The storm has been diagnosed as a Depression by IMD. This means that it is not formally named yet – the name “Nisarga” is not actually used yet on the IMD web site that I can see, even though it is all over the media already – but the formal track and intensity forecasts are now being produced.

Below are the current forecast maps. On the top one, the cone of uncertainty gives an indication of the range of possible tracks the center of the storm could take. On the bottom one, areas likely to be hit by winds of various strengths are indicated. To put these into words, the storm is forecast to pass almost directly over Mumbai with peak sustained surface winds of 60 knots, or about 110 kilometers per hour. This is a Severe Cyclonic Storm in the Indian scale, or a strong Tropical Storm (but not quite a Hurricane) on the US scale.

The track forecast is not good news. The Mumbai metropolitan area has 20 million people, is very vulnerable to cyclones at the best of times (not having had one ever, for all practical purposes), and is currently a hotspot of the coronavirus pandemic. On the other hand, the intensity forecast is good news, in that those models that yesterday were predicting much more powerful storms are no longer doing so. So this no longer looks like the worst case scenario, given current information. That said, a few things to keep in mind:

1) These forecasts are still just forecasts. Things could change.

2) Even 60 kt (110 kph) winds can cause significant damage. Winds will be stronger high above the ground, such as at the tops of tall buildings of which Mumbai has many.

3) There is still significant flood risk, both from rain, and possibly storm surge as well (particularly if the surge peaks at high tide). I have not seen any predictions of surge yet.

Please be prepared. Monitor the forecast and advice of local officials closely.

This post contains some remarks on the tropical cyclone, likely to soon become Cyclone Nisarga, currently forming over the southeast Arabian Sea. This storm poses a risk to the northwest coast of India, including the city of Mumbai. While it might well turn out to be a minor event, there is also some possibility it could be bad enough to cause a major disaster. Landfall is probably 3-4 days away; official cyclone forecasts usually give five days’ lead time, but haven’t been issued yet because the storm hasn’t officially formed yet. In this situation, where a storm forms a short time before landfall, it is challenging to communicate the nature of the risk. This post is an attempt to help do that.

A disclaimer: I am not issuing a forecast. The India Meteorological Department is the authoritative source of weather forecasts for India (and in the case of tropical cyclones, the entire North Indian Ocean basin), and their forecasts are state-of-the-art. I am just an academic and private citizen offering some interpretation. My comments are based on publicly available observations and numerical model predictions. I am writing this late in the evening in New York on May 31, 2020, which is early in the morning on June 1 in India. The information will become dated rapidly.

The latest IMD Tropical Weather Outlook is available here. As I write, it discusses two storms; the first is over far the western Arabian Sea, around the Oman and Yemen coasts. The second, forming over the southeast Arabian Sea, has not officially formed yet (meaning its winds have not yet reached the threshold necessary to qualify as a cyclone), but once it does, it will be named Nisarga. The IMD Outlook currently predicts of this storm that “IT IS VERY LIKELY TO CONCENTRATE INTO A DEPRESSION OVER EASTCENTRAL AND ADJOINING SOUTHEAST ARABIAN SEA DURING NEXT 12 HOURS AND A CYCLONIC STORM DURING THE SUBSEQUENT 23 HOURS. IT IS LIKELY TO REACH NORTH MAHARASHTRA AND GUJARAT COASTS BY 3^RD JUNE.” By this point, if this forecast is correct, the storm will have been formally given its name. The category “Cyclonic Storm” is a storm of relatively mild intensity (34-47 knots, 63-88 km/hour), equivalent to the US “Tropical Storm”. Such storms generally do not produce tremendous wind or storm surge damage (though they can still produce enough rain to cause heavy flooding).

The different numerical weather prediction models (run by different government agencies around the world) have been predicting a disturbance to form around this time for some days, but have come to agree in the last day or two that a storm of some intensity will form and make landfall somewhere on the northwest coast of India, either Gujarat or Maharashtra. Below (Figure 1) is a map of predictions of the storm center’s track from different models as of 12 UTC today.

Figure 1: Track predictions from an ensemble of different numerical weather prediction models. The different colors indicate different models, whose names are abbreviated in the legend at upper right. “Invest 93” is the name for this storm given by the U.S. Joint Typhoon Warning Center. Plot obtained here from the National Center for Atmospheric Research.

The models disagree not only on the exact landfall point, but also on the intensity of the storms at landfall. In many of the models, the storm remains relatively weak. But in a few of them, it becomes quite strong.

In the ECMWF model – arguably the best in the world (though that doesn’t make it the most correct on every occasion) – the storm becomes quite powerful and makes landfall just north of Mumbai. Below (Figure 2) is a map of the latest model run available as I write, showing the prediction for 12 UTC on Wednesday (5:30 PM IST); the storm is headed northeastward so that it comes ashore just north of Mumbai, likely exposing the city to the stronger winds on the right side of the storm and attendant storm surge. This scenario would be quite bad for the city, as it could generate a large storm surge and consequent flooding, especially if it occurs at high tide.

Figure 2. Plot of surface pressure (contours) and winds (barbs, kt) and vorticity (a measure of the winds’ rotation) at 850 hPa, around 1.5 km above the earth’s surface, from the European Centre for Medium Range Weather Forecasting (ECMWF) model. The model run was initialized on May 31, 2020 at 12 UTC, and the prediction shown is for June 3, 2020 at 12 UTC (530 IST).

The ECMWF could well be wrong, but there’s no way to be certain at this moment. So, as far as we know now, there is at least some risk of a substantial disaster in Mumbai, a city with a metropolitan area of 20 million, exposed to the sea, where an event like this has not occurred at least since the late 19^th century (perhaps centuries earlier). There have been major floods in recent years, 2005 especially, also 2017 and 2019, caused by rain from monsoon weather systems. These were not cyclones. A cyclone could be different, due to the possibility that it could produce very high winds and storm surge. It could possibly be even worse than these events.

In the models, landfall occurs on Wednesday or Thursday local time, three or four days from now as I write this. Normally – in India as in most of the world – official forecasts of a tropical cyclone’s track and intensity are issued for a period from now to five days into the future. However, these official forecasts don’t start being issued until the storm has officially formed and been given a name. So if a storm makes landfall less than 5 days after it forms, there is less advance notice than their normally would be. This is what is happening in this case.

The IMD’s Tropical Weather Outlooks, if one reads the text, have been discussing the possibility that a storm could form for a couple of days, and have been giving broad, general statements about its future path, as I quoted above. But no quantitative, specific forecasts of the storm’s track or intensity have been issued yet, because it hasn’t yet formed. One could make such forecasts based on the models, and particularly now, that might be a good idea, but that isn’t the practice. (In the US, the National Hurricane Center started a couple of years ago to issue Hurricane Warnings in such cases, where a Hurricane is expected to form and make landfall within a couple of days, if it poses a significant threat to land. But that is a new practice, not yet adopted worldwide.)

It could well be that the storm will end up being weak, or will only affect areas less populated than Mumbai, or both. But because Mumbai is so exposed and so vulnerable, even a small chance of a major tropical cyclone landfall there is worth being aware of, as far ahead as possible. There does seem to be, at present, such a chance.

Science can make predictions, but government has to act on them or they aren’t any use.

On Wednesday evening, Harvey was still just a tropical depression with winds of 35 miles per hour. The National Hurricane Center forecast that it would intensify to 40 miles per hour by Thursday morning, and reach its peak intensity two days later at 75 miles per hour – just barely crossing the threshold to be called a hurricane – just before landfall.

By 10 AM Central on Thursday, the picture had radically changed. The storm’s winds had already jumped up to 65 miles per hour, and the rapid intensification was forecast to continue, with Harvey reaching 115 miles per hour – category 3 on the Saffir-Simpson scale – before landfall. With these more intense winds came a prediction of 6-10 feet of storm surge for parts of the Texas coast, and an increase in forecast rainfall, with potential for up to two feet in some locations. Suddenly Harvey looked likely to produce a major disaster.

The machinery of government, media, and private sector began to engage, with emergency preparations and warnings ramping up to amplify the State of Disaster already issued Wednesday for 30 counties by Texas’ Governor.

Harvey was still just a tropical storm at this moment. All the emergency actions were predicated on forecasts of future changes in Harvey’s intensity. Intensity prediction is normally considered the weak point in the science of hurricane forecasting, and rapid intensification just before landfall, in particular, has often been called a “forecaster’s nightmare”. But the forecasters predicted it with confidence, and no one, as far as I know, questioned that prediction. The conditions were right, with sea surface temperatures in the Gulf very warm and vertical wind shear (the difference in the steering winds at different heights that can disrupt a hurricane’s circulation) very low. With landfall still somewhere between 36 and 48 hours away – not nearly as much lead time as would be ideal, but enough for many life- and property-saving actions – the NHC’s experts were able to sound the alarm effectively.

The forecasts quickly proved right. Harvey is closing in on the coast near Corpus Christi, and NHC a few hours ago declared it a category 3 Major Hurricane, with sustained winds now up to 125 miles per hour. A storm with winds of this magnitude has not made landfall in the US in over 11 years, but that “drought” will almost certainly end tomorrow.

These forecasts are a scientific success story, but their value depends on the actions of the state, local and federal governments.

Corpus Christi, a city of 325,000 people, is directly in Harvey’s sights. Much of the city – and all of its oil refineries – lie close to the water on a bay that can funnel and amplify the surge, now forecast to be between 6-12 feet (it’s already more than 4 feet as I write). The flooding will come with winds well in excess of hurricane force, and will likely be exacerbated and prolonged by heavy rains that could continue for days.

Yet on Thursday, the Mayor of Corpus Christi decided not to issue a mandatory evacuation order, but instead just to “encourage” evacuation of residents in the lowest-lying areas. I can’t understand this decision.

In a press conference this morning, the Mayor stood by it. When a reporter asked how many people might be left in the areas most at risk, the Mayor replied that he didn’t know, as there’s no way at this moment to take a survey. That’s probably true, but it would have been better if he could have said “we’re doing everything possible to get every one of those people out of harm’s way”. He couldn’t say that, because he hadn’t.

The motivation for avoiding a mandatory order appears to have been a fear of risking first responders’ lives by sending them in to enforce it. But there was no need to send them in. In the United States, a mandatory evacuation order is almost never truly mandatory in that sense. The Mayor could have simply said it was mandatory, and warned everyone in Harvey’s path that if they stayed in defiance of it, no one would be able to come and help them.

How many more people would have evacuated if the order had been mandatory? It’s impossible to know. Some won’t leave no matter what. But it’s hard to imagine that it wouldn’t have made a difference to some people if they’d heard their Mayor tell them they have to leave now, rather than just that it seems like a good idea.

By now, it is probably too late for anyone to get out who hasn’t already. I hope not many are left, and for those who are, I hope Harvey’s impacts turn out less catastrophic than predicted. Corpus Christi may get some luck yet – as I write, the latest forecasts have the center coming in very close to the city, which could put the strongest winds and surge just to the east. Several of the closest counties on that side have issued mandatory evacuation orders; some have not.

I recently gave this talk at Columbia’s Nevis Labs, in their evening public lecture series. The video is here – you can watch it all or in bite-size chunks. It’s about the relationship between extreme weather events and climate change: what we know and don’t know, what evidence we have so far, projections for the future, etc.

I have a post on the Oroville dam failure and what it does or doesn’t have to do with climate change here.

In this article, Andrew Freedman of Mashable argues that the Saffir-Simpson scale should be changed because it doesn’t measure rainfall, and yet rainfall causes disasters in many hurricanes – including Matthew in North Carolina just now. The Governor of North Carolina has been saying the same thing. The National Weather Service disagrees, and so do I.

In fact the Saffir-Simpson scale used to be based on not just wind, but also pressure and storm surge. It was simplified to just wind a while back because the different variables often wouldn’t match for a single storm – e.g., Sandy would have been a cat 1 due to wind but a cat 3 due to surge. There is no way to make a single number capture all the different hazards without being confusing. You wouldn’t know just from the number what the reason for the number was, and would still have to read the fine print in the forecast to find that out.