What a Weather Satellite Measures as a Bay of Bengal Cyclone Forms

India's weather satellites sit in two kinds of orbit, and the difference matters the moment a disturbance appears in the Bay. The INSAT-3D and INSAT-3DR series park in geostationary orbit roughly 36,000 kilometres above the equator, which means they hover over the same patch of ocean continuously. They cannot see fine detail, but they see everything, all the time, every 15 minutes. ISRO's scatterometer satellites in lower polar orbits pass over the Bay several times a day and measure something the geostationary birds cannot: actual wind speed and direction at the ocean surface, picked up by bouncing microwave pulses off wave patterns.

When the India Meteorological Department detects a low-pressure area forming, the earliest stage before anything is officially a depression, it is this continuous geostationary watch that catches the first organised cloud spiral. The satellite is not guessing. It is measuring outgoing longwave radiation, and where that radiation drops sharply, it means tall, cold cloud tops are blocking heat from escaping. That is where the convection is deepest. That is where the storm is trying to organise itself.

A cyclone draws its energy from warm ocean water. The threshold that atmospheric scientists treat as meaningful is 26 degrees Celsius, sustained to a depth of at least 50 metres. Below that depth and that temperature, a storm weakens. Above it, it intensifies. Satellites measure sea surface temperature using thermal infrared sensors, the ocean radiates heat at wavelengths the sensor captures, and the instrument converts that radiance into a temperature map updated every few hours.

The Bay of Bengal runs warmer than the Arabian Sea for most of the pre-monsoon and post-monsoon seasons, which is why it produces more intense storms more reliably. Before Cyclone Amphan made landfall in May 2020 as a super cyclonic storm, the Bay's surface temperatures in its path were running between 30 and 32 degrees Celsius across a wide area, satellite thermal data showed the fuel was deep and uninterrupted. The storm fed on it for days.

Satellites also track something called the warm core, a pocket of anomalously warm air at the storm's centre in the upper troposphere. A strengthening warm core, visible as a brightness temperature anomaly in infrared imagery, is one of the clearest satellite signals that a storm is intensifying rather than just maintaining itself.

The Dvorak technique, developed in the 1970s and still in operational use, uses the shape of the cloud spiral in satellite imagery to estimate a cyclone's intensity. Forecasters measure the curvature of the banding features, the size and temperature of the central dense overcast, and the presence or absence of an eye. Each pattern corresponds to a numerical intensity estimate. It sounds like art, but it is a calibrated method with decades of verification behind it.

Modern satellites go further. The INSAT-3DR carries a sounder, a separate instrument from the imager, that measures atmospheric temperature and humidity at multiple pressure levels simultaneously. This vertical profile tells forecasters how much moisture is available through the troposphere, whether a dry air layer is intruding at mid-levels (which can weaken a storm rapidly), and where the jet stream sits relative to the developing system. Wind shear, the difference in wind speed and direction between low and high altitudes, is the single biggest factor that determines whether a depression becomes a cyclone or falls apart. The sounder data feeds directly into the numerical models that produce the track and intensity forecasts the IMD publishes.

Scatterometer data from satellites such as ISRO's SCATSAT-1 and its successor adds the ocean-surface wind field: a map of arrows showing where the circulation is tightening, where the inflow is strongest, and whether the storm's centre of rotation is where the visible cloud suggests it is. Sometimes it is not. The wind centre and the cloud centre diverge in a weakly organised system, and knowing which one to track changes the forecast track by hundreds of kilometres.

Every measurement the satellite takes, infrared brightness temperature, water vapour channel imagery, sea surface temperature, sounder profiles, scatterometer wind vectors, is ingested into numerical weather prediction models within minutes of acquisition. India runs the Global Forecast System and its own high-resolution models at the National Centre for Medium Range Weather Forecasting in Noida. The satellite data is assimilated through a process that adjusts the model's initial state to match what the instruments actually observed, rather than what the model assumed the atmosphere looked like.

The 72-hour track forecast the IMD issues for a Bay of Bengal cyclone is not a single line. It is the median of an ensemble of model runs, each starting from a slightly different initial state. The cone of uncertainty around that line reflects how much the satellite data has constrained the initial conditions. A well-observed storm, with dense satellite coverage and recent scatterometer passes, produces a tighter cone. A storm that formed quickly over a data-sparse region produces a wider one. The width of that cone is itself a measurement, of how much the satellites have been able to tell us, and how much they have not.

The chain from a warming patch of Bay water to an evacuation order along the Odisha coast runs through instruments 36,000 kilometres overhead, reading heat and moisture and wind in wavelengths the eye cannot see, every quarter hour, without stopping.