yosemite to flood??

Search
Go

Discussion Topic

Return to Forum List
This thread has been locked
Messages 161 - 180 of total 202 in this topic << First  |  < Previous  |  Show All  |  Next >  |  Last >>
Ken M

Mountain climber
Los Angeles, Ca
Jan 11, 2017 - 11:03pm PT
Dingus mistakenly said NOAH's prior 60% inaccuracy was from before satellites and radar.. not true.,
and, they were also 60% inaccurate, again, last week.

You predicted 17, and it was 12. I don't think that's such hot sh*t.

You refused to predict monday, although you were going all "Chicken-Little" on us.

Your "prediction" of 17 was just a guess. You provide NO methodology. There was NO model. What brand of darts do you use?

The accuracy of the prediction a number of days out is always going to based on information that is incomplete. Why didn't you predict 17 a month ago???

YOU changed your prediction as the time got closer. YOU DID. Does that typify you as WRONG?? I certainly wouldn't think of it that way. One alters the prediction based upon the gradually increasing quality of the information. Their prediction just hours before the event was pretty spot on.

So mr. Clairvoyant, tell us where, when, and the size of the next earthquake.
neebee

Social climber
calif/texas
Jan 11, 2017 - 11:59pm PT
hey there, say, all...

wow, i am just impressed that we can at least get warnings, these days,
as to the weather... :)

being 'such a mommy' hen, i'd prefer to get my little chicks,
out of flood range, as best i could--due to warnings, even if
they might not be perfect, :)


or, blizzards, etc...

thanks to all that share here, it is very interesting, and
such...


i liked werner's comment, about the yosemite park, here:

And by the way, those predictions prevented the chaos that ensued during the 97 flood by the park planing preventative measures before the SHTF this time ......



WHEN i lived in south texas, in the hurricane landfall-prone areas,
man oh man, we got quite a few 'never hits' that nearly did...

many stores were 'bought out' and buildings boarded up,
but-- all it takes is JUST ONE TIME NOT TO DO IT, and:

you could regret it for a long long time... :(

we just did the best we could, each storm, and then--HOPE TO GOD that it
did NOT hit... :O


we were GLAD when it was a 'off target' a bit... :)


sadly, though, for others-- they got hit...
it was a WORSE 'off target' than a 'less of a flood' situation...


it is still really really amazing though,
what we can be WARNED about, these days...


BEING in this snow area now, and BEING able to watch that ol'
DOPPLER RADAR, as the snow comes near us, man oh man,
it is AMAZING to me, :O

always makes me REMEMBER THIS:
__https://en.wikipedia.org/wiki/Schoolhouse_Blizzard__

What made the storm so deadly was the timing (during work and school hours), the suddenness, and the brief spell of warmer weather that preceded it. In addition, the very strong wind fields behind the cold front and the powdery nature of the snow reduced visibilities on the open plains to zero. People ventured from the safety of their homes to do chores, go to town, attend school, or simply enjoy the relative warmth of the day. As a result, thousands of people—including many schoolchildren—got caught in the blizzard. The death toll was 235.[3] Teachers generally kept children in their schoolrooms. Exceptions nearly always resulted in disaster.[4]


wow, guys, thanks again, for sharing all this, whether they
all match up, with each other, and all that, it is:

still amazing, :O

Travel was severely impeded in the days following



EDIT:

WOW, I JUST SAW THAT THIS WAS FROM:
january TWELFTH of all things:

http://voices.washingtonpost.com/capitalweathergang/2011/01/freak_deadly_storm_childrens_b.html

One moment the air was clear, calm, with spring-like warmth. Then, in a period of just a few minutes the sky darkened and temperatures dropped 18 degrees, and vicious winds drove tiny snow flakes (described as "ice dust") which almost instantaneously created a whiteout with visibility near zero. Blizzard conditions continued until about midnight as temperatures fell precipitously to double digits below zero with a wind chill of -40. An estimated 4-5 feet of snow had fallen, although drifting undoubtedly made accurate measurements virtually impossible.

By the next morning (Jan. 13), hundreds were killed with a high proportion of children among the storm's victims as they attempted to return home from school.



oh my:

interesting quote, from writer of article:

Laskin is also a weather geek and explains the meteorology behind the event given the available, but relatively limited sources of weather observations and analyses. This includes describing the development and progression of lows and highs, fronts, jet streams and jet streaks, which no doubt reflect the assistance in such matters when writing his book from Dr. Louis Uccellini, Director of the National Center for Environmental Prediction (NCEP) and co-author (with Paul Kocin) of Northeast Snowstorms.

To say the least, the state of the science and art of forecasting in 1888 was in the early days of development. There were no satellite observations, limited surface data and no observations aloft, let alone computers and weather models for forecasters (like Wes Junker) to peruse and come up with the best forecasts possible days in advance of possible weather threats. There were some indications of a drop in temperature and snow from surface data available to forecasters at the time, but lack of timely and reliable communications kept word from getting out before it was far too late.

We'd like to believe that advances in atmospheric science, vast system of observations, global satellite coverage, computer models, and instant communication would preclude a storm of this magnitude striking without advance notice. But surprise snowstorms remain a possibility even at short ranges to the extent the threat is communicated as a simple deterministic yes or no. With today's tools though, outright surprises can be avoided when the forecast information also includes reasonable and reliable estimates of the degree of confidence in alternative scenarios (to yes/no).

By Steve Tracton




well, night night now...

prayers for all the folks, going through stuff, right now...

and for the 'avalanche' ski patrols and
park workers in flooded areas...
all up and or down, the calif, etc, areas...
(nevada, etc, etc) ...
i-b-goB

Social climber
Wise Acres
Jan 12, 2017 - 08:59am PT
Good photos...

https://www.facebook.com/YosemiteNPS/?hc_ref=NEWSFEED&fref=nf
mouse from merced

Trad climber
The finger of fate, my friends, is fickle.
Jan 12, 2017 - 09:07am PT
NutAgain!

Trad climber
South Pasadena, CA
Jan 12, 2017 - 11:32am PT
From that sequence, it looks like the model is fairly nonlinear, so that "small" changes in inputs can have a "big" change in outputs.

There you have a definition of chaos. Here is a link to a (the?) seminal paper in the field by Robert May from 1976:
"Simple Mathematical Models with Very Complicated Dynamics"
http://abel.harvard.edu/archive/118r_spring_05/docs/may.pdf


What is most remarkable to me is how you can take a simple equation from junior high math and find an INFINITE (but somehow still constrained!) level of rich varied behavior the more deeply you explore it, in terms of making infinitely small changes to a parameter and getting a dramatically different result. Perhaps it is similar to the beauty of Yosemite from the air, then a specific cliff, a specific climb, a specific pitch, a specific rock crystal, a specific attom, a specific proton, a specific quark... Mathemtically it is easier to drill down more and more deeply into a specific part of an equation in terms of what outputs are yielded for a change in input.

The part that interests me most is called "orbital analysis." You take an equation like F(x)= ax(1-x) with 'a' as some constant. Then you explore what happens when you take the answer and use it as input to the next iteration. For some values of 'a', looping through the equation will quickly blow up to infinity (or say a 100 foot flood warning). For other values of 'a', it will quickly converge to zero or another constant number and get stuck there. And then there is a value of 'a' where it "forks" to 2 different stable numbers that it toggles between. And as you change the 'a' value a tiny bit, it forks from 2, then to 4, then to 8, 16, etc... and pretty soon you just have a seemingly random cloud of possibilities that the outputs jump between. This is pictorially described in a very famous graph:


You may ave seen this more richly detailed version of a Bifurcation Diagram:


Connecting the dots back to weather... as you might imagine, the models that predict weather involve inputs of heat and moisture and pressure from thousands if not millions of point sensors, images of heat and humidity and pressure that are effectively big matrices of a bazillion data points, and they are fed into systems of equations for fluid dynamics. As you might guess, this is pretty vastly more complicated than using a single simple junior high equation.

Back to that junior high equation for a moment to graphically show HOW MUCH VARIATION is contained just in that simple expression. You have probably seen images that graphically depict what happens to an equation when you keep iterating it on itself (from stable point output to multiple stable points to chaos with seemingly infinite points but in a constrained pattern), but maybe you didn't know what you were looking at. Here are some examples:





Modern computing power has allowed everyone to see in almost inconceivably high detail the richness contained here. This is not an artists rendition of random patterns- THIS IS REAL MATHEMATICAL BEHAVIOR CONTAINED WITHIN A SINGLE SIMPLE EQUATION!!! Just hit the pause button anywhere a few minutes into this and look around for a moment before it zooms in more. It blows my mind.
[Click to View YouTube Video]

What does this mean for weather? Let's just say our weather could be predicted with that simple equation F(x) = ax (1-x), and all we had to do was get a very careful measurement of 'a'. Let's say 'a' is the temperature of a spot in the ocean somewhere. Now we get that measurement a= 7.65 degrees Celsius. Perhaps you plug that into the model and it blows up to infinity, showing a dark spot on this fractal graph. Huge flood coming. But oh wait, maybe that temperature was really a=7.648 degrees, and now the graph shows a stable result fixed at zero, so the weather report is no rain. Oh, but maybe a=7.6479 and now we get a different result. What this zoomed in graph shows you is how much the output can change as you keep zooming in, showing that for some natural systems, even a change of +/- .000000001 or more will have different results!!!! And it keeps going the deeper you look... a change of 0.000000000000000000000000000000000000001 or as far deeply as you care to look, we keep getting different results in certain critical ranges of the input variables. Perhaps the answers to life, the universe, and everything are contained herein?


So the fact that weather forecasting works at all for long range times (where you have to keep iterating the equations) is a truly remarkable human achievement. Perhaps not enough people appreciate this. Given these factors, it's pretty reasonable that weather reporters err on the side of dire forecasts if there is a risk. They just have to balance that against the risk of giving too many false-positive dire forecasts where people start ignoring them.

Edit: This might be the part where I sit down and the big kid mathematicians explain the distinction between Julia Sets and the Mandelbrot set, and fix whatever inaccuracies I stated in the process of trying to get the main point across in a simplistic way.
Clint Cummins

Trad climber
SF Bay area, CA
Jan 12, 2017 - 02:33pm PT
The variance is big because the critical event of the storm track hitting the small Merced watershed is hard to predict 100 hours in advance, when it's all far out in the Pacific.
The rain also has to arrive at a high enough temperature to have the interaction with the snowpack to yield the big flood like 1997.
The effect is nonlinear but bounded.
A lot like a drive-by shooting - if the bullet hits a wall, no big deal, but if it hits you in the head, that's often catastrophic.

It's not an unstable / chaotic interaction like x(t+1) = a*(1 - x(t)) .
It doesn't have a negative self feedback like that.
It has a small positive feedback, i.e. more runoff at time t yields more slightly more runoff at t+1 due to snow melted by the runoff.
Radish

Trad climber
SeKi, California
Jan 12, 2017 - 02:38pm PT
Thought this pic would be relative.
Clint Cummins

Trad climber
SF Bay area, CA
Jan 12, 2017 - 02:43pm PT
To be fair to Ed B, he did not forecast a point value of 17;
he said "personal prediction, not over 17 feet".
So he was right.
Ed H's graph does show that the NOAA model was heading in the direction of 17 and lower when Ed B discussed 17.
IntheFog

climber
Mostly the next place
Jan 12, 2017 - 03:45pm PT
To follow up on Clint's (and healyje's) point, that these storms are coming in from the ocean creates a bunch of problems for the models. Maybe the biggest problem is that there's much less data about conditions in the eastern Pacific. This makes it hard to predict the storm track precisely.

The reason is simple. In the jargon of forecasting, the sparser the data, the lower the resolution of the model. Lower resolution means the model isn't as precise about everything from the path of the storm to its speed to the freezing level. Any of these can make a big difference in the accuracy of the forecast at a particular point, such as the Valley.

This week's big snow in Portland shows how badly the lower resolution can screw up a forecast. The snow storm that hit Portland was part of a low pressure system centered in southern Oregon. Once the low hit land, it moved much more slowly than predicted. This made the showers over Portland move more slowly. Voila: Snowpocalypse!

The models missed the big dump because the low's speed depended on conditions in the ocean. In this case, the models needed more data to distinguish between a low that would move slowly and one would move quickly.

There's a nice discussion of "The Challenge of Pacific Northwest Weather Prediction" in Chapter 11 of Cliff Mass' "Pacific Northwest Weather." Much of what he says applies to northern California as well.
SteveW

Trad climber
The state of confusion
Jan 12, 2017 - 06:40pm PT

Tagging along with the tradster up above. . .

"NOAH, HOW LONG CAN YOU TREAD WATER?"
healyje

Trad climber
Portland, Oregon
Jan 12, 2017 - 06:49pm PT
Also keep in mind that we've only been collecting data on and attempting to model atmospheric rivers for a relatively short time.
IntheFog

climber
Mostly the next place
Jan 12, 2017 - 07:41pm PT
^^^^^
The lack of historical data means we don't have a lot of data for estimating parameters/calibrating the model, which means the the predictions will be off.

In other words: The lack of data affects what we know about both the structure of the model and the initial conditions for any particular day.

A true double whammy.

Add in that these systems come ashore in the coast ranges, where topography influences everything from wind speed and direction to moisture content, and you have a lot more whammies!

It's a wonder these models get anything right.
Majid_S

Mountain climber
Karkoekstan
Jan 12, 2017 - 09:51pm PT
Downtown San Jose main creek had easily 10 feet of water. Never seen it so high. probably homeless tents got washed out to bay and bodies will show up in Alviso
Ed Hartouni

Trad climber
Livermore, CA
Jan 13, 2017 - 01:18am PT
for those interested in pursuing the issues having to do with the uncertainty in the forecasts, the page to look at is here:

http://www.cnrfc.noaa.gov/ahps.php

especially follow the links:

http://www.cnrfc.noaa.gov/ensemble_theory.php
http://www.cnrfc.noaa.gov/esp_trace_analysis.php

which will give you an idea of what is being calculated.

For instance, I was able to make the plot forecasting the Pohono Bridge river flow from today through the end of the month:


but you can also make this plot:
this provides the estimate range for the forecast.

Seems to be a storm coming in on the19th-21st...

You can see the archived sequence of the forecast vs. measurement here:
http://www.cnrfc.noaa.gov/formHistGraphRVF_loop.php?id=POHC1&loopno=00

the sequence of plots:




shows that the precipitation input to the models at the peak fell below the actual precipitation. I would have been interesting to see the "expected value" plot for that time period next time (later this week).
healyje

Trad climber
Portland, Oregon
Jan 13, 2017 - 01:57am PT
Forecasts for the rest of the month sound like they generally revolve around a stronger polar vortex keeping arctic air from crossing very far south of the Canadian border, but they also voice a lot of uncertainty around exactly how strong or weak it will actually turn out to be. It's a tough game.
WBraun

climber
Jan 13, 2017 - 07:49am PT
Merry said at the time all the projected predictions were made the river will hit 13 feet.

She was only 3 inches off .......
rincon

climber
Coarsegold
Jan 14, 2017 - 08:05am PT
I put a 5 gallon bucket out before all the rain. It has 12 inches of water in it now!
Reilly

Mountain climber
The Other Monrovia- CA
Jan 14, 2017 - 08:14am PT
My bucket has 6" just this week here in Sin City. It is so much more scientific than my old weather rock.
Clint Cummins

Trad climber
SF Bay area, CA
Jan 14, 2017 - 12:06pm PT
Nice plots, Ed.
Especially that first one.
WBraun

climber
Feb 8, 2017 - 02:39pm PT
Oh oh .... 2 inches to flood stage today on Feb 8 2017

Messages 161 - 180 of total 202 in this topic << First  |  < Previous  |  Show All  |  Next >  |  Last >>
Return to Forum List
 
Our Guidebooks
spacerCheck 'em out!
SuperTopo Guidebooks

guidebook icon
Try a free sample topo!

 
SuperTopo on the Web

Recent Route Beta