary side panels, parcel trace, hodograph, and Skew-T | App-computed; stored ensemble payloads must be reingested after method changes. |
| SHOW, LIFT, SWET, KINX, TOTL, BRCH | The application does not recompute these index equations; values are read from BUFKIT station parameters. | Legacy instability and severe-weather guidance. | Derived Parameters when present | Upstream BUFKIT values |