|
Weather Perspectives
for the
Contiguous 48 States
FAQ's
Note:
WeatherPerspectives.com has been expanded and redesigned. This page will be
replaced or deleted in the near future.
-
Question/Comment:
LA Times: Does this mean that the red flags only apply to the six states
shown, or do you have data for the other western states as well?
note: reference was to an emailed copy of Red Flags for
data through September, 2006.
Response:
Data in the form of multi-yhear charts (477 of the things!) is online at http://www.weatherperspectives.com/
for each climate division in the 17 contiguous western states. The "Red
Flags" (for "data through September, 2006") just apply to
those areas that crossed a "drought" threshold for water year
2006, or are expected to do so in 2007 if it is a normal precipitation year.
Several divisions that are not shown in the "Red Flags" are very
close, but did not reach a threshold.
Please note that the Weather Perspectives charts are not direct
presentations of raw precipitation data. They use a multi-year smoothing
algorithm and departures from medians to estimate cumulative effects, and
the classes for each chart are calibrated to reflect expected return
intervals based on the record for the entire 20th century for that specific
area. Other than that calibration and the specific algorithm used, they are somewhat
similar in concept to the U. S. Drought Monitor's "experimental
long-term blends" (see the link below the current month's map and
legend at http://drought.unl.edu/dm/monitor.html
).
I have refrained from placing too many details about the specific
algorithm that I use on the WeatherPerspectives.com web site in order to
keep it proprietary (it is actually dirt simple) so that I can give it to a
501(c)(3) organization if it proves to have monetary value. But - I might
risk a few more details for a story as I am not getting a very large volume
of traffic at the WeatherPerspectives.com web site where I am
"acceptance testing" it. }:>))
By the way - Somebody who is interested in such things might want to look
into the humongous warming and drying trends shown on the charts (attachment
1, attachment 2) for Wyoming climate
division 1, the core area of Yellowstone National Park. Four possible
explanations (individually or in combination) for those trends come to
mind:
1. Numerous major errors in NOAA's online data (unlikely)
2. The area may be a "bellweather" location for reflecting global
warming.
3. The fires of 1988 may have altered the vegetation enough over most of the
climate division to alter the weather.
4. Increasing temperatures in the geothermal regime of the Yellowstone
super-volcano may be affecting the weather (I sent this speculation to the
producers of NOVA's super-volcano program, but no response).
-
Question/Comment:
PM: Where can readers find the annual precipitation received by a
location year to date (01 January of the current year to the current date)
in inches?
Response: A
good place to start for data on an individual weather station is at:
http://www.ncdc.noaa.gov/oa/climate/stationlocator.html
-
Question/Comment:
DK: Check it out. The graph on this site shows below average
precipitation for a span of several years. 1968-1969 was an El Nino year.
This is not reflected in the weather records on this web site. Somebody
needs to go back and look at the data, not argue about whether it was a
calendar year or a water year.
The record precipitation for SLO is from the 1968-69 water year. Check out http://www.slocountywater.org/weather/volunteer/pdf/179%20-%20Callies%
20Precipitation%20Data.pdf
Check out http://tornado.sfsu.edu/geosciences/elnino/elnino.html.
"For the eight well-documented Type 1 winter El Niño events since 1950
(rainfall seasons of 1951-52, 1957-58, 1965-66, 1968-69, 1972-73, 1977-78,
1982-83 and 1991-92), rainfall at San Francisco averaged about 37% greater
than normal, with a mean anomaly of about 7.5 inches at San Francisco1
(thus, implying an average rainfall for Type 1 El Niño events of around
28.5" compared to the 30 year average of 21" or so). Note also
that several Type 1 seasons were very wet (>170%), suggesting that there
should be a reasonable concern for such amounts in the winter of
1997-98."
TK: Yes that record seems out of whack. It does make you wonder about
the quality assurrance and quality control that went into these data. Thanks
for pointing that out. We are all so willing to make use of easy accessible
data and "believe in it".
Response: Thank you. The underlying data used for the division charts is a
simple extraction from the drd964x.pcp.txt file published online (see
http://www.ncdc.noaa.gov/oa/climate/onlineprod/drought/ftppage.html
) by the National Climate Data Center (NCDC) of the National Oceanic
& Atmospheric Administration (NOAA), (eg. the keepers of the record)
with their associated "quality assurrance and quality control".
Or, it can also be accessed in another format at http://lwf.ncdc.noaa.gov/oa/climate/onlineprod/drought/xmgrg3.html
by selecting the state, division, precipitation and then generating the GIF
where a link to the data table for that division will be found at the bottom
of the page.
So, I think that the observations or problems that you are expressing relate
more to (1) scale and (2) the interpretation and accuracy of calculation of
"cumulative effects of precipitation" than they do to the accuracy
of the fundamental data.
On the question of scale, there can be a great deal of variability among the
individual reporting stations within any climate division. That leaves it to
the user to determine whether any data at the climate division level is
adequate for their own particular uses. From the aspect of scale, the
Weather Perspectives charts are no worse and no better than anything else
aggregated to the climate division level.
On the question of interpretation, this is where I could see a legitimate
beef, and where the greatest possibility of error can be found. I do have
only minimum quality assurance and quality control capability available to
me which makes your feedback very valuable for identifying errors. Please
note, however, that the charts are not labeled "annual
precipitation" as that is not what they are intended to show. Instead,
they are labeled "cumulative effects of precipitation" as that is
what they are intended to show. I have not given the specific formulas or
algorithms used as I wish to keep those "proprietary", at least
for now. Instead, I will repost what I had to say in an earlier email that
may not have been forwarded to all recipients of this string of email
messages in order to get us all on the same page.
-----
"... the charts are labeled "cumulative effects" because they
seek to portray just that by using a multi-year exponential smoothing
algorithm as one year seldom a major drought makes, or breaks. Such
smoothing also makes it easier to identify possible cycles and trends.
As an aside, the "cumulative effects" are somewhat similar to the Climate
Prediction Center's "Experimental Long Term Blends"
though only precipitation is used, the weighting factor for prior years is
different, and the classification system (return intervals calibrated
against the 20th century rather than just a percentage of normal scale) is
different."
-----
With regards to San Luis Obispo for the 1969 water year (aka 1968-69 water
year) specifically, it appears that the extreme precipitation may have been
localized and was diluted or offset by drier readings from other reporting
stations within climate division 4 as the NCDC data (see attached [spreadsheet]
[chart]) does not
reflect such extremes for the climate division as a whole. If not, then I
suggest that you call the error to the attention of NCDC.
-
Question/Comment: I question the accuracy of the
information. The all time record rainfall for San Luis Obispo was in
1969--48.76 inches. The graph for California region 4 shows 1969 as a below
average year for precipitation.
Response: First, The NOAA division average may be considerably
different from the figures for a single weather station, particularly one
that is right on the coast subject to the dynamics of Pacific storms. Also,
annual precipitation (Jan. 1 - Dec. 31) is not the same as the water year
precipitation (Oct. 1 - Sept. 30) which is used in developing the
charts.
Second, the charts are labeled "cumulative effects" because they
seek to portray just that by using a multi-year exponential smoothing algorithm
as one year seldom a major drought makes, or brakes. Such smoothing also
makes it easier to identify possible cycles and trends.
As an aside, the "cumulative effects" are somewhat similar to the
Climate Prediction Center's "Experimental
Long Term Blends" though only precipitation is used, the weighting
factor for prior years is different, and the classification system (return
intervals calibrated against the 20th century rather than just a percentage
of normal scale) is different. Beyond that, I have refrained from giving
many specifics of how the "cumulative effects" are calculated on
this web site in order to protect a "trade secret", even though it
is a very small one.
-
Question/Comment: There appear to be too many years
in the "extreme drought" class.
Response: A calibration error was found . The threshold for
"Extreme Drought" was calculated for a 20-year event when it
should have been for a 25-year event. All of the charts that are online now
(as of March 16, 2006) have been corrected for a 25-year event threshold.
The net effect of this correction was to shift about 20% of the years
previously shown as "extreme drought" into the "severe
drought" class.
-
Question/Comment: Just wanted you to know, I am
finishing up a declaration on the North Sheep allotments in the SNRA and
your graph of central Idaho drought is in the
declaration and shows that over half the years since 1980 have been drought
with no attention by the Sawtooth NF. Valuable stuff…with that I can
demonstrate impairment of productivity and irreparable harm…
Response: Glad I could help.
-
Question/Comment: (1) You should know that for
folks who don't have dsl, the state pages load very slowly. What I can see
is that each graph is about 200-205 kb, and the typing is not that legible
on the small cells, so you can't tell the distinctions between the graph
data. Once they are loaded on the state page, individual graphs load very
quickly when you click on one of the small images. You can improve that
performance by reducing the load of the multiple images. One thing you might
do would be to put a label in each box rather than the entire graphic image.
Something like "Graph Cumulative effects of Precipitation - Division 1
period of record" would work as well. (2) I did look at the
index on the dial-up heading. It did load in a reasonable amount of time,
and the individual graphs loaded fairly well too. You could improve that
page by having the 20kb state maps next to the legend so people can see what
division lies where in the state.
Response: Thank you for the comments and suggestions.
(1) About 60% of U.S. internet users now have broadband connections
and, as you noted at (2), an alternative index for those using dialup is
provided. The reason that the full configuration state pages load so slowly
is that they preload all full-sized charts for the state being accessed
(about 2MB) rather than using thumbnails (Update Note: The state index
pages have now been redesigned so that they now only contain links without
the small chart images, and the separate "dial-up" state index
pages have been eliminated). Then, when you click on the link
for an individual chart, it displays instantly rather than having to be
downloaded over the web. It is a trade-off which should just about average
out if you are going to look at several full sized charts for any individual
state. While the small linking images do not provide readable text, they do
provide a quick summary overview of some major patterns and trends combined
on one page. (2) Adding even small map images to the dialup index
would make an already large index much larger. Instead, that page is
configured such that all maps and charts load in new windows. That allows
the map to be kept open in its own window while using the index to view
different charts. Another trade-off.
-
Question/Comment: I really like your
WeatherPerspecitves.com I can see where University professors, students, and
environmental research analyst, farmers, forest service professional, water
works and curious people like me will find it informative. Thank you for the
add placement it looks great. I noticed the hyperlink didnt hold to
AmericasBookShelf. Also when I went into the temperature charts they
were idle as well.
Response: Thank you for the compliment. The links to
Americas Bookshelf have been fixed. One set of temperature charts has been
developed as an example. That sample set can be accessed by clicking on the
"CLICK HERE" link in the heading for the temperature section.
Otherwise, that section is just a filler to show what the completed project
might look like if further developed. Such further development depends
entirely on how much user acceptance Weather Perspectives receives
which is not looking promising at this point.
-
Question/Comment: How are current year cumulative
effects predicted.
Response: Current year cumulative effects are projected rather than
predicted. What is shown is what cumulative effects will be if normal
precipitation is received for the remainder of the water year.
-
Question/Comment: Very interesting compilation of
climate data you are doing, especially the temp data. Does your source of
wind data as well? Here our glaciers and snowfields have been melting off at
a very rapid pace. Bu this year, in spite of a very low snow year and very
warm temps both summer and winter the glaciers had more snow left by late
Aug than anytime in the last 25 years. And where “permanent snowfields”
once had been and had disappeared we had snowfields this year. After
puzzling on this for some time I remembered a major 2 day wind storm in
January, so my theory is that the strong wind events lead to greater
deposition in collection areas, but with global warming peak sustained wind
speeds have decreased, causing less deposition. But I have not looked at
ridge top wind speed data to verify.
Response: Wind data is available online, but not in the form that I
am using for precipitation and temperature. Updating the charts as they are
now is a tedious pain with a lot of manual interface. It takes about 20
hours just to update the precipitation charts, and would take two or three
weeks to build a set of the temperature charts. This project is a test to
see if there is enough interest to invest in hiring a programmer to build a
program to fully automate the process. If there was, then it is my intention
to turn it over to [a certain 501(c)(3)] for use as a possible advertising
funding stream. So far, that is not looking promising as you will note if
you took the "reader survey" and noted how few have responded
since I brought the site up in May. The access statistics haven't been any
better with monthly visitors only in the thirties. This year, I'll try to do
more regular updates - but I don't think that I'll be able to handle more
than about bimonthly and I won't be adding the temperature set unless people
start beating my door down demanding it. Of course almost everybody is in
pretty good shape going into 2006 (some Wyoming divisions appear to be
exceptions) so things might change if another drought starts developing to
generate interest. Thanks for the feedback. Hope you recorded it online by
doing the "reader survey".
-
Question/Comment: This is wonderful to have.
Thanks.
Response: Thanks for the feedback, I hope you entered it in the "reader
survey". The precipitation charts take me about 20 hours just to update
using worksheet templates. I'll try to update them more regularly this year,
but probably won't try for more than bimonthly, and I don't plan to bring
the temperature sets up at all as it would take me most of 3 weeks just to
develop and initially populate the templates. If there were enough interest
to justify it (doesn't look promising at this point based on the reader
survey and web stats), a programmer could be hired (assuming a grant) to
automate the process and I could turn it over to [a certain 501(c)(3)] as a
possible advertising funding stream. The bright side of being an unfunded
show-of-one is that I can try just about anything that I am capable of doing
by myself without getting a by-your-leave from anybody (dictatorships can be
very efficient). The down side is that I get a lot of bombs, so I am used to
them and will not hesitate do deep-six Weather Perspectives if things don't
pick up over the next year.
-
Question/Comment: I have been downloading the page at http://www.weatherperspectives.com/Oregon/index.html
for about 15 minutes and I haven't gotten to Division 8 yet. If it takes this
long for others on dial-up, I doubt if most will use it. Is it possible to break
it up so each division loads individually, or something like that?
Response: Thanks for the feedback. I knew that they would be slow
on dialup, but couldn't check it since I don't have dialup even as a backup
anymore. I thought that they would be more in the 2-4 minute range though. The
reason that they are so slow is that the state index pages don't have thumbnails
on them. Those are actually the full graph squeezed down. I did that because
thumbnails do not automatically update and folks viewing at 800 to 1024
resolution would notice the disconnect for some graphs. I think that what I'll
do is to add a single page with all of the links and no graphics for the slow
people. It is still going to have something on the order of 300 links on it, so
it may be a while before I get it up. If the site doesn't start drawing traffic
and reader survey feedback over the next year - I'll be deep-sixing it anyway so
don't want to invest much more time in it other than trying to do bimonthly
updates (takes about 20 hours just for the precipitation). Thanks again, and
I'll put the link to the index for slow-moes up on the main page at http://weatherperspectives.com
when I get it developed.
|
