HFU HF Underground
General Category => Weather => Topic started by: EricPeterson on September 30, 2020, 0052 UTC
-
I read the HURDAT database with some python and looked for hurricanes that came close to land. To get figure out land, I used NOAA GLOBE which is linked on the page. The NOAA GLOBE data is a set of simple raster files, binary files with nothing but 16 bit integers representing meters of elevation for every square km of the globe (roughly square km for lower latitudes, varies at higher latitudes). If the meters of elevation is -500, then that's ocean. The code to read in the raster files and determine elevation for any lat/lon is very simple using the rasterio library. I think installing that sucked in GDAL which is a pretty large library, but I seem to have both on my mac now.
Anyway once I figured out how to determine if hurricanes came close to land, or not, then I assume that percentage should be constant. There's no reason for that percentage to change. I used it to correct the tropical storm and hurricane numbers in the HURDAT database. Web page with explanation, source code and instructions is here: https://virtualcoinclub.com/wx/count/count.html (https://virtualcoinclub.com/wx/count/count.html) There's a very striking increase in very strong hurricanes (high end cat 4 and cat 5):
(https://virtualcoinclub.com/wx/count/count-s.png)
Also tropical storms are increasing. Other hurricanes are flat. Some of the TS increase may be policy changes. I can't correct for those, only presumed observation changes based on the heuristic above. If you want to recreate the graphics you'll need to install matplotlib. It's a very nice package for simple bar graphs, line graphs and scatterplots.
-
Out of curiosity, what's the number of storms for each category? I assume far fewer CAT 4&5 vs the lower intensities?
-
The >=125 knot storms exceeded the weaker hurricanes in years like 2004. There were more tropical storms which is a different chart:
(https://virtualcoinclub.com/wx/count/count-w.png)
Really should not be surprising in those years where natural cycles and manmade warming allow strengthening of many storms, even the majority, but only if they can get past the tropical storm or weak hurricane (blue above) threshold. What happens then is two things. They strengthen and weaken, and the planes fly out and catch the strongest winds at peak strength.
I believe the increase in the most intense storms is a real effect of both ocean warming and better measurement. It's hard to separate those.
-
Oh another question, any chance there's artifacts due to the arbitrary selection of windspeed bins? Kind of like the number of 90 degree days vs 88 degree or 93 degree days.
-
Yes, binning is a problem. Even without binning, TS and hurricanes are relatively rare events that don't provide sufficient numbers to do statsitics on annual numbers. Perhaps grouping years would help with that. But as you can see on the lower charts where I plot the detection trends, there are lots of dots at zero, 50% and 100%. That means that in many years, with just one or two or three storms, it's likely that zero percent or 100% of storms went near land or stayed away from land out of happenstance. Lots of happenstance can affect the trend. I removed all the year-bins with zero storms in that bin. Also those are not in the trend calculation. Python makes some of those things very easy to do.
But yes, bins are discrete, their thresholds are arbitrary, "near land" is a binary choice, etc. Bins could be gaussian with overlap. Near land could be graded. There are ways to help deal with the small numbers problems that come with binning.