Weather satellites watch a cyclone's clouds day and night, and forecasters read the cloud pattern and cloud-top temperatures to estimate its position and strength.
Cyclones form over the open sea, far from weather stations. For most of a cyclone's life, satellites are the only eyes on it. The India Meteorological Department (IMD) uses imagery from India's geostationary weather satellites, the INSAT series and earlier Kalpana-1, to fix where a storm is and how strong it is.
The standard method is the Dvorak technique, developed by the US scientist Vernon Dvorak in the 1970s. NOAA describes it as a way to estimate a tropical cyclone's intensity from enhanced infrared and visible satellite images. The analyst sorts the storm into a pattern type, such as a curved band of cloud, a central dense overcast, or a clear eye surrounded by cold cloud, and measures it. The result is a T-number: the higher the number, the stronger the winds. NOAA gives the example that T3.5 corresponds to sustained winds of about 55 knots.
Infrared images matter most here, because they work at night and show how cold, and therefore how tall, the clouds around the centre are.
IMD scientists checked their satellite estimates against the final 'best track' records of cyclones over the north Indian Ocean. In a 2013 study, the average difference in position was about 39 km. The satellite intensity matched the best-track intensity in about 85.5% of cases. The biggest errors came when storms were near coasts or islands.
Different satellites add different pieces. INSAT-3DS's mission includes weather forecasting and disaster warning. Ocean satellites such as EOS-06 carry a scatterometer that measures winds over the sea surface; ISRO published EOS-06 images of Cyclone Mandous. Once a cyclone threatens the coast, warnings must reach people at sea: ISRO says IRNSS-1A sends alerts about cyclones, tsunamis and high waves to fishermen.
Find INSAT-3DS: from its fixed spot above the equator it keeps the Bay of Bengal and the Arabian Sea in view all the time. See the satellites that watch cyclones
Velden and colleagues (2006) describe how the Dvorak method turns a picture into a number.
1. Locate the centre and decide the scene type: eye, central dense overcast, embedded centre, shear, or curved band. 2. Measure. For a curved band, the analyst measures how far the curved cloud band wraps around the centre, using a logarithmic spiral overlay. For an eye pattern, infrared gives the temperature of the eye and of the surrounding eyewall cloud; a warmer eye and colder surrounding cloud mean a stronger storm. 3. Assign a T-number. One T-number unit represents a typical one-day change in intensity, and rules limit how fast the estimate may change between images. 4. Convert to a current intensity (CI) number, which can differ from the T-number for weakening storms, and then to a maximum wind speed with a standard table (NOAA's example: T3.5 ≈ 55 knots).
The same paper notes that half of Dvorak wind estimates fall within 5 knots of aircraft-aided best-track values.
To put the IMD study's numbers in context: a 39 km average position difference is about a third of a degree of latitude (1° ≈ 111 km). For a cyclone 300 km off the Odisha coast, that is small compared with the distance still to travel, but it matters at landfall, which is exactly where the study found the largest errors.
Intensity in knots can be converted to km/h by multiplying by 1.852: 55 knots ≈ 102 km/h.
Geostationary imagery suits this method because the satellite watches the same view continuously, so analysts can follow how the pattern evolves rather than relying on one snapshot.
A method of estimating a tropical cyclone's intensity from the pattern and temperature of its clouds in infrared and visible satellite images.
They work day and night and show how cold, and so how tall, the clouds around the centre are.
55 × 1.852 ≈ 102 km/h.