Sensorscope: Sensing the Alps - MICS

Sensorscope: Sensing the Alps - MICS

Sensorscope: Sensing the Alps School of Architecture, Civil and Environmental Engineering School of Computer Science and Communications Zurich, MICS ...

16MB Sizes 0 Downloads 4 Views

Recommend Documents

choosing the sample - UNICEF MICS
In the context of multiple-indicator surveys, sampling is a process for selecting ... systematic sampling, you must ensu

Hannibal Crossing the Alps
historian Polybius. Though neither was from the exact time of Hannibal, they both were able to write in detail bout the

processing the data - UNICEF MICS
a complete program for data entry, checking and analysis to calculate all the ... number of checks to these data, either

EDSC-MICS II - The DHS Program
LISTE DES TABLEAUX ET DES .... 2.3.1 Indice de bien-être économique . ...... Tableau 2.9.1 Niveau d'instruction de la po

The Alps 2007 - Alpine Journal
Atlantic Ocean Wall but a little less serious than Zenyatta Mondatta. Just up the valley stands the Torrione Moai (aka T

Alps Tour Membership - Alps Tour Golf
Tours*. Challenge. Tour. European. Tour. * 4 Satellite Tours: Alps Tour (West & South Europe) .... Our events are owned

O MICS - Era-Edta
G. Castellano, F. Pesce, G. Zaza, S. Simone, E. Ranieri, G. Grandaliano. Local Committee: M. Accetturo, G. Dalfino, M. G

O MICS - UniBa
Gianluigi Zaza. Italy ... Daniele Cusi, Bernd Mayer, Gert Mayer and Gianluigi Zaza ... G. Castellano, F. Pesce, G. Zaza,

MICS - MEDICA.de
Rounded blunt ends - perfect for blunt dissection to mobilize the artery, vein and bronchus for VATS procedures: • as

Bangladesh - UNICEF MICS
(BBS) in collaboration with UNICEF Bangladesh, as part of the global MICS ... Bangladesh Multiple Indicator Cluster Surv

Sensorscope: Sensing the Alps School of Architecture, Civil and Environmental Engineering School of Computer Science and Communications

Zurich, MICS „08

Karl Aberer, Sam Assouline, Guillermo Barrenetxea, Alexis Berne, Elie Bou-Zeid, Olivier Couach, Chris Duffy, Martin Froidevaux, Jacques Golay, Hendrik Huwald, François Ingelrest, Mounir Krichane, Michael Lehning, Vincent Luyet, Daniel Nadeau, Marc Parlange, Andrea Rinaldo, Pablo Ristori, Ilya Serikov, Valentin Simeonov, John Selker, Silvia Simoni, Cara Tobin, Scott Tyler, Nick van de Giesen, Martin Vetterli

Lake Geneva

We‟ve often been in search of flat fetch to help guide development of turbulence models in the context of Large Eddy Simulation (LES) …won‟t discuss per se today

Spatial turbulence measurements

Plaine Morte, 7 km length

energy balance….

Elie Bou-Zeid, Hendrick Huwald

Outline

1. Continued basic challenges in hydrology and modeling …new perspectives moving to the Alps

2. Wireless technology for environmental sensing …new communication technologies 3. Fall 2007 Campaign in the Drances Watershed

4. Development of a new lidar to probe the ABL …if time allows

1 Alpine perspectives

Regional hydrologic cycle – climate change – profound implications for the Alps water resources management, electricity, economics, hazards …

Water tower of Europe

1 Alpine perspectives

The landscape and meteorology is variable and typically we lack observations at the scale of interest, appropriate instrumentation

Plaine Morte, Olivier Couach

1 Alpine perspectives

Evaporation into the atmosphere

Penman-Monteith…tricky

rs

Qne   Ceu ( q2*  q2 ) E [   (1  rsCeu )]

?

At watershed scales? With which measurements? By chance somewhere in the catchment…..

Jura, CH

Martin Froidevaux

1 Alpine perspectives

Many basic open issues remain… Prediction of natural hazards – critical regions with poor „predictability‟ Measurement at appropriate spatial and temporal scales across the landscape and the lower atmosphere to improve predictions

Need for field measurements remains crucial in hydrology to test simulations and guide the design of new models used in warning networks…

Olivier Couach & Scott Tyler

1 Alpine perspectives

„Our ability to estimate the hydrologic cycle at regional scales generally remains behind societal needs‟….(streamflow, snow melt, land slides) Scott Tyler, 2006…somewhere on a run

…associated components – e.g. soil erosion

1 Alpine perspectives

Natural Hazards and Risk Management

Val d'illiez, Valais, CH

Drances, Valais, CH, Oct. 2007

1 Alpine perspectives

what can advance our current capability of managing natural hazards? Silvia Simoni meteo data precipitation, wind speed, air temperature and humidity, solar radiation ...

topographic data dem, remotely sensed images ...

soil data texture, strength parameters, soil moisture and soil suction measurements, soil temperature ...

geology hydrological characterization hydraulic conductivity , water table, discharge measurements ...

1 Alpine perspectives

Surveyed Landslides

7

Developing „maps‟ of hazards

November 2002 event

7

1 Alpine perspectives

For prediction hydrologic variables across watersheds computed at 2m, for 3 different days. head maps mm Volumetricnumerically water content Pressure

1 Alpine perspectives What Simoni and Rigon provide based on watershed model…

Triggering probabilities – different days

14

1 Alpine perspectives

Failure patterns varying in space and time according to meteorological and topographical conditions Testing such models problematic need spatial measurements

Silviaphd Simoni simoni, 2007, dissertation

2 SensorScope

SensorScope: An Environmental Monitoring Network Guillermo Barrenetxea, EPFL Vincent Luyet, M. Krichane, François Ingelrest, Olivier Couach, M. Vetterli, M.B. Parlange, J. Selker NSF Swiss National Center on Mobile Information and Communication Systems http://www.mics.org

Martin Vetterli and Guillermo Barrenetxea, Genepi

2 SensorScope Motivation

One of the primary limitations in hydrologic research today is the lack of simultaneous spatial and temporal observations Measurement systems today Few and often expensive (limited spatial coverage) Data collection: • Data logger (limited storage, time consuming) • One GPRS connection per station (expensive)

Sodar/rass at EPFL

Data storage and visualization • Data stored locally (difficult to access / share) • Data storage depends on the instrument

We need new measurement tools! Hendrick Huwald, Plaine Morte, 2007

2 SensorScope SensorScope: A New Scientific Instrument

SensorScope is a large-scale distributed environmental measurement system centered on a new generation wireless sensor network Main Characteristics • • • • •

High temporal and spatial density measures Easy to deploy and maintain High flexibility: add new measuring points, new sensors Real-time data Online tools to survey, analyze, and download data

Genepi, 2007

Based on multiple sensing stations …

Plaine Morte, 2007

2 SensorScope SensorScope: A New Scientific Instrument

SensorScope station modules Processing and Communication • Centered around TinyNode module – shockfish - Lausanne • Low power consumption • Long communication range (600 m) Energy management • Double buffer system: Ni-Mh + Li-Ion battery • Solar panel (1 W) • Withstand a one month “blackout”

2 SensorScope SensorScope System SensorScope: A New Scientific OverviewInstrument

Data Analysis

Data Path • Multi hop wireless routing • From the sensing stations to a Base Station • Base Station to the server • From the server to the user

Deployments

2 SensorScope SensorScope: A New Scientific Instrument

Wireless Routing Characteristics Dynamic: allows us to add/remove stations Load balanced: traffic distributed between nodes

Survey, analyze, and control the sensor deployment Real-Time module Geographical locations Shows the latest available data Alarm system (markers) Provides immediate feedback Warning about malfunctions Real time control

2 SensorScope SensorScope: A New Scientific Instrument

Evolution of deployments 1. LUCE : Lausanne Urban Canopy Experiment, 2006-2007 100 Weather stations Soil moisture patterns Local microclimate Single hop First weather station design EPFL Campus

Skin temperature

SensorScope station

Location of the 100 stations

John Selker

2 SensorScope SensorScope: A New Scientific Instrument

9h00

13h00

17h00

14h00

12h00

15h00

Air temperature distributions of EPFL campus from 9h00 to 17h00 November 29th 2006

2 SensorScope SensorScope: Aa New SensorScope in BoxScientific Instrument

2. SensorScope in a Box Plaine Morte Winter 2007 (~ 3000 m) Off-the-shelf wireless measurement system Simple and fast setup 13 SensorScope weather stations GPRS base station Data available online in real-time during the deployment Constantly learning New weather station design What Alpine deployments take Limitations to system

Surface temperature

2 SensorScope

How accurately do we measure air temperature?

VS

Rotronic MP100A (Pt-100)

Campbell Scientific CSAT3

Hendrick Huwald

March 2007, Plaine Morte

2 SensorScope SensorScope: A New Scientific Instrument

Air temperature – non-trivial to measure precisely!

3 Fall field campaign 07

Drances Watershed

Lake Geneva Rhone River

Partnership with Valais Cantonal responsible engineers and schools Often problems in the fall during extreme precipitation events

Le Génépi Study Site 1 Hazards: rock falls, land slides, mud flows

Grand-St-Bernard Study Site 2 Floods: poor predictability

3 Fall field campaign 07

Genepi rocky glacier (2600 m) Hostile enviroment: minimal local maintenance 16 Sensing stations + GPRS base station Emphasis on wind fields, temperature and precipitation

field campaign 07 Scientific Instrument 3 Fall SensorScope: A New

SensorScope in a Box: Genepi deployment Setup time: 2 days Max. number of hops: 3 (depending on weather conditions)

3 Fall field campaign 07

Deployment, Sept. 2007

Base camp

Olivier Couach, new system to attach stations

Maintaining strength

Vincent Luyet and team assemble platform

3 Fall field campaign 07 Deployment at Genepi

Alexis Berne

3 Fall field campaign 07

First Generation „Green Camera‟ sensorscope.epfl.ch

Sample view

Rock glacier 20 m of frozen soil under rocks

3 Fall field campaign 07 16 wireless weather stations • Air Temperature and relative Humidity : Sensirion SHT 75 • Surface Temperature : Zytem TN9 • Wind measurement : Davis anemometer • Rain meter : Davis • Solar radiation : Davis 6650

Role of rock glacier in local microclimate

Blue stations above „glacier‟

3 Fall field campaign 07 Weather stations measurements

Presence of rock glacier apparent, in blue

Period 20 to 31/10/2007 Air Temperature [ C]

Surface Temperature [ C]

3 Fall field campaign 07 Weather stations measurements

Wind variability over Genepi

Period 20 to 31/10/2007 Wind Speed [m/s]

Wind Direction [ ]

Spatial air temperature distribution [ C] over the Génépi rock glacier Period 28 to 30/10/2007

Wind Speed [m/s]

Spatial air temperature distribution [ C] and wind measurements – Génépi rock glacier Period 28 to 30/10/2007 October 29th – 12h00 LT Spatial Air temperature [ C] and wind measurements

Spatial wind speed [m/s] and wind measurements

3d Digital Elevation Model and air temperature distribution [ C] Period 28 to 30/10/2007 October 29th – 12h00 LT

3d Digital Elevation Model and air temperature distribution [ C] Period 22 to 23/10/2007 October 22nd – 18h00 LT

3 Fall field campaign 07

St. Bernard Pass Study site 2 St Bernard pass (2457m): Italy and Switzerland 25 Sensing stations: emphasis hydrologic response Setup time: 1.5 days Max number of hops: 4

3 Fall field campaign 07

Nick van de Giesen, John Selker July field selection

Hendrick Huwald Flux tower

Florian Habermacher, Cara Tobin, Jacques Golay, Echo Yue, Daniel Nadeau installing the optical fiber

3 Fall field campaign 07

Sensorscope stations along the optical fiber

First Analysis

Drainage toward Martingy

Hospice Outside Hospice

Comparison of stream temperatures and sensor scope stations, cold day with rain

Shallow section

Outside

Shallow

Comparison of stream temperatures and sensor scope stations, sunny day

3 Fall field campaign 07 Calculating solar radiation from 50 cm DEM, time of day,….

3 Fall field campaign 07

Comparison of measured and calculated solar radiation It pays to measure if you can

Grand St. Bernard

3 Fall field campaign 07 Wind patterns, 21 Oct. 2007

Clear channel - unlike Genepi

3 Fall field campaign 07

3 Fall field campaign 07 Calculated Latent Heat Fluxes using local stations, Oct 21, 2007

3 Fall field campaign 07 Calculated Sensible Heat Fluxes using local stations, Oct 21, 2007

3 Fall field campaign 07 Observing soil moisture pattern from SensorScope network Evaporation still important

Dry down Oct 21, 22

Rain, Oct. 19

3 Fall field campaign 07 New mesoscale pattern

Wind patterns, 24 Oct. 2007

3 Fall field campaign 07

One weather station can‟t tell the full story

3 Fall field campaign 07

3 Fall field campaign 07

analysis of soil samples d < 2 m d = 2-50 m d > 50 m

clay loam sand

texture (clay/loam) 80.0 silt-loam 70.0 st.3 60.0 loam 50.0

Loam

sandy-loam/loamy-sand 40.0

st.20

st.10

st.4 st.28

st.12

st.11

st.25 st.7

30.0

st.9 st.19

20.0

Series1

st.32 st.18

st.29

st.2

st.17

st.14 st.31

st.5

st.13

10.0

0.0 0.0

5.0

10.0

15.0 Clay

20.0

25.0

3 Fall field campaign 07

3 Fall field campaign 07

Soil Moisture

Station 3, Silt Loam

Station 5, Sandy Loam

3 Fall field campaign 07 Soil moisture patterns reflection of: evaporation, soils, upslope drainage….late in the season

3 Fall field campaign 07

Integrated Hydrologic Model

Chris Duffy

Qu and Duffy, WRR 2007

3 Fall field campaign 07

Nested Triangulation Seamless assimilation of forcing and parameters at different resolutions - sensorscope Combine large-scale simulations with nested mesoscale forecasts Gd St Bernard River Network

Cara Tobin

SensorScope Stations TIN Generation River Network

3 Fall field campaign 07

Environmental Education program to change the way people perceive the environment by „touching‟ the environment. Teaming up with schools in the watershed for the 2008 Drances deployment: Swiss Experiment, 1000 children 10 - 14

Environmental Education is a way to reach citizens from all socioeconomic backgrounds and involve the broad community

Vincent Luyet

4 LIDAR

b Position of the sounded volume: Retrieved with speed of light and time until pulse reflection comes back.

R R a

Signals proportional to 1/R2 P(R)

P  R  A P( R)  P0 2   R    R  exp  2    r  dr   R  R  0 

0

A Martin Froidevaux

4 LIDAR Atmospheric Boundary Layer Lidar Telescopes

Polychromators

Laser Acquisition

4 LIDAR

Four telescope design  nearly constant signal

1000

Simulation Measurements

900 800

0.8

Lidar return, mkV

[μV] Lidar returnlidar returns Normalized

Normalized lidar returns

1.0

700

0.1 m Ø 0.2 m Ø 0.2 m Ø 0.3 m SUM Ø

600 0.6 500

400 0.4 300

0.2 200 100

0.0 0 00

200200 250 300300 350 400400 450 500500 550 600 600 50100 100 150 Range [m] Range [m]m Range, Range, meters

4 LIDAR

Temperature and Humidity Temperature : Pure-rotational Raman branches of N2, O2 < 274 nm

Water vapor : Ro-vibrational Raman branches of H2O, N2, O2 > 274 nm

7

x 10

-30

Edge Filter Transmission 6

---- Pure Rotational Raman N2 & O2

Intensity [a.u.]

5

4

---- O2

Ro-vibrational Raman

---- N2

Ro-vibrational Raman

---- H2O Ro-vibrational Raman 3

Edge Filter

---- Elastic Line

2

1

0 264

266

268

270

272

274

276

278

280 282 284 Wavelength [nm]

286

288

290

292

294

296

298

300

4 LIDAR Filter – all 4 need to be identical to separate signals

By Alain Herzog

4 LIDAR

Water vapor polychromator Mixing ratio proportional to H2O / N2, O2, O3 correction in UV (finishing up coming weeks…)

H2 O 294.6 nm N2 283.6 nm O2 277.5 nm

4 LIDAR Pure rotational Raman spectrum of N2

T = 220 K T = 300 K

0

4.5

1.0 4.0

Raman ratio ratio Raman

0.8

Normalized line intensities

0.6 0.4 0.2

3.5

n R(T )  n

jlo

(T )

jhi

(T )

jlo

3.0

jhi

2.5

0.0 -200

-150

-100

-50

0

50

100

150

200

Frequency shift, cm-1 Optical scrambler Stage I

from Telescope

Lens

Diffraction Gratings

4x

Optical Fibers

PMT λelastic

Attenuator Stage II

Lens

Diffraction Gratings

Diffraction gratings: 600 gr/mm kdif = 10 αdif = 540 PMT Jlow

Lenses: f = 286 mm d = 132 mm

PMT Jhigh Attenuators

2.0 220

230

240

250

260

270

280

Temperature, Temperature [KK]

Frequency shift [cm-1]

Acquisition System

Normalized line intensities

re rotational Raman spectrum of N2Temperature polychromator

290

300

4 LIDAR

EPFL Lidar High temporal and spatial resolution Daytime measurement 4 Telescope design Temperature and water vapor simultaneously Scanning But it is not eye safe ! (<700m)

4 LIDAR

First test temperature

Variation de température, degrés

10 8 6 4 2 0 -2 -4 70

90

110

130

150

170

Distance, mètres

Water vapor

50 m

lidar

torche

Concentration H2O, unités arbitraires

1.00

0.95

0.90

0.85

0.80 70

90

110

130

Distance, mètres

150

170

4 LIDAR

Testing down the corridor, too short

Lidar

Sauna

S

R

R

Radiators

W

Buckets with boiling water

W

R

S

W

R

Temperature Water vapor

40

0

S

R3

Temperature, C

R2

40

30

30

20

20

10

10

Temperature and humidity sensors

0

0 0

10

20

30

40

50

Range, meters

60

70

80

Water vapor concentration, g/kg

R1

0m

4 LIDAR

Underground 450 m tunnel Closed area (safe) 6m Calibration chamber 5 points sensors

4 LIDAR

Temperature Calibration 38

Lidar Sensors

36 34

Temperature [ C]

0

Temperature, C

32 30 28 26 24 22 20 18 16 0

50

100

150

200

250

Distance, m

Range [m]

300

350

400

450

4 LIDAR

Vineyard experiment

4 LIDAR

Vineyard, September 2007

EPFL Lidar

4x Rotronic T & RH

Sonic anemometer

60 m

4 LIDAR

Along the fixed sensors

Vineyard, along the beam, not scanning profil MR num 16 avg 1 mn 06-Sep-2007 16:05:30

profil MR num 8 avg 1 mn 06-Sep-2007 15:57:30 7.5

7.5

lidar profil

sensor 38 at 135 m

sensor 38 at 135 m

sensor 58 at 169 m sensor 78 at 59 m

sensor 58 at 169 m sensor 78 at 59 m

sensor 101 at 97 m

7

Mixing Ratio [g/kg]

Mixing Ratio [g/kg]

7

lidar profil

6.5

6

5.5

sensor 101 at 97 m

6.5

6

0

20

40

60

80

100

range [m]

120

140

160

180

5.5

0

20

40

60

80

100

range [m]

120

140

160

180

4 LIDAR

Vineyard, 2007.09.11 10:20

16

Water vapor

14

Temperature [C]

20 19

12

18

10

17

8

16

6

15

4

14

Range [m]

2

13 20

70

120

170

220

270

320

370

420

470

Distance, m

Lidar Lidar

Grapes Grapes

Forest Forest

Low field Low field

0 520

Water vapor mixing ratio [g/kg]

Temperature

Water vapor mass mixing ratio, g/kg

18

21

4 LIDAR

Lidar-Sonic comparison

Vineyard experiment, 2007.09.14 12:47-13:18.

25.0

Temperature [ C]

0

Temperature, C

Lidar Sonic 24.5

24.0

23.5

12:50

12:55

13:00

13:05

Local time Local time

13:10

13:15

4 LIDAR

Vertical scans

4 LIDAR

Internal Boundary Layer from Lake Geneva

400

2007.02.21 sodar-rass Temperature 400

400 400

350

350

350 350

300

300

300 300

250

250

250 250

200

200

Range [m]

17:30-19:15

2007.02.21 sodar wind speed

2007.02.21 sodar wind direction

2007.02.21 sodarvertical wind velocity shear W [m/s] Sodar data, 21.02.07,

150

150

150 150

100

100

100 100

50

50

5050

0

0

350 300

17h30 18h00 18h30 19h

Range [m]

250 200

200 200

150 100

17h30 18h00 18h30 19h

50 0 0

5

10 15 Speed [m/s]

20

0

100 200 300 Diirection [deg]

8

8.5 9 9.5 Temperature [C]

10

00 0-3

-2 5 -1 0 10 1 speed W [m/s]

215

3

~ 710 m

4 LIDAR

Horizontal profiles

4 LIDAR

16:37

Fog and drainage flow, ABL mixing

16:45

16:53

17:02

17:10

17:18

17:27

17:35

4 LIDAR

Diurnal temperature profile over EPFL campus

Mean T. profiles, Nov. 28, 2007 10:30 11:00 11:30 12:00 12:30 13:00 13:30 14:00 14:30 15:00 15:30

450 400 350

Height, m

300 250 200

500

400 350 300 250 200

150

150

100

100

50

50

0

15:30 16:00 16:30 17:00 17:30 18:00 18:30 19:00 19:30 19:50

450

Height, m

500

0 -3 -2 -1 0

1

2

3

4

5 O

Temperature, C

6

7

8

-3 -2 -1 0

1

2

3

4

5 O

Temperature, C

6

7

8

4 LIDAR Comparison with tethersonde

Balloon 19:25-19:44 Balloon 19:45-19:57 300

Lidar 19:50-20:00

250

Height, m

200

150

100

50

0 1.0 By Alain Herzog

Martin Froidevaux

1.5

2.0

2.5

3.0 O

Temperature, C

3.5

4 LIDAR

Formation of the Stable Boundary Layer T @ 2007-11-28

DIR @ 2007-11-28

WIND SPEED @ 2007-11-28

W conponant @ 2007-11-28

250

250

250

250

200

200

200

200

150

150

150

150

Altitude [m]

Altitude [m]

Altitude [m]

Altitude [m]

Panorama of the campus, 17:00

100

100

100

100

50

50

50

50

17:00 19:30

0

0 0

2

4 T [C]

6

0 0

100

200 DIR [deg]

300

0 0

0.5

1 1.5 WIND [m/s]

2

2.5

-2

-1

0 [m/s]

1

2

Water vapor mixing ratio [g/kg]

16:00

16:30

17:00

17:30

18:00

18:30

19:00

19:30

20:00

17:30

18:00

18:30

19:00

19:30

20:00

Decrease of temperature [ C]

16:00

16:30

17:00

Summary

Sensorscope – development of distributed wireless sensor technology first field deployments, sensor scope in a box, active control reprogramming from distance, video rethinking how to model and predict risks field deployments over a wider spatial area with schools in 2008 Raman temperature, humidity lidar –a new tool for probing the lower atmosphere

thank you