On June 28, Delhi encountered, without warning, a record-breaking spell of rain that led to an absolute collapse of infrastructure in the city. The India Meteorological Department (IMD) described it as “almost a cloudburst”, with 91mm precipitation recorded in a single hour between 5am and 6am, and over 228mm in a 24-hour window. But IMD forecast for the day was nowhere close – it predicted “light to moderate” rains.

Days later, however, IMD officials said that the intense spell of rain was “impossible to forecast”, as it rained down over a highly confined spatial area at Safdarjung.

But this is where the problem lies.

Forget expertise, the city appears to be woefully short in terms of weather prediction equipment.

Two out of the three doppler weather radars (DWR) – used to track clouds, and thus predict rain – in the city are not functional.

Delhi has three DWRs – at Palam, Lodhi Road, and Ayanagar, which have ranges of 400km, 250km, 100km respectively. DWRs at Palam and Lodhi Road are currently down for maintenance and out of operation for at least a month, which is particularly problematic in the monsoon season, when they’re most required. This means that only the Ayanagar DWR is currently functional, leaving Delhi reliant on the radar with the smallest range for forecast.

Scientists are also forced to rely on the DWR in Patiala (with a range of 300km), nearly 250km away, for short-term forecasts.

This, experts said, is a big reason why IMD scientists struggle to give accurate rain predictions for Delhi.

A DWR sends radio waves from its antenna, which hits clouds and bounces back, allowing it to gauge how far rain clouds or rainfall activity actually is from the city. Based on the distance and trajectory of the clouds, IMD forecasts when rainfall will hit a particular point in the city, along with the intensity expected, issuing alerts for the public accordingly.

To be sure, the issue with the radars compounds the already poor prediction record of IMD. In June 2021, HT analysed IMD’s long-range forecasts and found that it gets them right only 50% of the time.

IMD scientists, however, insist they are working on expanding its weather network by adding three more DWRs and more rain gauges in Delhi-NCR to improve its forecasting.

IMD makes forecasts using weather models, satellite data and DWRs to issue both long-term forecasts that range from seven to 15 days, to short-term forecasts valid for the next 24 hours and even three-hourly updates, called nowcasts. The absence of two out of the three DWRs are further compounding Delhi’s woes in terms of short-term forecasting.

Misses, misses and more misses

Ever since the monsoon arrived, not only has IMD been getting its daily forecasts wrong for Delhi, its short-term forecasts – meant for the next four to six hours – are also impacted, with the Met department revising its daily weather bulletins for Delhi multiple times a single day.

IMD revised its daily weather forecast for Delhi twice during the day on July 1. The colour-coded alert – used to depict severity of weather events, was downgraded to “yellow” from “orange” at 9am, with light to moderate rain forecast for the rest of the day. Then at 12.30pm, this was upgraded to “orange” again, with forecasts stating moderate to heavy rain was possible on Monday. The forecast for July 1 on June 30 was again “moderate to heavy”, but Delhi received zero rainfall that day, making all three forecasts wrong.

On June 29, Delhi’s forecast for June 30 was moderate to heavy rainfall, but only 8.9mm – equivalent to ‘light’ rainfall, was recorded in 24 hours.

Weather experts believe Ayanagar, due to its short range, it unable to capture cloud movement or development, until it’s too late, raising questions at why the Met department was not prepared, ahead of the monsoon.

“Dopplers help look at cloud formation and accordingly, nowcasts get issued. This allows IMD to forecast accurately where it will rain in the next few hours. If the radius is too small, then one may see cloud formation too late. In case of Ayanagar, we are seeing nowcasts or alerts issued by the time it has already started raining, as it is not able to capture cloud movement from afar,” said Mahesh Palawat, vice president at Skymet meteorology, stating a combination of both weather models and short-term DWRs forecasts were not giving the desired results.

IMD director-general Mrutyunjay Mohapatra told HT that they were working to fix the problem, with at least one of Lodhi Road and Palam likely to be functional soon. Meanwhile, IMD is relying on the Patiala DWR (range of 300km) – nearly 250km away, to make forecasts for the Capital, he said.

Palawat, however, said even Patiala was a significant distance away and will not provide an accurate picture for Delhi. “If at least two of the three local DWRs are operational too, they will complement each other,” he said.

“Certainly, we are looking to fix DWRs and one of them will soon be functional,” Mohapatra said.

A second IMD official, on condition of anonymity, explained that parts of a DWR are often imported, which is where the delay in fixing these DWRs has occurred. “For some specific parts, the procurement bit takes time. The actual fixing of the DWR is not a lengthy process,” the official said.

Mohapatra had earlier said that the spell of rain on June 28 was highly localised and thus, difficult to predict. “There are some situations that seem just impossible. Our data shows some stations recorded up to 91 mm rain between 5am and 6am. This extreme rainfall event was quite confined, both spatially and temporally over some parts of the city, like Safdarjung for example. So, definitely it is a challenge but we are trying to address it.”

Madhavan Rajeevan, former secretary at the Union ministry of earth sciences (MoES), said metropolitan cities like Delhi and Mumbai needed a dense network of dopplers and rain gauges.

“Mumbai already has two DWRs and four more are being made operational soon. Mumbai also has a dense rain gauge network, which captures localized rainfall and Delhi needs that too, for better urban planning. Once we know which areas receive more rainfall and what the localized spread of rainfall is, one can make forecasts more accurately. At the same time, this data is shared with the public and it helps them to stay up to date as well,” he said, adding a 100km radius was still substantial for forecasting weather around Delhi, but with more DWRs, the forecasting can be more accurate.