Last time | Next time |
This morning (on Zoom) Kate and I discussed the shape of the logistic curves: this sketch might be helpful. (Thanks for your patience, you students who are helping me get used to Zoom, the iPad, Google Drive, and all this stuff....)
Please, please: put MAT375 (as one word (i.e. not "MAT 375"), as well as assignment info) in the subject line of your message -- I'll be routing all messages using the tag MAT375. It helps.
I didn't present the "seasonal" analysis last time, but two students (Jon and Proctor) had a look at the first three months of the year, looking at all extreme years, rather than a single time series of years.
The question is a little different, but the answer should be the same: if all decades are "equivalent", then we should see a uniform distribution of extremes per decade across the three months.
Jon and Proctor looked at the first three months of the year, which should give results very close to "Winter", below. In what follows I divy things up into the actual seasons (e.g. Dec 21-March 21, etc.).
Spring, Summer, Autumn, Winter |
I've added a "uniform" bar to each histogram - to show what
bin height you'd expect in the absence of any change over
decades (it's really just the average bin height).
Once again, we note the paucity of extreme years in the first four decades which we saw in the previous plots for each of the four extreme temperature readings' years. |
Spring, Summer, Autumn, Winter |
I've added a "uniform" bar to each histogram - to show what
bin height you'd expect in the absence of any change over
decades (it's really just the average bin height).
I want to emphasize again, after grading your homework and seeing folks talk about how "normal" the histograms look, the following: that
we don't expect normal distributions -- not at all! There's nothing special about the middle decades. They should be just as high as any others.
Notice that I've changed (reflected) the bottom plots, to better perceive the "direction" of the fit (or lack of fit) to a uniform distribution. I've made the empirical results look low -- lower than the straight line of the uniform distribution -- when there is a lack of data for a period. |
That means that we're asserting that climate has changed over the 138 years from 1854 to 1992 (even though we're suspicious of the validity of earlier data).
It turns out that there is good reason to be suspicious of the earlier data. More on that later....
Upon inspection of the graphs above (and those from our previous day's agenda), I went to Lyle Fletcher to see if he would reveal anything that could explain the graphs we're getting. He was cavorting with the angels, but took a few moments out to speak with me (through his book).
If you do this reading, you'll understand my cryptic remark "good reason to be suspicious" above. Email me the answer, to let me know that you're on to it. Careful reading will reveal the issue.
You may choose any paper from this conference proceedings ( Asymetric Change of Daily Temperature Range, Proceedings of the International MINIMAX WORKSHOP Held Under the Auspices of NOAA National Environmental Watch and the DOE Global Change Research Program).
Please submit a (typed) two-page "book report" about the implications of what we might expect to see in our maximum and minimum temperature data, based on what others have seen due to climate change. Base your conclusions on what you've read in your chosen article (due Monday, 3/30).
Some good links that I might recommend (a few of which we'll focus on):
In particular, we will implement The SIR Model for Spread of Disease - The Differential Equation Model in InsightMaker.
(An excellent introduction to SIR models, from both the infectious disease and mathematical sides)
Questions:
This on-line estimator (i.e., a model!) allows one to estimate deaths, as well as death by age-category.