Meteo 496 Independent Studies

crashclappergapSoftware and s/w Development

Dec 13, 2013 (3 years and 6 months ago)

60 views

Meteo 496 Independent Studies

2013
-
14

By

Richard H. Grumm

NWS State College

Paul Knight

The Pennsylvania
State University

Philosophy

(adapted from original)

The evolution of technological advancements is changing the forecast aspect
of meteorology. The accessibility of massive meteorological datasets is
leading to
an explosion of apps and bots
. The former facilitates the direct
involvement of users in weather decision making.


The future of forecasting and
decision making
based on forecasts is trending
toward automation. Similar to rapid financial transactions and automated
medical decision making; in the not to distant future; most forecast decisions
will be accomplished by BOTS.


Your
forecasting future requires high
-
tech skills
and sound meteorological
knowledge applied within
a highly
-
imagination enabling (
HIE
)

environment.


The suggested independent studies projects require a good base skill set,
which you are readily willing to expand and improve upon. The goal is to
position you to thrive in the unfolding automation of meteorological decision
making

in a world trending toward
STEM
.


Overview


Provide projects to expand
your

base knowledge


Apply automated techniques to existing datasets


Explore and expand forecast knowledge using existing
datasets


Highly innovated projects for innovated students



Prepare you for a more automated future


BOTS and APPS are the future


BOTS and then APPS will make decisions



Make you more marketable and innovative


STEM

Fields and STEM in education


Science, technology, engineering and
mathematics


STEM
Fields


Chemistry
, Computer and Information Technology Science,
Engineering, Geosciences, Life Sciences, Mathematical Sciences,
Physics, and STEM Education and Learning Research


STEM is



aimed to improving high technology skills


Critical for a rapidly evolving world


Critical for job positioning in the 21
st

Century


HIE:
Highly
-
Imagination Enabling


HIE
can be for a country or work environment


Going somewhere


moving forward


organizations
that nurture innovation and
innovators. People
can turn ideas into products,
services faster
and cheaper than ever.


LIE: low imagination
-
enabling
environment


Going no where


maintain the status quo


organizations
that do not
nurture innovation and
innovators. People
do what they have always done
and become irrelevant as technology changes

Adapted from
Thomas Friedman
H.I.E August 2013

Real World Examples


Apps on phones



Specifically weather Apps on phones


Decision making in the palm of your hand


Can have decentralizing effect



RadarScope



RadarScope the Phone App


View radar live


Make warning decisions for
friends and family


Draw threat areas with your
finger and post to any social
media site in seconds.

The future is here

we just take a decade or two to know it


Most mobile APPS require some knowledge


Online sites provide forecasts and decisions


Decision APPS will follow


Decisions beyond a severe storm or snow storm


Decisions to bring warm cloths on vacation or rain
gear, buy rock salt, gas for snow blowers, bread,
milk. The list goes on
.


Central forecasting and decision making are
becoming too costly and inefficient

Sample Projects


We have some sample projects


Not as complex as an APP but a start


Most require some skills but a good prototype is a
good start if it is useful and innovative


A new idea is as good as a planned project



Some skills (we will present skills first)


Are required or can be built upon and learned


Some skill sets on next slide



Useful Technical Skills

listed in order relevance


Programing skills of value


C/C++/C#, Java, Perl, Python


Mobile/APPS programming base knowledge



Plotting and rendering skills of value


GrADS, GEMPAK, Pygrib, and NCL



Database skills of value (SQL)


MySQL, MS
-
Access, Postgres, Oracle



Operating skills of value:


Mobile
apps/GIS


Linux operating and navigation skills



Applications software


Word and Excel and try to
avoid death by power point
!

Skills you may already have

but need to grow


Ability to use Excel or Word


Ability to manipulate and analyze data in Excel,
MATLAB, or SQL


Ability to ftp/wget data sets


Navigate in LINUX/UNIX


Ability to display data in NCL, GrADS or GEMPAK


Programming skills
(C,C++,Java,C#,F90,Perl,Python)

Skills you may need to develop

this list is just a start


Better use Excel and Word


Data analysis in Excel, MATLAB, or SQL


Accessing large data sets via the web


Navigate in LINUX/UNIX


Improved displaying and rendering of data in
NCL, GrADS or GEMPAK


Data manipulation in C, C++, Java, C#, F90,
Perl, Python


NAM or GFS Grid Point precipitation climatology. This allows the forecaster to know when the
forecast system is predicting a significant event within the model atmosphere
.
(B,C,E
)



Seasonal precipitation frequency for a station teleconnected to various indices and amounts of
liquid precipitation. Aid in long range forecast and planning on seasonal scales
.
(A,B,D)




A measure of atmospheric blocking patterns and its associated cycles (POLAR
-
EURASIAN INDEX).
Similar to the previous topic this is an aid in long range forecasting and planning.

(B,D,E
)



Develop an RTMA climatology (based on 7
-
10 years of data) and compare with NCDC hourly
climatology. Improves knowledge of datasets and could improve short
-
term forecasting and
climate model evaluations
. (B,C,E,F)



Objectively evaluate 45 day forecasts, such as the Accuweather 45
-
day forecast for period the
period of Oct 15
-
Nov 15 or evaluate the NCEP Climate Forecast System forecasts for the period
of October through December 2013 and project its skill into the winter months Jan
-
Feb 2014
.
(A,B,D,E
)



Case study of significant event with some emphasis on using technologies related to automated
techniques and
applications

(B,C,E,F)

SKILLS
KEY Decoder ring:

A = Facile with Excel with ability to import and
format data sets (text to columns, etc.)

B = Familiarity with ftp and downloading large
data sets (if needed)

C = Ability to write a simple program in either
Perl, Java, Python or C+

D = MATLAB experience or willingness to learn
as you go

E = MYSQL or MS
-
Access experience/learn as

F
= GrADS, GEMPAK, NCL experience

Potential Projects

See handout for complete list

Advanced topics


Touch screen technology to draw and compose
high impact weather threats



Integrated decision forecasts for the common
person


would require reading raw data files
(GRIB) and making logical decisions on the data.
Would likely involve climate data component.


….
these are all ideas we must start small with
something manageable within our evolving skill
set





Useful Scientific Skills


Knowledge of meteorology




Basic foresting


Advanced forecasting integrating statistics and
climatology



Knowledge of Statistics


Basics, probabilities and use of first 3 moments, and
leveraging the probability distribution function



Knowledge of big datasets and the desire to find
and exploit these big datasets


Find exploit and improve use even the CFSV2!

Example

Figure
6. NMQ estimated rainfall from gages and radar for the 6
-
hour period ending at 0900 UTC 28 August 2013. The inset shows the loca
lly heavy
rainfall in southwestern Pennsylvania in the same time window.
Return to text.


Rainfall verse historic returns and
could use Flash Flood guidance

Boswell,
PA
Somerset
Cty

Observed

1 Year PF
Estimate

10 Year PF
Estimate

100 Year
PF
Estimate

1000 Year
PF
Estimate

Time Window

1 Hour

3
-
4”

0.95

1.71

2.64

3.82

2200
-
0100 UTC 22
-
23 July 2013

3 Hour

3
-
4”

1.19

2.12

3.40

5.16

2100
-
0300 UTC 22
-
23 July 2013

Table. 1. Boswell Estimated 1 and 3 hourly rainfall amounts compared to the
NOAA Atlas
-
14

volume 2, 10, 100, and
1000 year return periods. The observations were estimated from NMQ Q2 data. The 3
-
hour window 2200 to 0100
UTC and the 6
-
hour window was 2100 to 0300 UTC. Return to text.

Could write an algorithm to alert when 80% of FFG and historic
thresholds


Generate warning at 90% of 10 year and FFG and urge issuance at
100%


Automation would know all data and always check distraction free

Summary


C
ourse about the future though the future is now


Prepare yourself for today and improve tomorrow


Projects should be


fun and innovative


And provide the basis for a more promising tomorrow


And of interest to
YOU!



Automation is the future and automaters are required.


Requires
fewer

people to automate
more

decisions faster.



Learn to innovate, explore and improve your tomorrow



Other technologies


R
Statistical Computing