Weather 101

Chart Guides

Plain-language guides to the technical charts on this site, from FOUS bulletins to skew-Ts and ensembles.

Short, plain-language guides to the more technical charts on this site. Each one covers what's on the axes, what the colors and numbers mean, and the mistakes people commonly make reading it. For one-line definitions, see the glossary.

FOUS bulletins MJO Hovmöller RMM phase diagram Skew-T Ensemble charts Outlook & departure maps

Model guidance

How to read a FOUS bulletin

FOUS bulletins are compact, fixed-width tables of raw model output for individual stations, every 6 hours out to 48 hours. NCEP produced them from the NGM and then the NAM. The NAM, and the official FOUS with it, was retired on November 3, 2026. The FOUS Guidance page rebuilds them in the same format from the HRRR, the RRFS and the REFS ensemble. Those bulletins are unofficial and HRRR- or RRFS-derived, but the coding follows NCEP's own NAM FOUS program (nam32tx), so they read exactly like the originals.

FOUS61 (UNOFFICIAL, MILTONCLIMATE.COM)
OUTPUT FROM HRRR 18Z NOV 01 26
TTPTTR1R2R3 VVVLI PSDDFF HHT1T3T5
ALB//573921 00110 322003 47110701
06000534218 -2609 301911 50130903
12000485325 -0108 281909 52121104

Each station gets one line per forecast hour (00, 06 … 48). On the first line the station ID (ALB//) replaces the hour and precipitation, since there's no precipitation total at the initial time. The groups are:

TT: Forecast hour: 06 is 6 hours after the run time (here 18Z, so 00Z).
PTT: Precipitation in the previous 6 hours, in hundredths of an inch: 000 = none, 012 = 0.12″.
R1 R2 R3: Relative humidity (%) in three layers: near the ground, through the low-to-mid troposphere and in the upper troposphere. 534218 = 53%, 42%, 18%.
VVV: 700-mb vertical velocity in tenths of a microbar per second. Positive means rising air and negative means sinking. -26 is sinking.
LI: Best (lowest-layer) lifted index in °C. Negative values are written as 100 plus the value, so 96 = −4, which is unstable. 09 is stable.
PS: Sea-level pressure, last two digits in mb. 30 = 1030 mb, and 98 = 998 mb.
DD FF: Low-level wind direction in tens of degrees and speed in knots. 1911 = from 190° at 11 kt.
HH: 1000–500 mb thickness, last two digits in decameters. 50 = 550 dam. Values around 540 are the old rain/snow rule of thumb.
T1 T3 T5: Temperatures (°C) at three low levels: about 18 mb above the ground, about 900 mb and about 800 mb. Negatives are written as 100 plus the value, so 98 = −2.

Decoding the 06-hour line: 06 000 534218 -2609 301911 50130903 means no precipitation in hours 0–6, RH 53/42/18%, weak sinking motion, LI +9, 1030 mb, wind 190° at 11 kt, thickness 550 dam, and temperatures of 13, 9 and 3 °C.

Tips & caveats

A wet column shows up as all three RH values in the 70s–90s with positive VVV, and that's when PTT climbs.
A low T1 with T3 or T5 above 0 °C points to a warm layer aloft (sleet or freezing rain risk). All three below 0 favors snow.
The NAM computed the layers on its own "sigma" surfaces, which are fractions of the surface pressure. Our version applies the same fractions to the newer model's surface pressure and interpolates from its pressure levels, so values can differ slightly from what the NAM would have produced. It's raw model output with no statistical correction, unlike MOS or the NBM.

Tropics

How to read the MJO Hovmöller

A Hovmöller diagram squeezes a map down to one strip and stacks the strips over time. The Teleconnections page averages outgoing longwave radiation (OLR) anomalies over the band from 15°S to 15°N. Longitude runs across the chart, and time runs downward, with the oldest day at the top and the newest at the bottom.

Schematic only. An eastward-moving signal slants down and to the right.

Blue = negative OLR anomaly, meaning cold, high cloud tops and enhanced thunderstorms. Orange = positive anomaly, meaning clearer skies and suppressed convection.
An active MJO appears as broad alternating bands that slant down and to the right, because the pattern drifts east at roughly 5 m/s, crossing the Indian Ocean and western Pacific in a few weeks. The steeper the slant, the slower the movement.
Narrower, flatter streaks moving east faster (15 m/s or more) are usually convectively coupled Kelvin waves. Streaks slanting down and to the left move west, like equatorial Rossby waves or tropical cyclone tracks.
Bands that run straight down without slanting are stationary features, such as a persistent La Niña or El Niño pattern. These aren't the MJO.

Caveats

The chart is smoothed over 3 days. Short-lived storms blur, and the latest few rows can change as new data arrive.
OLR is an observed proxy for convection. It shows where the MJO has been, not where it's going. For that, use the RMM forecast below.

Tropics

How to read an RMM phase diagram

The Real-time Multivariate MJO (RMM) index (Wheeler & Hendon, 2004) condenses the MJO into two numbers, RMM1 and RMM2, built from tropical OLR and upper- and lower-level winds. Each day is plotted as a point, with RMM1 across and RMM2 up.

The eight phases and where the MJO's enhanced convection sits in each. The orange arrow shows a typical counter-clockwise path.

Phase is the octant the point is in, which tells you where the enhanced convection is: phases 2–3 over the Indian Ocean, 4–5 over the Maritime Continent, 6–7 over the western Pacific, and 8–1 over the Western Hemisphere and Africa.
Amplitude is the distance from the center. Inside the circle (amplitude < 1) the MJO is weak or incoherent, and the phase doesn't mean much there.
A healthy MJO moves counter-clockwise, which is eastward propagation, taking about 30–60 days for a full lap.
On the Teleconnections page the solid line is observed (BoM, last 40 days) and the dashed line is the ECMWF ensemble-mean forecast, with rings showing member spread.

Caveats

An ensemble-mean forecast usually curls toward the center with time. That partly reflects members disagreeing and averaging out, not necessarily a weakening MJO.
U.S. impacts by phase depend on season and ENSO and usually lag by 1–3 weeks. Treat them as tendencies, not rules.

Upper air

How to read a skew-T

A skew-T log-P diagram plots a vertical profile of the atmosphere from a weather balloon or a model. It's the forecaster's go-to view of stability, moisture and wind with height.

Schematic only, not real data.

Height is shown as pressure, on a log scale up the side: 1000 mb near the ground, 500 mb around 18,000 ft and 250 mb near jet level.
Isotherms (lines of equal temperature) are skewed 45° up and to the right, so read a temperature by following its slanted line, not straight down.
The red line is temperature and the green line is dewpoint. Where they nearly touch, the air is saturated (cloud). A wide gap means dry air.
A full chart also has curving background lines (left out of this schematic): dry adiabats (the path of rising unsaturated air), moist adiabats (saturated air) and dashed mixing-ratio lines (water-vapor content).
A rising surface parcel follows a dry adiabat up to its LCL (cloud base), then a moist adiabat. Where the parcel is warmer than the environment, the area between them is CAPE, the fuel for storms. A layer where it's colder below that is CIN, the cap.
A temperature line that bends to the right with height is an inversion. Wind barbs on the right show direction and speed at each level. Winds that veer (turn clockwise) with height mean warm advection, and strong speed and direction change means shear.

Winter tip

Check whether any part of the red line crosses to the warm side of the 0 °C isotherm between the cloud and the ground. A warm nose aloft over a sub-freezing surface layer is the classic sleet and freezing-rain setup.

Model guidance

How to read ensemble charts

An ensemble runs the same model many times with slightly different starting conditions and physics. NOAA's GEFS has 31 members, and the ECMWF ensemble has 51. The spread among members shows how predictable the forecast is.

Schematic plume, like the AO/NAO/PNA charts on the Teleconnections page.

Plumes show one variable over time, with one thin line per member. Tight lines mean high confidence, and a wide fan means low confidence. The site's index charts add a shaded 10–90% range and a thick ensemble-mean line.
The ensemble mean is usually the best single forecast on average, but it smooths out extremes. A flat mean near zero can hide members split between strongly positive and strongly negative.
Spaghetti plots draw one contour (for example the 552-dam 500-mb height) from every member on a map. Where the lines bundle together the pattern is confident, and where they scatter it isn't.
Probabilities are the share of members meeting a threshold (say, 1″ of snow). They're raw model odds, not calibrated, and ensembles are often underdispersive, meaning too confident.
Look for clusters. Two distinct groups of members usually mean two plausible scenarios, and the mean sits awkwardly between them.

Climate

How to read outlook & departure maps

CPC outlooks, on the Climate page, show the probability of the most likely of three equally likely categories (above, near or below normal, based on 1991–2020). "40%" above normal still leaves a 60% chance of near or below. Darker shading means higher confidence, not a bigger departure. EC means equal chances.
Departure maps show what actually happened: degrees above or below normal, or precipitation as a percent of normal. Percent of normal can look extreme in climatologically dry places, where 300% of a very small normal is still very little rain.
Verification scores CPC's maps with the Heidke skill score: 0 is no better than chance and 100 is perfect. Precipitation outlooks usually score lower than temperature outlooks, and any single month can be negative.
Excessive Rainfall Outlooks, on Rivers & Flooding, are probabilities of rain exceeding flash flood guidance within 25 miles of a point. They aren't rain totals, so use the QPF for amounts.