New developments and operational experience with surface observation technology and national networks

yardbellΤεχνίτη Νοημοσύνη και Ρομποτική

14 Νοε 2013 (πριν από 3 χρόνια και 10 μήνες)

83 εμφανίσεις

New developments and operational
experience with surface observation
technology and national networks

Jitze P. van der Meulen,

Royal Netherlands Meteorological Institute

TECO 2006

2006
-
12
-
04

TECO
-
2006

2

Historic Background


1643:
Evangelista Torricelli

discovers
vacuum invents mercury barometer and
suggested




p




1645:René Descartes suggests
p

=
p
(
h
)




1650:
Blaise Pascal and Robert Boyle
,
by experiment:




p &
p

=
p
(
h
), using a
mercury barometer

All using calculus: hydrostatics & mathematics

2006
-
12
-
04

TECO
-
2006

3

Historic Background


1802: Alexander von Humboldt,
explorer to discover
environmental
phenomena:
t
=
t
(
h
), later on with
Joseph
-
Louis Gay
-
Lussac


< 1850:
phenomenological studies
with a deterministic approach,
e.g.
networks to measuring the intensity
of Earth's
magnetic

field


> 1850:
use of mathematics comes up and experiments produce
quantitative data to prove any theory

by calculus

Discovery of phenomena (qualitative), later on mathematics (quantitative)

2006
-
12
-
04

TECO
-
2006

4

Historic Background


> 1835:
networks

to
measure the geological
behavior of the earth electromagnetic field in
combination with
barometers

and
themometers
, to discover any correlation


Measurements of the variable behaviour of
atmospheric
pressure
: a relation with
wind

(speed and direction) was discovered


Tendency of pressure

useful to forecast
weather (
still popular today at home

)


Geospatial

relationship of pressure gradients
and the behaviour of wind

Discovery of phenomena

2006
-
12
-
04

TECO
-
2006

5

Historic Background


> 1845: introduction of

Telegraphy, providing
in "near real time“ and synoptic surface
observations to be plotted in synoptic charts

(a ‘
nearly instantaneous picture of the state of
the atmosphere’)




1900: postulation that hydrostatic formulas
could be used to describe the current state of
the atmosphere and for forecasts (‘calculating
the weather’)


> 1945 introduction of operational upper air
measurements and NWP


Use of geophysical data in weather charts,

calculus requires “
infinite amount” of

arithmetics

2006
-
12
-
04

TECO
-
2006

6

Historic Background

2006
-
12
-
04

TECO
-
2006

7

Historic Background


Today
:
B
oth science and operational services
(meteorology, climatology,
etc.
) are oriented
on both


discovery of weather phenomena

(the classic deterministic approach), and on


thermodynamics with plenty of mathematics

(used in NWP & NCP models).


This statement not only holds for
in situ

measurements
but also for
remote sensing

and especially for
observations from

satellites (Earth Observation).

2006
-
12
-
04

TECO
-
2006

8

Systems development and functional
requirements


functional specifications of networks and weather
stations have a complex background


A network should provide


present weather information expressed in pre
-
defined parameters, like type
-
of
-
cloud, special
phenomena, type
-
of
-
precipitation, icing


numerical data, i.e. quantitative physical variables


Complex? Simple compared to the present needs of
the multiform disciplines in meteorology and the
available technologies of observation, both in situ and
remotely sensed

2006
-
12
-
04

TECO
-
2006

9

Systems development and functional
requirements


Development of
observing systems,
in particular those
systems based on

satellite born remote sensing
(
'discovery'
entity
, but also producing quantities)


Geospatial differences in network densities


Differences in representativety (area averaged
vs.
point
measurements)


Differences in observation times and frequency

in situ

satellite

2006
-
12
-
04

TECO
-
2006

10

Systems development and functional
requirements


Development of
observing systems,
in particular those
systems based on

satellite born remote sensing
(
'discovery'
entity
, but also producing quantities)


Geospatial differences in network densities


Representativety (area averaged
vs.
point measurements)

wind derived from the ERS
-
2 scatterometer,

south
-

east of Greenland, 2006
-
10
-
15 21:30 UTC

2006
-
12
-
04

TECO
-
2006

11

Systems development and functional
requirements


The various disciplines in meteorology, climatology,
hydrology, etc. expressed an increasing demand on
observational data and with discipline specific
conditions.


Alternative or new observations, performance,
representativety and frequency of observations.


Complete automation (of subjective observations,
present weather, phenomena), modern AWS



Strong increasing diversity in both the users'
requirements and in observation technology

2006
-
12
-
04

TECO
-
2006

12

Systems development and functional
requirements

Cl

¿


¿


¿


INFO


INFO


INFO


INFO


USER

INFO


INFO


INFO


¿


¿


¿


Ag

Hy


Ae

Ma

AS


Sy

¿


SOURCE #1

SOURCE #2

SOURCE #3

SOURCE #N

SOURCE #N
-
1

OBSERVATIONS

[Sy] Synoptical Meteorology,

[Cl] Climatology,

[Hy] Hydrology,

[Ag] Agro
-
meteorology,

[Ae] Aeronautical meteorology,

[Ma] Marine meteorology,

[AS] Sciences of the atmosphere

.

2006
-
12
-
04

TECO
-
2006

13

Cl

¿


¿


¿


INFO


INFO


INFO


INFO


Observations

systems

USER

INFO


INFO


INFO


¿


¿


¿


Ag

Hy


Ae

Ma

AS


Sy

¿


SOURCE #1

SOURCE #2

SOURCE #3

SOURCE #N

SOURCE #N
-
1

database oriented
information generators


Systems development and functional
requirements

2006
-
12
-
04

TECO
-
2006

14

Systems development and functional
requirements


Defining functional specifications of physical variables to be
measured is relatively straightforward: measurement uncertainty
can be expressed quantitatively and references are clearly defined.


To specify quality figures for reporting weather events,
atmospheric conditions or qualitative descriptions (the
assessment of the state and development of the atmosphere, and
of significant weather) is complicated and affected by subjectivity:

E.g., s
hall we use performance indicators like


Skill scores


Success rates


False alarms rate, or


A mixture of these?

2006
-
12
-
04

TECO
-
2006

15

Systems development and functional
requirements


A contingency matrix is most applicable, but how to
interpret?





detector




yes

no


reality

yes

a

b



no

c

d


Probability of Detection:
a

/ (
a

+
b
)

False Alarm Ratio:
c

/ (
a

+
c
)

Equitable Skill Score:
a

/ (
a

+
b
)


c

/ (
c

+
d
)

BIAS =
(
a

+
b
) / (
a

+
c
)

[should be 1]

Event?

2006
-
12
-
04

TECO
-
2006

16

Performance indicators



detector




yes

no


reality

yes

15%

5%



no

5%

80%


POD = 75%

FAR = 25%

ESS = 69%

BIAS = 1



‘overall score’ ?



range ?:
e.g. ‘icing’
most relevant:
-
2
°
C <
T <
+2
°
C


2006
-
12
-
04

TECO
-
2006

17

Systems development and functional
requirements


New challenge: observations in combination with a
high resolution model:


using the actual NWP grid
-
point data as
background (
e.g.
11 km grid)


Using a detailed orographic database (relief
map)


Climate data for further fine
-
tuning +
knowledge of atmospheric behavior in land
-
water transitions


Now
-
casting technique with ‘downscaling’
capabilities

50 km

Virtual reality ?

2006
-
12
-
04

TECO
-
2006

18

Systems development and functional
requirements

Methods of observing Present Weather, also
virtual reality?


Determination based on empirical relations


Direct and indirect phenomenological en
physical relationships are applied


Climate data as filter tool


Algorithms, tuned for optimal performance

2006
-
12
-
04

TECO
-
2006

19

Systems development and functional
requirements

Methods of observing Present Weather, also
virtual reality?

Examples:


Undercooled precipitation using wet & dry
bulb temperature


Hail detection from doppler weather radars


Precipitation determination based on optical
characteristics of the hydrometeors in
combination with air temperature

2006
-
12
-
04

TECO
-
2006

20

Systems development and functional
requirements

Methods of observing Present Weather, also virtual
reality?


Data from the true world is not available, so
validation is a hazardous task


How to calibrate? What is a reliable reference?


Algorithms are not well documented and
published


A quality standard for such determination is
missing (may be in line with ISO 17025)

2006
-
12
-
04

TECO
-
2006

21

Development and new design of
instruments


Clustering of sensors/instruments,
e.g.
on one single
statue discarding the mutual impacts

2006
-
12
-
04

TECO
-
2006

22

Development and new design of
instruments


Clustering of sensors/instruments,
e.g.
on one single
statue discarding the mutual impacts


Micro
-
sized systems with 'multi
-
sensors on a single
chip'

wind sensor

2006
-
12
-
04

TECO
-
2006

23

Development and new design of
instruments


Clustering of sensors/instruments,
e.g.
on one single
statue discarding the mutual impacts


Micro
-
sized systems with 'multi
-
sensors on a single
chip'

Pressure
sensor

2006
-
12
-
04

TECO
-
2006

24

Development and new design of
instruments


Clustering of sensors/instruments,
e.g.
on one single
statue discarding the mutual impacts


Micro
-
sized systems with
'multi
-
sensors on a single
chip'



Introduction of ‘
alternative
’ techniques

holograms


monostatic radar:
drop size distribution

‘Hotplate snowgauge’

Acoustic detector

2006
-
12
-
04

TECO
-
2006

25

Trends and criteria for further
development, a conclusion


Ongoing development of new methods of observation
and measurement devices



Satellite born remote sensing techniques to observe
the surface of the whole earth



Ongoing automation of weather stations, automation
of observing the weather


All kind of observations to be converted into geo
-
physical quantities



Determination of weather parameters based on
phenomenological assumptions

2006
-
12
-
04

TECO
-
2006

26

Trends and criteria for further
development, a conclusion


Priority on efficiency and cost reduction


Observation systems and networks will be further
optimized; wish to integrate observing systems;
surface variables determined from
in situ
(surface)
and
remote sensing

(satellites) data


Automation of subjective observations to be
standardized or at least well documented in open
literature; requirements on skill scores


Focus on specific, dangerous weather events, and on
methods of observations to improve the performance
of NWP

2006
-
12
-
04

TECO
-
2006

27

Thank you