Soil moisture monitoring using cosmic-ray neutrons

Potential application to drought early warning

Marek Zreda, (University of Arizona)

San Pedro, 5 March 2010

0

5
40

-3

Soil moisture, m m

15

30

Depth, cm

10

20

25

20

10

30
0.1

0.2

0.3

Soil moisture, m3 m-3

Jul-2007

Jul-2008

Jul-2009

Jul-2010

Jul-2011

Developing a Coordinated National Soil Moisture Network, NIDIS-NOAA workshop, Kansas City, 13-14 November 2013

Elements of presentation

(1)

Brief description of purpose and applications of COSMOS

(2)

Extent (including map)

(3)

(4)

(i)

COSMOS-US

(ii)

Global COSMOS

Methods
(i)

Rationale and advantages

(ii)

Cosmic-ray neutron sensing

Proposal specific to NIDIS-NOAA drought monitoring initiative

Purpose and applications of COSMOS

To create a US network that will provide consistent soil moisture data

To expand globally

Applications include:

Hydrology (recharge, drought, floods, hydraulic properties, snow water equivalent)

Ecology (biomass, evapotranspiration)

Weather forecasting (model initialization, data assimilation)

Climate forecasting (seasonal, feedbacks, forcings, teleconnections)

Agriculture (irrigation)

Engineering (slope stability etc)

Satellite remote sensing (cal/val)

Extent of COSMOS
USA

Current: 56 stationary probes

Goal: 500 probes

Mobile sensing developed

Global

Installed: 122+ probes

New funded: 105+

Proposals pending: 125+

Mobile sensing: 2+

Measurement volume of cosmic-ray probe

Horizontal footprint
Circle, radius of ca. 300 m at sea level
Nearly independent of soil moisture
Increases with decreasing air density
Decreases with increasing air humidity
Measurement thickness
70 cm in dry soils (zero hydrogen)
Decreasing to 12 cm in wet soils
Independent of air density
Independent of atmospheric humidity

Rietholzbach, Switzerland

Variations in soil moisture: circle, 400 m diameter

Rietholzbach, 11 Apr 2011

0

Depth, cm

10

15

20

25

30
20

30

40

50

Soil moisture, vol. %

60

Rietholzbach, Switzerland
Footprint of the Rietholzbach COSMOS probe (cosmos.hwr.arizona.edu)

Variations in soil moisture: circle, 400 m diameter

San Pedro, 5 March 2010

0

Depth, cm

10

15

20

25

30
10

20

30

Soil moisture, vol. %

Zreda et al., 2012, HESS 16, 4079-4099.

Footprint of the San Pedro COSMOS probe (cosmos.hwr.arizona.edu)

Variations in soil moisture: circle, 400 m diameter

Rietholzbach, 11 Apr 2011
0

SMAP-OK
23 July 2010

10

10

15

20

Depth, cm

10

Depth, cm

Depth, cm

SMAP-OK
16 September 2010

15

20

15

20

25

25

30

25

30

20

30

40

50

30

60

10

Soil moisture, vol. %

San Pedro, 5 March 2010

0

15

20

25

10

15

20

Soil water content, wt. %

Soil water content, vol. %

Mana Road

Iowa Validation Site

17 June 2010

September 2010

5
5

5
10
10

15

15

Depth, cm

Depth, cm

Depth, cm

10

20

20

20

25

25

30

30

10

20

Soil moisture, vol. %

30

15

25

10

15

20

25

Water content, vol. %

30

35

20

25

30

35

40

Soil moisture, vol. %

45

50

Variations in soil moisture: 35 COSMOS sites

4

10

22 vol. %

16

20

40 vol. %

30

15

25

35 vol. %

15

35 vol. %

25

10

15

20 vol. %

10

20 vol. %

15

8 vol. %

12

15 wt. %

Depth, cm

Sterling
16 Sep 10

SMAP-OK
16 Sep 10

10

SMAP-OK
20 Jul 10

ARM-1
22 Jul 10

Santa Rita
10 Oct 10

SMAP-OK
23 Jul 10

20
Iowa
Sep 10
30

15

25

35 vol. %

10

30 vol. %

20

8 wt. %

25 vol. %

15

15

45 vol. %

30

vol. %

Depth, cm

Toulouse
2 Mar 11
10

Island
Dairy
15 Jun 10

Mana
Road
17 Jun 10

Manitou
26 Jul 10

Rancho
no tengo
22 Aug 10

20

Kendall
29 Aug 10

Santa Rita
6 Jan 11

30

10

15

20 wt. %

20

40 vol. %

30

10

20 vol. %

15

20

40 vol. %

30

30

50 vol. %

40

30

35

40

45 vol. %

20

30

40

50 vol. %

Depth, cm

0
Rietholzbach
11 Apr 11

Desert
Chaparral
8 Mar 11

10
Marshall
23 Oct 09

Morgan
Monroe
24 Mar 11

Mozark
18 Apr 11

Coastal
Sage
9 Mar 11

20

Neb 3
23 Apr 11

30

6 wt. %

10

30 wt. %

20

10

15

20 wt. %

8 wt. %

15

wt. % 25

25 vol. %

15

10

15

20 vol. %

Depth, cm

San Pedro
7 Jul 07

San Pedro
3 Apr 10

San Pedro
12 Nov 09

10

San Pedro
9 Aug 07

20

San Pedro
6 Dec 08
San Pedro
12 Feb 10

San Pedro
5 Mar 10

30

15

25 wt. %

10

15

20 vol. %

15

20

25

30 wt. %

20

30

40

50 vol. %

10

20

30

40 vol. %

20

30

40 vol. %

35

40

45 vol. %

Depth, cm

10

20

Park Falls
20 Jul 11

Harvard
2 May 11
Tonzi
11 May 11

Metolius
14 Jun 11
Howland
4 May 11

Chestnut
21 Mar 11

Bondville
25 Mar 11

30

Zreda et al., 2013.

Fraction of variance: 400 m vs. continental USA

Soil moisture, vol. %

50
All COSMOS sites
Soils at one COSMOS site

40
30
20
10
0
Jul
140

Aug

Sep

Oct

Nov

120

total

100

soils

80

60
40
20
0
Jul
1.0

Aug

Sep

Oct

Nov

Aug

Sep

Oct

Nov

0.6
0.4

2
soils/ total

0.8

0.2
0.0
Jul

Zreda et al., 2013.

Conclusions regarding horizontal variations in soil moisture

San Pedro, 5 March 2010

0

Soil moisture is highly variable in space.

Shapes of the profiles are as variable.

Neither single measurements nor profiles are

representative of the surrounding area.
Depth, cm

Small-scale variability makes up a large fraction

of the continental variability.

10

15

Therefore

20

Point measurements of soil moisture are

unreliable.

25

Need to shift to area-average measurements.

30
10

20

Soil moisture, vol. %

30

Cosmic-ray neutrons above the surface

600

July - August 1964

April - May 1965

Height in air (meters)

500

400

dry earth
300

water
200

100

0
1

10

Neutrons (10-7 cm-2 sec-1 eV-1)

Hendrick and Edge, 1966.

Production and removal of fast neutrons

Space:
incoming high-energy
cosmic-ray protons

Primary - mostly protons and alphas

Interact with magnetic field

Atmosphere:

- intensity depends on solar activity

and geomagnetic latitude

generation of
secondary cosmic rays

Interact with atmospheric nuclei

Produce secondary particles - cascade
- intensity depends on barometric
pressure
Produce fast neutrons
Remove fast neutrons
- slowing down by elastic collisions
- leads to thermalization

Ground:

- and then absorption

scattering
thermalization
absorption

The last three processes depend on the

chemical composition of the medium, in
particular on its hydrogen content.
Zreda et al., 2012.

Moderating power of rocks & neutron flux

100

Water content: 0 kg/kg

Granite
Basalt
Limestone
Quartzite

80
60
40
20
0
O

Si

Na

Ca

Al

Fe

Mg

Fe

Mg

Fe

Mg

Fe

Mg

100

Water content: 0.01 kg/kg

Percentage of total stopping power

80
60
40
20
0
O

Si

Na

Ca

Al

100

Water content: 0.03 kg/kg

80
60
40
20
0
O

Si

Na

Ca

Al

100

Water content: 0.10 kg/kg

80
60
40
20
0
O

Zreda et al., 2012.

Si

Na

Ca

Al

Zreda et al., 2008.

Moderating power and neutron intensity

(E) =

Q
E [ N H H H + ( N )]

(E) - flux of neutrons of energy E

Q - strength of source function
N - number of atoms of an element
sc - scattering cross section for an element
- log decrement of energy per collision
- slowing down power for an element
NH H H + (N ) - slowing-down power of
the medium, separated into:
(N ) - slowing-down power of the medium
minus hydrogen
NH H H - slowing-down power of hydrogen

Cosmic-ray probe

Marshall Lake, Colorado

Zreda et al., 2012.

Testing calibration: repeated ground truthing

San Pedro River valley, AZ

30

Soil moisture, m m

-3

40

20

10

Jul-2007

Jul-2008

Jul-2009

Jul-2010

Jul-2011

Zreda et al., 2012

COSMOS: Cosmic-ray soil moisture observing system

Latency: 1 hour

Zreda et al., 2012.

COSMOS sites

Main network:

54 probes in the 48 states

Additional probes:

2 in Hawaii and 5 abroad

Reliability: 95% uptime; 1 probe bad (due to lightning damage)

COSMOS web site: probe data

COSMOS public server: cosmos.hwr.arizona.edu

Mobile cosmic-ray probe (COSMOS rover)

Measures cosmic-ray intensity continuously using a probe in a vehicle.

Two modes:
(1) stop and measure (SAM);
(2) drive and measure (DAM).
Can do 1-D transects.
Can do 2-D maps (from multiple transects or multiple stops).
Has multiple large tubes for high count rates (fast measurements).
Has GPS to record positions.

Advantages:
(1) Fast mapping of soil moisture over large areas
and over variable terrains
(2) Possible with different moving platforms:
cars (done, working well)
aircraft (in progress, likely to work)
on foot (backpack rover, should work)
trains (future, likely to work)

COSMOS rover: transect, Hawaii

Desilets et al., 2010

COSMOS rover: soil moisture in Tucson Basin

40 km

25 km

COSMOS rover: soil moisture in Tucson Basin

40 km

25 km

COSMOS rover: soil moisture in Tucson Basin

Possible use of COSMOS probes for drought monitoring

Produce time series of area-average soil moisture

data at COSMOS sites

Generate soil moisture maps using stationary

COSMOS probes

Connect COSMOS sites using COSMOS rover to

enhance resolution and coverage

Produce high-resolution mesoscale maps of soil

moisture
Use COSMOS rovers on trains to produce daily
continental-scale soil-moisture transects

Determine soil properties from neutron data

Compute water mass balance
Separate evapotranspiration and biomass seasonal
changes from neutron data
Extract snow water equivalent from neutron data

Correlation: soil moisture, 0-5 cm vs. 0-30 cm

Soil samples from cores at COSMOS sites in the continental USA

Raw (unweighted)

COSMOS-weighted (for depth)

50

50

Soil samples
Regression, slope=1.03, R2=0.98
95% confidence

avg all vs avg top 5

Regression, slope = 1.06, R2 = 0.92
95% confidence limits

40

Average, 0-5 cm

Average, 0-5 cm

40

30

20

10

30

20

10

0
0

10

20

30

40

50

Average, 0-30 cm

10

20

30

Average, 0-30 cm

Zreda et al., 2013.

40

50

Conclusions regarding vertical variations in soil moisture

Individual profiles are highly variable and unpredictable.

But area-average profiles are less variable and more predictable.
Strong correlation (R2=0.98) between 0-5 cm and 0-30 cm COSMOS-weighted soil moisture.
Therefore
COSMOS data can be used to establish area-average soil moisture profiles for drought
monitoring.

Future prospects: big picture

Big change: single-probe data vs. network data

Research to date focused on single-probe data, such as

soil moisture or biomass at the probe footprint (image)
That research was necessary to build confidence in
cosmic-ray measurements
Armed with that confidence, we are now ready to
study processes that operate at mesoscale and larger
scales
Small-scale research at single sites will be
transformed into large-scale research that uses the
whole network, guided by the principle that the whole is
greater than the sum of its parts
The current network of 56 probes will be enlarged
COSMOS probes will be collocated with other
instruments and data analyzed jointly
New applications will be developed

Growth of global COSMOS

TERENO

COSMOS
Other

CosmOz
Network

Current
(installed)

New
(funded)

New
(proposals)

COSMOS (US)
CosmOz (Australia)
TERENO (Germany)
CEH (UK)
Saudi COSMOS
Other

61
13
17
5
-->26

---->61
40
--4

--several
----125
hundreds

TOTAL

>122

>105

>125

Year 2008:
Probes: 1
Year 2013:
Existing probes: 122+
Probes funded: 105
Probes proposed: 125+

Area-average hydraulic conductivity derived from COSMOS data