Not a lot of major disasters lately (although the small ones obviously hurt the folks in the way), and I’ve been busy doing a climate analysis project for the UK DFID. The real time site always has the latest earthquakes, volcanoes, and hurricanes/tropical cyclones situation.
The recent snowstorms across the US have triggered the usual snarky comments from those who don’t accept the *fact* of anthropogenic climate change. But anyone who has lived in the far north or arctic knows that it has to be “warm” to get snow.
You read that right, but “warm” is a relative term. Take a look at this graph, from Danial Cobb, science officer of the NWS/WFO in Caribou, Maine (click to embiggen):
Not the biggest/fluffiest/deepest snow rates will be between -14 and -18 C. OK, -16C (3.2F for you folks stuck in the 17th Century) isn’t warm if you are from the south, but it’s actually pretty “warm” for places that routinely see temperatures below zero F. The other big issue is that the colder it gets, the less moisture the air can hold. So colder air is naturally drier, and can’t hold as much snow. So the “optimum” for both quantity and depth is somewhere between freezing and zero F. Ironically, in the arctic, to get a blizzard, it has to warm up! The other big issue is transporting moist air into the colder areas. The models have predicted (and we seem to be seeing) more “latitudinal” (eg north-south) flow, which would cause more stormy and snowy weather – even though the overall temperatures are actually “higher” in relative terms.
Climate change is a very complex phenomena, intermixing human and natural changes and activities. Far too many people on both sides of the debate do the discussion a disservice by simplistic, apocalyptic explanations.
The track has shifted a bit from yesterday’s forecast, and the storm is now predicted by the Joint Typhoon Warning Center to track directly across Japan. Here’s the wind swath using my TARU model:
Impact forecast at $1.5 to $2 Billion USD, not a huge event for Japan, but given fragile economic conditions unwelcome. The big research question is the impact on Fukushima, and how much trouble the rain and runoff will aggravate the radiation washout situation.
We are in the middle of the Atlantic Hurricane season, but not much is going on. This is a water vapor band image from this morning:
The red areas are dry, the blue areas are wet. The blue blob in the upper right corner is AL90, an “invest” area that will probably not develop (NHC gives it a 10% chance). See the red swath between it and the Caribbean and US? That is very dry air. It will literally suck the life out of the system and kill off any chance of development. This time of year the powerful “Cape Verde” storms that come off the coast of Africa just don’t have much of a chance to develop with all that dry air out there. The Caribbean is “wet” enough for storm to form, but surface pressures are high (you need a low for a storm), and conditions are not otherwise favorable.
The second model suite is showing lower damage estimates than the initial model run. The “Primary” model is indicating $1.7 Billion, with a range of around $400 Million to $2 Billion, average right at $1 Billion. These may be a bit high given the event happened on a Sunday morning and the economic impact side is probably a little high. We’ll see as the damage reports come in.
Update: Earthquake damage estimates are highly dependent on what is called the “attenuation function” or AF, also called the ground motion function. Simplified a bit, this is the rate at which the ground shaking decreases as the distance from the fault rupture increases. For this earthquake it is especially tricky. First, there are two main “families” of AF, for large and small earthquakes. This earthquake falls right on the borderline of those two families. Second, California is geologically complex and, finally, parts are virtually uninhabited right next to densely populated areas, small differences in position and depth make a big difference in projected damage. The full range of AF that we have run so far this morning (14 models) gives a range of from $400 Million to nearly $10 Billion in impacts! The best estimate is between $1 and $3 Billion at the moment. As more seismic data is processed, and actual damage observations come in we can refine those estimates.
9:30am ET Update: With the updated magnitude (6.05), depth, and position, the average of the 15 models that make sense to run for this event are hovering at around $3 Billion in impacts, although the range is quite large, from just under $1 Billion to over $10 Billion, and our “primary” single model is at over $6 Billion (which, again, maybe be ‘hot’ due to the time of day and that it is a Sunday).
The Icelandic Met Office has detected a small lava eruption under the Dyngjujökull glacier. If it melts through to the surface (it is under between 150 and 400 meters of ice), an eruption could spew steam and ash high enough to disrupt transatlantic and European air traffic. If it blows in the next few hours (around Noon US time), here is a forecast of where the ash from an eruption might go over the next 48 hours:
I ran this forecast using a modified version of the Advanced Research WRF model (ARW); the graphics were generated by the VAPOR package. And, yes , there is an animation . . .
The decaying Hurricane Iselle should be making landfall on the big island of Hawai’i in just under two days. By then it should be a tropical storm, and impacts are forecast to be in the $10 to $20 Million range. Here’s the forecast wind swath, using my Taru model and the Central Pacific Hurricane Center forecast track and intensity:
Next up is Hurricane Julio, which is about five days out. It is forecast to pass north of the islands, and impacts should be light if that track holds up:
Having two or more hits in the same general area is not at all unusual. Statistically, if you are hit by a hurricane or tropical storm, there is a one in four chance you will be hit by another tropical system that same year. Why is that? Hurricanes are “steered” by middle and upper level winds. While these winds vary from day to day, they tend to follow a pattern in a given year due to the interactions of various climate cycles like the El Nino/Southern Oscillation, the Quasi-Biennial Oscillation, and others. So the pattern that steered one storm over you tends to push any others storms that form in that same are later in the year over similar tracks. On the plus side, when a storm passes over the ocean it cools the water, both by rain and by churning up deeper, cooler water. So if the second storm passes too close behind the first, before the ocean has a chance to reset, the second storm will tend to be weaker because the ocean will not have as much energy to feed it.
Typhoon (now a transitioning tropical storm) Neoguri is moving north off the east coast of Japan, dumping a lot of rain, over 30 cm (1 foot) in places). There have been landslides and flooding, but at least the winds have died down a lot. One of the concerns is for the tsunami damaged reactors at the Fukushima Daiichi Nuclear Power Station, and in particular the hundreds of storage tanks of contaminated water on site. Here’s a recent satellite view of the plant:
They shouldn’t experience severe wind or waves. One question is how much of a problem the rain itself will be. In looking at the data, it doesn’t seem like that much unless something breaks that shouldn’t (which, given the track record for the site, isn’t really comforting). Let’s take a look at the “data”. I have “data” in quotes, because this comes from Tokyo Electric Power Company (TEPCO), and I’m honestly not sure I trust it very far. For one thing, they keep shifting sample locations, techniques, and intervals, and to top it off the English vs. Japanese sites seem to have different data sets. This is using the data file from the Japanese site, for a location in front of the “impermeable wall” (which is obviously at least somewhat permeable) near the Unit1 water intact channel. This plot is looking at one of the by-products of fission, Cesium-137. Cs-137 is a serious health concern because it is water soluble and concentrates in soft tissues. The WHO guidance for “drinking water” quality is 10 Bq/L, USA EPA limits are 7.41 Bq/L, the Japanese regulators apparently allow 90 Bq/L. Here are the numbers for the last year:
So is there a relation between rainfall and the Cs137 levels? There doesn’t seem to be much of one. The raw correlation for changes peaks on day one after the rain event, and where there was more than 5mm of rain is about 0.2. That means at best only 4 percent of the variation is explainable by rainfall alone. If all observations are included, there is essentially no correlation. Here is a plot of the change is Cs137 readings vs. rainfall:
Lots of “gotcha’s” to this quick look analysis. Note the biggest rainfall events in the last year were only in the 6-8cm range. Since the tsunami, the peak rainfall event seems to have been 9.18 cm. Neoguri could dump two or three times that amount.
I’m still looking at other measurement locations, as well as trying to make sense of their hashwork of measurements. To paraphrase Mr. Scott, analysis is easy; reading Japanese, now that’s difficult.
The latest track and forecast guidance is more tightly clustered as a direct hit on the Island of Kyushu, with the storm continuing up the center of Honshu before exiting the islands:
The impact estimates using the Joint Typhoon Warning Center (JTWC) forecast are still the most enthusiastic, at $24 Billion in this run. Areas along the Inland Sea could see storm surges of over 3 meters. Will publish a detailed surge map when the storm gets closer. Here’s the wind swath based on the JTWC forecast:
The HWRF model, by contrast, crashes the storm intensity to minimal hurricane/tropical storm status, with damage of “only” $2.7 Billion USD. The impact estimate based on the Japan Meteorological Agency (JMA) track, is also much lower than the JTWC track, in the $1-$2 Billion USD range.
So what about the Fukushima site? If the JTWC track and intensity forecast are correct, the site should experience mainly offshore winds, so storm surge flooding is not a problem. Sustained winds are likely to at 55mph/88kph, with gusts to 70mph/112kph possible. The biggest issue is likely to be rain. While the storm will be picking up speed as it moves north, our latest run showed about 15cm (6″) along the coast. That’s not great for a site that already has water management problems.
I’ve already seen some apocalyptic claims what Neoguri might do to Fukushima, including carrying radiation to the west coast of the US. That’s 99% “garbage”, to be polite. Sure, some radioactive particles will be swept along with the storm. But just because something is detectible doesn’t mean it’s a problem. If there is increased spillage, even some compromised containers, it will more than likely be a local problem, or a regional (eg Japan) problem at the worst, and even that is more of a chronic problem such as increases in the rate of long term cancers or birth defects. Tragic as that is, talk of a global catastrophe or acute affects is just scare mongering.
Neoguri is about 3 days (72 hours) away from the island of Kyushu. First, here is the Joint Typhoon Warning Center wind swath. On this track and intensity it would cause about $30 Billion in damage:
Here is the Japan Meteorological Agency Ensemble forecast. It makes landfall at the same location, albeit about 12 hours later, and significantly weaker – a 90 knot storm as opposed to JTWC’s 105 knots:
The estimate based on the JMAE track is “only” 3 Billion dollars -a factor of 10 difference!
Here are a few views of Arthur this afternoon – it’s starting to look like a real storm. Current forecasts are for it to continue tracking towards the Outer Banks of North Carolina, and conditions remain favorable for it to become a hurricane. First, here is a visual satellite image. You can clearly see an eye trying to form:
Next, here is the view from the Melbourne, Florida, Radar as of 4pm. Again you can see the classic spiral structure of a maturing storm:
Here is a 3d view from a computer model, based on data from this morning, showing that the storm is trying to develop vertical structure and outflow:
And, finally, for those of you who might care where the storm is going to go next :), here are the latest (3pm) track forecasts, and the 2pm position fixes (that the models use for the 5pm official forecast):