For India Meteorological Department, it has been a frenetic week. Tracking a tropical cyclone that behaved differently was not easy but for the national forecaster, Amphan was like clockwork. The prediction was spot on but required IMD DG Dr Mrutyunjay Mohapatra and his team to burn midnight oil to keep an eye on the storm's movement every hour of the day. Shifting through a dozens of models, understanding their variations was necessary because his forecast shaped the strategies of government agencies and save hundreds of lives. The top weather scientist from Odisha – known as India's Cyclone Man - spoke to Asish Mehta about how the storm was tracked, challenges IMD faced and the weather agency's ascendancy as a top forecaster of the world.
Q: The India Meteorological Department was once again extremely accurate with its forecast on Amphan. How has IMD emerged into such a top weather forecasting agency in last 10 years?
A: There has been a paradigm shift in the role of IMD in the field of weather forecasting in past 10 years. As per the Vision-2020 & Vision-2030 of Ministry of earth Sciences (MoES), IMD continuously upgrades it’s observational, communicational, modeling and forecasting systems. Take the example, the modernisation of observational network. The number of Doppler Weather Radars has increased from 4 ( in 2006) to 27 (2019), Automatic Weather Stations from 107 (2006) to 711 (2019) and Automatic Rain Gauge Stations to 1,350. The satellite based observations also increased significantly with launching of INSAT- 3D, 3DR and SCAT SAT) in recent years.
There has been significant improvement in forecasting system also. The numerical weather modelling has improved manifold in recent years in terms of model resolution, amount of data assimilation in models’ lead period of forecast and accuracy. The Ministry of Earth Sciences (MoES) now runs seven global and regional models in-house against only one regional model 10 years back. The resolution of global models has increased 10 times from 120 km to 12 km and that of regional models three times from 27 km to 3 km in this period.
Similarly, MoES now runs specific models for different hazards like cyclones. The amount of data being assimilated has increased to 60 GB for each cycle of running of models. All these have been possible due to establishment of high power computing system in MoES. The lead period of the models has also increased from 48 hours to 10 days forecast in the interval of three hours.
As a result, there has been improvement in forecast accuracy of various severe weather events including tropical cyclones and associated severe weather. The cyclone track forecast error during 2015-19 was 81, 126, 171 km against 107, 165, 230 km during 2010-14 for the 24, 48 and 72 hours lead period respectively. Track forecast skill and landfall point forecast error have seen huge improvement too. Landfall time forecast error had been 3.0 and 5.4 hours during 2015-19 as against 3.4 and 4.3 hours during 2010-14.
Q: Amphan was a strange system. Within 30 hours, it turned into a Super Cyclone and then weakened over the sea. Why did this happen?
A: Yes, it was a cyclonic storm on May 17 morning and became Super Cyclonic Storm on May 18 noon within 30 hours. It maintained the intensity of Super cyclone till noon of May 19. Thereafter, it started weakening over sea and crossed coast as a very severe cyclonic storm with wind speed of 155-165 kmph gusting to 185 kmph.
During initial stages of its development (May 16-18), various intensification features like sea surface temperature (SST), ocean heat content (OHC) and vertical wind shear (VWS), low level vorticity (a measure of rotation) etc were in a favourable state over south and southern part of central Bay of Bengal and unfavourable over northern part of Bay of Bengal. The OHC was quite high over south and west-central Bay of Bengal. It decreased towards the north Bay. Similarly, the SST was quite high (around 32⁰ C) over west-central Bay becoming 27-28⁰C over north region. The VWS was low to moderate over south and central Bay and high over north Bay. Low to moderate VWS is favourable for intensification and high VWS does not support it. The system also moved slowly during May 16-18 exposing itself to an area with favourable higher SST, higher OHC and lower VWS for a longer time. Thus, it enabled the system to attain its peak intensity over west-central BoB. When it moved towards northern parts of Bay of Bengal, it encountered lower SST, OHC and higher VWS which contributed in weakening of the system to a very severe cyclonic storm.
Q: Was tracking and forecasting Amphan a challenge for IMD?
A: There were four main challenges while monitoring Amphan. It originated from a low pressure area which developed over south Andaman Sea around April 25. It persisted over the same region for about five days and became less marked. However,
the remnant cyclonic circulation meandered over the region for a long time upto May 12. It again reappeared as low pressure area on May 13 over southeast Bay of Bengal.
Considering the model guidance, there was a false alarm from April 25 onwards about its genesis over the Bay and its landfall over different coasts like Bangladesh, Myanmar, Andaman & Nicobar Islands, AP, Odisha and West Bengal. It was a challenge to predict the area, date of genesis of cyclone and movement towards a particular coast.
Amphan moved very slowly during initial two days with a speed of 4-5 kmph and very fast during last two days prior to landfall with wind speed of about 20-30 kmph. Thus, the movement of the system was not uniform. It was very challenging to determine the speed of movement in different phases accurately and thus determining the landfall time correctly.
The challenge was more severe considering the numerical model guidance about the possible track of the cyclone. We usually examine about 12 global and regional models including six models run by MoES and six international models. The
model guidance with respect to track was highly inconsistent with variation from day to day and also from morning to evening. There was a large spread in the tracks suggested by different models even two days before the landfall. So developing a consensus based on these models was very challenging.
Amphan underwent rapid intensification (increase in intensity by 30 knots (55 kmph) or more in 24 hours) from May 17 morning to May 18 early morning which is one of the rare record of intensification in this ocean basin. A cyclone (85 kmph)
becoming a super cyclone (235 kmph) in about 30 hours is a rare event (increase in wind speed almost three times).
However, IMD utilised all the objective tools to compare, comprehend and analyse the NWP model guidance and the observations from various satellite, buoys, ships, and coastal observations to correctly predict the genesis of low pressure area on May 13. Also, IMD issued the first press release and special informatory message to the state and central government agencies about its possible intensification into a cyclone and movement towards north Bay of Bengal following a re-curving track.
On May 16, it was indicated that the system would intensify into a cyclone and move towards West Bengal-Bangladesh coasts. Next morning, it was indicated that the system would cross West Bengal-Bangladesh coast between Sagar Island and Hatia
(Bangladesh). The information was further upgraded and the landfall was precisely indicated over West Bengal-Bangladesh coasts between Digha (West Bengal) and Hatiya (Bangladesh) across Sundarbans since May 18 morning. Actually, the system crossed West Bengal coast between Digha (West Bengal) and Hatiya (Bangladesh) across Sundarbans with a wind speed of 155-165 kmph gusting to 185 kmph.
Q: Reduced particulate matter emissions during the lockdown meant fewer aerosols, such as black carbon, that are known to reflect sunlight and heat away from the surface. Did this elevate the ocean temperature?
A: Amphan was a large scale system with a size of about 700 km diameter. The SST is one of the important factors but not the only factor. Thus rapid intensification of the system cannot be attributed to SST alone. However, the response of ocean to the reduced aerosols during lock down period especially during the life period of Amphan needs to be investigated and I hope that the ocean and environment experts will carry out this investigation.
Q: In last two years, we have seen two summer cyclones. First Fani, now Amphan. Are we looking at a pattern? What could be the reason?
A: Climatologically, Bay of Bengal and Arabian Sea normally experience five cyclones annually – one over Arabian Sea and four over the Bay of Bengal. Considering the seasonal frequency, it has got a bimodal characteristic for genesis
with primary maxima of frequency of cyclones during April-June and secondary maxima during Oct-Dec. Out of five cyclones, one develops during pre-monsoon season and four during post monsoon season. Hence the occurrence of Fani and Amphan is a near normal phenomenon.
Q: Bay of Bengal generates most weather systems during June-October period. The cyclones come in September-October. Has Bay of Bengal become unpredictable?
A: No, with respect to cyclogenesis, IMD has demonstrated its forecasting capabilities in recent years. We are second to none in the field of cyclone forecasting. There has been paradigm shift in tropical cyclones forecasting over north Indian Ocean. Accuracy has significantly improved in recent years as is evident from decreased number of deaths due to cyclones in recent years. The number of deaths has been around six during Nilam (2012), 22 during Phailin (2013), 46 during Hudhud (2014), 26 during Vardah (2016), 26 in Socotra & Oman due to Mekunu (2018), 89 due to Titli (2018) and 64 due to Fani (2019). IMD earned appreciation from WMO, United Nations, global & national scientific community, Government of India for accuracy in tropical cyclone forecasting.
However, the accurate prediction of heavy rainfall with a lead period of five days like that of cyclone is still challenging due to its smaller local scale occurrence and limited life period. The predictability of very heavy rainfall events due to low pressure systems during monsoon season (June-September) developing over the north Bay of Bengal is decreasing as an impact of climate change.
Q: Are we looking at climate change impacting oceanic currents and SST leading to such unpredictability of weather systems?
A: The SST is one of the most important factors for cyclogenesis. Considering unpredictability, the past studies indicate that as an impact of climate change, the predictability of heavy rainfall warning during monsoon season is decreasing. However, IMD has increased its forecast accuracy in recent years with respect to monsoonal heavy rainfall by increasing the observational network and enhancing the numerical modeling capabilities in collaboration with Indian Institute of Tropical Meteorology, Pune and National Centre for Medium Range Weather Forecasting, Noida.
The rapid intensification of cyclones in the deep sea is a global problem as it has low predictability across the ocean basins in the worlds. The R&D continues worldwide to address the above issue.
Q: Do these systems pose fresh challenge to IMD?
A: The heavy rainfall forecasting is more challenging in view of the impact of climate change. To address the issue, IMD and MoES including IITM, NCMRWF and INCOIS are improving their capabilities time to time by enhancing their observational and modeling capacities and introducing impact based forecasting. How do vulnerable states like Odisha, WB and Andhra Pradesh face this challenge?
Q: How do vulnerable states like Odisha, WB and Andhra Pradesh face this challenge?
A: In these states, the National Cyclone Risk Mitigation Project is being implemented by National Disaster Management Authority and states concerned. In this initiative, the vulnerability and risk due to cyclone in terms of heavy rain, wind and storm surge can be assessed with the IMD’s prediction of track and intensity of cyclone and associated severe weather. It will help in providing impact based forecast and risk based warning at local level.