-
Notifications
You must be signed in to change notification settings - Fork 16
/
TODO.txt
57 lines (44 loc) · 3.4 KB
/
TODO.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
TODO
* add weighted average smoother
* add smoother based on median (as an alternative to the default mean)
* distinction between baselines and climatologies (currently a climatology is calculated by default, and there is now an option to add an independent baseline); there should be added an option to provide a 365-day (or 366...how to handle leap years?) long daily climatology that had been independently calculated as a custom climatology against which the events can be detected. In order to understand how this works, I need to properly define what I mean by baseline and climatology as they are different things in my mind...
* more climatology and baseline options
* output option: summary statistics of built-in climatology
* output option: summary plot of built-in climatology
* define events and climatologies as S3 classes
* maybe change the names 'smooth_percentile_width' and 'smooth_percentile' as they actually have nothing to do with percentiles; in the documentation, make clear how the new names used relate back to the same (but differently named) arguments in the python code
* Perhaps a function for real time monitoring use
* (from Hobday et al. in review) Based on this criterion, MHW duration, mean intensity, rate of onset and rate of decline are not suitable measures, as they are not known until the MHW has concluded.
* (from Hobday et al. in review) Thus, MHW area is not a suitable measure, as it would apply singularly to all locations identified within the boundary of the MHW in question.
* (from Hobday et al. in review) Thus, measures related to MHW intensity, such as ‘I’, ‘Icum’ and ‘Imax’ are all suitable candidates (see below for details).
* Add block_average metric() 'intensity_max_max'
* Possibly create a mhw_block() function
* If so, also give it a qualitative score as in the python version
Friday, May 11th, 2018 Skype meeting:
* A code paper that uses various user requests over the years as the case study
* Also play around with ALL of the different arguments
** Show what the sensitivities are to these changes
* Also provide options and advice about what to do given certain limitations
** Such as NA values
** Short time series
** Predictable NA time periods (e.g. sea ice regions)
* The alternative baseline and climatology options need to be demonstrated, too
* It needs to be insured that all core options are available across both languages
** And that the results are exactly the same
** And that the names of the arguments etc. are exactly the same
* Base period choice needs to be investigated, too
** Try to offer how to choose the best base period
* Best practices, broadly, need to be demonstrated and discussed
** This consists of the technical issues
** As well as the scientific ones
*** Including how to determine the correct baselines and climatologies to use
* Round it out with case studies and pitfalls
Correct maxPadLength to match Python
Wednesday, August 22nd, 2018 additional MHW indeces ideas from Russo et al. 2015
* The Europeans here use the same HW definition as Sarah
* The magnitude is calculated differently
** Uses the 25th and 75th percentiles (inter-qaurtile range; IQR) in the calculation
** So magnitude of 5 means that the HW on a certain day was five times the IQR of the base period
* Uses self-made 'extRemes' package
* Specifically investigated/visualised HWs covered in newspaper articles
* Percent of land area over a certain magnitude seems like a good idea